ASHARE Transactions-2012 PDF

ASHRAE
TRANSACTIONS
ASHRAE ANNUAL CONFERENCE

June 23–27, 2012
San Antonio, Texas
2012 Volume 118

PART 2
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Copyright American Society of Heating, Refrigerating and Air-Conditioning Engine

Provided by IHS Markit under license with ASHRAE Licensee=Hong Kong Polytechnic University/9976803100, User=WONG, Kit Chun
No reproduction or networking permitted without license from IHS Not for Resale, 12/29/2018 20:39:41 MST
ASHRAE Transactions and HVAC&R Research are the archival publications of ASHRAE for unsolicited research papers and
Society-sponsored research in the HVAC&R technical areas. ASHRAE Transactions is the official record of the research pre-
sented at the ASHRAE Annual and Winter Conferences. For more information on submitting to HVAC&R Research,
ASHRAE’s prestigious peer-reviewed bimonthly journal offering the most comprehensive reporting of research that is original
and of lasting value, visit www.ashrae.org/hvacr-research.
ASHRAE Research: Improving the Quality of Life

ASHRAE is the world’s foremost technical society in the fields of heating, ventilation, air conditioning, and refrigeration. Its
53,500 members worldwide are individuals who share ideas, identify the need for and support research, and write the
industry’s standards for testing and practice. The result of these efforts is that engineers are better able to keep indoor
environments safe and productive while protecting and preserving the outdoors for generations to come.
One of the ways that ASHRAE supports its members’ and the industry’s need for information is through ASHRAE Research.
Thousands of individuals and companies support ASHRAE Research annually, enabling ASHRAE to report new data about
material properties and building physics and to promote the application of innovative technologies.
Much of the content of ASHRAE Transactions is made possible through ASHRAE Research. Papers reporting on work funded
by ASHRAE Research are given the RP designation on the first page of the paper.
For information about ASHRAE Research or to become a member of ASHRAE, contact ASHRAE, 1791 Tullie Circle, NE,
Atlanta, GA 30329-2305 USA; telephone: 404-636-8400; www.ashrae.org.
© 2012 ASHRAE
1791 Tullie Circle, NE
Atlanta, Georgia 30329-2305
All rights reserved.
Printed in the United States of America.
Product code: 82122
ISSN 0001-2505 ISBN 978-1-936504-35-0
ASHRAE is a registered trademark of the American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc.
ASHRAE has compiled this publication with care, but ASHRAE has not investigated, and ASHRAE expressly disclaims any
duty to investigate, any product, service, process, procedure, design, or the like that may be described herein. The appearance
of any technical data or editorial material in this publication does not constitute endorsement, warranty, or guaranty by
ASHRAE of any product, service, process, procedure, design, or the like. ASHRAE does not warrant that the information in
this publication is free of errors, and ASHRAE does not necessarily agree with any statement or opinion in this publication.
The entire risk of the use of any information in this publication is assumed by the user.
No part of this publication may be reproduced without permission in writing from ASHRAE, except by a reviewer who may
quote brief passages or reproduce illustrations in a review with appropriate credit, nor may any part of this publication be repro-
duced, stored in a retrieval system, or transmitted in any way or by any means—electronic, photocopying, recording, or other—
without permission in writing from ASHRAE. Requests for permission should be submitted at www.ashrae.org/permissions.
Any updates/errata to this publication will be posted on the

ASHRAE Web site at www.ashrae.org/publicationupdates.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
RON JARNAGIN
President
2011–2012

ASHRAE TRANSACTIONS
Purpose: One of the purposes of ASHRAE is to disseminate technical information of permanent interest through
the publication of ASHRAE Transactions. Following are guidelines for submission of Conference Papers
and Technical Papers.
Information for Authors

Conference Papers
Submission: An abstract of a proposed paper should be submitted through the ASHRAE Conference Web site for
review by the conference organizers. After abstract acceptance, papers are due within three months and
should be submitted through the ASHRAE Conference Web site.
Review Process: A paper based on an accepted abstract is put into a conference paper template available at the ASHRAE
Conference Web site and is submitted as a PDF file for review through the ASHRAE Conference Web
site. The names of the reviewers are withheld from the author to ensure objective comments. The submit-
ted paper should not exceed 8 single-spaced pages using a font no smaller than 10 points. Graphs, photo-
graphs, charts, and drawings are all included in the 8-page limit. All papers undergo review by at least
two reviewers to ensure high quality of content. The ASHRAE Conferences and Expositions Committee
decides, based on these reviews, whether papers are acceptable, need revision, or should be rejected.
Final Format: The final manuscript is put into a conference paper template available at the ASHRAE Conference Web
site and is submitted as a PDF file through the ASHRAE Conference Web site. Manuscript pages should
not be numbered and the paper should not exceed 8 single-spaced pages using a font no smaller than 10
points. Graphs, photographs, charts, and drawings are all included in the 8-page limit. The first page
should include the title; the authors’ names in desired order, principal author first, with (if applicable)
ASHRAE membership grade, academic degrees at the doctoral level (e.g., PhD, DSc), and professional
certification (e.g., PE) for all authors; and an author note with the job title, affiliation, and address (city,
state, country) for each author. See the sample at the ASHRAE Conference Web site.
Units: ASHRAE policy calls for dual units of measurement with Inch-Pound (I-P) units first and International
System (SI) units in parentheses. In cases where the work being reported was calculated in SI, the SI
units may come first. For currency, include the monetary unit of the country where the research is based
followed by the US dollar conversion in parentheses. NOTE: ASHRAE requires dual units of measure-
ment in figures and tables as well as text.
Nomenclature: A special section defining symbols should be included if the number of equations is extensive. If not,
nomenclature may follow each equation.
Tables: Vertical lines should not be used in tables. Horizontal lines may be used to set off headings and at the
bottom of the table. Label each column and include units of measurement and other necessary qualify-
ing information. Units should be set off with a comma (e.g., “Velocity, m/s”). For complex tables, use
two tables, labeled “a” and “b”—one in I-P units and one in SI units.
References: The method of citation is the author-date method. The author’s last name and the year of publication are
inserted in the text in parentheses. The list of references at the end of the paper is alphabetized.
Commercialism: Commercialism in ASHRAE publications is strictly prohibited. Commercialism is deemed to be inclu-
sion of references to an organization for the promotion or commercial advantage of that organization.
Please review the ASHRAE Commercialism Policy on the ASHRAE Web site, www.ashrae.org, under
Events then Papers and Programs.
Abstracted in: ASHRAE Abstract Center; Air Infiltration & Ventilation Center's AIRBASE; American Chemical Soci-
ety's CAS (Chemical Abstracts Service) and STN (Scientific and Technical Information Network); and
Building Services Research & Information Association's Information Centre Quarterly, and IBSA
(International Building Services Abstracts).
Cited/Indexed in: Cengage Gale's Academic OneFile, InfoTrac, and Small Business Collection; Engineering Information's Ei
Compendex and Engineering Index; International Institute of Refrigeration's FRIDOC; ProQuest Technol-
ogy Research Database's CSA Materials Research Database with METADEX, CSA Engineering Research
Database, and CSA High Technology Research Database with Aerospace; SciVerse Scopus's Scopus and
Compendex; and Thomson Reuters (formerly Institute for Scientific Information) Web of Knowledge's
Conference Proceedings Citation Index (CPCI) and Index of Scientific & Technical Proceedings.
Further Information: Address questions to Conference Program Administrator, ASHRAE, 1791 Tullie Circle NE, Atlanta,
GA 30329-2305; Telephone: (404) 636-8400; Fax: (404) 321-5478.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
iv
ASHRAE Transactions
Technical Papers
Submission: An electronic version of the manuscript should be submitted through ASHRAE Manuscript Central,
http://mc.manuscriptcentral.com/ashrae.
Review Process: All papers undergo critical review by at least three reviewers to ensure high quality of content. The
ASHRAE Conferences and Expositions Committee decides, based on these reviews, whether papers are
acceptable, need revision, or should be rejected. The name of the author is withheld from the reviewers and
the names of reviewers are withheld from the author and from each other to ensure objective comments.
Format: Manuscripts should be prepared according to the guidelines in ASHRAE’s Author’s Manual and
Author’s Template (these instructions are posted in the “Instructions & Forms” section of ASHRAE
Manuscript Central). Use a font size no smaller than 10 points. The pages should be numbered and the
paper should not exceed 30 pages. When possible, line numbers should be added to the text on each
page. A maximum of 12 figures may be included with each paper. Graphs, photographs, charts, and
drawings are all designated as figures. NOTE: Authors should not identify themselves in manuscript
files to enable reviewers to maintain a blind review.
Author Listing: A separate cover sheet in Microsoft Word format should include the title; the authors’ names in desired
order, principal author first, with (if applicable) ASHRAE membership grade, academic degrees at the doc-
toral level (e.g., PhD, DSc), and professional certification (e.g., PE) for all authors; an author note with the
job title, affiliation, and address (city, state, country) for each author; and complete contact information for
the corresponding author, including mailing address, phone and fax numbers, and e-mail address.
Units: ASHRAE policy calls for dual units of measurement with Inch-Pound (I-P) units first and International
System (SI) units in parentheses. In cases where the work being reported was calculated in SI, the SI
units may come first. For currency, include the monetary unit of the country where the research is based
followed by the US dollar conversion in parentheses.
Nomenclature: A special section defining symbols should be included if the number of equations is extensive. If not,
nomenclature may follow each equation.
Figures: Artwork for figures should be submitted via ASHRAE Manuscript Central as separate tiff or eps files
saved in either CMYK mode (for color) or Grayscale mode (for black and white) at high resolution (600
ppi or greater). Do not insert figures into the Word file for final submission. A procedure to convert
graphics from Word or Excel to tiff format is available by going to the “Instructions & Forms” section of
ASHRAE Manuscript Central. Then, go to the Transactions Papers Page and the procedure can be found
in the Author’s Manual. NOTE: ASHRAE requires dual units of measurement in figures as well as text.
Tables: Tables may be inserted into the text as long as they are alterable tables. DO NOT submit tables as
images. If using Microsoft Word, use the Insert Table function from the toolbar to create tables. If not
using Microsoft Word, separate columns using single tabs. Vertical lines should not be used in tables.
Horizontal lines may be used to set off headings and at the bottom of the table. Label each column and
include units of measurement and other necessary qualifying information. Units should be set off with a
comma (e.g., “Velocity, m/s”). NOTE: ASHRAE requires dual units of measurement in tables as well as
text. For complex tables, use two tables, labeled “a” and “b”—one in I-P units and one in SI units).
References: The method of citation is the author-date method. The author’s last name and the year of publication are
inserted in the text in parentheses. The list of references at the end of the paper is alphabetized.
Commercialism: Commercialism in ASHRAE publications is strictly prohibited. Commercialism is deemed to be inclu-
sion of references to an organization for the promotion or commercial advantage of that organization.
Please review the ASHRAE Commercialism Policy on the ASHRAE Web site, www.ashrae.org, under
Events then Papers and Programs.
Abstracted in: ASHRAE Abstract Center; Air Infiltration & Ventilation Center's AIRBASE; American Chemical Soci-
ety's CAS (Chemical Abstracts Service) and STN (Scientific and Technical Information Network); and
Building Services Research & Information Association's Information Centre Quarterly, and IBSA
(International Building Services Abstracts).
Cited/Indexed in: Cengage Gale's Academic OneFile, InfoTrac, and Small Business Collection; Engineering Information's Ei
Compendex and Engineering Index; International Institute of Refrigeration's FRIDOC; ProQuest Technol-
ogy Research Database's CSA Materials Research Database with METADEX, CSA Engineering Research
Database, and CSA High Technology Research Database with Aerospace; SciVerse Scopus's Scopus and
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Compendex; and Thomson Reuters (formerly Institute for Scientific Information) Web of Knowledge's
Conference Proceedings Citation Index (CPCI) and Index of Scientific & Technical Proceedings.
Further Information: Address questions to Conference Program Administrator, ASHRAE, 1791 Tullie Circle NE, Atlanta,
GA 30329-2305; Telephone: (404) 636-8400; Fax: (404) 321-5478.
© 2012 ASHRAE
v
CONTENTS
Host Chapter and Conferences and Expositions Committee .................................................................................. xii
State of the Society Address, Ron Jarnagin, 2011–2012 ASHRAE President ...................................................... xiii
Inaugural Address, Thomas E. Watson, 2012–2013 ASHRAE President ............................................................... xv
Technical Program ................................................................................................................................................. xix
CONFERENCE PAPERS
SA-12-C001 Achieving Deep Energy Savings in Existing Buildings through Integrated Design
Dave Moser, Guopeng Liu, Weimin Wang, and Jian Zhang ......................................................3
SA-12-C002 Right Steps for Retrofits: Byron G. Rogers Federal Office Building Case Study
Kendra Tupper, Nicole Hammer, Rick Osbaugh, and Michelle Swanson ................................11
SA-12-C003 Current Performance of Ground Source Heat Pumps for Space Conditioning 
and for Water Heating under Simulated Occupancy Conditions
Moonis R. Ally, Jeffrey D. Munk, Van D. Baxter, and Anthony C. Gehl....................................19
SA-12-C004 Measured Impact on Space Conditioning Energy Use in a Residence 

Due to Operating a Heat Pump Water Heater inside the Conditioned Space
Jeffrey D. Munk, Moonis R. Ally, and Van D. Baxter ................................................................27
SA-12-C005 Thermal Storage Properties of a Hybrid Ground Source Heat Pump

Amanda Pertzborn, Greg Nellis, and Sanford Klein .................................................................34
SA-12-C006 Field Trials of a Waterless Home Heating and Humidification Technology

Dexin Wang, Shawn Scott, Ainan Bao, and William Liss .........................................................42
SA-12-C007 Environmental Noise Case Studies: Air-Cooled Refrigeration Chiller Installations 

near Residential Structures
Jack B. Evans, Chad N. Himmel, and Joshua D. Leasure .......................................................50
SA-12-C008 Field-Test of the New ASHRAE/CIBSE/USGBC Performance Measurement Protocols: 

Intermediate and Advanced Level Indoor Environmental Quality Protocols
Hyojin Kim and Jeff S. Haberl...................................................................................................58
SA-12-C009 An Experimental and Simulation Study of Lighting Performance 

in Offices with Automated Roller Shades
Hui Shen and Athanasios Tzempelikos....................................................................................66
SA-12-C010 Impacts of Air Filters on Energy Consumption in Typical HVAC Systems

Nabil Nassif ..............................................................................................................................74
SA-12-C011 Design and Analysis of an Integrated Heat and Energy Recovery Ventilation System 
with Economizer Control for Net-Zero Energy Solar Houses
Carl Yu Chen, Jiten Mistry, Alan S. Fung, Wey H. Leong, and Sumeet Jhingan .....................82
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
SA-12-C012 Validation Methodology to Allow Simulated Peak Reduction 

and Energy Performance Analysis of Residential Building Envelope 
with Phase Change Materials
Paulo Cesar Tabares-Velasco, Craig Christensen, and Marcus V.A. Bianchi .........................90
vi
ASHRAE Transactions
SA-12-C013 A Comparison of Thermal Comfort between the Archetype Sustainable Homes 
at the Kortright Centre for Conservation
Agatha Pyrka and Alan S. Fung ...............................................................................................98
SA-12-C014 Long-term Evaluation of Internal Thermal Comfort with Different Kinds of Glazing Systems 
and Window Sizes: from Energetic Considerations to Users’ Comfort
A. Gasparella, F. Cappelletti, G. Pernigotto, and P. Romagnoni............................................106
SA-12-C015 Occupant Perceptions of an Indoor Thermal Environment in a Naturally Ventilated Building

Salman Ilyas, Ashley Emery, Judith Heerwagen, and Dean Heerwagen...............................114
SA-12-C016 What We Have Learned From 20 Years of HVAC System Commissioning

E. Thomas Lillie ......................................................................................................................122
SA-12-C017 Modeling and Optimization of HVAC Systems Using Artificial Intelligence Approaches
Nabil Nassif ............................................................................................................................133
SA-12-C018 Beyond the Shoebox: Thermal Zoning Approaches for Complex Building Shapes
Lillian Smith ............................................................................................................................141
SA-12-C019 A Simulation and Experimental Study of the Impact of Passive and Active Façade Systems 
on the Energy Performance of Building Perimeter Zones
Ying-Chieh Chan and Athanasios Tzempelikos .....................................................................149
SA-12-C020 Uncertainties in Achieving Energy Savings from HVAC Maintenance Measures in the Field
Kristin Heinemeier, Marshall Hunt, Marc Hoeschele, Elizabeth Weitzel, and Brett Close......157
SA-12-C021 Laboratory Measurements of HVAC Installation and Maintenance Faults

Robert Mowris, Ean Jones, and Robert Eshom .....................................................................165
SA-12-C022 Energy Modeling Basics: A Brave New World for Young Building Modeling Professionals
Jared A. Higgins .....................................................................................................................173
SA-12-C023 DOE Commercial Building Energy Asset Rating: An Application of Centralized Modeling Tool
Na Wang, Will Gorrissen, Viraj Srivastava, and Cody Taylor .................................................178
SA-12-C024 Calibration of a Building Energy Model Considering Parametric Uncertainty

Zheng O’Neill, Bryan Eisenhower, Vladimir Fonoberov, and Trevor Bailey ...........................189
SA-12-C025 Use of Flexible Research Platforms (FRP) for BIM and Energy Modeling Research
Piljae Im and Mahabir Bhandari .............................................................................................197
SA-12-C026 Using Energy Models to Verify Utility Consumption

Jared A. Higgins, Stephen D. Foster, and J. Ryan Bailey ......................................................205
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
SA-12-C027 Investigating the Performance of Two Types of Solar Domestic Water Heating (SDWH) Systems
with Drain Water Heat Recovery through Computer Simulation and Experimental Analysis
Kamyar Tanha, Alan S. Fung, and Wey H. Leong .................................................................214
SA-12-C028 Decision Making for HVAC&R System Selection for a Typical Office Building in the UK
Mehdi Shahrestani, Runming Yao, and Geoffrey K Cook ......................................................222
SA-12-C029 Laboratory Evaluation of Aftermarket Boiler Control System

Patricia Rowley and Paul Glanville.........................................................................................230
© 2012 ASHRAE
vii
SA-12-C030 Using Steady State Predictions to Improve the Transient Response 
of a Water to Air Heat Exchanger
David A. Hodgson, Peter M. Young, Charles W. Anderson, 
Douglas C. Hittle, William S. Duff, and Daniel B. Olsen .........................................................237
SA-12-C031 A High-Efficiency, Reduced-Emissions, Combustion Control System 

for Commercial and Industrial Boilers
Guido Poncia, Christoph Haugstetter, David Liscinsky, Junqiang Fan, 
Joseph Mantese, Meredith Colket, Heidi Hollick, Ricardo Garvey, Yan Zhang, 
Kenneth Swanson, Michael McCarron, and Duffy Parlett ......................................................245
SA-12-C032 Hybrid Solar Thermal and Ground Source Heat Pump System
Farzin M. Rad, Alan S. Fung, and Wey H. Leong ..................................................................253
SA-12-C033 Zero Refrigeration for Data Centres in the USA

Robert Tozer and Sophia Flucker...........................................................................................261
SA-12-C034 The Challenge of Low Energy Design in Low Energy Cost States
Stephen W. Duda ...................................................................................................................269
SA-12-C035 Targeting a Net-Zero Energy Student Center: Part 1—Thermal and Electrical Load Minimization
Trevor Caldwell, Kelton Friedrich, Brad Gregus, Ryan Verschuere, 
Jordan Anderson, Aaron Murenbeeld, Kaveh Arfaei, Mahsa Boroumand, 
David S. Wilkinson, and James S. Cotton ..............................................................................276
SA-12-C036 Targeting a Net-Zero Energy Student Center: Part 2—Systems to Meet Building Loads
Kelton Friedrich, Trevor Caldwell, Brad Gregus, Ryan Verschuere, 
Jordan Anderson, Aaron Murenbeeld, Kaveh Arfaei, Mahsa Boroumand, 
David S. Wilkinson, and James S. Cotton ..............................................................................284
SA-12-C037 Contaminant Removal Effectiveness of Displacement Ventilation Systems 

During Heating Season; Summary Results from Three Field Studies
Boualem Ouazia, Iain Macdonald, Michel Tardif, 
Alexandra Thompson, and Daniel Booth................................................................................292
SA-12-C038 Supply Air CO2-Based Demand-Controlled Ventilation for Multi-Zone HVAC Systems
Nabil Nassif ............................................................................................................................300
SA-12-C039 Modeling and Analysis for Mixed-Mode Cooling of Buildings

Jianjun Hu and Panagiota Karava ..........................................................................................308
SA-12-C040 Contaminants in Hotel Room Exhaust Air

W. Brad M. Stanley and Bryan K. Ligman ..............................................................................316
SA-12-C041 Evaluation of Thermal Comfort in Low Load Homes Supplied by High Sidewall Air Jets
El Hassan Ridouane...............................................................................................................322
SA-12-C042 Experimental Study on Ceiling Radiant Cooling Panel System with Personal Floor Diffuser
Sei Ito, Minoru Kawashima, Yoshito Arai, Michiya Suzuki, 
Koji Murakami, and Tatsuo Nobe ...........................................................................................330
SA-12-C043 Determining the Cost-Effectiveness of Various Sub-Metering Strategies

Christine C. Maurer, Peter Rojeski, Jr., and Matt Davey........................................................338
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
viii
ASHRAE Transactions
TECHNICAL PAPERS
SA-12-001 Simulations of Chilled Water Cooling Coil Delta-T Characteristics

Zhiqin Zhang, Hui Li, and Jingjing Liu ....................................................................................349
SA-12-002 Using Energy Flow Analysis to Assess Energy Savings from Recirculation Controls 
in Multi-Family Central Domestic Hot Water Systems
Yanda Zhang, Charlotte Bonneville, Stephen Wilson, Michael Maroney, 
Jeff Staller, and Julianna Yun Wei .........................................................................................357
SA-12-003 Bayesian Analysis of Savings from Retrofit Projects

John A. Shonder and Piljae Im .............................................................................................. 367
SA-12-004 Energy Savings Potential of Variable Condenser Water Flow Systems

Zhiqin Zhang and Jingjing Liu ................................................................................................380
SA-12-005 Method for Estimating Energy Savings Potential of Chilled-Water Plant Retro-Commissioning
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Zhiqin Zhang and William D. Turner.......................................................................................388
SA-12-006 ASHRAE/IES Standard 90.1 Metal Building U-factors—Part 5: Mathematical Modeling 

of Wall Assemblies and Validation by Calibrated Hot Box Measurements
Manoj K. Choudhary, Chris P. Kasprzak, David E. Musick, 
Mark J. Henry, and N. Douglass Fast ....................................................................................400
SA-12-007 Review and Economic Feasibility Study 

of the Currently Practiced New Housing Constructions in Ontario
Aya Dembo and Alan S. Fung ................................................................................................409
SA-12-008 Reliability Modeling for Data Center Cooling Systems

Bob Schuerger, Gardson Githu, Bob Arno, and Christopher Kurkjian....................................427
SA-12-009 The Development of Simplified Rack Boundary Conditions 

for Numerical Data Center Models (RP-1487)
John Z. Zhai, Knud A. Hermansen, and Salah Al-Saadi ........................................................436
SA-12-010 Simulating Central Chiller Plant Loop Delta-T with Quadratic Regression Models
Zhiqin Zhang and Jingjing Liu.................................................................................................450
SA-12-011 Calculation of Evaporation from Indoor Swimming Pools: Further Development of Formulas
Mirza Mohammed Shah .........................................................................................................460
SA-12-012 Short-Term Curtailment of HVAC Loads in Buildings (RP-1390)

Lixing Gu and Richard Raustad..............................................................................................467
SA-12-013 Development of a Tool to Improve the Energy Efficiency 

of Existing Commercial and Institutional Buildings
Angela Lewis ..........................................................................................................................475
SA-12-014 Laboratory Testing of a Fabric Air Dispersion System

D. Kulkarni, A.N. Nalla, S. Idem, and K. Gebke .....................................................................484
SA-12-015 Ventilation Requirements for Refrigerating Machinery Rooms (RP-1448)

Scot K. Waye, Ronald L. Petersen, and Anke Beyer-Lout .....................................................491
© 2012 ASHRAE
ix
SA-12-016 Secondary Control Strategies for Cycling of Air Conditioning and Refrigeration Systems
Swarooph Nirmal Seshadri and Bryan P. Rasmussen ...........................................................508
SA-12-017 Withdrawn
SA-12-018 Development of High Efficiency Home Heating and Humidification Technology

Dexin Wang, Ainan Bao, and William Liss .............................................................................523
SA-12-019 Selection of Desiccant Equipment at Altitude (RP-1339)

Nelson Fumo, Pedro J. Mago, and Eric Kozubal....................................................................531
SA-12-020 Thermal Performance and Charge Control Strategy of a Ventilated Concrete Slab (VCS) 
with Active Cooling Using Outdoor Air
Yuxiang Chen, A.K. Athienitis, and Khaled Galal ...................................................................556
SA-12-021 Thermal Performance of Building Envelope Details for Mid- and High-Rise Buildings (RP-1365)
Patrick Roppel, Mark Lawton, and Neil Norris........................................................................569
SA-12-022 Parametric Investigation of PCM Thermal Properties on Temperature of Buildings in Toronto

M. Ebrahim Poulad and Alan S. Fung ....................................................................................585
SA-12-023 Field Measurements of Thermal Conditions During Surgical Procedures 

for the Development of CFD Boundary Conditions (RP-1397)
James S. McNeill, Zhiqiang (John) Zhai, and Jean Hertzberg ...............................................596
SA-12-024 Modeling Phase Change Materials with a Building Simulation Code Developed in MATLAB
Sridhar Sadasivam, Dahai Zhang, Alan S. Fung, and Fabio Almeida....................................610
SA-12-025 Occupancy Simulation in Three Residential Research Houses

Philip R. Boudreaux, Anthony C. Gehl, and Jeffrey E. Christian ............................................625
Study of the Degradation of Typical HVAC Materials, Filters and Components 

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
SA-12-026
Irradiated by UVC Energy—Part I: Literature Search (RP-1509)
Robert E. Kauffman ................................................................................................................638
SA-12-027 Study of the Degradation of Typical HVAC Materials, Filters and Components 
Irradiated by UVC Energy— Part II: Polymers (RP-1509)
Robert E. Kauffman and J. Douglas Wolf...............................................................................648
SA-12-028 Design of Roofs for Increased Solar Potential BIPV/T Systems 

and Their Applications to Housing Units
Caroline Hachem, Andreas Athienitis, and Paul Fazio...........................................................660
SA-12-029 Achieving Net-Zero Energy Conditions in Retrofit—

A Case Study for a Performing Arts Building Using an Hourly Simulation Approach
D.J. O’Connor, C.C. Yavuzturk, and J.E. Fuller .....................................................................677
SA-12-030 HVAC Systems Commissioning in a Manufacturing Plant

Yunhua Li, Bei Zhang, Mingsheng Liu, Lixia Wu, Jinrong Wang, and Tom Lewis .................689
SA-12-031 Investigation of CO2 Tracer Gas-Based Calibration of Multi-Zone Airflow Models

Steven C. Snyder, T. Agami Reddy, and William P. Bahnfleth ..............................................698
x
ASHRAE Transactions
SOCIETY BUSINESS
2011–2012 ASHRAE Officers, Directors, Staff, and Council and Committee Members .......................................713
2011–2012 ASHRAE Chapter Officers..................................................................................................................718
2011–2012 ASHRAE Technical Committees, Task Groups, Technical Resource Groups, and 
Multidisciplinary Task Groups...........................................................................................................722
2011–2012 ASHRAE Standard and Guideline Project Committees......................................................................758
ASHRAE Past Meetings ........................................................................................................................................772
Society Presidents .................................................................................................................................................774
ASHRAE Honors and Awards ...............................................................................................................................776
2011–2012 ASHRAE Intersociety Representatives...............................................................................................794
ASHRAE Associate Societies................................................................................................................................795
In Memoriam, July 2011 to June 2012...................................................................................................................799
Index of Conference and Technical Papers, Volume 118, Part 2..........................................................................803

--`,`,``,,`,`,```,`,,``,`,``,
© 2012 ASHRAE
xi
HOST CHAPTER AND
CONFERENCES AND EXPOSITIONS COMMITTEE
SAN ANTONIO, TEXAS
ALAMO CHAPTER OFFICERS 2011–2012 ASHRAE CONFERENCES

President AND EXPOSITIONS COMMITTEE
Jennifer King Chair
President-Elect Dennis J. Wessel
Rocky Ormand Vice-Chair
Secretary Ben A. Leppard
Rey Gutierrez Members
Treasurer Pamela L. Androff
Shannon Novak Robert G. Baker
Walid Chakroun
SAN ANTONIO HOST COMMITTEE Jon J. Cohen
Wade H. Conlan
General Chair
Kelley P. Cramm
Fred Gleeson
K. William Dean
Honorary Chair
Howard Decker A. Damon Gowan
Entertainment Charles E. Henck
Jim Rodriguez Mohammad H. Hosni
Sessions Yunho Hwang
Bill Klock Julia A. Keen
Ken Graham Dustan L. Macauley III
Hospitality Sarah E. Maston
Golda Weir Michael J. McDermott
Tours Keith C. Newcomer
Rocky Ormand Robert B. Risley
Madison Bryer Monte G. Troutman
Sustainability
Joe Brooks
Information/Publicity
Shannon Novak
Daniel Nuckolls
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
xii
ASHRAE Transactions
STATE OF THE SOCIETY ADDRESS
Sustaining ASHRAE through Leadership
Ron Jarnagin, 2011–2012 ASHRAE President
Thank you very much. This has been a great year—a great I also had the opportunity to visit a number of the student
year for me, a great year for the organization. I want to thank branches, and I really love the student branches. I think it’s just
everyone in this room for the opportunity to serve in this posi- a lot of fun doing that. I visited student branches at the Univer-
tion. It’s a marvelous job. It keeps you busy. It keeps you on sity of Central Florida and at the University of South Florida
your toes, and you need a lot of help to do it. I’ve got a great in Tampa because they were jealous that I was at the University
Executive Committee. I’ve got a great Board, a great staff to of Central Florida.
help in all this. We simply couldn’t do it without them. I also got to visit with the students at Kuwait University.
I’d like to remind you of the presidential theme this year, And when I showed up at Kuwait University, those students
Sustaining ASHRAE through Leadership. We focused on had made name tags out of the new ASHRAE logo, which had
three simple bullet points: been introduced just several months before—I have mine still.
It was lovely. They had cupcakes with the new logo on them.
• developing future leadership, They had cakes with the new logo on them, and they had candy
• building our global presence, and with the new logo on it. And I can’t imagine anybody more
• strengthening our industry. excited about ASHRAE than this student branch there.
I got to go to a very good friend of mine’s favorite loca-
I’m just going to touch on each of those in the context of
tion, the college he graduated from—Ben Leppard from
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
some of the places I’ve been as a way of helping you under-
Clemson. We chartered a student branch there. I told Ben, “If
stand not so much where I travel but more about ASHRAE
you get that Clemson branch”—he told me he’d been working
members in different parts of the world and how they respond
for years to do it—“I’ll come charter it.” I did, and I got the
and what they’re doing, because I think that’s really what
double benefit of being able to have the folks from the Green-
counts.
ville Chapter come, and we had a nice presentation and we
I want to tell you that in the area of future leadership, we
chartered the branch.
implemented, through the Young Engineers in ASHRAE
(YEA) Committee, the YEA Leadership University. At the I got to meet with students at Notre Dame University in
2012 Winter Conference in Chicago, we had our first four Beirut. They have a huge student branch over there. The
young members taking part in YEA Leadership University. faculty advisor for that happens to be the head of the depart-
We brought these members to the Society conference, and we ment of mechanical engineering who also happens to be the
matched them up with the vice presidents. We brought them next incoming president of the Lebanese chapter in Beirut.
into meetings of consequence, such as the meetings of the The young lady who introduced me spoke like an orator
Executive Committee and the Board, and we let them start to when she read my introduction. She was so poised. She’d done
understand the scope, the reach, the breadth, and the issues of such a marvelous job with organizing the event it just stunned
ASHRAE—as well as the strengths. me. The capability of these students always stuns me. From
We’ve got four Leadership U participants here: Morgan just the few vignettes of the student branches that I met with
Stephens, Frank Rivera, Tiffany Bates, and Margaret Mathi- and the student groups I met with, I’d say our future is in really,
son. I encourage you to seek some of these people out and talk really good shape.
to them. They really are great people. One of them, on the Let’s move to building our global presence. We are a
application for the program, said that their goal in life was to global organization and we always have been, but in the last
make it to the ASHRAE presidency, and I’m telling you, I’m few years, we have put a lot of emphasis in this particular area.
going to try to help that person do it when I get out of the job. If you read our mission statement, it says “advancing the arts
There’s also another end of the spectrum if you look at the and sciences of HVAC&R.” It doesn’t say doing that in North
students. I had the opportunity to travel to Region XII this year America. It doesn’t say doing that in the United States. It just
and was introduced to a young man by the name of Adam, who says advancing those arts and sciences. The global community
I think is about 12 years old. It was down in Fort Myers, Flor- is very important to us—very important for our future, our
ida. This is an amazing young man. He came up to me, intro- future growth, our reach, and our influence.
duced himself to me. I found out that he had won a science fair During my visit with our good friends over at The Char-
project in HVAC&R. I found out that he was interested in our tered Institution of Building Services Engineers (CIBSE), I
industry. I found out that he was being mentored by one of the and the president of CIBSE signed an agreement to collaborate
people from the chapter down there, and I found out that on building energy labeling in the Middle East. CIBSE had
somebody gave him a volume of the ASHRAE Handbook to go some targets of opportunity, and we have subsequently devel-
home and read, and he’s figured out how to do load calcula- oped some other ones.
tions in his house at 12 years old. This is the kind of people that During discussions while in Bangalore, India, at the
we want to have in our organization. ACREX conference, we determined that it looks like
© 2012 ASHRAE
xiii
ASHRAE’s going to have an opportunity to have a bigger they’re known. They are very, very interested in having
influence, perhaps, and a bigger role in their expo. I’m pretty ASHRAE participate and help them in a wide array of activ-
excited about that, as it’s growing, it’s vibrant, and it’s really ities, including building labeling, energy efficiency projects,
a lot of fun. and sustainability. And we had the opportunity to sign a reso-
I also went to Beijing for the Chinese Association of lution. A great thing about it was that I took the president of
Refrigeration expo and had discussions with the head of that the Oryx Chapter from Qatar and told him, “I want you to be
organization about cosponsoring some specialty conferences a signatory to this because if we start doing some of this work
on hot, humid climates. We’re starting to pick up not only with their resources, they’re going to need a lot of training. The
strength with our international partners and our international next time you offer a technical session, we may double or triple
colleagues, but we’re also starting to be able to move our tech- your attendance.” There’s a great way to spread the word,
nical information and our education to a broader environment, spread ASHRAE’s influence, strengthen the industry with
and that’s pretty exciting from my perspective. knowledge, and get our local chapters involved. They’re very
We also visited Timisoara, Romania, where we had excited the local chapter is there. And so the local chapter pres-
discussions with REHVA, the European Federation of HVAC ident signed that agreement with myself and the executive
societies. We talked a lot about some of the points in my pres- director.
idential theme, particularly partnerships. I told them Europe
And finally, we’re working with the U.S. Green Building
has the same basic problem as the United States in that we are
Council and the International Code Council (ICC) about how
using an awful lot of energy in the building sector. That energy
we might pursue a consolidated green building standard. As
and the money we spend on energy saps away the money we
you probably know, now there’s the International Green
could use for investments in other areas, and we’ve got to do
Construction Code, in which Standard 189.1 is included as a
something about it.
compliance option. One of the discussions we’ve been having
I told him I don’t think that REHVA can solve the problem
is: what would it take and what would it look like if we had a
by itself in Europe, and I don’t think that the U.S. can neces-
consolidated code? We had some meetings in Washington,
sarily solve the whole problem by themselves. But I’m pretty
DC, in which the idea emerged that in the future ASHRAE
darn sure we have a much better chance of solving the problem
standards could be adopted as part of the International Green
if we work together. We spent a lot of time talking about how
Construction Code in their entirety. We haven’t crossed all the
we could work together, and I came away from that meeting
hurdles that we need to on that, but we had a marvelous conver-
feeling that a new relationship may be blossoming with
sation. I left that meeting thinking there’s a chance here that we
REHVA.
can make a big breakthrough, increase our influence, and also
Now let’s turn to my final presidential theme speech
improve our relationship with the ICC.
point—strengthening our industry. We’ve continued to work
with our industry partners. We told you in the past that we’re So with all those trips and the 295,000 miles around there,
working trying to bring some of the people in the industry one of the things I can assure you of was that every day and
together in some quality initiatives. We’ve worked pretty hard. every hour I was out there, I thought about what was good for
It’s been an uphill battle with some of the big players. We may ASHRAE and what was good for ASHRAE members. I hope
be refocusing that a little bit, but we’re certainly not giving up the next president does that. I’m sure that he will and all who
on that activity. When I was in Qatar we met with the Qatar follow will, but I hope that when you do your work, you
Energy and Environmental Research Institute, or QEERI, as remember that as well. Thank you very much.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
xiv
ASHRAE Transactions
INAUGURAL ADDRESS
Broadening ASHRAE’s Horizons
Thomas E. Watson, 2012–2013 ASHRAE President
The theme I have chosen for this Society year is Broad- OUR HERITAGE
ening ASHRAE’s Horizons. What does that really mean? It “We have problems relating to air pollution, water pollu-
means providing solutions for issues that affect people— tion, thermal pollution, and noise pollution among other
people in the communities in which ASHRAE members live things, but air, water, heat, and noise are just the things
and work. ASHRAE members cut their teeth on. We cannot simply sit on
You, the ASHRAE member, are a problem solver. You’ve the sidelines. We have so much talent that it is our responsi-
been a problem solver since you started in the industry. Now bility, indeed our duty as professionals, to lend our aid.” This
we need to apply these problem-solving skills to a wider group was said by ASHRAE President Stan Gilman in 1972, the year
of people in order to serve communities throughout the world. I joined ASHRAE.
The emphasis this year is about giving you more opportunities It reminds me that ASHRAE has been involved with
and more ways to fulfill our mission by the application of tech- sustainability—real sustainability from a technical basis—for
nology to better serve humanity and promote a sustainable many decades. The sustainability we are working on today is
world. ASHRAE’s heritage. It is nothing new. It is not a fad. It is here
to stay. It is important and it has been an ongoing process for
ASHRAE is a very diverse society with more than 53,500 well over 40 years.
members in 132 countries. We are a global society, and we
need to serve the members throughout the world, not just those TECHNOLOGY
in North America. So what do I mean by Broadening ASHRAE’s Horizons?
It is about ASHRAE members taking simple and easy-to-use
MY PERSONAL JOURNEY technology and applying it throughout the globe. You cannot
separate the technology and the applications of technology
Broadening ASHRAE’s Horizons is based on my experi- from the people of ASHRAE and their communities. So many
ences with ASHRAE volunteers. Leaders of ASHRAE exist at members serving on technical committees and on standards
every level: the chapter level; the regional level; and at the committees have developed the technology of our industry
Society level on technical committees, standards committees, through the years. Now we must broaden its application as we
and policy committees. My role is to support you, the volun- strive to serve additional communities.
teers of ASHRAE. As I have been involved over the years in At the 2012 Winter Conference in Chicago, ASHRAE
various ASHRAE capacities, I have looked at what our volun- rolled out a new ASHRAE logo. The logo represents who we
teers have accomplished. I have learned from everything you were, who we are, and where we want to be. ASHRAE shapes
have done. This year’s theme is really about that interaction. tomorrow’s built environment today—not just HVAC&R. We
I am grateful to have had the opportunity to learn so much from are involved in the sum of built environment technologies. We
so many of you. need to reach out and collaborate beyond our traditional
I grew up in a very small community, one in which I still boundaries. We need to expand how we think about ourselves.
live. Like many small towns, the railroad and train station were Our technology needs to consider many issues: climate,
central to its development. My father started taking me to culture, how people think about air conditioning, how they use
watch steam engines at the age of two. I was fascinated by it, and the economic environment. In my hometown, and in
machinery. By the age of six, I knew I wanted to be a mechan- every one of your communities, there are economic issues. We
ical engineer. That’s not usual in today’s environment, but need to consider our resources, the technical education of the
we’ll come back later and discuss how to get young people people, and the infrastructure available to us. These are the
challenges we face in balancing our technology. In many
involved in engineering.
cases, we have not yet achieved the balance needed to meet the
I started with an extremely narrow vision. When I joined needs of all the communities that ASHRAE serves.
ASHRAE, I was primarily involved in technical committees to When it comes to technology, we need to keep the users in
learn the technology. The ASHRAE Handbook was a valuable mind. We need to keep it simple. Do we really need those latest
resource to learn about our industry. I really wasn’t focused on technologies in all cases? Do we need super-sophisticated solu-
a greater good. For me, it was all about education. But through tions? Do we need complicated buildings? We need to focus on
working with you, my fellow members, ASHRAE has given impact, making sure the advanced technology is used through-
me the opportunity to increase my technical knowledge, meet out the life of the building, not just installed as a showpiece to
the best and brightest people in our industry, and solve prob- win an award then not used. We need to have buildings that
lems with people from all parts of the globe. It’s been a great remain viable for years to come. We need to use innovation that
experience. works. An important issue for our industry is providing simple,
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
© 2012 ASHRAE
xv
affordable solutions. We need to use global expertise to meet We need to explore cold climate applications, such as
local needs. those being discussed at a cold climate conference coming up
in Calgary, Canada, in November. We are cosponsoring the
APPLICATIONS conference with REHVA (Federation of European Heating
and Air Conditioning Associations) and SCANVAC (Scan-
Our control systems are too complicated. I remember 30 dinavian Federation of Heating, Ventilation and Sanitary
years ago when I became involved with a wonderful project. Engineering Associations). We are also looking at a tropical
I was so proud of it—a triple cascade system that took water climate conference in Shanghai with the Chinese Association
from 45°F to 200°F. It had three different machines with of Refrigeration.
different refrigerants. Wonderful... but the controls were so
complicated that only two people in the world—I happened to As I grew in ASHRAE, one of my focus areas was
be one of them—could start the machines. It was great tech- refrigerants. Maybe 400 or 500 people can relate to what I do.
nology, but it was too difficult to operate, and I had a lot of air Earlier in the year, I thought “how do I make a theme out of
miles before it was over. Those are exactly the types of solu- that?” I didn’t even try. However, I have to be true to my past,
tions that we do not need. so we’re working to reduce our impact on the environment.
Kirk Mescher, an incoming ASHRAE director-at-large, This means that we need to consider the use of flammable
tells me there are geothermal units overpumping with too refrigerants, including hydrofluorocarbons such as HFOs.
much water flow. Why is that? Because the controls are too Also, we have to look at the management of refrigerants to
complicated and the operators don’t understand the impact of ensure we have good stewardship of refrigerants.
the settings they use. We have variable-frequency drives In the fall of this year in Gaithersburg, Maryland, we are
(VFDs) that people have installed that are running at a constant having a conference with NIST (National Institute of Stan-
speed on compressors and fans. Why wouldn’t you have them dards and Technology) that is focused on moving toward
control the speed to match the demand? It’s that simple. If sustainability. The idea is to provide an opportunity for people
you’re going to install a VFD, make sure you design the from all facets of refrigerant technology, natural refrigerants,
system to match the demand and train the operators to main- ammonia, CO2, and hydrofluorocarbons to get together and
tain the design. It seems fundamental, but it does not happen have meaningful discussions for better refrigerant applica-
often enough. tions. We need to encourage more opportunities around the
We have complicated thermostats. How many of you need world for these serious technical discussions.
to program thermostats for relatives or friends? It is important
that we develop interfaces for controls that can be used by a PEOPLE
typical user. We all understand that we have to advance the
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
technology to improve the built environment. However, the Ashok Virmani is the Director and Chair of ASHRAE’s
operation and maintenance of these advanced systems must be Region-at-Large, serving members stretching from Europe to
simple and easy to use to be successful. India. He knows the importance and challenges of meeting
ASHRAE members need to give guidance for other local needs. Next January in Dallas at ASHRAE’s Winter
applications, including historic buildings. That is a pet project Conference, we are having a workshop for cost-effective,
of mine. There are many historic buildings in my community; simple, and efficient systems for economically challenged
how do we save them? I have asked Technology Council to communities. That could be in any number of U.S. cities and
help. We have met with AICARR (Italian Association of Air- it could be anywhere in the world, including in your commu-
Conditioning, Heating and Refrigeration), one of our valued nity.
Associate Society Alliance members. We are considering a Ashok tells us that ASHRAE’s global initiative is bridg-
conference in Italy to provide technical guidance for historic ing technologies across nations, languages, and cultures: over
buildings. We just can’t give up on using historic buildings. 85% of global population and 58% of land mass. ASHRAE’s
They are too valuable, and leave too large an environmental rationalization of membership fees and its economical pricing
footprint, to be neglected or abandoned. of standards and publications for developing nations have
We need to look at collaborating with lenders on energy- resulted in exponential growth in its member base. ASHRAE’s
efficiency retrofits. How can ASHRAE provide credibility so participation in international expositions and seminars across
financial institutions will know that when they lend money the developing world has led to wide dissemination of knowl-
there is a good probability they are going to get it back and the edge and promotion of regionally suited eco-friendly technol-
energy retrofits will actually pay back in a reasonable amount ogies.
of time? A beginning has been made in making large numbers of
We should be talking about “net present value” or the certified professionals available in Asia by opening several
value going forward. How do we get that money in the hands examination centers. A vast pool of potential talent awaits,
of the people who need it? It is a serious issue right now ready to be trained and absorbed. ASHRAE’s Learning Insti-
throughout the world. tute must play a leading role in this movement.
xvi
ASHRAE Transactions
ASHRAE’s continuing involvement in developing econ- The other program described by Joel Primeau is for
omies will enable it to tap into vast unexplored opportunities kindergarten through 12th grade STEM. The goal of this
for training and utilizing thousands of eager young minds. program is to work with teachers so they can help us educate
Not only ASHRAE members but also Associate Society students on the value of engineering and the wonderful oppor-
Alliance members need to collaborate to extend our industry’s tunities for careers in engineering. One thing I learned when
impact beyond what it is today. We have to work with other I attended a seminar on STEM was that parents in disadvan-
societies throughout the globe to do this. We cannot do it all taged families do not understand what engineering is and how
by ourselves. It is very important that we continue the collab- they can get their children involved. I had the advantage of
oration that has been started. parents who knew how to support me from the age of two. Not
all children have that. It is critical that we help parents guide
We need to be more inclusive to accomplish our vision.
their children, get involved, and help organize coordination
Last year’s ASHRAE President, Ron Jarnagin, challenged us
and collaboration with other societies.
to play a leadership role by bringing members and our industry
together. We need to bring together owners, operators, archi- We have, in every city, in every community, many coordi-
tects, contractors, service and maintenance personnel, engi- nated or collaborative efforts in which we could participate
neers, manufacturer’s reps, commissioning personnel, and with other organizations, such as companies and other techni-
other affected parties. We need to look at everybody involved cal societies. If you watched the U.S. Golf Open, you saw that
and bring them together so we have a solution. Everyone needs Chevron had many ads for STEM. Many companies are doing
to be involved in these integrated designs. this with support from mathematics and scientific societies
and engineering societies. All want to promote the concepts of
STEM, but sometimes their efforts in a given community are
STUDENTS AND STEM
not coordinated so there is a duplication of efforts or there are
Let me share some information about a new program. We no effective efforts. We want to get involved with helping
have a Student Design Competition that’s an applied engineer- coordinate the effort. That is something we can do to leverage
ing challenge. Joel Primeau, our 2011–12 Student Activities our work with other societies.
Committee chair, has been leading the efforts to get this
started. YOUNG ENGINEERS IN ASHRAE
According to Joel, Student Activities Committee is reach- Another exciting initiative that will help broaden
ing out to a greater audience. ASHRAE has consistently done ASHRAE’s horizons is the first Young Engineers in ASHRAE
great work with colleges and university students from all over (YEA) Leadership International weekend, which will be held
the world, with a steady stream of new student branches. Our this August in Sri Lanka. We have had these YEA leadership
expanded Student Program at our Winter Conferences now weekends for years in Atlanta and on the west coast of the
provides three days of activities for our students. ASHRAE’s United States, but this is the first to be held outside of North
Student Design Competition continues to be successfully inte- America. This serves as an excellent way we can develop lead-
grated into HVAC&R courses, and the number of entries ership, and it is being extended to many other YEA members.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
continues to grow year after year. ASHRAE’s Grants Program
also sees more applicants each year. COMMUNITY SUSTAINABILITY
This year, we are reaching out to other engineering and Another activity that is very exciting is the Community
technical students with a new design competition. This new Sustainability Project. The inspiration to create this program
design challenge is aimed at students who are not traditionally came from something I saw at the local chapter level.
involved with ASHRAE and also at traditional schools that According to Dan Pettway, a member of the Board of
have existing engineering programs. The challenge focuses on Governors of the Hampton Roads Chapter and an incoming
both refrigeration and renewable energy and will require Society vice president, the Hampton Roads Chapter of
teams of engineering students to design a portable, one-cubic- ASHRAE has teamed up with the Hampton Roads Chapter of
foot refrigeration unit powered entirely by renewable energy. the International Facilities Management Association (IFMA)
You can find the details of this design competition on the to help local not-for-profit organizations help children and
ASHRAE Student Zone website (www.ashrae.org/student their families. We are calling the program E4K, Energy for
zone). Kids. The formula is very simple, and one that any chapter can
Additionally, we are taking on big socioeconomic chal- replicate to benefit their own local community. ASHRAE and
lenges as we reach out to other like-minded societies to rein- IFMA members have teamed up to design and install energy-
vigorate the STEM coalition (STEM stands for science, saving HVAC&R equipment and ductwork so that organiza-
technology, engineering, math) to coordinate our efforts to get tions can spend more of their valuable money accomplishing
more K-12 students interested in careers in STEM. Chuck their mission and less for the cost of utilities.
Curlin, K-12 activities champion extraordinaire, will head a To date, their partnership has installed a solar hot-water
new ad hoc committee on STEM. ASHRAE will assume a system in an infant’s home and installed four new systems in
leadership role in getting more young people into engineering. a brand new Habitat for Humanity quadruplex, and they are
© 2012 ASHRAE
xvii
currently redesigning duct systems and installing new systems education for code officials, translation of our technology into
for a YWCA battered family shelter. In all of these projects, languages other than English, and training for code enforce-
engineering know-how by ASHRAE members and facility ment agencies. And once again, we are looking at STEM. In
management specialties from IFMA members have combined this case, we want to go to local school boards to look at
to provide these organizations with the most efficient, up-to- curricula to ensure it involves sufficient STEM education. If
date systems available today. The two organizations also part- we do not have STEM, all the research and development
ner to raise the revenue for these projects. Through an annual ASHRAE is doing and all the research ASHRAE is sponsor-
ASHRAE/IFMA golf tournament and an oyster and barbecue ing will not benefit our communities. We have to educate our
event, members of both organizations, who always enjoy golf young people so they have challenging, fulfilling, and produc-
and eating, enjoy it even more knowing they’re contributing to tive jobs going forward. It is vital for our communities.
these organizations and helping the children and the families
of the Hampton Roads community. ENGINEERING FOR CHANGE
To date, the organizations have combined to earn reve-
nues of upwards of $40,000 to accomplish this work, and This is another opportunity: Engineering for Change, a
we’ve only just started. We’re only in our fourth year. The nongovernmental group of engineering societies founded by
Hampton Roads Chapters of ASHRAE and IFMA invite you ASME (American Society of Mechanical Engineers), IEEE
to take their example and replicate these programs and their (Institute of Electrical and Electronics Engineers), and Engi-
benefits in your communities. neers Without Borders USA. We are partnering with these
If you are on the Internet, search www.ashrae.org/ societies so that ASHRAE members, chapters, sections, and
community, and you will find more information about the regions can work on their projects. A lot of these projects
ASHRAE Community Sustainability Project, including a really need ASHRAE technology and knowledge. This effort
toolkit. There is also a quick-start guide, a Frequently Asked is about solving challenges. Does any group do this better than
Questions document, and contacts for people who can help as ASHRAE? I think not. This effort is about improving the qual-
well as case studies and a PowerPoint presentation for train- ity of life for people around the world.
ing. The program that the Hampton Roads Chapter has devel-
oped is tailored for smaller and mid-sized chapters. IF NOT NOW, WHEN? IF NOT YOU, WHO?
This program is similar to the ASHRAE Sustainable
I’ve mentioned a lot of new initiatives to you. By introduc-
Footprint Project started by the Utah Chapter as part of
ing these new programs, we hope to provide more options for
ASHRAE Annual Conferences to leave a legacy representing
participation as an ASHRAE member. Altogether, the choices
our commitment to sustainability and to offset the environ-
might seem a bit overwhelming, but it’s about finding the
mental impact of holding the Annual Conference. There is a
program that speaks to you and will allow you to make your
footprint project right here in San Antonio, and there was one
greatest impact.
in Montreal for the 2011 Annual Conference.
So what can you, the ASHRAE member, do? We need to
GOVERNMENT INTERACTION focus on our greatest impact. We need to match the technology
I would like to share another quote: “As a society we to the need. We need affordable regional technologies, and we
should be interested in legislation for the protection of health need to benefit our communities and ourselves. By doing these
and comfort.” This comes from Walter Timmis, ASHVE pres- things, you really benefit yourself as well. It is a great payback.
ident in 1919. I know you have thought of some things you can do. Act on
So what are we doing today that is different, since we have those thoughts.
talked about this for many years? We are considering a grass- When you are looking at our technology and its applica-
roots government activities committee, which Jeff Gatlin, tions, ask how you can get involved. If you are already
Region VII Director and Regional Chair, is spearheading involved, get someone else involved; get someone to help you.
along with member Eileen Jensen. If you’re not involved, get involved. Don’t wait as long as I did.
We are looking to governments in our local communities Today is the day to get started. Our community—your
and at state levels and we are considering codes and standards community—is counting on and depends on you.
xviii
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---

ASHRAE Transactions
TECHNICAL PROGRAM
2012 ASHRAE Annual Conference
San Antonio, Texas
4. Using Energy Flow Analysis to Assess Seminar 1 (Intermediate)

Sunday Energy Savings from Recirculation Con-
trols in Multi-Family Central Domestic Hot Breaking Boundaries in
Water Systems (SA-12-002) Building Controls
June 24, 2012 Yanda Zhang, Ph.D., Member, Charlotte Integration
8:00 a.m.–9:30 a.m. Bonneville, Stephen Wilson, Michael
Maroney, P.E., Jeff Staller and Julianna Yun
Track: Integrated Building Controls
Room: 001A
Wei, Heschong Mahone Group, Inc., Gold
Sponsor: 07.06 Building Energy Perfor-
Technical Paper Session 1 River, CA
mance
(Intermediate) 5. Bayesian Analysis of Savings from Ret- Chair: Kimberly A. Barker, Member, Sie-
Case Studies of Efficiency rofit Projects (SA-12-003) mens Building Technologies, Inc., Buffalo
Grove, IL
Improvements in Chilled John A. Shonder, Member and Piljae Im, Oak
Advanced, integrated BAS connects
Water, Domestic Hot Water Ridge National Laboratory, Oak Ridge, TN
building systems in ways that pay off for
and Formula Standardization building owners. They optimize building sys-
Conference Paper Session 1 tems (HVAC, lighting, elevators, irrigation,
for Estimating Savings etc.) based on occupancy (e.g., access con-
Uncertainties (Intermediate)
trol, occupancy sensors) and weather condi-
Low Energy Design and tions. Building occupants interact with the
Track: HVAC&R Fundamentals and building using dashboards to extend occu-
Applications Integrated Energy System pancy conditions in their office to seeing the
Room: 007B buildings carbon-footprint. They allow
Track: Integrated Energy Systems remote monitoring and control from outside
Chair: M. Ginger Scoggins, P.E., Member,
Engineered Designs Inc., Raleigh, NC Room: 007A the building, empowering those who manage
Sponsor: 07.01 Integrated Building Design, portfolios to even coordinating consumption
This session groups a set of technical in a sustainable city.
papers that all relate to expanding our knowl- 07.04 Exergy Analysis for Sustainable Build-
edge of energy savings potential by using ings 1. Enterprise Portfolio Management
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
case studies which focus on more efficient Chair: Tom Meyer, Member, National Envi- James P. McClendon, P.E., Member,
chiller plants, domestic hot water loops in ronmental Balancing Bureau, Gaithersburg, Walmart Stores Inc., Bentonville, AR
multi-family dwelling units, and the use of MD 2. Integrated Automation in a Net Zero
Bayesian formula for estimating savings This conference paper session explores Energy Laboratory
uncertainties in retrofit projects. low energy design with integrated energy Joseph Kilcoyne, Member, SC Engineers,
systems including applications in existing Inc. , San Diego, CA
1. Simulations of Chilled Water Cooling
Coil Delta-T Characteristics (SA-12-001) buildings and retrofits of a federal office build- 3. Sustainable Cities Need Smart Integra-
ing. tion
Zhiqin Zhang, Ph.D.1, Hui Li, Ph.D., P.E.2
Pornsak Songkakul, Member, Siemens
and Jingjing Liu1, (1)Nexant, Inc., San Fran- 1. Achieving Deep Energy Savings in Building Technologies, Buffalo Grove, IL
cisco, CA, (2)Texas A & M University, College Existing Buildings Through Integrated
Station, TX Design (SA-12-C001) Seminar 2 (Advanced)
2. Energy Savings Potential of Variable Dave Moser, P.E., Member1 and Guopeng
Condenser Water Flow Systems (SA-12- Liu, P.E.2, (1)PECI, Portland, OR, (2)Pacific Cleanroom Mechanical
004) Northwest National Laboratory, Richland, System Design and
Zhiqin Zhang, Ph.D. and Jingjing Liu, Nexant, WA Operation Impact on Energy
Inc., San Francisco, CA Usage
2. Right Steps for Retrofits: Byron G. Rog-
3. Method for Estimating Energy Savings ers Federal Office Building Case Study
Track: HVAC&R Fundamentals and
Potential of Chilled Water Plant Retro- (SA-12-C002)
Applications
Commissioning (SA-12-005) Kendra Tupper, P.E., Member1, Nicole Ham-
Zhiqin Zhang, Ph.D.1 and W. Dan Turner, mer, P.E.1, Richard Osbaugh, P.E., Member2 Room: 007D
Ph.D., P.E.2, (1)Nexant, Inc., San Francisco, and Michelle Swanson, P.E., Member2, Sponsor: 09.11 Clean Spaces
CA, (2)Energy Systems Laboratory, College (1)Rocky Mountain Institute, Boulder, CO, Chair: Vincent A. Sakraida, P.E., Member,
Station, TX (2)RMH Group, Lakewood, CO Engineered Air, Denver, CO
© 2012 ASHRAE
xix
Cleanrooms typically have very high air 3. Potential Impact of An HFC Phase- and guarantee thermal comfort and good air
change rates and energy usage per square Down On the HVAC&R Industry quality for the employees. However commer-
foot, making cleanroom air very expensive Rajan Rajendran, Ph.D., Associate Member, cial kitchens also have spaces for food prep-
air. This seminar evaluates design and oper- Emerson Climate Technologies, Inc., Sidney, aration and dishwashing that experience
ational topics that can have a substantial OH different thermal environments. The existing
affect on a cleanroom’s energy usage. The knowledge of evaluation of thermal comfort
first topic evaluates the potential energy sav- Seminar 4 (Advanced) may be inadequate and unsuitable for practi-
ings and technical challenges of implement- cal application in environments like commer-
ing a cleanroom variable air flow control cial kitchens. The purpose of the ASHRAE
system. The second topic evaluates the dif-
Power Trends of Information (RP 1469) sponsored study “Thermal Com-
ferent mechanical system types that can be Technology Equipment and fort in Commercial Kitchens” is to determine
implemented for a specific cleanliness clas- the Usage of ASHRAE’s the occupant thermal comfort range for work-
sification and their respective energy usage. ers in commercial kitchens’ environments
The third topic presents research results on Thermal Guidelines in Data and to subsequently provide baseline data
airlock impact on cleanroom performance. Center Design and Operation for the design community to evaluate new
technologies for sustainable kitchen heating,
1. Evaluate Potential Energy Savings and Track: HVAC&R Fundamentals and ventilation, and air conditioning (HVAC)
Technical Challenges Implementing a Applications designs.
Cleanroom Variable Air Flow Control Sys-
Room: 103A 1. Thermal Comfort In Commercial Kitch-
tem
Sponsor: 09.09 Mission Critical Facilities, ens
Vincent A. Sakraida, P.E., Member, Engi- Technology Spaces and Electronic Equip- John Stoops, Member, KEMA, Inc., San
neered Air, Denver, CO ment Francisco, CA
Chair: Robin A. Steinbrecher, Member, Intel
2. Evaluate the Different Mechanical Sys- Corp., Dupont, WA 2. Thermal Comfort In Kitchens: Proce-
tem Types That Can Be Implemented for a Datacom (data processing and telecom- dure, Instrumentation and Surveys for
Specific Cleanliness Classification and munications) equipment advances quickly Documentation of the Thermal Environ-
Their Respective Energy Usage resulting in short product cycles and rapid ment In Kitchens
Kevin Breslin, Merrick and Co., Aurora, CO upgrade frequency. Based on the latest infor- Angela Simone, Ph.D., Member, Denmark
mation from the leading datacom equipment Technical University (KEMA, Inc.), San Fran-
3. Functionality and Performance of Air- manufacturers, this seminar summarizes the cisco, CA
locks, Anterooms and Vestibules, Impor- latest datacom equipment power trends doc-
tant Findings From Recent ASHRAE 3. Thermal Comfort In Kitchens: Evalua-
umented in the 3rd edition of the Datacom tion of Thermal Comfort In Kitchens
Research (RP-1431) Power Trends and Cooling Applications, Based On Physical Measurements and
Wei Sun, P.E., Member, Engsysco Inc., Ann authored by ASHRAE TC 9.9. Optimization Subjective Results
Arbor, MI of data center cooling may be misinterpreted Bjarne W. Olesen, SCANVAC, Lyngby, Den-
as simply increasing the inlet temperature to mark
the ITE. However, energy savings and opti-
Seminar 3 (Basic) mization must be viewed as a carefully con-
sidered and methodical process including all Seminar 6 (Intermediate)
Climate Change Basics aspects of data center optimization. This
seminar presents that process and its appli- ASHRAE Guideline 20 and
Track: HVAC&R Fundamentals and cation to all types of data centers. Better HVAC Software: A BIM
Applications 1. Datacom Component Power Trends Perspective
Room: 001B Robin A. Steinbrecher, Member, Intel Corp.,
Dupont, WA Track: Building Modeling Applications
Sponsor: 02.05 Global Climate Change
Room: 103B
Chair: Van D. Baxter, P.E., Fellow ASHRAE, 2. Datacom Equipment Power Trends
Sponsor: 01.05 Computer Applications,
Oak Ridge National Laboratory, Oak Ridge, Roger R. Schmidt, Ph.D., P.E., Member, IBM
07.01 Integrated Building Design
TN Corporation, Poughkeepsie, NY
Chair: Stephen Roth, P.E., Member, Carmel
This session discusses the basics of the 3. A Methodology for Determining the Software Corporation, San Rafael, CA
science of climate change and stratospheric Optimum Data Center Temperature for ASHRAE GPC-20 (Guideline 20) is titled
ozone depletion, the International and U.S. Energy Efficiency “Documenting HVAC&R Work Processes
regulatory directions, and the potential Michael K. Patterson, Ph.D., P.E., Member, and Data Exchange Requirements.” Its pur-
impact of these issues on the HVAC&R Intel Corp., Dupont, WA pose is to define procedures for documenting
industry. The first speaker discusses the work processes (Use Cases) and data
basics of ozone depletion and climate exchange requirements for specific HVAC&R
change science and how it drives the Kyoto
Seminar 5 (Basic) activities. For example, this guideline can be
and Montreal Protocol regulations. Speaker used by software designers to develop better
two addresses the current status of the inter-
Thermal Comfort in software for HVAC engineers, energy model-
national regulations (Kyoto and Montreal Commercial Kitchens ers, and architects. This seminar provides an
Protocols) as well as potential US domestic overview of the Guideline and how it can help
legislation and/or regulatory action. The final Track: Indoor Environmental HVAC practitioners improve their daily work-
speaker discusses what the potential future Applications flow efficiency. It also describes a real-world
regulatory actions/directions might mean for Room: 007C example of how a use case is being used to
the HVAC&R industry. Sponsor: 05.10 Kitchen Ventilation improve HVAC load calculation work-flows.
Chair: Greg DuChane, Member, Trane, 1. Overview of Guideline 20: Document-
1. Basics of Climate Change and Ozone Columbus, OH
Depletion Science ing Work Processes and Data Exchange
The commercial kitchen is a unique space Requirements
Donald Brundage, P.E., Member, Southern where many different HVAC applications Robert J. Hitchcock, Member, Hitchcock
Company Services, Atlanta, GA take place within a single environment Consulting, Kelsey, CA
(exhaust, supply, transfer, air conditioning,
2. Montreal and Kyoto Protocol Status etc.). Obviously, the main activity in the com- 2. A Real-World Use Case: Building Load
and Decisions mercial kitchen is the cooking process that is Calculations
Steven H. Bernhardt, Ph.D., Member, Hon- generating heat and effluents which must be Charles S. Barnaby, Member, Wrightsoft
eywell International, Morristown, NJ captured and exhausted in order to control Corporation, Lexington, MA
--`,`,``,,`,`,```,`,,``,`
xx
ASHRAE Transactions
3. Create a Sample Guideline 20 Use Case 1. ASHRAE Standard 90.1 Metal Building Laboratory, Madison, WI, (2)University of
Benjamin Welle, P.E., Member, Center for U-Factors-Part 5: Mathematical Modeling Wisconsin - Madison, Madison, WI, (3)Uni-
Integrated Facility Engineering (CIFE), Stan- of Wall Assemblies (SA-12-006) versity of Wisconsin, Madison, WI
ford University, Stanford, CA Chris Kasprzak, Associate Member1, David
4. Field Trials of a Waterless Home Heat-
Musick1, Mark Henry2, Doug Fast, P.E.3 and
ing and Humidification Technology (SA-
Manoj K. Choudhary4, (1)Owens Corning,
12-C006)
Toledo, OH, (2)Butler Manufacturing, Grand-
Dexin Wang, Ph.D., Shawn Scott, Ainan Bao,
Sunday view, MO, (3)Johns Manville, Littleton, CO,
(4)Owens Corning Center of Science and
Technology, Granville, OH
Ph.D and William Liss, Gas Technology Insti-
tute, Des Plaines, IL
June 24, 2012 2. Review and Economic Feasibility Study Conference Paper Session 3
of the Currently Practiced New Housing
9:45 a.m.–10:45 a.m. Construction in Ontario (SA-12-007) (Intermediate)
Alan Fung1 and Aya Dembo, Student
Member2, (1)Ryerson University, Toronto, Performance and
Technical Plenary (Intermediate) ON, Canada, (2)Ryerson University MIE, Operational Measurement
Toronto, ON, Canada
Track: Indoor Environmental
Interoperability of Smart
Conference Paper Session 2 (Basic) Applications
Building Systems and Smart Room: 007A
Grid Improving System Chair: Gregory Dobbs, Ph.D., Member,
Pennsylvania State University, Philadelphia,
Track: Integrated Energy Systems Efficiency: How Measured PA
Room: 001A Performance and Analysis It is increasingly important to be able to
Sponsor: Conferences and Expositions Can Yield a Low Energy measure the performance of buildings. The
data accumulated will allow us to better man-
Committee Solution age and predict energy use as well as evalu-
Chair: Dunstan Macauley, P.E., Member, TAI ate the effectiveness of new design,
Engineers, Owings Mills, MD Track: HVAC&R Fundamentals and
construction and operational strategies. This
As the industry moves toward smart build- Applications session reports several new performance
ings systems, the interoperability of smart Room: 007B measurement techniques.
buildings and the power grid becomes para- Chair: Carl F. Huber, WaterFurnace Interna-
mount. On the forefront of this activity of 1. Environmental Noise Case Studies:
tional. Inc., Roanoke, IN
smart grids is Lawrence E. Jones, Ph.D., Air-Cooled Refrigeration Chiller Installa-
This conference session will cover a
Director, Regulatory Affairs, Policy & Indus- tions near Residential Structures (SA-12-
range of high efficiency options for residen-
trial Relations at Alstom Grid Inc. Dr. Jones C007)
tial applications that have the potential to
has over 20 years of experience in the Jack B. Evans, P.E., Member, Chad N. Him-
lower energy consumption and/or the
energy industry with expertise in, power sys- mel, P.E., Member and Joshua D. Leasure,
installed cost. Ground source heat pumps for
tems engineering, communications and con- P.E., JEAcoustics / Engineered Vibration
space conditioning and water heating shar-
trol technologies in power grid and electricity Acoustic & Noise Solutions, Austin, TX
ing a common foundation loop, the impact of
market operations, smart grid technologies a heat pump water heater within the condi- 2. Field-Test of the New ASHRAE/CIBSE/
including renewable energy integration. Dr. tioned space, a hybrid ground source heat USGBC Performance Measurement Pro-
Jones will speak on the interoperability of pump that uses the ground as a thermal stor- tocols for Commercial Buildings: Part I
smart building systems and smart grid. age medium and humidification via a trans- Intermediate and Advanced Level Indoor
port membrane in a gas furnace will be Environmental Quality Protocols (SA-12-
1. Interoperability of Smart Building Sys- presented. The field sites, equipment, data C008)
tems and Smart Grid collection methods and analysis techniques Hyojin Kim, Student Member and Jeff S.
Lawrence Jones, Ph.D., Alstom Grid Inc., for these experimental applications will be Haberl, Fellow ASHRAE, Texas A&M Univer-
Washington, DC examined. sity, College Station, TX
1. Measured Performance and Analysis of 3. An Experimental and Simulation Study
Ground Source Heat Pumps for Space of Lighting Performance in Offices with
Conditioning and for Water Heating in a Automated Roller Shades (SA-12-C009)
Sunday Low-Energy Test House Operated Under
Simulated Occupancy Conditions (SA-12-
Hui Shen, Student Member and Athanasios
Tzempelikos, Purdue University, West Lafay-
June 24, 2012 C003) ette, IN
Moonis R. Ally, Ph.D., Member, Jeffrey D.
11:00 a.m.–12:30 p.m. Munk, Van D. Baxter, P.E., Fellow ASHRAE
and Anthony C. Gehl, Oak Ridge National
Seminar 7 (Intermediate)
Laboratory, Oak Ridge, TN
Effective Calibration of
Technical Paper Session 2 2. Measured Impact on Space Condition- Building Energy Modeling
ing Energy Use in a Residence Due to
(Intermediate) Operating a Heat Pump Water Heater Using Measured Data
Inside the Conditioned Space (SA-12- Track: Building Modeling Applications
Advanced Technologies in C004)
Room: 103B
Building Envelope Jeffrey D. Munk, Moonis R. Ally, Ph.D., Mem-
Sponsor: 04.07 Energy Calculations
ber and Van D. Baxter, P.E., Fellow ASHRAE,
Construction Oak Ridge National Laboratory, Oak Ridge, Chair: Keith R. Cockerham, P.E., Member,
TN DLB Associates, Eatontown, NJ
Track: HVAC&R Fundamentals and Building energy modeling consistency
Applications 3. Thermal Storage Properties of a Hybrid has and will continue to be an area of con-
Ground Source Heat Pump (SA-12-C005) cern for the BEM community. As BEM is
Room: 007D Amanda Pertzborn, Student Member1, G.F. relied on more and more as part of the United
Chair: Andrew Price, P.E., Member, Stanley Nellis2 and Sanford A. Klein, Ph.D.3, (1)Uni- States’ and ASHRAE’s commitment to
Consultants, Muscatine, IA versity of Wisconsin - Madison, Solar Energy reduce energy consumption, more accurate
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
© 2012 ASHRAE
xxi
and better calibrated output from the models 1. Predicted Thermal Comfort and Venti- Seminar 10 (Basic)
is critical to this energy saving effort. Con- lation Performance In Combined
tinuing with the recent successful session Hydronic Floor and Displacement Ventila- Some Acoustical Impacts of
from the 2012 Winter Conference on the tion Systems
same topic, this session shows best-practice Michel Tardif, P.Eng., Member, CanmetEN-
Poor Aerodynamic
methods to identify plausible causes of dis- ERGY Natural Resources Canada, Ottawa, Conditions in HVAC Systems
crepancies between measured and modeled ON, Canada
performance based on utility-level data, iden- Track: HVAC&R Systems & Equipment
tifying causes of discrepancies, and under- 2. Room Air Stratification and Ventilation Room: 001B
standing the challenges encountered. A Performance In Combined Chilled Ceiling Sponsor: 02.06 Sound and Vibration Con-
number of case studies from various building and Thermal Displacement Ventilation trol
types are presented. Some of these studies Systems
also compare post ECM utility data with the Chair: Mark E. Schaffer, P.E., Member,
Stefano Schiavon, Ph.D., Member, University Schaffer Acoustics Inc., Pacific Palisades,
original simulation data. of California, Berkeley, CA CA
1. Lessons Learned with Calibrating An The noise ratings for the equipment used
3. Applied Chilled Sails and Thermal Dis-
Energy Simulation Model of a High Perfor- in air distribution systems are determined
placement Ventilation
mance Hi-Rise Office Building from laboratory testing that is done with opti-
Jerry Sipes, Ph.D., P.E., Member, Price mal airflow aerodynamics at the equipment’s
Tom Webster, P.E., Member, Center for the Industries Inc., Suwanee, GA
Built Environment (CBE), University of Cali- inlet and discharge connections. Real-world
HVAC systems often do not have these opti-
fornia, Berkeley, CA 4. DV in Healthcare mal aerodynamic conditions. This seminar
2. BESTEST-EX: A Method for Testing Bob Gulick, P.E., Member, Mazzetti Nash shows how less-than-optimal airflow aerody-
Model Calibration Procedures Lipsey Burch, Portland, OR namics in and near air handling units, termi-
nal boxes and diffusers can cause the
Ron Judkoff, Ph.D., Member, National systems to produce unexpected, excessive
Renewable Energy Laboratory, Golden, CO Seminar 9 (Intermediate) noise.
3. Case Study of Calibrating Multiple 1. The Effects of Poor Aerodynamics On
Energy Simulation Studies for Hospital
Practical Applications for Diffuser Noise
and Datacenter Clients As Part of a Utility DDC Dashboards and Green Zaccary Poots, Member, Nailor Industries,
Incentive Program Kiosks Houston, TX
Nick Gmitter, Associate Member, DLB Asso-
ciates, Eatontown, NJ Track: Integrated Building Controls 2. Aerodynamic Effect of Terminal Unit
Room: 001A Noise
Seminar 8 (Intermediate) Sponsor: 01.04 Control Theory and Applica- Patrick Oliver, P.Eng., Member, E.H. Price,
tion, 07.03 Operation and Maintenance Man- Ltd., Winnipeg, MB, Canada
Exploring Indoor agement, 07.01 Integrated Building Design 3. The Effect of Poor Aerodynamics On
Environmental Applications Chair: Frank Shadpour, P.E., Fellow Air Handling Unit Noise
ASHRAE, SC Engineers, Inc., San Diego, Mark E. Schaffer, P.E., Member, Schaffer
with Displacement CA Acoustics Inc., Pacific Palisades, CA
Ventilation and Radiant What is the most useful information to
Cooling and Heating motivate occupants to reduce building Seminar 11 (Intermediate)
energy use? As owners, facility operators,
Systems and occupants become more interested in
the health and energy performance of their Standards Development for
Track: Indoor Environmental buildings, access to real-time data has 2L Flammable Refrigerants
Applications become a priority in many high performance
buildings. The displays, which were once Track: Refrigeration Applications
Room: 103A
locked away in mechanical rooms, are now Room: 007C
Sponsor: 05.03 Room Air Distribution, proudly exhibited in lobbies. This seminar
06.01 Hydronic and Steam Equipmment and Sponsor: 03.01 Refrigerants and Second-
demonstrates practical applications to inte- ary Coolants
Systems, 06.05 Radiant Heating and Cool- grate DDC and lighting controls, web ser-
ing vices, and metered energy data into effective Chair: William Walter, Member, Carrier Cor-
and eye-catching facility dashboards. Enable poration, Syracuse, NY
Chair: Mike McDermott, Member, Grumman
Butkus Associates, Evanston, IL and empower building stakeholders to be Environmental considerations are reduc-
more energy efficient by becoming better ing the number of chemicals that can be used
Radiant chilled ceilings (CC) with thermal as refrigerants. The HFCs were introduced to
informed and ultimately make better deci-
displacement ventilation (TDV) represent a replace ozone depleting CFCs and HCFCs,
sions.
promising integrated system design that but they are greenhouse gases and are sub-
combines the energy efficiency of both sub- 1. Intelligent System Integration for Facil- ject to use restrictions in some countries. As
systems with the opportunity for improved ity Dashboards and Green Kiosks a result the industry is evaluating chemicals
ventilation performance resulting from the with very low global warming potential. Some
thermally stratified environment of DV sys- Ron Burnstein, Echelon Corporation, Encini-
of these chemicals are mildly flammable and
tems. Presenters cover recent indoor envi- tas, CA
have been assigned to a new flammability
ronmental developments in TDV systems safety classification, 2L, in ASHRAE Stan-
2. Teaching Through Touch
combined with radiant cooling and heating dard 34. This seminar provides information
hydronic systems. We present appropriate Nathan Bailey, Automated Logic, Pasadena, on the changes to standards that specify
applications for schools, offices, laboratories TX requirements for the use of 2L flammable
and large volume spaces. We explain how refrigerants.
ASHRAE Standards 55-2010 and 62.1-2010 3. An Owner’s Perspective On Leveraging
impact system design of these integrated air Dashboard Information to Reduce Energy 1. ASHRAE Standard 15: Proposed
and hydronic systems. This presentation Use Changes to Incorporate 2L Refrigerants
introduces a new design method based on David Umstot, P.E., San Diego Community Phillip Johnson, P.E., Member, McQuay
laboratory experiments. College District, San Diego, CA International, Staunton, VA
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
xxii
ASHRAE Transactions
2. EN 378, ISO 5149, IEC 60335-2-40: Pro- In most data centers both the IT equip- 3. How Hood Selection, Exhaust, and
posed Changes to Incorporate 2L Refrig- ment and the data center itself are highly Makeup Air Systems Contribute to a High
erants managed to optimize power, cooling and per- Performance Building Design
Els Baert, Daikin Europe N.V., Ostende, Bel- formance. In the past that optimization did Richard T. Swierczyna, Associate Member,
gium not include simultaneous optimization of the
Food Service Technology Center, San
overall entity. This seminar explores resulting
3. Ul’s Effort to Harmonize Product Safety non-optimal design characteristics and the Ramon, CA
Requirements for A2L, A2, and A3 Refrig- methods used to operate both the IT equip-
erants 4. Considerations When Designing High
ment and the data center to solve cooling Performance Exhaust Systems
Brian Rodgers, Underwriters Laboratories, problems and better optimize its capability.
Northbrook, IL John Clark, P.E., Member, Karges-Faulcon-
The interaction of containment systems on IT
equipment is covered along with the opportu- bridge, Inc, Minneapolis, MN
nity created using thermal sensing and sim-
ulation to manage the cooling. Techniques Seminar 14 (Intermediate)
Sunday for avoiding data center limitation and
stranded capacity are addressed as they Energy Efficient Snow Melt
June 24, 2012 relate to both the ITE and the data center
cooling infrastructure.
System Design
1:30 p.m.–3:00 p.m. 1. Data Center Containment and Control:
Track: HVAC&R Systems & Equipment
Room: 007B
Data Center and ITE Interactions
Sponsor: 06.05 Radiant Heating and Cool-
David Moss1 and Robin A. Steinbrecher,
Conference Paper Session 4 Member2, (1)Dell, Inc, Austin, TX, (2)Intel
ing
(Intermediate) Chair: Devin A. Abellon, P.E., Member,
Corp., Dupont, WA
Uponor, Phoenix, AZ
Energy Conservation 2. Optimization of Containment Control Pedestrian and transportation safety,
Strategies Using Wireless Technology security, and building accessibility may
Strategies for HVAC Systems Michael K. Patterson, Ph.D., P.E., Member1 require snow and ice control afforded by
Track: HVAC&R Systems & Equipment and Raju Pandey, Ph.D.2, (1)Intel Corp., snow melting systems. Balancing energy
Room: 001B Dupont, WA, (2)University of California, and environmental issues with client expec-
Davis, Folsom, CA
Chair: Andrew Rhodes, Southland Indus- tations has never been more important,
tries, Dulles, VA 3. Data Center Operational Management: requiring greater analysis of system designs
1. Impacts of Air Filters on Energy Con- Evaluating Data Centers and Avoiding and equipment selection. The seminar
sumption in Typical HVAC Systems (SA- Stranded Capacity reviews research leading to the development
12-C010) Mark Seymour, Member, Future Facilities, of the ASHRAE Transient Snow-melting
Nabil Nassif, Ph.D., P.E., Associate Member, London, United Kingdom Analysis Software, examples of load calcula-
North Carolina A&T State University, tions based on storm characteristics using
Greensboro, NC Seminar 13 (Intermediate) the program, conversion of loads into system
2. Design and Analysis of an Integrated designs and equipment selections for vari-
Heat and Energy Recovery Ventilation Designing Commercial ous applications.
System with Economizer Control for Net- Kitchen Ventilation for High 1. Snow Melt Research and Analysis of
Zero Energy Solar Houses (SA-12-C011)
Jiten Mistry1, Carl Yu Chen, Student Performance, Low Energy Design Alternatives Using the ASHRAE
Snow Melt Software Tool
Member1, Sumeet Jhingan2, Wey Leong, Buildings
Ph.D.1, Alan Fung1, Seyed Pishavee1 and Jeffrey Spitler, Ph.D., P.E., Oklahoma State
Saman Shaban Nejad3, (1)Ryerson Univer- Track: Integrated Energy Systems University, Stillwater, OK
sity, Toronto, ON, Canada, (2)Morrison Her- Room: 007C
shfield, Brampton, ON, Canada, (3)Ryerson 2. Thermal to Hydraulic Calculation Pro-
University, North York, ON, Canada Sponsor: 05.10 Kitchen Ventilation, cedure for Snow Melting Systems
ASHRAE Std 90.1 Robert Bean PL(Eng.), Member, Healthy
3. Potential Peak Reduction and Energy Chair: Francis Kohout, P.E., Member, Heating, Calgary, AB, Canada
Savings of Innovative Building Envelope McDonald’s Corp, Oak Brook, IL
Technologies Using Phase Change Mate- A commercial kitchen is an energy inten- 3. Optimizing Circulator Selections for
rials (SA-12-C012) sive manufacturing facility. It is the largest Temperature Extremes In Glycol Mixes
Paulo Cesar Tabares-Velasco, Ph.D., Asso- consumer of energy per unit of floor area Hans Brink Hansen M.Sc., Member, Grund-
ciate Member, Craig Christensen and Mar- when compared to other commercial or insti- fos A/S, Bjerringbro, Denmark
cus Bianchi, National Renewable Energy tutional occupancies. Given this condition,
Laboratory, Golden, CO there exists opportunities to design ventila-
tion systems which minimize both energy Seminar 15 (Advanced)
Seminar 12 (Advanced) use and environmental impact, all while still Exploiting the Advances in
addressing large internal heat gains and high
Data Center Containment, ventilation requirements. To achieve these Absorption Fundamentals
Control and Modeling: goals, this seminar provides guidance on for Transformative Impact
general design approaches as well as spe-
Interactions and cific design examples. Track: HVAC&R Systems & Equipment
Room: 007A
Optimization Opportunities
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
1. How Does Standard 189 Impact My Sponsor: 08.03 Absorption and Heat Oper-
Track: HVAC&R Fundamentals and Kitchen Ventilation Design? ated Machines
Francis Kohout, P.E., Member, McDonald’s Chair: Ersin Gercek, P.E., Associate Mem-
Applications Corp, Oak Brook, IL
Room: 103A ber, Concord Engineering Group, Voorhees,
Sponsor: 09.09 Mission Critical Facilities, 2. CKV DCV, Energy Considerations, NJ
Technology Spaces and Electronic Equip- Codes, and Standard 90.1 In this session, three different research
ment Jeff Stein, P.E., Member, Taylor Engineering, facets towards improving performance and
Chair: David Moss, Dell, Inc, Austin, TX Alameda, CA efficiency of absorption cooling technologies
© 2012 ASHRAE
xxiii
are evaluated. First author presents method Seminar 17 (Intermediate) presents simple operation and maintenance
of utilizing absorption cooling system measures guided by trained O&M staff to
assisted with a vapor compression system to Practical Applications with optimize demand with building automation,
improve system performance at high ambi- Storage and Treatment with load shaving technologies and collaboration
ent temperatures. Second author deals with Using Reclaimed and with other parts of organizations to integrate
the application and evaluation of micro-chan-
Harvested Waters in Your HVAC&R systems with other building sys-
nel heat exchangers in absorbers. Finally, tems to enhance the effectiveness of inte-
the last authors present alternative fluids for Water Conservation Efforts
grated operational and maintenance
absorbers to improve system resistance to Track: HVAC&R Fundamentals and practices.
crystallization issues. Applications
Room: 007D 1. An Approach to Facilities Operational
1. Water/LiBr Absorption System
Sponsor: 03.06 Water Treatment Improvements: Train O&M Staff and Pro-
Assisted Vapor Compression System for
Chair: Scott Mayes, LAKOS, Olathe, KS vide Tools to Measure, Track and Adjust
High Ambient Temperatures
Yunho Hwang, Ph.D., Member, University of Everyone wants to save water, but how to Equipment and Systems
Maryland, College Park, MD do it. This seminar addresses the practical Om Taneja, Ph.D., P.E., Member, US, Gen-
realities encountered using reclaimed and eral Services Administration, Manhattan, NY
2. Physical Properties of the Lithium Bro-
harvested waters by presenting real life 2. Green Buildings Can Waste Energy
mide + 1, 2-Propanediol Aqueous Solu-
experiences with water conservation efforts, Efficiently
tion
including: 1) logistics of reclaimed/harvested
Omar Abdelaziz, Ph.D., Associate Member, Larry Spielvogel, P.E., Fellow Life Member,
Oak Ridge National Laboratory, Oak Ridge, water storage, 2) getting reclaimed water to
Consulting Engineer, Bala Cynwyd, PA
TN storage/treatment areas, 3) practical uses for
reclaimed/harvested waters, 4) expected 3. The Commissioning OPR, BOD and
3. High Efficiency Energy Recovery Utiliz- treatment requirements for reclaimed/har- Systems Manual: Make Them Work for
ing Micro-Channel Heat & Mass Exchang- vested waters, and 5) a comparison of issues You
ers Absorption System with reclaimed vs. harvested waters. Ronald Wilkinson, P.E., Member, Consultant,
Ebrahim Al-Hajri, Ph.D., Member, Petroleum New York, NY
Institute, Abu Dhabi, United Arab Emirates 1. Practical Applications with Storage
and Treatment with Using Reclaimed and
Seminar 16 (Intermediate) Harvested Waters
Modeling the Performance of

Dan Weimar, Member, Chem-Aqua, Talla-
hassee, FL
Sunday
VRF Systems June 24, 2012
2. Using Reclaimed Water for Boiler Water
Track: Building Modeling Applications Make Up
Room: 103B Mike Adams, Member, Garrett Calahan,
3:15 p.m.–4:45 p.m.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Sponsor: 08.07 Variable Refrigerant Flow Atlanta, GA
Chair: Dominic Kolandayan, Member, Daikin Seminar 19 (Intermediate)
3. Water Treatment of Harvested Waters
AC (Americas), Inc., Carrollton, TX
Robert Walki, Ph.D., Associate Member, When Integrated Controls
Expectations are at an all-time high for
Nalco, Naperville, IL
energy models to match the mechanical Are Not Integrated
design exactly. However, many energy mod-
Seminar 18 (Intermediate) Track: Integrated Building Controls
eling programs do not explicitly model Vari-
Room: 007D
able Refrigerant Flow (VRF) systems. This Strategies for
seminar presents methods that utilize con- Sponsor: 09.01 Large Building Air-Condi-
trolled laboratory data in energy simulation
Transformation of Facilities tioning Systems
software with the goal of improving existing Management to Foster Chair: John Kuempel Jr., P.E., Member,
and future VRF energy models. Techniques Economically, DeBra-Kuempel, Mechanical/Electrical, Cin-
are presented for modeling VRF systems Environmentally and Socially cinnati, OH
using eQUEST which is based on the DOE- Sustainable Buildings Three case studies are shared where
2.2 modeling engine. Finally, the differences integrated controls did not work and lessons
Track: Integrated Building Controls learned presented. These examples will help
in results among three energy modeling pro-
Room: 001A designers improve project delivery by learn-
grams based on an example building are
Sponsor: 07.03 Operation and Maintenance ing from others’ mistakes.
explored.
Management, 02.08 Building Environmental
1. Data for Publicly Available VRF Sys- Impacts and Sustainability 1. Case Studies: How We Got This Stu-
tems Chair: Om Taneja, Ph.D., P.E., Member, US, dent Union Building, Unified
Ronald Domitrovic, Ph.D., Associate Mem- General Services Administration, Manhat- Drew Rimmer, P.E., Member, Henderson
ber, Electric Power Research Institute, Knox- tan, NY Engineering, Lenexa, KS
ville, TN Buildings have more impact on the natural 2. Case Studies: City Hall Out of Control-
environment than any other sector of the Overview
2. Modeling VRF Systems Using eQUEST
economy. The future of facilities is going to
Jaeyoon Koh, Ph.D., Member, LG Electron- Howard J. McKew, P.E., Member, Build-
be governed by how effectively we can trans-
ics U.S.A., Inc., Alpharetta, GA ingSmartSoftware, Inc., North Andover, MA
form facilities management towards lowering
3. Interpreting the Variations In Results utilities and operating costs, minimizing 3. Case Studies: Commissioning Finds
When Modeling VRF Systems waste, and improving work space based on Project without Control
Pam Androff, Associate Member, Mitsubishi measurements, commissioning and retro- Wade Conlan, P.E., Member, Exp, Maitland,
Electric, Atlanta, GA commissioning of buildings. This seminar FL
xxiv
ASHRAE Transactions
Member, University of Colorado, Boulder, tion at the occupant’s interface makes it sim-
CO ple to use a room with sophisticated
Monday 3. Simulating Central Chiller Plant Loop
capabilities.
Delta-T with Quadratic Regression Mod- 1. Room Automation in Today’s Office
June 25, 2012 els (SA-12-010) Building
Zhiqin Zhang, Ph.D. and Jingjing Liu, Nexant, Markus Kissel, Siemens Industry Inc, Gubel-
8:00 a.m.–9:30 a.m. Inc., San Francisco, CA strasse , Switzerland
4. Calculation of Evaporation from Indoor 2. Specifying Integrated Lighting and
Technical Paper Session 3 Swimming Pools: Further Development HVAC Controls
of Formulas (SA-12-011) Ronald Poskevich, Blue Ridge Technologies,
(Intermediate) M. Mohammed Shah, Ph.D., P.E., Fellow Atlanta, GA
ASHRAE, Consultant, New York, NY
Modeling and Simulation for 3. Integrated BAS from the Hospital Bed
Robert Dubiel, Member, Mayo Clinic Health
Data Center Cooling, Rack Conference Paper Session 5 (Basic) System, Eau Claire, WI
Boundary Conditions,
Central Chiller Plant Loop Thermal Comfort Seminar 21 (Basic)
Delta T and Indoor Track: Indoor Environmental
Can I Determine My Loads
Swimming Pool Evaporation Applications
with My Energy Modeling
Room: 007A
Track: HVAC&R Systems & Equipment Chair: Bjarne W. Olesen, SCANVAC, Lyn- Program?
Room: 007C gby, Denmark
Track: Building Modeling Applications
Chair: Nick Gangemi, P.E., Member, Facility The basic requirement of a properly
Gateway Corp., Madison, WI designed and managed indoor environment Room: 103B
is to maintain an acceptable level of thermal Sponsor: 04.01 Load Calculation Data and
The use of probabilistic methods of pre-
comfort. This session explores various Procedures, 04.07 Energy Calculations
dicting failure in Reliability Engineering has
approaches and their impact on thermal Chair: Glenn Friedman, P.E., Member, Taylor
greatly assisted the design of reliable electri-
comfort. Engineering, Alameda, CA
cal power distribution systems for data cen-
Load calculations and energy modeling
ters. We will discuss this same proven 1. A Comparison of Thermal Comfort tools are based on the same fundamental
technique as applied to the mechanical cool- Between the Archetype Sustainable heat transfer theory. At present, the methods
ing system in a data center. As cloud comput- Homes at the Kortright Centre for Conser- are fairly similar, yet there are some key dif-
ing and computational needs grow, data vation (SA-12-C013) ferences in assumptions, especially with
centers will continue to be a large part of our Agatha E. Pyrka and Alan Fung, Ryerson regards to the weather data. This seminar
energy load. Proper design and layout of the University, Toronto, ON, Canada looks at the differences between the meth-
data center are critical to efficient energy use
ods using the ASHRAE Headquarters Build-
within the data center. Rather than just mod- 2. Long-Term Evaluation of Internal Ther-
ing. The results from both the design load
eling racks as a black box, we will validate mal Comfort with Different Kinds of Glaz-
calculation approach and the energy calcula-
them against experimental temperature and ing Systems and Window Sizes: From
tion approach are compared to actual mea-
velocity data. Chilled water loop Delta “T” is a Energetic Considerations to Users’ Com-
sured data.
critical parameter for the simulation and opti- fort (SA-12-C014)
mization of a central chilled water system. A Francesca Cappelletti, Ph.D.1, Andrea Gas- 1. Overview of Cooling Load Calculation
quadratic regression model will be used to parella, Ph.D.2, Giovanni Pernigotto, Dr.Ing.3 and Energy Modeling Procedures
model the loop Delta “T”. The modeling pro- and Piercarlo Romagnoni, Ph.D., Member1, Jeffrey Spitler, Ph.D., P.E., Oklahoma State
cedure will be divided into three steps: vari- (1)University IUAV of Venice, Venezia, Italy, University, Stillwater, OK
ables selection and filtering, best regression (2)Free University of Bolzano, Bolzano, Italy,
screening, and model adequacy check. (3)University of Padova, Vicenza, Italy 2. Comparison of Results From Cooling
Accurate calculation of evaporation from Load Calculation and Energy Modeling
swimming pools is needed to ensure proper 3. Occupant Perceptions of an Indoor Programs
sizing of HVAC equipment as well as for the Thermal Environment in a Naturally Venti- Steve Bruning, P.E., Fellow ASHRAE, New-
estimation of energy consumption. Previ- lated Building (SA-12-C015) comb & Boyd, Atlanta, GA
ously, formulas for evaporation from occu- Salman Ilyas, Member1, Ashley Emery, Fel-
low ASHRAE2, Dean Heerwagen, Life 3. Comparison of Calculated and Mea-
pied and unoccupied indoor swimming pools sured Loads for the ASHRAE HQ Building
have been published and shown to be in Member2 and Judith Heerwagen, Ph.D.3,
(1)Arup, Los Angeles, CA, (2)University of Daniel Fisher, Oklahoma State University,
good agreement with test data. However it Stillwater, OK
was applicable only to positive density differ- Washington, Seattle, WA, (3)J.H. Heerwa-
ence. This paper now extends them to nega- gen & Associates, Inc., Seattle, WA
tive density differences and includes the Seminar 22 (Intermediate)
case when air is forced over the surface of Seminar 20 (Intermediate)
the pool to remove off gases. Combined Heat and Power as
BAS Integration in the a Policy Tool to Meet State
1. Reliability Modeling for Data Center
Cooling Systems (SA-12-008) Occupied Space Mandates to Provide Clean,
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Bob Schuerger, P.E.1, Gardson Githu, P.E.1, Track: Integrated Building Controls Reliable Cost Effective
Bob Arno2 and Christopher W. Kurkjian, P.E., Power
Member3, (1)HP Critical Facility Services, Room: 001A
Los Angeles, CA, (2)HP Critical Facility Ser- Sponsor: 01.04 Control Theory and Applica- Track: Integrated Energy Systems
vices, Frankfort, NY, (3)kW Mission Critical tion
Engineering , Troy, NY Chair: James Coogan, P.E., Member, Sie- Room: 007B
mens, Buffalo Grove, IL Sponsor: 01.10 Cogeneration Systems
2. The Development of Simplified Rack Integrated BAS is not confined to equip- Chair: James Freihaut, Ph.D., Member, The
Boundary Conditions for Numerical Data ment rooms. In today’s automated room, Pennsylvania State University, State College,
Center Models (SA-12-009) lighting, shading HVAC and natural ventila- PA
John Zhai, Ph.D., Member, Knud A. Herman- tion systems work together to serve the Twenty-two states adopted Energy Effi-
sen, Student Member and Salah Al-Saadi, occupant and maximize efficiency. Integra- ciency Resources Standards (EERS)
© 2012 ASHRAE
xxv
between 2007 and 2010, passing the tipping Seminar 24 (Intermediate) 1. Designing the Indoor Environment for
point so that now more than half of all states Health and Comfort
have EERS in place for electricity, natural Proper Sizing of Copper Tube Robert Bean PL(Eng.), Member, Healthy
gas, or both. Utilities have been responding Heating, Calgary, AB, Canada
to this new policy environment by adding and and Fittings for High
2. Radiant Cooling: The Bangkok Airport
developing programs, efficient technologies, Pressure Refrigerant Peter Simmonds, Ph.D., Fellow Member, IBE
market segmentation strategies, program
approaches, and program designs. Certain Applications Consulting Engineers, Sherman Oaks, CA
states have recognized the importance of 3. Radiant Cooling: Clemson University
Combined Heat and Power technologies as
Track: Refrigeration Applications
Michael Talbot, P.E., Talbot & Associates
an important tool in economically achieving Room: 001B Consulting Engineers, Inc., Charlotte, NC
energy efficiency, reliability and carbon Sponsor: 10.03 Refrigerant Piping, 03.08
reduction. This seminar presents three stra- Refrigerant Containment
tegic arguments for increasing the use of Chair: John J. Sluga, Member, Hansen
CHP focusing on clean power production,
reliable energy supply and economics of
Technologies Corporation, Bolingbrook, IL
The purpose of the session is to address
Monday
energy supply.
inconsistencies in current practices for pre- June 25, 2012
1. Clean Power dicting failure in pressure piping. This is par-
Richard Sweetser, Member, Exergy Partners ticularly important to engineers using R410a 9:45 a.m.–10:45 a.m.
Corp., Herndon, VA and CO2 in their refrigeration designs. Pre-
sentations show hoop strain is a better pre-
2. Reliable Power dictor of failure than wall thickness for Conference Paper Session 6
Bruce Hedman, Ph.D., ICF, Arlington, VA pressure piping; compare hoop strain pre-
dicted performance to other performance (Advanced)
3. Cost Effective Power predictors; and lastly describe what changes
Gearoid Foley, Member, Integrated CHP to ASME and ASTM codes are under devel- Integrated Building Controls
Systems Corp., Princeton Junction, NJ opment with this knowledge. Track: Integrated Building Controls
1. Development and Evaluation of Copper Room: 001A
Seminar 23 (Basic) Tube and Fittings Used In R-410A and Chair: Michael R. Brambley, Ph.D., Fellow
CO2 Applications ASHRAE, Pacific Northwest National Labo-
New Acoustical Criteria and Chris Mueller, Member, Mueller Industries, ratory, Richland, WA
Methods of Measuring HVAC Memphis, TN This conference paper session shares
lessons learned in 20 years of experience in
Sound 2. Evaluation of Predicted Failure Using HVAC system commissioning services and
Dimensional Criteria In Current Stan- recommendations for solving common
Track: HVAC&R Fundamentals and dards issues found in commissioning. It then exam-
Applications Charles Stout, P.E., Member, Mueller Indus- ines a method for reducing energy consump-
Room: 007D tries, Memphis, TN tion in HVAC systems using model-based
optimization that could be integrated into
Sponsor: 02.06 Sound and Vibration Con- energy management and control systems.
3. Proposed Changes to Regulations
trol This optimization process uses evolutionary
Governing Pressure Piping
Chair: Karl L. Peterman, P.E., Member, algorithms and artificial neural networks to
Charles Stout, P.E., Member, Mueller Indus-
Vibro-Acoustics, Markham, ON, Canada implement the optimization process.
tries, Memphis, TN
Noise problems in buildings remains one
of the top causes of occupant complaints. 1. What We Have Learned from 20 Years
Recognizing this, many of the high perfor- Seminar 25 (Intermediate) of HVAC System Commissioning (SA-12-
mance building programs now require a min- C016)
E. Thomas Lillie, P.E., Member, Horizon
imum standard of acoustical performance - Radiant Cooling in High Engineering Associates, LLP, St. Louis, MO
though the method to determine compliance
has not yet been formalized. Recent Humidity Environments
2. Modeling and Optimization of HVAC
ASHRAE-sponsored research and influence Systems Using Artificial Intelligence
from international standards have led the
Track: HVAC&R Fundamentals and
Approaches (SA-12-C017)
technical committee on sound and vibration Applications Nabil Nassif, Ph.D., P.E., Associate Member,
control to change the acoustical criteria table Room: 103A North Carolina A&T State University,
in the ASHRAE Applications Handbook. This Greensboro, NC
Sponsor: 06.05 Radiant Heating and Cool-
session reviews the new table, sound criteria
ing
methods, new guides such as the IgCC and
ASHRAE 189, and a sound measurement Chair: Devin A. Abellon, P.E., Member, Conference Paper Session 7 (Basic)
methodology currently under development. Uponor, Phoenix, AZ
The use of radiant cooling systems has Optimal Zoning Strategies to
1. Standardized Methodology for Sound
Level Measurement
been proven as an alternative strategy to Maximize Building
provide superior thermal comfort while signif-
Karl L. Peterman, P.E., Member, Vibro- icantly reducing overall building energy Performance
Acoustics, Markham, ON, Canada usage. However, its adoption in many Track: Building Modeling Applications
regions has been slow due to concerns over
2. Multi-Cultural Acoustical Criteria and surface condensation. This seminar explores Room: 007A
the New ASHRAE Criteria the conditions which define an indoor envi- Chair: Megan LaHiff, P.E., Associate Mem-
Jason Swan, Member, Sandy Brown Associ- ronment conducive to human health and ber, Stan Weaver Company, Orlando, FL
ates, LLP, London, United Kingdom comfort, and show that a radiant cooling sys- 1. Beyond the Shoebox: Zoning
tem designed to such guidelines would not Approaches for Unusual Building Shapes
3. Uncertainties in Measuring HVAC
be at risk for condensation. This seminar also (SA-12-C018)
Noise
includes case studies showing how radiant Lillian Smith, Autodesk, Waltham, MA
Richard Peppin, Member, Scantek, Inc., cooling systems have been used as part of
Columbia, MD an energy efficient design solution in high 2. A Simulation and Experimental Study
humidity environments. of the Impact of Passive and Active
xxvi --`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---

ASHRAE Transactions
Façade Systems on the Energy Perfor- turbine inlet cooling, ice storage, and geoex- 1. The Need for Standards for Upper Air
mance of Building Perimeter Zones (SA- change systems. UV Applications
12-C019) Richard L. Vincent, Member, Mount Sinai
1. Chilled Water Thermal Energy Storage:
Ying-Chieh Chan, Student Member and School of Medicine, New York, NY
Case Studies
Athanasios Tzempelikos, Purdue University,
Guy S. Frankenfield, P.E., Member, Natgun 2. System Design for Upper Air Disinfec-
West Lafayette, IN
Corporation, Grand Prairie, TX tion
Chuck Dunn, Member, Lumalier Corporation,
Seminar 26 (Intermediate) 2. Integration of Energy Storage with
Memphis, TN
GeoExchange Systems
Laboratories and Ed Lohrenz, Member, Geo-Xergy Systems,
Environmental Performance Inc., Winnipeg, MB, Canada Forum 1 (Advanced)
Criteria IBD, IPD - We OK? When
Track: HVAC&R Fundamentals and Seminar 28 (Intermediate) Performance Is the Goal,
Applications Advanced Techniques for What Should Be In the Deal?
Room: 103B Editing Building Energy Track: Integrated Energy Systems
Sponsor: 09.10 Laboratory Systems Models Room: 007B
Chair: Michael Ratcliff, Ph.D., P.E., Member, Sponsor: 07.01 Integrated Building Design,
RWDI, Redlands, CA Track: Building Modeling Applications
01.07 Business, Management & General
Laboratories have special requirements Room: 103A
Legal Education
and characteristics compared to regular Chair: Molly Curtz, P.E., Member, Arup,
Chair: E. Mitchell Swann, P.E., Member,
buildings. The Labs21 Environmental Perfor- Seattle, WA
MDC Systems, Paoli, PA
mance Criteria was an attempt to establish a In this practical seminar, attendees will
IBD and IPD have become common
LEED type program to reduce the environ- learn to work more efficiently and accurately
terms in project execution today. In many of
mental impacts for laboratories. Unfortu- with large, complex building energy model
these projects, high performance is the goal.
nately the LEED organization has decided to input files. Techniques presented will include
But perceptions of performance can be sub-
not pursue a LEED for labs program. This both direct manipulation of text-based input
jective. How do you know when it is good
seminar discusses the current Labs21 crite- files and approaches for efficient editing
enough? This forum captures the experi-
ria and how such a program may advance in within the eQuest GUI (Graphical User Inter-
ences and expectations of ASHRAE mem-
the future. face), including regular expressions, BDL
bers on IBD, IPD and performance-based
expressions, and user-defined defaults. We
1. What Gives with the Labs 21 Environ- projects with an aim to help inform the indus-
will focus primarily on working with DOE-2.2
mental Performance Criteria? try about what works and doesn’t work so
(eQuest) input files, but the core techniques
Philip Wirdzek, International Institute of Sus- well in the contractual agreements covering
are also useful in other modeling software.
tainable Laboratories, Arlington, VA their projects. The feedback collected will
This seminar will help experienced modelers
help to develop a future joint guidance publi-
learn to work faster and smarter, avoiding
2. A Planned Sustainability Improvement cation from TC 1.7 and 7.1. Forewarned is
repetitive and error-prone manual editing.
Evaluation System for Labs forearmed!
Gordon Sharp, Ph.D., P.E., Member, Aircuity, 1. Advanced Techniques for Editing
Inc., Newton, MA Building Energy Models Forum 2 (Intermediate)
Molly Curtz, P.E., Member, Arup, Seattle, WA
Seminar 27 (Intermediate) Will Standard 90.1 Help or
Seminar 29 (Intermediate) Hurt the Data Center
Thermal Energy Storage: Industry?
From Concept to Completion Current Design Standards
for Upper Air UV Disinfection Track: HVAC&R Fundamentals and
Track: HVAC&R Systems & Equipment Applications
Room: 007C
Installations: Bridging Room: 001B
Sponsor: 06.09 Thermal Storage, 01.10 Global Gaps Sponsor: 09.09 Mission Critical Facilities,
Cogeneration Systems Track: HVAC&R Fundamentals and Technology Spaces and Electronic Equip-
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Chair: Robert Willis, Member, Ice Energy, Applications ment
Fort Collins, CO Room: 007D Chair: David Quirk, P.E., Member, Verizon
Thermal energy storage (TES) is a proven Sponsor: 02.09 Ultraviolet Air and Surface Wireless, Basking Ridge, NJ
technology, historically providing benefits to Treatment ANSI/ASHRAE/IESNA Standard-90.1,
system owners and utilities by shifting electric Chair: Sam Guzman, Member, American Energy Standard for Buildings, has been
loads from peak to off-peak periods. In addi- Ultraviolet Company, Schooleys Mountain, NJ amended to include data centers within the
tion, integration of TES into the design of var- This seminar discusses the need for a set of scope of the 2010 edition, following approval
ious mechanical systems can reduce cost, standards regarding the proper application of of Addendum aq. Since this change to Stan-
improve efficiency and reliability, and enable upper air UV disinfection units and how it dard 90.1 there have been many industry let-
more effective operating strategies. This ses- impacts the transmission of airborne diseases ters, public review comments, articles, and
sion provides an in-depth look at several sup- worldwide. The lack of standards and the vari- even presentations indicating concerns and
ply and demand side applications of TES ation in the application of UV wall and ceiling disagreement over the changes. This forum
demonstrating these benefits. Design and mounted fixtures has led to resistance to the reviews some of the current requirements
modeling considerations are covered along use of UV technology as a solution for trans- and get industry feedback on whether the
with performance results of the installations. mission control. The design and installation of current requirements will hurt the data center
Case studies presented feature a range of the UV fixtures has a dramatic impact on the industry and possibly require steps back-
technologies including chilled water storage, effectiveness and safety of their use. wards for innovation in data center design.
© 2012 ASHRAE
xxvii
Seminar 30 (Intermediate) Seminar 32 (Basic)
Monday Case Studies in Controls Ramification of Aisle
June 25, 2012 Track: Integrated Building Controls
Containment and Server
Airflow Design in New and
Room: 001A
11:00 a.m.–12:00 noon Sponsor: 01.04 Control Theory and Applica-
Existing Data Centers
tion Track: HVAC&R Fundamentals and
Chair: David Kahn, P.E., Member, RMH
Conference Paper Session 8 Applications
Group, Lakewood, CO
(Intermediate) This session examines two differing Room: 103A
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
approaches to energy savings: a case study Sponsor: 09.09 Mission Critical Facilities,
Field and Laboratory Results of a retrofit and a programming strategy, and Technology Spaces and Electronic Equip-
performance verification for VAV terminals. ment
from HVAC Maintenance The case study is an energy retrofit of a Salt Chair: Nick Gangemi, P.E., Member, Facility
Programs Lake office building resulting in 50% energy Gateway Corp., Madison, WI
reduction. Controls were replaced, Cold and hot aisle containment has
Track: HVAC&R Systems & Equipment enhanced monitoring installed and solar PV become a widely accepted method of
and domestic water heating employed. Strat- increasing the efficiency of data center oper-
Room: 007A
egies for controlling terminal equipment to ation. The concept of physically separating
Sponsor: 06.03 Central Forced Air Heating minimize energy usage while remaining in
and Cooling Systems cold and hot air paths in the data center
compliance of ventilation requirements are makes logical sense, and aisle containment
Chair: Judith Jennings, P.E., Member, presented. Specialized database optimized provides a simple method for accomplishing
Pacific Gas and Electric, San Francisco, CA for the intake, management, analysis and this, that on the surface appears to remove
HVAC Maintenance Measures have been presentation of large volumes of “time series” the chance of any airflow mixing. However,
a central part of utility HVAC energy effi- data from building systems, sensors, and like most data center operation concepts, the
ciency efforts in recent years, yet the savings controls is described. true nature of containment is more complex.
reported using different metrics vary widely 1. Over 50% Energy Reduction with Ret- This presentation explores the theoretical
and are subject to a wide range of factors. rofits Plus Solar: An Office Building Case application of aisle containment, and the
HVAC installation and maintenance faults, Study reality of server airflow leakage that results in
weather dependency, refrigerant charge bypass air that is at the heart of the potential
method, system characteristics such as Gaylen Atkinson, Member, Atkinson Elec- inefficiencies that may be found. Methods of
metering device, line length, unit location, tronics, Salt Lake City, UT eliminating this bypass air are presented.
individual vs. combined measures, and a 2. Dynamic Reset of VAV System Duct
host of other variables affect the outcome of 1. Ramification of Aisle Containment and
Pressure, Ventilation Airflows and Con- Server Airflow Design In New and Exist-
maintenance efforts. Utility programs, con- current Compliance/Performance Verifi-
tractors and researchers have spent consid- ing Data Centers
cation
erable effort trying to understand the key Daniel B. Kennedy, Associate Member, Tate
elements and to determine the most suc- Matt Dugan, Member, DLV Automation, Inc, Jessup, MD
cessful approach to achieving savings. Atlanta, GA
EM&V efforts have been limited and chal-
lenging. The three papers in this session pro- Seminar 33 (Intermediate)
vide three perspectives to broaden our
understanding this complex problem. Heine- Review of the Updated AHRI
Commissioning the Building
meier provides analysis of the sources of Standards 260 and 880
uncertainty in delivering and measuring Envelope
these programs. Stoops will focus on the Track: HVAC&R Systems & Equipment
challenges and uncertainties faced in con- Track: HVAC&R Fundamentals and
ducting evaluations of HVAC tune-up and Applications Room: 007C
refrigerant charge correction programs. Sponsor: 02.06 Sound and Vibration Con-
Mowris will provide laboratory test results of Room: 001B trol
HVAC diagnostic fault detection and the Sponsor: 07.09 Building Commissioning Chair: Chris Papadimos, Member, Papadi-
impacts of installation and maintenance Chair: Mike Eardley, P.E., Member, Cannon mos Group, San Francisco, CA
faults on the measured application Energy Design, Boston, MA More than ever the need for valid sound
Efficiency Ratio (EER) and Seasonal Energy This seminar focuses on the complexities data for ducted equipment and air terminals
Efficiency Ratio (SEER). of building envelope commissioning, specifi- is required to be subsequently used in accu-
cally examining the differences and interac- rately predicting sound levels inside the
1. Uncertainties in Achieving Energy Sav- tions between typical mechanical and building and developing appropriate sound
ings from HVAC Maintenance Measures in building envelope commissioning from pre- treatment options that have direct bearing on
the Field (SA-12-C020) design through post occupancy. Various costs. This session discusses at length the
Kristin Heinemeier, Ph.D., P.E., Member1, commissioning techniques and testing pro- recently updated AHRI Standards 260 and
Marshall B. Hunt, P.E.2, Marc A. Hoeschele, cedures utilized by the presenters are dis- 880 for testing and rating ducted air handling
P.E.3 and Elizabeth Weitzel, Affiliate3, cussed. equipment and air terminals, respectively.
(1)Western Cooling Efficiency Center, Davis,
CA, (2)Pacific Gas & Electric, Davis, CA, 1. Leading the Commissioning Process 1. Update On AHRI Standard 260, Sound
(3)Davis Energy Group, Davis, CA for the Building Envelope Rating of Ducted Equipment
David Cantrill, P.E., Associate Member, Curt Eichelberger, P.E., Member, Johnson
2. Laboratory Measurements of HVAC Commissioning & Green Build Solutions Inc., Controls, York, PA
Installation and Maintenance Faults (SA- Duluth, GA
12-C021) 2. Consequences of Recent Updates to
Robert Mowris, P.E., Member1, Robert 2. Design and Field Testing Consider- AHRI Standard 880 Performance Rating of
Eshom2 and Ean Jones2, (1)Verified Inc., ations for the Building Envelope Air Terminals
Olympic Valley, CA, (2)Verified Inc., Truckee, Fiona Aldous, Wiss, Janney, Elstner Associ- Patrick Oliver, P.Eng., Member, E.H. Price,
CA ates, Inc., Irving, TX Ltd., Winnipeg, MB, Canada
xxviii
ASHRAE Transactions
Seminar 34 (Intermediate) for updating Chapter 50 of the HVAC Appli- practices at various stages are explored
cations Handbook. including ways to setup projects for success.
Update on Evaporative Case studies are presented providing les-
Cooling Technologies: Forum 4 (Intermediate) sons learned that have helped develop and
Simple Keeps Getting Better improve this non-traditional approach.
Track: HVAC&R Fundamentals and Optimum Design
Verification: Is It Really 1. Mechanical System Commissioning of
Applications
Optimum? Mission Critical Facilities
Room: 103B
Justin Seter, Member, DLB Associates,
Sponsor: 05.07 Evaporative Cooling Track: Building Modeling Applications
Atlanta, GA
Chair: Leon Shapiro, J.D., Member, VRTX Room: 007D
Technologies, Oak Park, CA Sponsor: TG1 Optimization 2. Controls and Automation for the Com-
Evaporative cooling technologies have Chair: Stephen J. Treado, Ph.D., P.E., Mem- missioning Process
been successfully utilized for years to pro- ber, The Pennsylvania State University, Uni- Sean Graham, P.E., Member, DLB Associ-
vide energy efficient cooling. Advances in the versity Park, PA ates, Atlanta, GA
design of indirect evaporative cooling equip- Optimization is frequently misused and
ment allow for hybrid systems that improve misunderstood. Part of the reason is that
efficiencies and indoor air quality. This pro- optimization means different things to differ-
gram provides insights in their use in high
performance and sustainable data centers
ent people, ranging from a general assess-
ment to a rigorous mathematical procedure.
Monday
and big box retail buildings. When applied to the design process, optimi- June 25, 2012
zation may require certain idealizations and
1. Indirect Evaporative Cooling Offers
assumptions in the modeling process to facil- 3:30 p.m.–4:30 p.m.
Advantages for Data Center Recovery
itate a timely solution or to account for
from Power Interruption
unknown parameters. As a result, there
Nicholas H. DesChamps, Ph.D., P.E., Fellow
exists lingering questions regarding the
Life Member, eForay Consulting LLC, Las
Vegas, NV
accuracy or realism of the optimization Control Specification
results. The purpose of this forum is to pro- Fundamentals: How to Get
2. Field Results for Retrofit Hybridization vide an opportunity for designers to discuss
their use of optimization and their experi-
What You Really Want
of High Performance Evaporative Cooling
In Big Box Retail ences with verifying their results, and identify Track: Integrated Building Controls
Steve Slayzak, Associate Member, Cool- needs for optimum design verification. Room: 007D
erado Corporation, Denver, CO Sponsor: 01.04 Control Theory and Applica-
tion
Forum 3 (Basic)
How Could the Service Water
Monday Chair: Larry J. Fisher, Member, ECT Ser-
vices, Louisville, KY
Heating Handbook Chapter June 25, 2012 Over the past 30 years controls have
advanced from nearly 100% pneumatic con-
Better Assist You? 2:15 p.m.–3:15 p.m. trols (which were interoperable), to proprie-
Track: HVAC&R Fundamentals and tary DDC controls, to the current day of
Applications integrated building systems that use open
Seminar 35 (Intermediate) protocols (BACnet, LON, MODBUS). This
Room: 007B
Sponsor: 06.06 Service Water Heating Sys- Integrating Design, has presented a challenge to the engineer-
ing community. Specifications today must not
tems Commissioning and BMS only address factory and field mounted con-
Chair: Amin Delagah, Associate Member, Controls to Deliver Mission trol systems; they must provide sequences of
PG&E Food Service Technology Center, San Critical Success operation for traditional HVAC systems, new
Ramon, CA; Jim Lutz, Ph.D., Member, Law-
Track: Integrated Building Controls green technologies, building energy use opti-
rence Berkeley National Laboratory, Berke-
Room: 007D mization, and predictive control strategies;
ley, CA
Sponsor: 07.09 Building Commissioning they must specify graphical user interfaces,
The water heater is the last major energy
including dashboards and educational
using appliance to be examined by energy Chair: Gerald J. Kettler, P.E., Life Member,
kiosks; they must address network commu-
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
efficiency professionals. Hot water systems AIR Engineering and Testing, Carrollton, TX
nication and response speeds; and they
have been evolving in the last five years, with Mission critical facilities require a unique
must span the specification sections within
many new water heating products and distri- holistic approach to design, controls and
the plan and spec documents that include
bution systems architectures introduced. commissioning to ensure maximum system
the building systems integrated into the build-
Research of domestic hot water systems reliability. Going beyond the traditional
ing automation system.
has gained steam and there has been great approach and establishing a more unified
emphasis to reduce the hot water use of fix- automation project team provides the best 1. Controls Specification 101
tures and equipment. Unfortunately, some of opportunity to meet the client owner’s project Chariti Young, Member, Automated Logic
the reference documents that support sev- requirements (OPR) while minimizing total Corp., Kennesaw, GA
eral key sections of the current Service cost of ownership (TCO). Integration of these
Water Heating chapter are outdated, dating critical project team members throughout all 2. DDC: How to Get What You Really
back to 1969 for commercial and 1985 for project phases from planning through opera- Want?
residential facilities. This forum is an oppor- tions is critical to optimizing a facility that is Frank Shadpour, P.E., Fellow ASHRAE, SC
tunity for the Society to help build a roadmap reliable and easy to own and operate. Best Engineers, Inc., San Diego, CA
© 2012 ASHRAE
xxix
energy and cost saving benefits associated Chair: Andrew Persily, Ph.D., Fellow
Tuesday with UFAD through the use of proper Cx

techniques. Real examples of successes
ASHRAE, NIST, Gaithersburg, MD
Among the many indoor air quality issues
and lessons learned will be used throughout faced by building designers, engineers and
June 26, 2012 the session including data collected to dem- occupants, indoor moisture is one of the
onstrate compliance of system operation and
8:00 a.m.–9:30 a.m. performance. The seminar also highlights
most important. While not a contaminant
itself, excessive moisture levels increase the
typical misunderstandings and misconcep- likelihood for microbial growth and the result-
Technical Paper Session 4 tions of UFAD operation and the resolutions ing health affects of allergy, asthma and other
(Intermediate) developed during design, construction, and respiratory impacts. In addition, relative
operation of UFAD systems. humidity is a key factor in determining occu-
Improvements in HVAC pant comfort, and there are significant ques-
1. Design Phase Commissioning of Under
System Efficiency Floor Air Distribution Systems
tions as to the impacts of dry air on both
health and comfort. This two part seminar will
Track: HVAC&R Fundamentals and Dennis Jones, P.E., Member, GROUP14
present the latest research findings on health
Applications ENGINEERING, INC., Denver, CO
effects associated with indoor moisture and
Room: 007B provide practical perspective on how to bet-
2. UFAD System Forensics and Trouble-
Chair: Dharam V. Punwani, Avalon Consult- ter manage indoor moisture through design,
shooting Challenges
ing, Naperville, IL operation and maintenance.
Jim Megerson, P.E., Member, Aviations &
1. Short-Term Curtailment of HVAC Loads Facilities, M.E. GROUP, Overland Park, KS 1. Indoor Moisture and Viruses: Will
in Buildings (1390-RP) (SA-12-012) Damp Air Give You a Cold or Not?
3. Proper Testing and Balancing of UFAD
Lixing Gu, Ph.D., P.E., Member and Richard Yuguo Li, Ph.D., Fellow ASHRAE, Hong
Systems
Raustad, Florida Solar Energy Center, Kong University, Hong Kong, China
Donald Hill, P.E., Member, Accutec Service,
Cocoa, FL
Inc., Lee’s Summit, MO 2. Is There a Connection Between House
2. Development of a Tool to Improve the Dust Mites and the Sensation of Dryness?
Energy Efficiency of Existing Commercial Seminar 38 (Intermediate) Jan Sundell, M.D., Fellow ASHRAE, Tsing-
and Institutional Buildings (SA-12-013)
hua University, Bejing, China
Angela Lewis, P.E., Student Member, Univer- Case Studies In Engineering
sity of Reading, Alexandria, VA Ethics, Part 1 3. Indoor Moisture In ASHRAE Standards:
What’s a Designer to Do!
3. Laboratory Testing of a Fabric Air Track: HVAC&R Fundamentals and
Dipersion System (SA-12-014) Andrew Persily, Ph.D., Fellow ASHRAE,
Applications NIST, Gaithersburg, MD
Stephen A. Idem, Ph.D.1, D. Kulkarni1, Room: 007C
Ananth Nalla2 and Kevin J. Gebke, Sponsor: 01.07 Business, Management &
Member3, (1)Tennessee Tech University, General Legal Education Seminar 40 (Intermediate)
Cookeville, TN, (2)Mechanical Engineering,
Tennessee Tech University, Cookeville, TN,
Chair: Michael Connor, P.E., Member, Con- Market Transformation of
nor Engineering Solutions, Alpharetta, GA
(3)DuctSox Corporation, Peosta, IA DCV Systems for Kitchen
This is an interactive session where par-
4. Ventilation Requirements for Refriger- ticipants will breakup into small groups and
Ventilation
ating Machinery Rooms (1448-RP) (SA- discuss an ethics case ajudicated by the Track: Integrated Energy Systems
12-015) NSPE. Test your ethics IQ against an actual Room: 007A
Scot Waye, Ph.D., P.E., Member, Ronald case decided by a board of your peers and Sponsor: 05.10 Kitchen Ventilation
Petersen, Ph.D., Member and Anke Beyer- obtain ethics continuing education credits in
Chair: Derek W. Schrock, Member, Halton
Lout, CPP Inc., Fort Collins, CO the process. Three cases are presented
Co., Scottsville, KY
along with the final outcome.
Commercial kitchens have the largest
Seminar 37 (Intermediate) 1. Don’t Burn the Fire Code! energy intensity (usage per square foot) and
Michael Bilderbeck, P.E., Member, Pickering, can exceed any other building segment by an
Commissioning of Under order of magnitude. Of the largest compo-
Inc., Memphis, TN
Floor Air Distribution nents of the energy use in a kitchen is the
Systems 2. Trespassing On Intellectual Property
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
HVAC component which consists of the large
Warren Hahn, P.E., Member, Hahn Engineer- exhaust hoods along with the energy
Track: HVAC&R Systems & Equipment ing, Tampa, FL required to treat the replacement air. By
Room: 103B
applying demand control ventilation (DCV)
Sponsor: TRG7 Underfloor Air Distribution, 3. Rush to Judgment
systems on the cooking equipment lineup
05.03 Room Air Distribution Michael Connor, P.E., Member, Connor Engi-
with the exhaust hoods, large energy savings
Chair: Michael McQueeny, P.E., Member, neering Solutions, Alpharetta, GA
can be obtained. This seminar presents an
AirFixture, Kansas City, KS overview of DCV systems, a case study of
Presenters cover the differences in com- Seminar 39 (Intermediate) energy savings that can be obtained with
missioning (Cx), troubleshooting, testing and DCV systems, a description of a field proto-
balancing underfloor air distribution (UFAD) Indoor Moisture and Health: col for evaluating the performance of DCV
systems as compared to traditional ducted The Ins and Outs of the Wet systems, and field challenges that these sys-
systems. Presentation content includes the and Dry, Part 1 tems encounter.
idiosyncrasies associated with UFAD that
engineers, contractors and operators need to Track: Indoor Environmental 1. DCV Technologies for Commercial
be aware of so that obstacles to a success- Applications Kitchens: Past, Present and Future!
fully operating system can be avoided. The Room: 001A Don Fisher, Food Service Technology Cen-
audience will learn how to obtain the IAQ, Sponsor: Environmental Health Committee ter, San Ramon, CA
xxx
ASHRAE Transactions
2. A Field Test Method to Determine the Seminar 42 (Intermediate) 1. Experimental Validation of the Multiple
Performance of DCV Systems Zones Procedure of ANSI/ASHRAE Stan-
Richard T. Swierczyna, Associate Member, Sustainable Chiller Plant dard 62.1
Food Service Technology Center, San Efficiency: GPC 22 In Grenville K. Yuill, Ph.D., Fellow Life Member,
Ramon, CA Application Retired – University of Nebraska – Lincoln,
Placitas, NM
3. M & V for Kitchen Ventilation DCV Con- Track: HVAC&R Systems & Equipment
trols 2. CO2-Based Demand Controlled Venti-
Room: 001B lation for Multiple Zone HVAC Systems
Vernon A Smith, P.E., J.D., Associate Mem-
Sponsor: 08.02 Centrifugal Machines, Josephine Lau, Ph.D., Associate Member,
ber, Smith Energy Engineers, Niwot, CO
09.01 Large Building Air-Conditioning Sys- University of Nebraska – Lincoln, Omaha,
4. Field Challenges with DCV Systems tems NE
Andrey Livchak, Ph.D.1 and Derek W. Chair: Ray Good, P.E., Associate Member, 3. Standard 62.1: One Dynamic Reset
Schrock, Member2, (1)Halton Group Ameri- McQuay International, Staunton, VA Approach
cas, Bowling Green, KY, (2)Halton Co., Monitoring the performance of chilled- Dennis Stanke, Member, Trane, La Crosse,
Scottsville, KY water plants is of growing importance in WI
these times of rising energy costs and
Seminar 41 (Intermediate) increased environmental awareness. This
session explores the role that ASHRAE
New Research In Ground Guideline 22-2008 can play in guiding the
designer and end user to obtain better per-
Tuesday
Source Heat Pumps
formance data for the purpose of monitoring June 26, 2012
Track: HVAC&R Systems & Equipment the efficiency of chilled-water plants. First an
Room: 007D overview of Guideline 22 is given with recom-
9:45 a.m.–10:45 a.m.
Sponsor: 08.04 Air-to-Refrigerant Heat mendations for its proper interpretation and
application. Specific experiences and recom-
Transfer Equipment, Publishing & Education
mendations are shared from the perspec-
Conference Paper Session 9 (Basic)
Council
tives of both a consultant and end user who Energy Modeling Basics
Chair: Reinhard Radermacher, Ph.D., Fel- have applied Guidelines 22 in real world sit-
low ASHRAE, University of Maryland, Col- uations. Track: Building Modeling Applications
lege Park, MD
Room: 103A
This session features two recently pub- 1. Overview of Guideline 22 Chair: Yunho Hwang, Ph.D., Member, Uni-
lished papers on advances in heat transfer Mick Schwedler, P.E., Member, Trane, La versity of Maryland, College Park, MD
research, in ASHRAE’s HVAC&R Research Crosse, WI 1. Energy Modeling Basics: A Brave New
Journal. Ground coupled heat pumps are World for Young Building Modeling Pro-
gaining considerable attention. To shed addi- 2. Use and Specification of Guideline 22: fessionals (SA-12-C022)
tional light on the challenges, opportunities A Consulting Engineer’s Perspective Jared A. Higgins, P.E., Member, Parkhill,
and latest research for such systems, Stephen W. Duda, P.E., Member, Ross & Smith, & Cooper, Inc., Lubbock, TX
HVAC&R Research prepared a topical issue Baruzzini, Inc., St. Louis, MO 2. DOE Commercial Building Asset Rat-
on these systems under the guidance of ing: An Application of Centralized Model-
guest editors Jeff Spitler and Michel Bernier. 3. Experiences Utilizing Guideline 22: A ing Tools (SA-12-C023)
Facility Director’s Perspective Na Wang, Ph.D.1, Viraj Srivastava, Ph.D.,
1. Multipole Method to Calculate Bore- Member1, Willy Gorrissen1 and Cody
John I. Vucci, Member, University of Mary-
hole Resistances In Borehole Heat Taylor2, (1)Pacific Northwest National Labo-
land, College Park, MD
Exchangers ratory, Richland, WA, (2)Department of
Johan Claesson, Ph.D., Member1 and Göran Energy, Washington, DC
Hellström, Ph.D., Member2, (1)Chalmers Seminar 43 (Intermediate) 3. Calibration of a Building Energy Model
University of Technology, Gothenburg, Swe- Considering Parametric Uncertainty (SA-
den, (2)Lund University, Lund, Sweden Standard 62.1: Multiple-Zone 12-C024)
Systems, Demand Control Zheng O’Neill, Ph.D., P.E., Member1, Bryan
2. New Analytical Solution for Sizing Ver- and Energy Eisenhower, Ph.D., Member2, Vladimir
tical Borehole Ground Heat Exchangers Fonoberov, Ph.D.3 and Trevor E. Bailey,
In Environments with Significant Ground- Track: Indoor Environmental Ph.D.1, (1)United Technologies Research
water Flow: Parameter Estimation from Applications Center, East Hartford, CT, (2)UCSB-Center
Thermal Response Test Data for Energy Efficient Design, Santa Barbara,
Room: 103A CA, (3)Aimdyn, Inc., Santa Barbara, CA
Andrew Chiasson, Ph.D., P.E., Member, Uni- Sponsor: 04.03 Ventilation Requirements
versity of Dayton, Dayton, OH and Infiltration Seminar 44 (Intermediate)
3. A Semi-Analytical Model for Serpentine Chair: John J. Carter, Member, CPP, Inc.,
Horizontal Ground Heat Exchangers Fort Collins, CO Advances in Heat Transfer
Mikael Philippe, Ph.D., Member1, Michel Standard 62.1-2010 has added additional Research
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Bernier, Ph.D., Fellow ASHRAE2, Domi- requirements to the design of demand con-
Track: HVAC&R Systems & Equipment
nique Marchio, Ph.D.3 and Simon Lopez, trolled ventilation. This seminar provides an
Room: 007D
Ph.D., Member1, (1)Bureau des Recherches update on the validity of the multiple-zone
Sponsor: 08.04 Air-to-Refrigerant Heat
Géologiques et Minières, Département Géo- systems calculations, presents several differ- Transfer Equipment, Publishing & Education
thermie, Orléans, France, (2)Ecole Polytech- ent schemes for implementing demand con- Council
nique De Montreal, Monreal, QC, Canada, trol ventilation in these systems, and Chair: Reinhard Radermacher, Ph.D., Fel-
(3)Mines ParisTech, Centre Energétique et provides analysis of the potential for reducing low ASHRAE, University of Maryland, Col-
Procédés, Paris, France energy consumption. lege Park, MD
© 2012 ASHRAE
xxxi
This session features three recently pub- Seminar 47 (Intermediate) 1. VFD Efficiency: Part-Load Variability
lished papers on advances in heat transfer and Effects On Air-Handling System
research in ASHRAE’s HVAC&R Research Indoor Moisture and Health: Energy Consumption
Journal. Andrea Krukowski, Brown University/ Law-
The Ins and Outs of the Wet rence Berkeley National Laboratory, New
1. Model-Based Robust Temperature and Dry, Part 2 York, NY
Control for VAV Air-Conditioning System
Gongsheng Huang, Ph.D., Member, Hong Track: Indoor Environmental 2. AHRI Standard 1210-2011, Perfor-
Kong Polytechnic University, Hong Kong, mance Rating of Variable Frequency
Applications
Hong Kong Drives
Room: 001A Rupal Choksi, Associate Member, AHRI,
2. Experimental Measurement And Sponsor: Environmental Health Committee Arlington, VA
Numerical Simulation to Determine Char- Chair: Andrew Persily, Ph.D., Fellow
acteristics of an Infrared Tube System ASHRAE, NIST, Gaithersburg, MD
Samer Hassan, Ph.D., Member, Schwank Among the many indoor air quality issues
Forum 5 (Intermediate)
Ltd.., Ontario, ON, Canada faced by building designers, engineers and
occupants, indoor moisture is one of the
Safety Requirements for
most important. While not a contaminant Class 2L Refrigerants
Seminar 45 (Intermediate) itself, excessive moisture levels increase the
likelihood for microbial growth and the result- Track: Refrigeration Applications
Case Studies in Engineering ing health affects of allergy, asthma and other Room: 007A
Ethics II respiratory impacts. In addition, relative Sponsor: 03.01 Refrigerants and Second-
humidity is a key factor in determining occu- ary Coolants
Track: HVAC&R Fundamentals and pant comfort, and there are significant ques- Chair: Christopher Seeton, Ph.D., Member,
Applications tions as to the impacts of dry air on both Honeywell, Buffalo, NY
health and comfort. This two part seminar will In 2010, ASHRAE Standard 34 added an
Room: 007C optional subclass 2L to the existing Class 2
present the latest research findings on health
Sponsor: 01.07 Business, Management & effects associated with indoor moisture and flammability classification of refrigerants.
General Legal Education provide practical perspective on how to bet- Subsequently Standard 34-2010 “addendum
Chair: Michael Connor, P.E., Member, Con- ter manage indoor moisture through design, h” reclassified several refrigerants from
nor Engineering Solutions, Alpharetta, GA operation and maintenance. Class 2 to Class 2L, including R-32, R-717,
This is an interactive session where par- R-143a, & R-1234yf, and “addendum i” des-
ticipants will breakup into small groups and 1. How Building Operators Avoid Soggy ignated R-1234ze(E) as Class 2L. This forum
discuss an ethics case adjudicated by the Facilities discusses the approaches to development of
NSPE. Test your ethics IQ against an actual Lawrence Schoen, P.E., Fellow ASHRAE, safety requirements for Class 2L refrigerants,
case decided by a board of your peers and Schoen Engineering Inc., Columbia, MD to define the safe use and application of
obtain ethics continuing education credits in 2. Humidification Systems As Potential HVAC&R equipment using these refriger-
the process. Three cases are presented Sources of Indoor Contaminants ants. Standard 15-2010 currently defines
along with the final outcome. safety requirements for Class 1, 2, and 3
Wayne Thomann, Ph.D., Associate Member,
Duke University/Medical Center, Durham, refrigerants, pertaining to restrictions on
1. In Plain Sight refrigerant use, installation restrictions,
Michael Bilderbeck, P.E., Member, Pickering, NC
design and construction of equipment and
Inc., Memphis, TN 3. ASHRAE’s Newly Revised Position systems, and operation & testing. This forum
2. Thanks for the Help Document On Mold: What We Know and will provide an opportunity for interested par-
Don’t Know about Mold In Buildings ties to discuss the appropriate use of Class
Warren Hahn, P.E., Member, Hahn Engineer-
Lew Harriman III, Fellow ASHRAE, Mason 2L refrigerants.
ing, Tampa, FL
Grant, Portsmouth, NH
4. Moisture In Residences: From Wet Forum 6 (Intermediate)
Basements to Leaky Roofs
William B. Rose, Member, University of Illi- Standards As They Relate to
Energy Conservation and nois, Champaign, IL Cooling Tower Operations
Fan Energy Efficiency Grade
and Design: Too Much, Too
Track: HVAC&R Systems & Equipment Seminar 48 (Intermediate) Little, Or Just Right?
Room: 103B
Sponsor: 05.01 Fans, 05.09 Enclosed
Variable Frequency Drive Track: HVAC&R Systems & Equipment
Vehicular Facilities Performance: What Is It and Room: 007B
Chair: Asesh Raychaudhuri, P.E., Member, How Does It Relate to AHRI Sponsor: 08.06 Cooling Towers and Evapo-
US Dept. of Veterans Affairs, Washington, rative Condensers
DC
1210-2011? Chair: Frank Morrison, Member, Baltimore
Aircoil Company, Baltimore, MD
Minimum fan efficiency grade is under Track: HVAC&R Systems & Equipment ASHRAE Standards (such as Standards
consideration in several national standard. Room: 001B
This seminar discusses the definition and 90.1, 189.1, and proposed Standard 191) as
Sponsor: 01.11 Electric Motors and Motor well as State and local Standards (such as
meaning of fan energy efficiency grade. The
Control California’s Title 24) have been used as an
impact of the efficiency grade on the fan
Chair: Derrick Vigil, Member, Baldor Electric effective means to achieve greater energy
selection and technologies also is discussed.
Company, Greenville, SC and water efficiencies. Proposed revisions to
1. Fan Efficiency Grade Classification for Variable frequency drive (VFD) usage is some of these Standards seek to push effi-
Fans on the rise in HVAC&R systems. Can VFD ciency requirements for many HVAC&R sys-
Michael Brendel, Ph.D., Member, Lau Indus- performance have a significant effect on systems and equipment further still. This forum
tries/Ruskin Company, Dayton, OH tem performance? How would you know how explores whether current Standards and pro-
your VFD performs? This seminar explores posed revisions do not do enough to maxi-
2. Impact of Minimum Fan Efficiency how VFD performance can effect system mize energy/water efficiencies; or do they go
Grade performance and provide an overview of so far as to render some equipment non-
John Cermak, Ph.D., Member, ACME Engi- AHRI Standard 1210-2011, Performance competitive in the marketplace; or are they,
neering & Manufacturing Corp., Tulsa, OK Rating of Variable Frequency Drives. as Goldilocks would say, “just right”?
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
xxxii
ASHRAE Transactions
3. Investigating the Performance of Two Compressors, 10.07 Commercial Food and
Types of Solar Domestic Water Heating Beverage Cooling Display and Storage
Tuesday (SDWH) Systems with Drain Water Heat
Recovery Through Computer Simulation
Chair: Georgi S. Kazachki, Ph.D. , Fellow
ASHRAE , DRS Technologies, Florence , KY
June 26, 2012 and Experimental Analysis (SA-12-C027) The properties of the Low-GWP refriger-
Kamyar Tanha, Student Member, Alan Fung
11:00 a.m.–12:30 p.m. and Wey Leong, Ph.D., Ryerson University,
ants are in most cases different from the
properties of refrigerants to be replaced.
Toronto, ON, Canada
These differences have an impact on the
Technical Paper Session 5 4. Decision Making for HVAC&R System design, performance and efficiency of the
Selection for a Typical Office Building in systems and their components. The purpose
(Advanced) the UK (SA-12-C028) of the session is to explain and illustrate the
Mehdi Shahrestani, Student Member, Run-
Innovative Strategies to ming Yao, Ph.D., Member and Geoffrey
evolution of the design process around the
specific properties of the Low-GWP refriger-
Improve Equipment Cook, Ph.D., School of Construction Man-
ants that are already in an advanced stage of
agement and Engineering, University of
Efficiency Reading, Reading, United Kingdom development.
Track: HVAC&R Systems & Equipment 1. Advances In Low-GWP Refrigerants for
Room: 007C Seminar 49 (Intermediate) Stationary Systems
Chair: Jon J. Cohen, H-O-H Water Technol- Brett L. Van Horn, Ph.D., Member, Arkema,
ogy, Inc, Palatine, IL Net Zero Energy with Large King of Prussia, PA
1. Secondary Control Strategies for Radiant Surface Heat
2. Analysis of Low-GWP Refrigerants In
Cycling Air Conditioning and Refrigera- Exchangers Existing Compressor Technologies for
tion Systems (SA-12-016)
Bryan Rasmussen, Ph.D., Member1 and Track: HVAC&R Systems & Equipment Refrigeration Systems
Swarooph Seshadri2, (1)Texas A&M Univer- Room: 103A Samuel Yana Motta, P.Eng., Member, Hon-
sity, College Station, TX, (2)The Mathworks Sponsor: 06.05 Radiant Heating and Cool- eywell - Buffalo Research Laboratory, Buf-
Inc., MA ing falo, NY
2. New Correlation Equations for Ammo- Chair: Devin A. Abellon, P.E., Member,
nia Water Vapor-Liquid Equilibrium (VLE) Uponor, Phoenix, AZ 3. Transient Performance Evaluation of
Thermodynamic Properties (SA-12-017) Radiant heating and cooling systems, uti- Automotive Secondary Loop Systems
Syed Said, M. A. El-Shaarawi, Ph.D. and lizing “low temperature heating and high tem- with Low-GWP Fluids
Muhammad Umar Siddiqui, King Fahd Uni- perature cooling” allow system components, Yunho Hwang, Ph.D., Member, University of
versity of Petroleum and Minerals, Dhahran, such as chillers, boilers, heat pumps and Maryland, College Park, MD
Saudi Arabia renewable sources, to operate within their
maximum engineered performance range. 4. Prospective of Low-GWP Refrigerants
3. Development of a High Efficiency This seminar covers design and control strat-
Home Heating and Humidification Tech- J. Steven Brown, Ph.D., Member, The Cath-
egies for embedded-tube radiant-based olic University of America, Washington, DC
nology (SA-12-018) HVAC systems for maximum energy effi-
Dexin Wang, Ph.D., Ainan Bao, Ph.D. and ciency and enhanced thermal human com-
William Liss, Gas Technology Institute, Des fort. A case study of an ambitious project in Seminar 51 (Basic)
Plaines, IL San Francisco, California will illustrate how
4. Selection of Desiccant Equipment at radiant based systems can used as part of Existing Building
Altitude (RP-1339) (SA-12-019) an energy efficient design solution on a pro- Commissioning Process:
Nelson Fumo, Ph.D., Member1 and Pedro jected LEED Platinum, Net-Zero project. Best Practices
Mago, Ph.D., Member2, (1)The University of 1. Designing for Net-Zero Energy Usage Track: HVAC&R Fundamentals and
Texas at Tyler, Tyler, TX, (2)Mississippi State
University, Mississippi State, MS
Using Radiant-Based Heating and Cool- Applications
ing Systems Room: 103B
Peter Simmonds, Ph.D., Fellow Member, IBE Sponsor: 07.09 Building Commissioning
Conference Paper Session 10 Consulting Engineers, Sherman Oaks, CA
Chair: Mark “Dusty” Wheeler Jr., Member,
(Basic) 2. Simplified Control and Piping Options Honeywell, Washington, DC
for Low Temperature Heating with High This session outlines the activities
Optimizing Energy Modeling Temperature Cooling
involved in performing the Existing Building
Strategies to Improve Energy Robert Bean, Healthy Heating, Calgary, AB,
Commissioning Process (EBCxP) by pre-
Usage Forecasting Canada
senting best practices according to ASHRAE
3. Case Study: Exploratorium Museum Guidelines as well as case studies.
Track: Building Modeling Applications San Francisco
Room: 001B Peter H. Rumsey, Member, Rumsey Engi- 1. Existing Building Commissioning Pro-
Chair: Michelle Contri, P.E., Member, DLB neers, Inc, Oakland, CA cess (EBCxP): Best Practices
Associates, Eatontown, NJ
Bill Dean, National Research Council of Can-
1. Use of Flexible Research Platforms Seminar 50 (Advanced) ada, Saskatoon, SK, Canada
(FRP) for BIM and Energy Modeling
2. Existing Building Commissioning
Research (SA-12-C025) Design, Performance and
Piljae Im and Mahabir Bhandari, Member, (EBCx): Case Studies
Oak Ridge National Laboratory, Oak Ridge, Efficiency Impact of Low- Holly Townes, P.E., Member, Puget Sound
TN GWP Refrigerants on Energy, Bellevue, WA
2. Using Measured Utility Data to Verify Systems and Components 3. On-Going Commissioning for Existing
Energy Models (SA-12-C026)
Jared A. Higgins, P.E., Member, Stephen D. Track: HVAC&R Systems & Equipment Buildings: Case Studies
Foster and J. Ryan Bailey, Associate Mem- Room: 007D David E. Claridge, Ph.D., P.E., Fellow
ber, Parkhill, Smith, & Cooper, Inc., Lubbock, Sponsor: 03.01 Refrigerants and Second- ASHRAE, Texas A & M University, College
TX ary Coolants, TC 8.01 Positive Displacement Station, TX
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
© 2012 ASHRAE
xxxiii
Seminar 52 (Basic) as well as existing buildings that are reno-
Federal, State and Local

vated and desire to be considered green.
Site energy, energy cost, primary energy, Tuesday
Grassroots Engagement: and greenhouse gas emissions calculation
June 26, 2012
options will be discussed. Pathways to imple-
How Helping ASHRAE Helps ment these methods will also be identified,
You including mandatory, prescriptive, and per-
5:30 p.m.–6:30 p.m.
Track: HVAC&R Fundamentals and formance pathways, as well as an outcome-
Applications based pathway. Major distinctions among Forum (Intermediate)
Room: 007A approaches and pathways will be illustrated,
Sponsor: Advocacy Committee as well as the benefits and challenges of the Comparing HVAC Efficiency:
different options.
Chair: Mark Wills, ASHRAE, Washington, Why Ground Source Heat
DC 1. ASHRAE Standard 189.1 and bEQ Pump Systems Are Superior
Federal, state, and local governments Baseline Energy Performance Calcula-
propose thousands of ordinances, regulations Track: HVAC&R Systems & Equipment
tions, and bills each year on building codes Neil P. Leslie, P.E., Member, Gas Technology Room: 001B
and standards; engineering licensure; sus- Institute, Des Plaines, IL Sponsor: 06.08 Geothermal Heat Pumps
tainable buildings; and many other issues and Energy Recovery Applications
impacting ASHRAE members’ work. Conse- 2. 2012 International Green Construction Chair: Keith Swilley, Member, Gulf Power
quently, member involvement in the policy Code Energy Performance Methodology Company, Pensacola, FL
process is crucial to achieving the Society’s R. Christopher Mathis, Member, MC2 Mathis OPEN SESSION: no badge required; no
overarching mission and goals. This session Consulting Company, Asheville, NC PDHs awarded; presented during the TC’s
addresses the current status of key policy ini- meeting. Make no mistake, when you need
3. Outcome-Based Energy Compliance space conditioning the most, part load or full
tiatives, and provides members guidance for
Pathway for Green Building Designs load, ground source heat pumps provide the
effectively combining technical expertise with
David Conover, Member, Pacific Northwest overall best efficiency. Metered and manu-
policymaking opportunities to become a facturer data will be shared comparing differ-
National Laboratory, Washington, DC
more effective advocate for themselves, the ent types of mechanical equipment to ground
Society, chapters, and the profession, thus source heat pump equipment and their sys-
enabling them to become a more “complete” Seminar 54 (Intermediate) tem efficiencies.
ASHRAE member.
How to Efficiently Serve Low
1. State and Local Energy Codes: Devel- Dewpoint Applications
opment, Adoption, Implementation and
Compliance Track: HVAC&R Fundamentals and Wednesday
Maureen Guttman, Building Codes Assis- Applications
tance Project, Washington, DC Room: 007B June 27, 2012
2. Challenges to Engineer Licensure
Sponsor: 08.12 Desiccant Dehumidification 8:00 a.m.–9:30 a.m.
Equipment and Components
Richard B. Hayter, Ph.D., P.E., Kansas State
Chair: Andrew Lowenstein, Ph.D., Member,
University, Manhattan, KS
AIL Research, Inc., Princeton, NJ Technical Paper Session 6
3. Key ASHRAE Issues and Priorities In The supply of low dewpoint air often is (Advanced)
Congress and Federal Agencies essential to the successful functioning of a
Mark R. Ames, ASHRAE, Washington, DC building or industrial process. Supplying air Improvements in Building
at dewpoints below 40F can be particularly Thermal Performance
Seminar 53 (Intermediate) challenging. This seminar will present case Track: HVAC&R Fundamentals and
studies where desiccant systems success- Applications
Green Building Energy fully served applications that demanded air Room: 007B
Performance Pathways and with dewpoints below 40F and as low as -
Chair: Charles E. Henck, Whitman,
55F. The potential for energy savings will be
Evaluation Approaches highlighted.
Requardt & Associates LLP,
Baltimore, MD
Track: Integrated Energy Systems 1. Desiccants for Low Dewpoint In Hospi-
Room: 001A tal and Archival Applications 1. Thermal Performance and Charge Con-
Sponsor: 02.08 Building Environmental trol Strategy of a Ventilated Concrete Slab
Impacts and Sustainability, 07.06 Building Scott McGinnis, Member, Munters Corpora-
(VCS) with Active Cooling Using Outdoor
Energy Performance tion, Selma, TX
Air (SA-12-020)
Chair: David Ellis, P.E., Member, HDR Archi-
2. Minimize Cooling System Refrigerant Yuxiang Chen, Student Member1, Andreas
tecture Inc., Alexandria, VA
Several government and private sector Charge Using Liquid Desiccant Dehumid- Athienitis, Ph.D., P.E., Member2 and Khaled
organizations, including ASHRAE, publish ification/Cooling Equipment Galal, Ph.D., P.E.3, (1)Concordia University,
model codes, standards, and guidelines for Michael Harvey, Kathabar Dehumidification Montreal, ON, Canada, (2)Concordia Univer-
use in the U.S. as well as other countries. Systems Div. of Niagara Blower Company, sity, West Montreal, QC, Canada, (3)Concor-
These documents address the design, con- Buffalo, NY dia University, Montreal, QC, Canada
struction, commissioning, and operation of 3. Keeping Dewpoints Low In Supermar-
green buildings and are intended for volun- 2. Thermal Performance of Building Enve-
kets - Costs and Savings
tary as well as mandatory adoption. This lope Details for Mid- and High-Rise Build-
Jeffrey Halley, J & J Mechanical, Brea, CA
seminar will describe different approaches ings (1365-RP) (SA-12-021)
used in model codes and standards to deter- 4. Application of Desiccant Dehumidifica- Patrick Roppel, P.Eng., Mark Lawton, P.Eng.,
mine and compare the energy and environ- tion to a Foundry Member and Neil Norris, Morrison Hersh-
mental performance of new green buildings Chris Bogart, BRY Air, Sunbury, OH field, Vancouver, BC, Canada
xxxiv --`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---

ASHRAE Transactions
3. Parametric Investigation of PCM Ther- Fan, Ph.D.1, Joseph Mantese, Ph.D.1, Mer- contribute to the energy consumption. How-
mal Properties on Energy Demand and edith Colket, Ph.D.1, Heidi Hollick1, Ricardo ever, air filters play a critical role in keeping
Temperature of Buildings in Toronto (SA- Garvey1, Kenneth Swanson2, Mike the HVAC system clean and contribute to
12-022) McCarron2, Alan Bukofske2, Duffy Parlett2 good IAQ within the occupied spaces. This
M. Ebrahim Poulad, Student Member and and Mauro Atalla, Ph.D.1, (1)United Technol- seminar combines the lessons from recent
Alan Fung, Ryerson University, Toronto, ON, ogies Research Center, East Hartford, CT,
ASHRAE-funded research with other emerg-
Canada (2)Fireye Inc., Derry, NH
ing research to help building owners and
4. Hybrid Solar Thermal and Ground operators understand the role that air filter
Technical Paper Session 7 Source Heat Pump System (SA-12-C032) pressure drop plays in building energy con-
(Advanced) Farzin Rad, P.Eng., Member, Alan S. Fung, sumption.
Dr.Ing., P.Eng., Member and Wey H.Leong,
Modeling Indoor Thermal Dr.Ing., P.Eng., Member, Ryerson University, 1. Energy Implications of Residential
Toronto, ON, Canada HVAC Filters
Performance
Jeffrey Siegel, Ph.D., Member, University of
Track: Building Modeling Applications Seminar 55 (Intermediate) Texas, Austin, TX
Room: 007C
Chair: Michael Meteyer, P.E., Member, Cog- bEQ In Operation Rating: An 2. Filter Pressure Drop Matters: Results
dell Spencer ERDMAN, Madison, WI Overview of Field Reductions In Static Pressure On
Air Conditioner Performance
1. Field Measurements of Thermal Condi- Track: Integrated Energy Systems
tions During Surgical Procedures for the John Proctor, P.E., Proctor Engineering, San
Development of CFD Boundary Condi- Room: 001A Rafael, CA
tions (SA-12-023) Sponsor: bEQ Committee
John Zhai, Ph.D., Member1, James S. Chair: Amy Musser, Ph.D., P.E., Member, 3. Selecting Air Filters with Lowest Life
McNeill, Student Member1 and Jean Hertz- Vandemusser Design, PLLC, Asheville, NC Cycle Cost Using Field Measurements of
berg, Ph.D.1, (1)University of Colorado, This seminar provides an overview of the Air Filter Performance and Fan Energy
Boulder, CO bEQ In Operation building rating overview, Measurements
describing what is involved in the rating and
2. Modeling Phase Change Materials with the how the rating is conducted. Also, it Donald D. Thornburg, Member, Camfil-Farr,
a Building Simulation Code Developed in addresses the BEAP certification and energy Riverdale, NJ
MATLAB (SA-12-024) auditing, which will cover the basics of get-
Dahai Zhang, Student Member1, Alan ting certified, explaining why and how an Seminar 57 (Basic)
Fung1, Fabio Almeida, Ph.D., Student auditor would become a certified BEAP.
Member1 and Sridhar Sadasivam, Ph.D.2, Ways to use certification to enhance your
(1)Ryerson University, Toronto, ON, Canada, consulting business and market your com- Methods and Tools to
(2)Purdue University, West Lafayette, IN pany’s services will be discussed. Finally, a Support the Modeling
review of an applied Building Energy Assess- Process
3. Occupancy Simulation in Three Resi- ment answers questions, such as what is
dential Research Houses (SA-12-025) required for building energy assessment and
Philip R. Boudreaux, Anthony C. Gehl and Track: Building Modeling Applications
how is it applied in the field and why does
Jeffrey E. Christian, Associate Member, Oak assessment go beyond energy and look at Room: 103B
Ridge National Laboratory, Oak Ridge, TN other aspects of the building? Sponsor: 04.07 Energy Calculations
Chair: Amir Roth, United States Department
Conference Paper Session 11 1. bEQ In Operation Rating of Energy, Washington, D.C.
Thomas H. Phoenix, SPC, Greensboro, NC
(Intermediate) Today’s burgeoning market for building
2. BEAP Certification and Energy Audit- energy modeling services presents great
Optimizing HVAC System ing opportunities and challenges for practitio-
John Dunlap, Dunlap and Partners, Rich- ners. Challenges exist due to the expansive
Performance mond, VA knowledge required to deliver services, the
lack of standardized methods and the
Track: HVAC&R Systems & Equipment 3. Applied Building Energy Assessment absence of building energy modeling (BEM)
Room: 007A Hoy R. Bohanon, P.E., Member, Working tools that support the modeling process. This
Chair: Henry A. Becker, Member, H-O-H Buildings, Winston-Salem, NC session reviews key BEM tasks and identi-
Water Technology, Inc, Palatine, IL fies those that would benefit from being auto-
1. Laboratory Evaluation of Aftermarket Seminar 56 (Intermediate) mated in software tools. The session
Boiler Control Systems (SA-12-C029) provides examples of practical user-devel-
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Patricia F. Rowley1 and Paul Glanville, P.E., How Air Filter Selection oped BEM tools that support delivering and
streamlining services. It introduces efforts to
Associate Member1, (1)Gas Technology Affects Energy Consumption develop modular, standardized whole-build-
Institute, Des Plaines, IL
Track: Indoor Environmental ing simulation components, automated work-
2. Using Steady State Predictions to flow procedures and an open source energy
Improve the Transient Response of a Applications modeling platform.
Water to Air Heat Exchanger (SA-12- Room: 007D
C030) Sponsor: 02.04 Particulate Air Contami- 1. Supporting the Modeling Process
David A. Hodgson, Ph.D., Member1, Peter nants and Particulate Contaminant Removal Ellen Franconi, Rocky Mountain Institute,
M. Young, Ph.D.2, Charles W. Anderson, Equipment Boulder, CO
Ph.D.2, Douglas C. Hittle, Ph.D.2, William S. Chair: Philip Winters, P.E., Member, Filtra-
Duff, Ph.D.2 and Daniel B. Olsen, Ph.D.2, tion Group, Joliet, IL 2. Supporting BEM: The Practitioner Per-
(1)Union College, Schenectady, NY, (2)Colo- Approximately half of all energy con- spective
rado State University, Fort Collins, CO sumed in commercial and residential build- Kate Turpin, P.E., Member, SERA Architects,
ings is used for the HVAC systems and a Portland, OR
3. A High-Efficiency, Reduced-Emissions
Combustion Control System for Commer- significant contributor to that energy con- 3. Supporting BEM: The “Technicians”
cial and Industrial Boilers (SA-12-C031) sumption is the fan power required to move Perspective
Guido Poncia, Ph.D.1, Christoph air throughout the building. Air filters create Jesse Dean, National Renewable Energy
Haugstetter1, David Liscinsky1, Junqiang resistance to that air flow and as a result, Laboratory, Golden, CO
© 2012 ASHRAE
xxxv
Seminar 58 (Intermediate) energy-efficient buildings which utilize natu- Chair: Mike McDermott, Member, Grumman
ral ventilation. Butkus Associates, Evanston, IL
Mold, Moisture, and Damp 1. How to Define Flow Boundary Condi- 1. Design of Roofs for Increased Solar
Buildings: New tions for Natural Ventilation Through Potential of BIPV/T Systems and their
Understanding and Large Openings by CFD Applications to Housing Units (SA-12-
Qingyan Chen, Ph.D., Purdue University, 028)
Guidelines for Reducing the West Lafayette, IN Caroline Hachem, Ph.D., Student Member,
Health Effects 2. Wind Driven Natural Ventilation When Andreas Athienitis, Ph.D., P.E., Member and
There Are No Windward Openings Paul Fazio, Ph.D., P.Eng., Member, Concor-
Track: Indoor Environmental dia University, West Montreal, QC, Canada
David Banks, Ph.D., Member, CPP Wind
Applications
Engineering and Air Quality Consultants, Ft. 2. Achieving Net-Zero Energy Conditions
Room: 103A Collins, CO in Retrofit: A Case Study for a Perfor-
Sponsor: 01.12 Moisture Management in
3. Impact of Indoor Buoyancy On Outdoor mance Arts Building Using an Hourly
Buildings, TC 8.10,TC 8.12, and SSPC 62.1
Wind Driven Natural Ventilation Simulation Approach (SA-12-029)
Chair: Steve Cornick, Member, National
Research Council Canada, Ottawa, ON, Atila Novoselac, Ph.D., Associate, University Dr. Cenk Cy Yavuzturk, Ph.D., SBA
Canada of Texas at Austin, Austin, TX Member1, J.E. Fuller2 and Dennis J. O’Con-
For years, damp buildings were thought to nor, Student Member3, (1)Mechanical Engi-
4. Modeling the Effects of Outdoor Gaso- neering Dept, University of Hartford, West
generate negative health effects. Until line Powered Generator Use On Indoor
recently, the magnitude of health effects has Hartford, CT, (2)University of Hartford, West
Carbon Monoxide Exposures Hartford, CT, (3)University of Hartford, Hart-
not been quantified. The mechanisms which Liangzhu (Leon) Wang1 and Steven Emm-
lead to negative effects were not understood. ford, CT
erich, Member2, (1)Concordia University,
Research advances have provided a clearer Montreal, QC, Canada, (2)National Institute
path towards understanding what levels of of Standards and Technology, Gaithersburg, Technical Paper Session 10
dampness are problematic, what the specific MD
links are between asthma and dampness, (Intermediate)
and what strategies are effective in reducing
moisture accumulation. The presentations Commission and Calibration
describe results of research as well as new
guidance for architects, mechanical system
designers and owners. The information is
Wednesday of HVAC Systems to Improve
Efficiency
also of interest to occupants of buildings June 27, 2012 Track: HVAC&R Fundamentals and
which are perceived to be damp.
1. The Revised ASHRAE Position Docu-
9:45 a.m.–10:45 a.m. Applications
ment On Indoor Mold and Moisture Man- Room: 001B
agement Chair: Charles E. Henck, Whitman,
Lew Harriman III, Fellow ASHRAE, Mason
Technical Paper Session 8 Requardt & Associates LLP, Baltimore, MD
Grant, Portsmouth, NH (Advanced) This session will present a case study of
2. Observations and Recommendations energy saving opportunities implemented in
Based On NIOSH Investigations of Damp
Study the Degradation of an existing building and then commissioning
Buildings and Their Health Effects Typical HVAC Materials, the systems to perform at their optimal levels.
A savings for over $500,000 was realized
Jean Cox-Ganser, Ph.D. and Kay Kreiss, Filters and Components with the improvements which is a 15% elec-
M.D., CDC-NIOSH, Morgantown, WV
Irradiated by UVC Energy trical reduction and a 7% gas reduction. The
3. New U.S. EPA Guidelines for Moisture second paper will present a methodology to
Management In Buildings Track: HVAC&R Systems & Equipment calibrate the multi-zone model CONTAM to
Terry Brennan, Camroden Associate, West- Room: 007A two buildings. An air flow sensitivity analysis
moreland, NY Chair: Richard L. Vincent, Member, Mount identifies influential systems and a tracer gas
Sinai School of Medicine, New York, NY analysis identifies macro zones which are
used to tune the model parameters.
Seminar 59 (Intermediate) 1. Study the Degradation of Typical HVAC
Materials, Filters, and Components Irradi- 1. HVAC Systems Commissioning in a
The Effect of the Outdoor ated by UVC Energy. Part I. Literature Manufacturing Plant (SA-12-030)
Environment on Indoor Search (1509-RP) (SA-12-026) Yunhua Li, Student Member1, Bei Zhang,
Robert E. Kauffman, Ph.D., University of Student Member1, Mingsheng Liu, Ph.D.,
Environment Modeling
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Dayton Research Institute, Dayton, OH P.E., Member2, Lixia Wu, Ph.D., Member3,
Track: Indoor Environmental Jinrong Wang, Member4 and Tom Lewis,
2. Study the Degradation of Typical HVAC
Member4, (1)University of Nebraska-Lin-
Applications Materials, Filters, and Components Irradi-
coln, Lincoln, NE, (2)DTL Controls, LLC,
Room: 001B ated by UVC Energy. Part II. Polymers
Omaha, NE, (3)Bes-Tech, Inc., Philadelphia,
Sponsor: 04.10 Indoor Environmental Mod- (1509-RP) (SA-12-027)
PA, (4)Omaha Public Power District, Omaha,
eling Robert E. Kauffman, Ph.D. and J. Douglas NE
Chair: James VanGilder, P.E., Member, APC Wolf, University of Dayton Research Insti-
by Schneider Electric, Billerica, MA tute, Dayton, OH 2. Investigation of CO2 Tracer Gas-Based
While it is common practice to represent Calibration of Multi-Zone Airflow Models
the effects of the outside world simplistically Technical Paper Session 9 (SA-12-031)
when performing an indoor environment (Advanced) T. Agami Reddy, Ph.D., P.E., Fellow
analysis, this approach may not always yield ASHRAE1, Steven Synder, Associate
realistic predictions. This seminar considers Member2 and William P. Bahnfleth, Ph.D.,
the coupled nature of the indoor and outdoor
Design Applications for High P.E., Fellow ASHRAE3, (1)The Design
environments with examples including indoor Performing Buildings School/ The School of Sustainability, Tempe,
CO exposure rates from outside generators. AZ, (2)Johnson Controls Inc., Philadelphia,
Concepts discussed here are directly appli- Track: Integrated Energy Systems PA, (3)Pennsylvania State University, Univer-
cable to the design of high-performance, Room: 001A sity Park, PA
xxxvi
ASHRAE Transactions
Seminar 60 (Intermediate) nar for the last five years. The seminar Forum 8 (Intermediate)
assists engineers, practitioners, and opera-
Datacom Space Technology tors in the design and operation of this type of If Hydronic HVAC Systems
equipment.
Advances; From Are So Great, Why Doesn’t
Containment Fire Protection 1. Energy Efficiency and Heat Exchanger Everyone Have One?
Compaction with the Next Generation
to Oil Immersion Cooling Micro Channel Heat Exchangers Track: HVAC&R Systems & Equipment
Track: HVAC&R Fundamentals and Michael M. Ohadi, Ph.D., Fellow ASHRAE, Room: 007B
University of Maryland, College Park, MD Sponsor: 06.01 Hydronic and Steam Equip-
Applications
ment and Systems, 06.05 Radiant Heating
Room: 103A 2. Ammonia/Carbon Dioxide Cascade and Cooling
Sponsor: 09.09 Mission Critical Facilities, Condensers Using Plate Exchangers Chair: Paul A. Torcellini, Ph.D., Member,
Technology Spaces and Electronic Equip- Zahid Ayub, Ph.D., P.E., Fellow ASHRAE, National Renewable Energy Laboratory,
ment Isotherm, Inc., Arlington, TX Golden, CO
Chair: Michael K. Patterson, Ph.D., P.E., Radiant heating/cooling and hydronic
Member, Intel Corp., Dupont, WA systems have been identified as underuti-
IT equipment and facilities are in a con- Seminar 62 (Intermediate) lized technologies and a low-energy HVAC
stant state of technology advancement; from system that can reach a 50% net site energy
airflow management to alternative cooling, Smart Grid in Texas, What’s savings goal. So, if radiant heating and cool-
such as oil immersion cooling. This seminar Happening? ing systems are so great, why are they not
covers specific topics supporting these “front of mind” for new construction and ret-
advances. Airflow containment is now com- Track: Integrated Energy Systems rofits? Is it the costs? Is it a lack of under-
monly used in the datacom space to improve standing of designs? Is it lack of knowledge
airflow management, helping drive energy Room: 007D in proper installation and commissioning?
efficiency improvements. One of the main Sponsor: 07.05 Smart Building Systems Come prepared to discuss the barriers and
challenges in containment implementation is Chair: Joshua D. Rhodes, Student Member, your methods to overcome them as we strive
fire protection design. This is covered in the The University of Texas at Austin, Austin, TX for low energy buildings to help create
first presentation. In the second, a side-by- research and implementation agendas for
The current energy grid is antiquated and
side evaluation of air and oil immersion cool- this technology.
is losing the ability to meet the demand
ing is presented. Energy and thermal perfor-
required of it. To meet rapidly growing
mance is shared from identical racks of IT
demand, the current grid is being modern-
equipment across a range of workloads.
ized not only with the capacity to move more
1. Datacom Aisle Containment Systems
and Associated Fire Protection Require-
resources, but to do so in a smarter way. This
Grid 2.0, or Smart Grid is evolving to not only
Wednesday
ments move more energy and water, but to move June 27, 2012
David Quirk, P.E., Member, Verizon Wireless, information as well. This track discusses the
Basking Ridge, NJ development and build-out of smart grid 11:00 a.m.–12:30 p.m.
infrastructure in Texas.
2. Oil Immersion Cooling and Reductions
In Data Center Energy Use 1. Making the “Smart Grid” a “Green Conference Paper Session 12
Michael K. Patterson, Ph.D., P.E., Member1 Grid”
and Christiaan Best2, (1)Intel Corp., Dupont, Colin Meehan, Environmental Defense (Intermediate)
WA, (2)Green Revolution Cooling, Austin, TX Fund, Austin, TX
Low Energy Design and
Seminar 61 (Intermediate) 2. The Pecan Street Smart Grid Demon- Integrated Energy Systems
stration Project
Chris Holcomb , Pecan Street, Austin, TX Track: Integrated Energy Systems
Next Generation of Heat Room: 001B
Exchangers and Energy Sponsor: 01.10 Cogeneration Systems,
Systems with Sustainability Forum 7 (Intermediate) 07.01 Integrated Building Design
Chair: Dharam V. Punwani, Avalon Consult-
in Mind Accurately and Correctly ing, Naperville, IL
Track: HVAC&R Systems & Equipment Modeling Air-to-Air Energy This conference paper session explores
low energy design with integrated energy
Room: 103B Recovery Technologies systems including collection of condensate
Sponsor: 01.03 Heat Transfer and Fluid water, combined heat/power systems, tri-
Flow, 08.05 Liquid-to-Refrigerant Heat Track: Building Modeling Applications generation systems and commissioning and
Exchangers Room: 007C energy management.
Chair: Amir Jokar, Ph.D., P.E., Member,
Exponent Inc. Thermal Sciences Practice, Sponsor: 05.05 Air-to-Air Energy Recovery 1. Zero Refrigeration in the USA (SA-12-
Los Angeles, CA Chair: Carol E. Marriott, P.Eng., Member, C033)
Design, optimization and maintenance of Carol Marriott Consulting, Maple Grove, MN Robert Tozer, Ph.D., Member, Operational
micro/mini-channel compact heat exchang- To help determine and isolate issues in cor- Intelligence Ltd, Kingston upon Thames,
ers are critical in energy efficiency of rectly modeling AAERV systems in the most United Kingdom
HVAC&R systems. This seminar discusses popular modeling programs used by archi- 2. The Challenge of Low Energy Design in
new technologies for heat exchangers that tects and engineers. How is the technology Low Energy Cost States (SA-12-C034)
aim to significantly improve energy efficien- properly quantified with respect to perfor- Stephen W. Duda, P.E., Member, Ross &
cies while preserving economic and environ- mance? Are controls sequences properly Baruzzini, Inc., St. Louis, MO
mental sustainability of the systems. These identified when estimating savings in build-
topics are critical to current and future ings? Should an algorithm be developed to 3. Targeting a Net-Zero Energy Student
ASHRAE research, as indicated in the support these initiatives? Is there/are there Center: Part 1 – Thermal and Electrical
ASHRAE 2010-2015 strategic plans. TC 1.3 metrics that can be used with existing pro- Load Minimization (SA-12-C035)
and TC 8.5 have sponsored and monitored grams? The results of this forum will fuel Trevor Caldwell, Kelton Friedrich, Brad Gre-
several ASHRAE funded research projects future Research and Program initiatives by gus, Ryan Verschuere, Jordan Anderson,
related to the topics presented in this semi- TC 5.5. Aaron Murenbeeld, Kaveh Arfaei, Mahsa
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
© 2012 ASHRAE
xxxvii
Boroum and James S. Cotton, Ph.D., P.E., Conference Paper Session 14 Seminar 64 (Intermediate)
Member, McMaster University, Hamilton,
(Basic)
ON, Canada How Smoke Control Systems
4. Targeting a Net-Zero Energy Student
Energy Efficiency Strategies Contribute to Sustainability
Center: Part 2 – Systems to Meet Building Track: HVAC&R Fundamentals and Goals
Loads (SA-12-C036) Applications Track: HVAC&R Fundamentals and
Kelton Friedrich, Trevor Caldwell, Brad Gre- Room: 007B Applications
gus, Ryan Verschuere, Jordan Anderson, Chair: Michael Deru, Ph.D., Member, Room: 007C
Aaron Murenbeeld, Kaveh Arfaei, Mahsa National Renewable Energy Laboratory,
Sponsor: 05.06 Control of Fire and Smoke
Boroum and James S. Cotton, Ph.D., P.E., Golden, CO
Chair: Paul Turnbull, Member, Siemens
Member, McMaster University, Hamilton, Building Technologies, Inc, Buffalo Grove, IL
1. Evaluation of Thermal Comfort in Low
ON, Canada Load Homes Supplied by High Sidewall Smoke control systems require coordi-
Air Jets (SA-12-C041) nated integration between the building’s fire
Conference Paper Session 13 El Hassan Ridouane, Ph.D., Member, alarm system and its mechanical equipment
National Renewable Energy Laboratory, controls to provide the intended life safety
(Intermediate) Golden, CO functions. Some building owners view smoke
control systems as an unavoidable cost. This
2. Experimental Study on Ceiling Radiant
Contamination and Cooling Panel System with Personal
seminar explains how smoke control sys-
tems can support a building owner’s sustain-
Environment Control Floor Diffuser (SA-12-C042) ability goals. Presentations describe how
Sei Ito1, Minoru Kawashima, Ph.D.,
Applications Member1, Yoshito Arai1, Michiya Suzuki,
different choices can lead to more sustain-
able smoke control system designs, and how
Ph.D.1, Koji Murakami1 and Tatsuo Nobe, operation of a smoke control system mini-
Track: Indoor Environmental Ph.D., Member2, (1)Shimizu Corporation, mizes the environmental impact during a fire
Applications Tokyo, Japan, (2)Kogakuin University, Tokyo, emergency. Case studies are presented
Japan showing how smoke control systems helped
Room: 007A
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
3. Determining the Cost-Effectiveness of achieve LEED innovation credits and LEED

Chair: Jeff J. Traylor, P.E., Member, EMCOR
Various Energy Sub-Metering Strategies energy usage credits.
Facilities Services, Durham, NC
(SA-12-C043) 1. How Smoke Control Contributes to
Increasingly, we are looking to advanced Christine Maurer, P.E., Member1, Peter Sustainability
technologies and automated control systems Rojeski, Ph.D., P.E., Life Member2 and Matt William A. Webb, P.E., Fellow ASHRAE,
and strategies to provide an acceptable Davey1, (1)Advanced Energy, Raleigh, NC, WEBB FIRE Protection Consulting, LLC,
indoor environment in a cost effective man- (2)M and V Services, Pleasant Garden, NC Brooksville, FL
ner. This session features several papers
that explore our indoor environment and how Seminar 63 (Intermediate) 2. An Overview of Sustainability and
we can maintain improved control over its Smoke Control
quality and acceptability for occupants. Designing Heat Recovery John H. Klote, Ph.D., P.E., Fellow ASHRAE,
Fire and Smoke Consulting, Leesburg, VA
1. Contaminant Removal Effectiveness of
Heat Pumps to Meet Today’s
3. Case Studies: Using Fire Engineering
Displacement Ventilation Systems During Requirements for High to Achieve Sustainable Designs
Heating Season; Summary Results from Performance Buildings Jeffrey Tubbs, P.E., Member, Arup, Cam-
Three Field Studies (SA-12-C037) bridge, MA
Track: Integrated Energy Systems
Boualem Ouazia, Ph.D., Member1, Iain Mac-
Room: 103B
donald, Ph.D.2 and Michel Tardif, P.Eng.,
Sponsor: 06.08 Geothermal Heat Pumps
Member3, (1)The Institute for Research in
and Energy Recovery Applications
Construction (IRC) / National Research Chair: Cary Smith, Associate Member, Impacts of Environmental
Council Canada (NRC), Ottawa, ON, Can- Energy Center of Wisconsin, Madison, WI Change on Building Design
ada, (2)National Research Council Canada, Heat pumps and chillers are increasingly
Montreal, QC, Canada, (3)CanmetENERGY
and Their HVAC Systems
an effective option for producing useful heat-
Natural Resources Canada, Ottawa, ON, ing in addition to their normal cooling produc- Track: HVAC&R Fundamentals and
Canada tion. Dedicated heat recovery, geothermal Applications
applications, and heat recovery chiller sys-
2. A Robust CO2-Based Demand-Con- tems are some different configurations that Room: 103A
trolled Ventilation Control Strategy for include this technology. But there are many Sponsor: 04.02 Climatic Information, 02.05
ways to configure and control a heat recov- Global Climate Change
Multi-Zone HVAC Systems (SA-12-C038)
ery chiller or heat pump; this seminar dis- Chair: Steve Cornick, Member, National
Nabil Nassif, Ph.D., P.E., Associate Member, Research Council Canada, Ottawa, ON,
cusses some of the more effective
North Carolina A&T State University, approaches. Canada
Greensboro, NC The environment is changing. The con-
1. From Hockey Rink to Ground Loop sensus is that the climate is getting warmer.
3. Model-Predictive Controls for Efficient Ed Lohrenz, Member, Geo-Xergy Systems, How are engineers and architects going to
Mixed-Mode Cooling of Buildings (SA-12- Inc., Winnipeg, MB, Canada design buildings to cope with a changing
C039) 2. Using Multiple Chillers In a Cascading environment? Global Climate and Regional
Jianjun Hu, Student Member and Panagiota Operation with a Geothermal Loop Climate models can make predictions of
Mike Filler, Member, Trane Company, future climate while Canyon Air Temperature
Karava, Ph.D., Associate Member, Purdue
Pueblo, CO models are capable of predicting tempera-
University, West Lafayette, IN tures in a street canyon for extended periods.
3. Using Heat Recovery Heat Pumps to This seminar will show how a changing envi-
4. Contaminants in Hotel Room Exhaust
Recover Heat from a Chiller to Heat Ser- ronment can be accounted for in the design
Air (SA-12-C040) vice Water of buildings either by making provisions in
W. Brad M. Stanley and Bryan K. Ligman, Frank Pucciano, Member, Sabot 6, Atlanta, advance through design or mitigating the
AAF International, Doraville, GA GA impact after the fact.
xxxviii
ASHRAE Transactions
1. Can I Predict the Future? Generating Large commercial and industrial refriger- Seminar 67 (Intermediate)
Future Climatic Data ation systems found in supermarkets and
Drury Crawley, Ph.D., Member, Bentley Sys- refrigerated warehouses are coming under U.S. General Services
tems, Inc, Washington, DC the umbrella of energy efficiency regulation. Administration’s New
The process started and continues to evolve Performance-based
2. Climate Change and Its Effect On
Design Load Calculations
in California through their Title 24 Energy Facilities Standards
Efficiency Standard. Similar measures are
Charles S. Barnaby, Member, Wrightsoft being explored and adapted at the national
Track: Integrated Energy Systems
Corporation, Lexington, MA level. Even ASHRAE’s Standard 90.1 is Room: 001A
Sponsor: MTG.BPM Building Performance
exploring what steps can be taken in these Metrics
3. Climate Change Adaptation – Will
areas. This seminar is designed to update Chair: Martin Weiland, P.E., Member, Gen-
There Be Any Consequences for the Built
the audience on the current advances in eral Services Administration, Washington,
Environment? DC
energy efficiency regulation for the super-
Kelly Kalvelage, Iowa State University, Ames, GSA’s PBS-P100, “Facilities Standards
market and refrigerated warehouse indus-
IA for the Public Buildings Service”, has pre-
tries so they can make informed commentary scribed better-than-code buildings for
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
4. Accounting for Urban Microclimate In to the ASHRAE 90.1 committee. decades. Recently, in response to the need
Computer Simulation of Building Energy to allow designers more flexibility to meet
1. 2013 California Title 24 Building Energy escalating goals, with robust means to verify
Performance
Efficiency Standards for Supermarkets the results, it was transformed into a perfor-
Evyatar Erell, Ph.D., Member, Ben Gurion Brian Dobbs, Associate Member, VaCom mance-based standard. The new standard
University of the Negev, Beer-Sheva, Israel Technologies, San Luis Obispo, CA incorporates four performance tiers, from
better-than-code to true high performance. A
2. 2013 California Title 24 Building Energy new process, incorporating a breadth of pro-
Seminar 66 (Intermediate) Efficiency Standards for Refrigerated fessionals in technical committees, keeps its
development and revision continuous, with
Refrigerated Warehouse Warehouses quarterly published updates. Issues encoun-
Efficiency Regulations for Brian Dobbs, Associate Member, VaCom tered in the development of the standard,
Technologies, San Luis Obispo, CA and its implementation in a large organiza-
CA and Beyond tion, will be covered.
3. ASHRAE Standard 90.1 and Commer-
Track: Refrigeration Applications 1. GSA’s New Performance-Based Facili-
cial Refrigeration
Room: 007D ties Standard
James P. McClendon, P.E., Member, Martin Weiland, P.E., Member, General Ser-
Sponsor: 10.05 Refrigerated Distribution
Walmart Stores Inc., Bentonville, AR vices Administration, Washington, DC
and Storage Facilities
2. Performance-Based HVAC and Light-
Chair: Daniel J. Dettmers, Member, Indus- ing Requirements
trial Refrigeration Consortium-U.W. Madi- Bose Thomas, P.E., Member, US General
son, Madison, WI Services Administration, Washington, DC
© 2012 ASHRAE
xxxix
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---

CONFERENCE PAPERS
Presented at the
June 23–27, 2012
San Antonio, Texas
Conference papers report on research in progress, applications, case studies, and other topics in
HVAC&R technical areas.
Conference papers are first submitted as abstracts for review. Papers based off accepted
abstracts undergo review by at least two reviewers. The ASHRAE Conferences and Expositions
Committee decides, based on these reviews, whether papers are acceptable, need revision, or
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
should be rejected. The names of the reviewers are withheld from the author to ensure objective
comments, which are then transmitted to the author.
The procedure for review of Conference papers for ASHRAE Transactions is as follows:
1. Author submits an abstract of the proposed paper through the ASHRAE Conference Web site
for review by the conference organizers.
2. After abstract acceptance, author submits an electronic version of his/her paper through the
ASHRAE Conference Web site.
3. Two qualified peer reviewers review the manuscript, deciding on acceptance, minor
changes, mandatory changes, or rejection.
4. Unless a manuscript is immediately accepted or rejected, the author responds to reviewers’
comments and submits a revised manuscript.
5. A second review is conducted to confirm changes in response to comments; papers requir-
ing a third review or extensive changes are rejected. The goal is to have both reviewers
accept the paper for publication. The ASHRAE Conferences and Expositions Committee
will make final decisions on acceptance or rejection of manuscripts.
6. Authors whose papers are approved are invited to present their papers at an ASHRAE
conference, and papers are published in the Transactions of that conference.

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---

SA-12-C001
Achieving Deep Energy Savings in Existing

Buildings through Integrated Design
Dave Moser, PE Guopeng Liu, PhD Weimin Wang, PhD Jian Zhang, PhD
Member ASHRAE Member ASHRAE Member ASHRAE Member ASHRAE
ABSTRACT HEADING
Commercial buildings consume a large amount of energy – 35% of the nation’s electricity, and 13% of the nation’s natural gas. And the
majority of this existing commercial building stock will still exist over 20 years from now. Such a large stock of existing buildings presents a good
opportunity for decreasing our national energy demand through energy retrofits. This paper will present key concepts for cost-effective retrofit projects
with deep energy savings (45% and above) in existing buildings. Achieving deep savings typically requires an integrated design process, where all
building systems and assemblies are evaluated in a holistic manner. With this process, for example, HVAC equipment replacement measures are
evaluated in consideration of load reduction measures. The paper will present examples and considerations related to the integrated design process.
The upfront cost of a deep retrofit project may be difficult to justify on the basis of energy savings alone. However, the business case is much easier
to make when planned upgrades and equipment replacements are taken into account. The paper will present methods and considerations related to
planning for retrofit projects, and evaluating and ranking deep retrofit options. The Department of Energy has recognized the energy savings
potential in existing commercial buildings, and has recently published two Advanced Energy Retrofit Guides that target methods for achieving
various levels of energy reduction in existing buildings. The guides address the office and retail building sectors. As representative examples of
content included in each of the guides, the paper will include technical details from the guides related to specific energy reduction measures. The
general project planning considerations presented in the guides are applicable nationwide, while the energy and cost savings estimates for
recommended energy efficiency measures have been developed based on energy simulations and cost estimates tailored to five distinct climate zones.
The results of these analyses are presented for individual retrofit measures, as well as a package of recommended measures for three project types:
operations and maintenance measures implemented through the existing building commissioning process, standard retrofits, and deep retrofits. The
paper will present these measures and packages, to show examples of retrofits that can be implemented in existing commercial buildings to achieve
various levels of cost-effective energy reduction.
INTRODUCTION
Industry leaders have long recognized the role that energy retrofits can play in reducing operating costs and increasing
asset value. Opportunities for improved energy performance exist in nearly every building. These opportunities come in
many forms, including improved operational and maintenance practices, equipment retrofits, occupant behavioral changes,
and building envelope modifications, to name just a few. Over the life of a building, different opportunities will be available
at different times, depending on the changing usage of a building, remaining life of the equipment and assemblies, and
availability of improved technologies in the market.
While the opportunities for energy improvements in existing buildings are significant, the process of identifying,
Dave Moser, PE is a Senior Engineer at PECI, Portland, Oregon.
Guopeng Liu, Weimin Wang and Jian Zhang are Senior Staff Engineers at Pacific Northwest National Laboratory, Richland, Washington.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

analyzing, and implementing those improvements is not always straightforward. This paper presents key concepts for cost-
effective retrofit projects with deep energy savings (45% and above) in existing buildings.
TYPES OF RETROFIT PROJECTS

Retrofit projects can be categorized into three types of projects: (1) Operations and maintenance (O&M)
improvements implemented through the existing building commissioning (EBCx) process, (2) standard retrofits, and (3)
deep retrofits. Energy savings typically increase in magnitude as you move from EBCx to deep retrofit. Regardless of the
approach, after implementing retrofits, it’s important to verify that the systems are installed properly and operating correctly
in order to achieve the maximum energy savings potential of the retrofit. (Liu 2011a, Liu 2011b)
Significant energy savings can often be achieved with minimal risk and capital outlay by improving building operations
and restructuring maintenance procedures, which is a key component of EBCx. EBCx is a quality-oriented process for
investigating and optimizing the performance of a facility and its systems to meet the current needs of the facility. Research
has shown 16% median energy savings among hundreds of EBCx projects across the country (Mills 2009).
Standard retrofit measures provide cost-effective and low-risk efficiency upgrade options for building owners who are
limited to making incremental capital upgrades to their building. They are often conducted as like-for-like retrofits, meaning
that the new equipment has same capacity as old equipment, but with higher energy efficiency. Standard retrofits are often
implemented in a staged process, with one retrofit implemented after another. The sequencing of standard retrofit measures
is important, as the impact of some retrofits may have an impact on other systems. For example, a lighting retrofit may
reduce the cooling load on the HVAC system, which could allow for downsizing the cooling system. Accounting for these
interactions may yield greater savings than just like-for-like retrofits.
A deep retrofit is a building energy retrofit approach that uses an integrated design process to improve the economics
of efficiency and achieve much larger energy savings than conventional energy retrofits. A deep retrofit project provides an
opportunity for a building owner to reduce energy consumption significantly beyond the savings from O&M and standard
retrofit measures. In deep retrofit projects, all systems are evaluated concurrently, as an overall system, and the project may
involve the simultaneous retrofit of multiple systems. While deep retrofits can achieve deep energy savings, they may
require a larger upfront investment and may have longer payback periods than O&M or standard retrofit measures. They
are usually conducted as part of a major building remodel.
KEY CONCEPTS FOR DEEP RETROFITS

Deep retrofit projects combine many O&M and capital retrofit measures in an integrated whole-building design
process to achieve energy savings of 45% or more from a building’s current usage. The integrated design process enables a
deep retrofit project to achieve more than just a simple sum of the O&M and retrofit measures. Since these projects affect
multiple building systems and assemblies (e.g., envelope, lighting, and HVAC), the retrofit of each system and assembly
should be designed in close consideration of the other retrofits for maximum benefits to be realized. For a deep retrofit
project that is a major building remodel, the design process can look similar to that of a new construction project.
Deep retrofit projects are rarely conducted on their own for the sole purpose of energy savings. Typically, they are
considered and implemented in conjunction with other needed work at a facility that is being performed for non-energy
reasons, such as major envelope upgrades, equipment at the end of its life and in need of replacement, or major occupancy
changes. Deep retrofit projects are most cost-effective when piggybacked on other major work happening in the facility.
INTEGRATED DESIGN PROCESS

The integrated design approach to energy retrofits focuses on the simultaneous or staged retrofit of multiple building
systems, with a package of measures of varying complexities and financial benefits. For example, a building owner may
complete a lighting retrofit at the same time as increasing the amount of roof insulation and replacing the HVAC system.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
4
ASHRAE Transactions
The integrated approach is well-suited to building owners who either have ambitious energy savings goals to be met in
a short period of time, or have the opportunity to install deep retrofit measures due to planned changes in a building’s
systems, such as those that occur when a building is repurposed or undergoes a major renovation. From a financial
perspective, implementing multiple measures simultaneously has two distinct benefits:
1. The overall economics of the project are often improved. Cumulative project costs can be reduced compared to
the staged approach (implementing retrofits one after the other), due to efficiencies from installing multiple
measures at once. Lifecycle benefits may be increased, as energy savings begin at a high level rather than increasing
over time as retrofits are gradually implemented.
2. The integrated approach allows for the optimization of equipment sizes when multiple building systems and
assemblies are replaced. For example, if lighting and HVAC systems are replaced, the HVAC system designer can
take into account the reduced cooling load from the lighting retrofit, resulting in a smaller cooling system. Though
this can also occur in the staged approach, the integrated approach is generally more conducive to identifying such
opportunities.
The integrated design approach typically involves architects, design engineers, and potentially commissioning
providers working together as part of an integrated design process, where the various design disciplines coordinate closely
to design and specify systems and assemblies that will meet the owner’s needs as well as result in minimal energy use
(Energy Design Resources 2002). Retrofit systems are designed in concert, rather than as a sum of individual parts, and the
final design is evaluated using lifecycle economics. The design team develops multiple retrofit options, and ranks these
options against energy reduction goals, cost constraints, and other goals and considerations before deciding on a final
design with the Owner.
The integrated design of lighting, envelope, HVAC, and plug load systems calls for a design team with special
capabilities. Chief among these capabilities is that of open communication among team members. To foster open
communication, the integrated design process uses a different team structure than a traditional design process. It is more of
a team-based approach, with more collaboration between the various disciplines than the hierarchical owner-architect-
engineer organization found in the traditional design process. (ASHRAE 2011)
Investing in greater efficiency and load reduction can significantly reduce costs through downsizing, or even
eliminating, HVAC systems. This is a key feature of deep retrofits, but it cannot be achieved without thoughtful, integrated
design. The following, step by step approach for designing a deep retrofit project will lead to maximum benefits:
1. Define the needs and services required by the building occupants in a current facility requirements document. Start
from the desired outcomes. This means identifying a purpose, such as cooling, instead of going directly to a
solution, such as chillers.
2. Understand the existing building structure and systems. What needs are not being met? Why not?
3. Understand the scope and costs of planned or needed renovations. What systems or components require
replacement or renovation for non-energy reasons? What costs and interruptions to service or occupancy do those
renovations entail?
4. Reduce loads. Select measures to reduce loads:
x First, through passive means such as increased insulation
x Then, by specifying the most efficient non-HVAC equipment and fixtures
5. Select appropriate and efficient HVAC systems. After reducing loads as much as possible, consider what HVAC
system types and sizes are most appropriate to handle the reduced loads.
6. Find synergies between systems and measures. Find opportunities to recover and reuse energy waste streams. This
exercise will often identify multiple benefits that arise from a single expenditure.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

7. Optimize controls. After the most appropriate and efficient technologies have been selected, the focus should shift
to optimizing the control strategies for the HVAC, lighting, and plug load systems.
8. Realize the intended design. Conduct initial and ongoing commissioning to ensure continued realization of the
intended design and its benefits.
For deep retrofits, it’s important that the design team consider the building’s various systems and components as an
integrated system. Members of the project team must coordinate to minimize the expected energy usage of the building and
meet the owner’s specific design goals. Because of the complex interaction between systems, a whole-building energy
modeling software program is often required for the integrated approach.
PLANNING FOR DEEP RETROFIT PROJECTS

The upfront cost of a deep retrofit project may be difficult to justify on the basis of energy and maintenance cost
savings alone. The business case is much easier to make when planned upgrades and the avoided costs of equipment and
assembly replacements are taken into account. Many building upgrades occur throughout the life of a building, and these
planned capital improvements represent opportunities to perform a cost-effective deep retrofit. Table 1 shows examples of
key opportunities for embarking on a deep retrofit project.
Table 1. Opportunities in a Building’s Life to Perform a Deep Retrofit
Building Event Opportunity

End (or near end) of Major equipment replacements provide an opportunity to also address the envelope and other
life major equipment building systems. After reducing thermal and electrical loads, the marginal cost of replacing the
replacement major equipment with smaller equipment, or no equipment at all, can be negative.
Planned roof, window and siding replacements provide opportunities for significant
Roof, window and
improvements in daylighting and envelope thermal performance at small incremental costs. These
siding replacement
improvements in turn allow for reduced artificial lighting and a smaller HVAC system.
A major occupancy change presents a prime opportunity for a deep retrofit. Owners may be able
Major occupancy
to leverage tenant investment in the fit out. Also, major retrofits are best done when the space is
change
vacant, before the new tenants move in.
A desire to achieve green building or energy certification may require significant work on the
Building greening
building and its systems, which may then make a deep retrofit economical.
When building owners are aware of these opportunities, they can engage the integrated design process and make a
planned component replacement grow into a deep retrofit. In some cases, the opportunity is obvious. For example, if the
roof must be replaced, insulation can be added to the new roof. But other opportunities are less straightforward. For
instance, if a building’s roof needs replacement in five years but the boiler is slated for replacement now, it probably makes
most economic sense to move that roof replacement up, and add insulation to reduce the heating load and the size and cost
of the boiler. This latter example highlights how a basic understanding of the deep retrofit process can help building owners
reap greater rewards from their investments.
Another major consideration is the deep retrofit project’s impact on the building occupants during the construction
phase. If the building will be occupied during the project, this will limit the types of measures that can be considered. For
example, a complete replacement of the building envelope is not a realistic option if the building will remain occupied
during construction.
Once motivated to embark on a deep retrofit project, building owners need to develop a project-specific business case
that will ensure that the project meets long-term cost-effectiveness requirements. There are many financial analysis methods
available, including simple payback, net present value, internal rate of return, and life cycle cost, to name a few. When
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
6
ASHRAE Transactions
reviewing the deep retrofit options developed by the design team, it’s important to evaluate the financial impacts in a
consistent and appropriate manner. A project can initially appear to be unattractive when viewed through the lens of simple
payback period, while a more complete economic analysis such as net present value reveals it to be a profitable investment.
Life-cycle cost analysis is more effective at identifying the best project option, once the costs and benefits of each
alternative are carefully analyzed and expressed in present value terms.
When performing the financial analysis for deep retrofit options, the incremental costs and benefits should be used,
comparing the costs and benefits of the options to the costs and benefits related to just a newer version of the current
equipment. Also, programs such as utility on-bill financing, utility incentive programs, and tax relief can be an effective way
to reduce a project’s total cost. These should be factored into the estimated costs of deep retrofit options when evaluating
alternatives.
EXAMPLE DEEP RETROFIT PACKAGES

Example deep retrofit measure packages for office and retail buildings are provided in the Advanced Energy Retrofit
Guides (“guides”) available from the U.S. Department of Energy (DOE) (Liu 2011a, Liu 2011b). The energy savings and
financial performance of the packages, which include a mix of measures, are analyzed for a representative retail and office
building (Deru 2011) in five primary climate regions in the U.S.: hot & humid, hot & dry, marine, cold, and very cold.
There are a number of O&M and capital retrofit measures that could be included as part of a deep retrofit package,
depending on the goals of the project and the outcomes of the integrated design process. The measures included in the
guides’ deep retrofit packages, shown in Tables 2 and 4, are representative examples of deep retrofit projects with a bundle
of capital and O&M measures. They may not be applicable to some retail and office buildings, and there may be other
measures that are applicable but aren’t included in the list. The measures listed in these packages are applicable to a
reference building that has characteristics similar to most standalone retail buildings and large office buildings in the U.S. of
pre-1980 construction. (Deru 2011)
In the guides, the analysis of the deep retrofit packages assumes that O&M measures are implemented first, as part of
an EBCx process, followed by the deep retrofit measures. This is estimated to result in savings of over 45% of site energy
usage in the reference buildings, as shown in Tables 3 and 5, based on an analysis of the measures in the packages using
EnergyPlus, a building energy simulation software program. Each climate zone shows significant energy savings, with slight
variations between the climate zones. The analysis accounts for interactive effects between building systems and measures.
The estimated financial metrics associated with the deep retrofit packages in each climate zone are shown in Tables 3
and 5. These metrics include implementation of the measures shown in Tables 2 and 4. The costs and savings shown in
these tables are incremental costs and savings for the retrofit measures, since it is assumed that the equipment is at the end
of its useful life and is in need of replacement. The incremental costs and savings are based on the difference between
similar standard efficiency equipment and an energy efficient option. Full costs were assumed for measures that only added
functionality to the existing system.
Retail
The measures included in the 24,695 ft2 (2,294 m2) retail building’s deep retrofit packages in the guide are shown in
Table 2 (Liu 2011a). The measures range from the addition of simple controls functionality (re-circuit and schedule lighting
system), to significant changes to the building’s systems (replace HVAC system). Most of the measures relate to the
building’s lighting system, since this is the end use that uses the greatest amount of energy for the retail building. The
technical details related to the measures included in the deep retrofit packages can be found in the guide (Liu 2011a).
Two of the measures listed in Table 2 apply to a specific type of HVAC system commonly found in standalone retail
buildings: single-zone packaged rooftop units. While this is probably the most common type of HVAC system found in
existing standalone retail buildings, these two HVAC measures may not apply to other HVAC system types. However, the
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

concepts can be applied to other HVAC system types: increase the efficiency of the existing system’s cooling and heating
sections, and utilize energy recovery.
Table 2. Retail Deep Retrofit Package Measures
System Measure Description Climate Zone

O&M measures: Calibrate exterior lighting photocells, reduce envelope leakage,
replace worn out weather stripping at exterior doors, clean cooling and heating coils
Lighting,
and comb heat exchanger fins, revise air filtration system, add equipment lockouts
Envelope,
based on outside air temperature, reprogram HVAC timeclocks to minimize run time, All
HVAC, Service
optimize outdoor air damper control, repair airside economizer, increase deadband
Hot Water
between heating and cooling setpoints, replace plumbing fixture faucets with low flow
faucets with sensor control.
x Add daylight harvesting
x Re-circuit and schedule lighting system by end use
Lighting x Retrofit interior fixtures to reduce lighting power density by 58% All
x Install skylights and daylight harvesting
x Retrofit exterior fixtures to reduce lighting power density, and add exterior
lighting control
Envelope Add roof insulation Hot & Humid
x Replace RTUs with higher efficiency units
HVAC All
x Remove heat from front entry
Marine, Cold,
HVAC Replace HVAC system with a dedicated outdoor air system
Very cold
As shown in Table 3, the deep retrofit packages have a five-to-six year payback and positive net present value, making
them a cost-effective method of achieving significant energy savings.
Table 3. Retail Deep Retrofit Savings and Financial Analysis
Site Energy Use Intensity

(EUI), kBtu/sf/yr Financial Analysis
(kWh/m2/yr)
Total Total Simple
Post-Deep Site EUI Net Present
Baseline Package Annual Payback,
Retrofit Reduction Value
Cost Savings years
Hot & Humid 107 (338) 44 (139) 59% $161,000 $27,800 5.8 $4,860
Hot & Dry 103 (325) 47 (148) 54% $129,000 $25,300 5.1 $37,800
Marine 90 (284) 38 (120) 58% $124,000 $22,600 5.5 $19,900
Cold 100 (315) 43 (136) 57% $139,000 $29,600 4.7 $61,300
Very Cold 102 (322) 46 (145) 55% $130,000 $24,400 5.3 $30,400
Average 100 (315) 44 (139) 56% $137,000 $25,900 5.3 $30,900
Office
The measures included in the 200,000 ft2 (18,581 m2)office building’s deep retrofit packages in the guide are shown in
Table 4 (Liu 2011b). The measures range from the addition of simple controls functionality (occupancy sensor control of
--`,`,``,,`,`,```,`,,``,
8
ASHRAE Transactions
lighting), to significant changes to the building’s HVAC systems (replace boilers and convert cooling plant pumping
system). Most of the measures relate to the building’s HVAC system, since this is the system with the greatest opportunity
for optimization. The technical details related to the measures included in the deep retrofit packages can be found in the
guide (Liu 2011b).
The measures were also chosen based on their impact on the tenants. The reference building is a multi-tenant
commercial office building, whose owner likely cannot afford to ask the tenants to leave for a few years during the remodel.
The measures included in the deep retrofit packages listed in Table 4 were selected based on this constraint.
Some of the measures listed in Table 4 apply to a specific type of HVAC system commonly found in large multi-
tenant commercial office buildings: a multi-zone VAV air system with a water-cooled chilled water plant and a heating
water system. These HVAC measures may not directly apply to other HVAC system types. However, the concepts can be
applied to other HVAC system types: add functionality to adjust the system operating parameters during part load
operation to realize energy savings and extended equipment life.
Table 4. Office Deep Retrofit Package Measures
System Measure Description Climate Zone

O&M measures: Reduce envelope leakage, revise air filtration system, calibrate air and
Lighting, water sensors, re-enable supply air temperature setpoint reset, reduce HVAC
Envelope, equipment runtime, close outside air damper during unoccupied periods, reduce All
HVAC economizer damper leakage (except Hot & Humid), shut down cooling plant when
there’s no cooling load.
x Retrofit interior fixtures to reduce lighting power density by 11%
x Install occupancy sensors to control interior lighting
Lighting x Add daylight harvesting All
x Retrofit exterior fixtures to reduce lighting power density, and add exterior
lighting control
All, except Hot &
Envelope Add roof insulation
Humid
x Widen zone temperature deadband, add conference room standby control
(upgrade to DDC zone control)
x Lower VAV box minimum flow setpoints, reset duct static pressure (upgrade to
HVAC All
DDC zone control)
x Add demand-controlled ventilation
x Replace supply fan motor and VFD
x Shut down heating plant when there's no heating load
x Increase efficiency of condenser water pumping system Hot & Humid,
HVAC
x Change cooling plant pumping system to variable primary Hot & Dry
x Add a VFD to one chiller
HVAC Increase efficiency of condenser water system Hot & Dry
Marine, Cold,
HVAC Replace boilers and change heating plant pumping system to variable flow primary
Very cold
As shown in Table 5, the deep retrofit packages have a five-to-six year payback and positive net present value, making
them a cost-effective method of achieving significant energy savings.
$6+5$(
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---

Table 5. Office Deep Retrofit Savings and Financial Analysis
Site Energy Use Intensity
(EUI), kBtu/sf/yr Financial Analysis
(kWh/m2/yr)
Total Total Simple
Post-Deep Site EUI Net Present
Baseline Package Annual Payback,
Retrofit Reduction Value
Cost Savings years
Hot & Humid 88 (278) 48 (151) 45% $697,000 $117,000 6.0 $227,000
Hot & Dry 97 (306) 46 (145) 52% $890,000 $161,000 5.5 $422,000
Marine 94 (297) 44 (139) 53% $918,000 $162,000 5.7 $369,000
Cold 86 (271) 44 (139) 48% $885,000 $153,000 5.8 $302,000
Very Cold 91 (287) 47 (148) 49% $837,000 $137,000 6.1 $211,000
Average 91 (287) 46 (145) 50% $845,000 $146,000 5.8 $306,000
CONCLUSION
Existing office and retail buildings consume 17% and 13% of total commercial building energy use respectively, the
top two energy use of any sector, and these buildings contain ample opportunity for energy improvements. The majority of
existing office and retail buildings were built before 1980, and the equipment in those buildings is inefficient when
compared to newer technologies. Deep retrofit projects are a cost-effective opportunity for substantially reducing energy
usage, through the implementation of a package of O&M and capital measures identified with an integrated design process.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
ADDITIONAL INFORMATION
The U.S. Department of Energy’s (DOE) Advanced Energy Retrofit Guides present general project planning guidance as
well as financial payback metrics for energy efficiency measures. Emphasis is put on actionable information, practical
methodologies, diverse case studies, and objective evaluations of promising retrofit measures. The guides are available
through the DOE’s Office of Energy Efficiency and Renewable Energy (EERE) at http://tinyurl.com/DOE-EERE.
Rocky Mountain Institute’s RetroFit Depot (www.retrofitdepot.org) offers case studies, free energy modeling tools and a
step-by-step overview of the deep retrofit process to help owners and energy service practitioners pursue deep retrofits.
ACKNOWLEDGMENTS
The Advanced Energy Retrofit Guides, which were the main source for the information presented in this paper, are the
result of numerous peoples’ efforts. The authors wish to acknowledge the valuable assistance provided by the staff of the
DOE and its Office of EERE. We’d also like to thank Weimin Wang, Jian Zhang and Bing Liu at Pacific Northwest
National Laboratory, and Eliot Crowe, Nick Bengtson, Mark Effinger, and Lia Webster at PECI for their contributions.
REFERENCES
ASHRAE. (2011). Advanced Energy Design Guide for Small and Medium Office Buildings. American Society of Heating,
Refrigerating and Air-Conditioning Engineers, Inc., Atlanta, GA.
Deru, M., et al. (2011). U.S. Department of Energy Commercial Reference Building Models of the National Building Stock. National
Renewable Energy Laboratory, Golden, CO.
Energy Design Resources (2002). Design Brief: Integrated Building Design. California Public Utilities Commission.
Liu, G., et al. (2011a). Advanced Energy Retrofit Guide for Retail Buildings. Pacific Northwest National Laboratory, Richland,
WA.
Liu, G., et al. (2011b). Advanced Energy Retrofit Guide for Office Buildings. Pacific Northwest National Laboratory, Richland,
WA.
Mills, E. (2009). Building Commissioning: A Golden Opportunity for Reducing Energy Costs and Greenhouse Gas Emissions. Lawrence
Berkeley National Laboratory, Berkeley, CA.
10
ASHRAE Transactions
SA-12-C002
Right Steps for Retrofits: Byron G. Rogers

Federal Office Building Case Study
Kendra Tupper, PE Nicole Hammer, PE Rick Osbaugh, PE Michelle Swanson, PE

ASHRAE Member Non-Member ASHRAE Member ASHRAE Member
ABSTRACT
Achieving deep energy savings is especially challenging when you’re faced with an existing structure with poor orientation, historic
preservation requirements and little space for renewables. This paper outlines how the design team used the right steps in the right order to achieve
a predicted 60-70% annual energy cost savings in a building with all of these challenges – the Byron G. Rogers Federal Office Building in
Denver, Colorado. The Byron Rogers building is currently undergoing a deep energy retrofit, to be completed in 2013, and upon completion it will
consume less than half the energy than it did pre-renovation. The building was constructed in the 1960s and is just under 500,000 square feet.
Upon its fiftieth birthday, it will officially be recognized as a historic building.
INTRODUCTION
The Byron Rogers Federal Office Building, just under 500,000sf and located in downtown Denver, CO, is pursuing
impressive sustainability targets above and beyond federal requirements. Built in the 1960s, this building not only has to
achieve federal NetZero goals of the future, but as a historic building it must also reflect respectfully on its past and
preserve architectural elements characteristic of the era in which it was built. Byron Rogers is pursuing LEED-NC Platinum
certification, which will require full participation of its 10+ tenants – all government agencies. When the American
Reinvestment and Recovery Act (ARRA) was created in 2009, this building was selected to receive funding for a complete
retrofit due to its age and condition. The design team also received several additional pockets of ARRA funding strictly for
the incorporation of emerging energy efficiency technologies such as LED lighting into the design.
Using whole building DOE-2.2 energy simulations, the design team predicts this building will save between 60-70% in
annual energy costs as compared to the pre-renovation utility bills. Prior to the renovation, the building’s energy use
intensity (EUI) in 2009 was 91.8 kBtu/ft2-yr (290 kWh/m2-yr). Post renovation, we expect the EUI of the building to be
between 28-38 kBtu/ft2-yr (88-120 kWh/m2-yr). These energy reductions can be achieved with a net present value of
$556,700 over a 20 year period.
APPLYING THE RIGHT STEPS IN THE RIGHT ORDER

Too often, design teams start designing systems and selecting equipment before optimizing the loads and function of
the building. In the deep energy retrofit process, it is important to identify the right steps to take, and equally important to
perform these steps in the right order. Following this process helps realize the most cost effective energy reductions and
produce multiple benefits from single expenditures.
Kendra Tupper and Nicole LeClaire are senior consultants at Rocky Mountain Institute, Boulder, CO. Rick Osbaugh and Michelle Swanson are
mechanical engineers at the RMH Group Inc., Lakewood, CO. --`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

Table 1. The Right Steps in the Right Order
1. Set Quantifiable Goals

6. Find Synergies
2. Define End-User Needs
7. Optimize Controls
3. Understand Existing Conditions
8. Incorporate Renewables
4. Reduce Loads
9. Realize the Intended Design
5. Select Appropriate and Efficient Systems
SET QUANTIFIABLE GOALS

At the beginning of a project, it is important to identify clear and quantifiable project goals. It is particularly effective
to hold a Goal Setting Workshop with all stakeholders to achieve commitment for these goals, and to assign responsible
parties. Goals can be classified into the following broad categories:
Table 2. Types of Quantifiable Goals

Goal Type Description of Goal Specific Goals for Byron Rogers
Specifies required overall energy or cost effectiveness 28 kBtu/ft2-yr (assumes best case for operation
Overall
metrics, such as EUI or a positive net present value schedules and plug loads);
Target Values
(NPV). Net zero operating energy by 2030
Specified in relation to some baseline - compares ≥55% energy cost savings compared with
Comparative
against similar building types or current consumption. ASHRAE 90.1-2007
Certification
A specific type of rating or level of certification LEED Platinum
or Rating
End-Use Specifies a concrete goal at the end-use level – useful 100% of the domestic hot water needs will be met
Specific in assigning responsibility among the design team. with on-site solar thermal
In order to set aggressive goals, it is important to understand what is technically feasible before applying constraints
related to budget, time, and user experience. To do this, we conduct a technical potential exercise during the Goal Setting
Workshop. The technical potential is the minimum level of energy consumption possible for the existing building, given
today’s technology (excluding renewables). This is an unambiguous, rigorous scientific basis for energy efficiency potential.
Byron Rogers Example: Technical Potential

Often, design teams limit their ideas based on cost, constructability or project schedule. For the Byron Rogers project,
the design team used the technical potential as a starting point for design. Using a technical potential exercise relieved the
design team of those typical constraints, helped to identify more creative solutions, and forced contributors to question why
particular efficiency measures may or may not be feasible. This process was crucial to get all stakeholders to first
acknowledge and consider the aggressive and forward-thinking possibilities for the project and then to determine what level
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
of energy efficiency is implementable by systematically reintroducing project constraints.
DEFINE END-USER NEEDS

The second step in reducing the energy consumption of an existing building is to define the needs and services that
end users require, rather than jumping right to the equipment or capacity needed to provide it. Here is a simple example:
depending on the task at hand, different building occupants require different visual environments. Instead of approaching
the design from the standpoint that “we need electric lighting”, consider that “we need even light distribution, luminance
12
ASHRAE Transactions
levels in a range of x-y foot-candles, and to eliminate glare from computer screens and other equipment, etc.” By
questioning and clearly defining the needs of the user, we can determine the reliable passive strategies that best meet those
criteria before assuming the need for powered technology.
Byron Rogers Example: Task/Ambient Lighting Solution

The lighting in the Byron Rogers office building will be a task-ambient system using LED type luminaires in
conjunction with local task lights within the furnishings. The spaces will have active daylight harvesting and occupancy
controls. The ambient light maintains an average illumination of 30 footcandles from permanently fixed luminaires across
the viewplane, while local task lighting provides supplemental illumination up to the required task lighting levels. This
approach conserves energy and allows occupants to fine-tune the amount of light in their work area.
Byron Rogers Example: Plug Loads

In extremely efficient office buildings, plug loads (computers, copiers, AV equipment, space heaters, appliances, etc.)
become a proportionally significant contributor to overall energy consumption. In the current energy model for the Byron
Rogers building, plug loads are projected to account for over 38% of the total energy use if no improvements are made. As
the design team has drastically improved the building envelope and completely redesigned the lighting and mechanical
systems, plug loads remain the one area in which energy savings have yet to be realized. Plug loads are often the most
difficult end use to address, as the type of equipment that each tenant brings into their space is often out of the design
team’s control. The Byron Rogers federal office building is no exception, as the building owner cannot mandate their
tenants to replace current plug load equipment with more efficient technologies.
In addition to creating tenant sustainability guidelines to help the tenants understand project sustainability goals and
strategies, the design team is currently conducting a plug load energy audit of the tenant spaces to gain an understanding of
the tenants’ current plug load energy use. The design team will use the information collected to create a more accurate
model of the building's energy use, and to identify opportunities for replacement with more efficient equipment.
Additionally, more aggressive plug load controls strategies will be studied and recommended.
Impact of Plug Load Efficiency and Operating Schedules
Aggressive Equipment; Aggressive Schedule

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
ENERGY STAR Equipment; Aggressive

Schedule
ENERGY STAR Equipment; Current

Schedule
Current Equipment; Aggressive Schedule
Current Equipment; Current Schedule
0 5 10 15 20 25 30 35 40 45
Energy Use Intensity (kBtu/ft2-yr)
Figure 1: Preliminary studies comparing estimated existing plug load energy use and operation schedules with more
aggressive equipment and operation schedules show that this factor affects the overall building EUI by over 36%!
$6+5$(

UNDERSTAND EXISTING CONDITIONS AND PLANS
Next, the design team should assess the current state of the existing facility. What needs are not being met? Why not?
What systems or components require replacement or renovation for non-energy reasons? What are the costs or
interruptions to service or occupancy? By identifying these planned renovations early, it may be possible to combine these
with a desired energy retrofit to optimize the total return on investment.
Byron Rogers Example: Historic Preservation and Building Envelope/Windows

The retrofit of the Byron Rogers building provides an opportunity to implement window and envelope technologies
that meet or exceed modern energy performance standards. Due to historic preservation constraints on the exterior
building façade, all upgrades are being made from the inside of the building.
The retrofit design strategy for the windows is to leave the existing punched window frames in place while minimizing
the conductive path for unwanted heat and cold into the building that they provide. This will be done by replacing the
existing one-inch insulated glazing unit (with a single air cavity and a grey tint) with a single pane of glass, similarly grey in
tint to comply with historic preservation criteria. A new ultra-insulated operable glazing unit (without tinted glass) is being
installed on the interior side of the frame, and will maintain a thermal break along with the proposed building envelope
upgrades. The new glazing assembly will provide a higher transmittance value for visible light and dramatically improved
solar heat gain performance than the existing windows. The operable assembly will accommodate regular cleaning needs.
The existing building envelope will be freed of any residual asbestos-containing material contamination from the
inside of the building, and new insulation will be added. The new exterior wall system has the potential to achieve a
composite R-value of approximately 20 Btu/°F-ft2-hr (113 W/m2-K). In addition, each floor slab will receive a continuous
insulation wrap on the bottom, and a thermal break material will be applied to strategic portions of the top of the slab.
REDUCE LOADS
Before considering how to satisfy the heating and cooling loads of a building, efforts should focus on reducing those
loads as much as possible, first through passive means and then by using the most efficient end use equipment. Sources of
heating and cooling loads can be classified as follows:
Table 3. Sources of Heating and Cooling Loads

Building Envelope Internal Gains Outside Air
Walls Lighting Infiltration
Roofs People Required ventilation Air
Windows Equipment
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Other (e.g. indoor pool)
Internal gains only generate cooling loads, and they actually help provide heat to the space when the building is in
heating mode. The building envelope and outside air can exert either heating or cooling loads on a building, depending on
the interior and exterior conditions. Loads from the building envelope come from solar heat gain as well as conductive,
radiative, and convective heat transfer. Loads from building occupants, outside air, and some types of equipment have
latent cooling loads as well as sensible cooling loads.
Reduce Loads through Passive Means (Requires No Power)

The more that energy demand can be met intrinsically by the passive design of the building, envelope, and site plan
alone, the less the burden on mechanical systems, renewables, and local infrastructure. Early on, the design team should
investigate ways to retrofit the existing building to have a more climate-responsive form and envelope that capitalizes on
the site’s climatic conditions. While it is difficult to change floorplate design, ceiling height and orientation (in this case, 45
14
ASHRAE Transactions
degrees off north) within existing buildings, it is possible to dramatically improve the insulation levels, add radiant barriers,
and block unwanted solar heat gain. For the Byron Rogers building, the focus on passively reducing loads was two-fold:
Table 4. Byron Rogers Passive Load Reduction Strategies

Design Out Space Heating Loads Optimize Solar Heat Gain
Super-insulate the walls and roof. Capitalize on the non-ideal orientation of the building
Add insulating windows to the inside of the and use it to capture heat for use in the building.
existing window boxes to create an Choose new windows that drastically reduce solar heat
uninterrupted thermal break. gain. The windows facing southwest have a lower solar
Employ heat recovery on incoming outside air. heat gain coefficient than those facing northeast.
The Byron Rogers building essentially has three separate heating and cooling zones—two with competing solar
exposures, and one interior zone. It is not uncommon to have one of the exterior zones in cooling mode while the other is
in heating mode. To take advantage of this, the mechanical system was designed to move energy between these zones and
store low temperature heat for times when the building is largely unoccupied but still requires heat (see Figure 1).
Reduce Loads by Using the Most Efficient End-use Equipment (Powered)

As cooling loads within office buildings are dominated by internal gains, the most significant reductions can be
achieved by specifying extremely efficient interior lighting and plug load equipment. When buildings are retrofitted to be
extremely efficient, plug loads and process equipment became a very significant overall end use, and one of the most
challenging to reduce. At Byron Rogers, given that tenant plug loads were largely out of the design team’s control, the team
first focused on electric lighting. ASHRAE 90.1 2007 Building Area Method Table 9.5.1 mandates a lighting power density
(LPD) of 1.0 W/ft2 (11 W/m2) for office buildings. The average LPD for Byron Rogers will be 0.65 W/sf (7 W/m2), using
100% LED lighting. In addition, the lighting control system accommodates daylight dimming in perimeter zones, meaning
the ambient lighting will automatically be dimmed or turned off entirely for the majority of daylight hours. Automated
dimming also provides the ability to maintain illumination levels over the life of the LED light sources, thus providing
further energy savings.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
SELECT APPROPRIATE AND EFFICIENT SYSTEMS

After reducing internal loads as much as possible, the next step is to consider which HVAC system types and sizes are
most appropriate to handle the drastically reduced loads, and then to select the most efficient technologies available for that
system. At this step, the heating and cooling load requirements at Byron Rogers had been drastically reduced through
efficiency efforts in other systems. A greater number of system solutions such as chilled beams, radiant cooled surfaces,
and wall diffuser displacement ventilation were thus available for consideration. Where possible, the design team attempted
to passively design out the building envelope heating load and handle the remaining service water heating and ventilation air
heating with a combination of solar thermal and recovered waste heat.
Byron Rogers Example: Evaluating HVAC Systems Selections and Efficiencies

With energy demands minimized, the goal for the HVAC system at Byron Rogers was to design an efficient, low
energy mechanical system that takes advantage of Denver's dry climate without sacrificing occupant comfort. The most
significant challenge was to reduce the heating energy requirement for the building. Geo-exchange and water source heat
pump systems were first considered due to their ability to transfer heat, simplicity of design and proven ability in the field.
These systems transfer energy back and forth to the ground using a large bore field, and they often use an average of forty
percent less energy than traditional HVAC systems. Due to site constraints in Byron Rogers’ urban setting and a lack of
space to locate the 430 heat pumps that would have been required, this type of system was not feasible.
$6+5$(

However, after a detailed evaluation of the Byron Rogers building’s physical orientation, the design team determined
that energy (heat) transfer could be accomplished in a manner similar to geo-exchange by using the building itself, along
with thermal storage tanks, as the geo-exchange bore field, and the chiller as a central heat pump. Because significant heat
is generated inside the building, this heat could be used to offset winter and night time heating needs.
Heat Reclaim/ Hybrid Heat Pump System
Thermal Storage
SW in sun; NE in shade Heating Mode Cooling Mode

Figure 1: The building’s poor solar orientation provided an opportunity to use heat reclaim and thermal storage in
combination with the hybrid heat pump system, requiring the boiler to be used only on extreme cold days.
An active chilled beam (ACB) system was selected for use at Byron Rogers to optimize the transfer of energy between
the two solar exposure zones and the interior zone, while still storing low temperature heat for times when the building is
largely unoccupied but still requires heat. The ACB system presented a significant energy reduction opportunity because it
provides a method for transferring energy between zones both efficiently and economically. The primary energy savings
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
comes from the reduction in required cooling—active chilled beams use 58°F (14°C) water for cooling as compared to
45°F (7.2°C) water for typical air-based cooling systems. Additional energy is saved since the cooling towers in this climate
can easily produce 58°F (14°C) water without chiller operation for a large portion of the year. Heating mode is also more
efficient because it uses low temperature water. Rated heat output in heating mode is only 88°F (31°C), making energy
transfer between building zones feasible.
Due to the size of the building, energy-efficient centrifugal chillers were the best choice for producing chilled water.
The active chilled beam system will use heat generated by the chiller to heat the building. The centrifugal chiller will
transfer low grade heat as required by the active chilled beam system, using ninety percent less energy than heating the
building with a high efficiency boiler. In the heat pump mode the system uses less chiller energy than the system in a tower
free cooling mode alone.
Heat can also be captured during the day when there is an abundance of available solar heat energy and internal heat
gains, and stored for use at night and on weekends. Similarly, solar energy that enters the building through the windows on
the weekends can also be stored and used at night without using much additional energy to move it from zone to zone.
Once the most appropriate mechanical and lighting systems were selected, the design team specified the most cost-
effectively efficient technologies within that system type. The heat pump chiller chosen for Byron Rogers was a magnetic
bearing chiller to optimize both full load and part load performance. This chiller is capable of transferring heat between the
16
ASHRAE Transactions
chilled water and heating water systems at a part load COP of 12 and a full load COP of 9. In addition, 95% efficient
condensing boilers are used when heat generated inside the building from lights, equipment, and people is not adequate to
maintain building temperature.
FIND SYNERGIES
To achieve deep energy savings, it is important to capitalize upon synergies across disciplines and find opportunities to
recover and reuse waste streams. Through these efforts, we can often realize multiple benefits from a single design decision.
Byron Rogers Example: Heat Recovery and Thermal Storage Design

Ventilation air heat recovery is accomplished through an enthalpy wheel in dedicated outside air units designed to
recover 80% of the heat being exhausted from the building. In addition, thermal storage tanks are utilized to store heat
generated during the day to supply building heating needs at nights and on the weekends. Through energy analysis, it was
determined that the building would consume approximately 15 million BTU's (15.8 kJ) during a typical winter weekend in
heating energy. To provide this heat, a 50,000-gallon (189 m3) thermal storage tank is located in the basement of the
building. By using heat generated during the day to charge the tanks, this heat can be used at night via the heat pump
chillers to boost heat into the building at a high COP.
OPTIMIZE CONTROLS
After appropriate and efficient technologies are selected, the focus should shift to optimizing control strategies.
Byron Rogers Example: Lighting and HVAC Controls

For the Byron Rogers project, the team focused on optimizing lighting and HVAC controls. Individual groups of
luminaires will be controlled by addressable modules connected to the lighting control system. The lighting control system
will integrate time-based and sensor-based lighting controls with a network of sensors, power packs, photocells, and wall
switches, with distributed intelligence. This innovative system provides global control via web-based lighting management
software, and it will enable zones of devices to self-commission and function independently. The lighting control system
will provide dimming, occupancy, and on/off control and will interface with the building automation system (BAS) for
zone control of the HVAC system. Utilizing signals from occupancy sensors, the lighting control system will be capable of
notifying the BAS which zones are unoccupied. In response, the BAS will shut down the chilled beam hydronic
heating/cooling supply. Ventilation air will continue to be supplied to the unoccupied zones, providing them with a small
amount of heating or cooling.
INCORPORATE RENEWABLES AND DEMAND SIDE MANAGEMENT

After the more cost effective efficiency measures have been exhausted, the team should consider incorporating on-site
renewable energy and demand side management strategies to meet the now greatly reduced energy demand. When
investigating these options, major considerations should include:
x How can the location of renewable energy systems be optimized within the site layout? How can they be integrated
with other design measures?
x What are the most site and climate appropriate renewable energy systems?
x What supply-side choices would provide the most benefit to the utility and the grid, accounting for the ability of that
choice to be predictable, provide flexibility, reduce peak demand, etc?
x How can energy management be used to maximize value for both the program participant and for the utility? That is,
what is the role of demand response, smart appliances, or thermal storage in shifting load off peak? What is the
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

appropriate level of interface between the utility and the program participant?
Byron Rogers Example: Renewable Energy

Renewable energy strategies will be utilized to meet a portion of the building’s remaining energy demand.
Technologies such as building-integrated photovoltaics (PV) and solar thermal hot water can further reduce the building’s
fossil fuel consumption and carbon footprint. Flat plate solar thermal collectors will supply 100% of the domestic hot
water demand in the building. A 90 kW roof-mounted PV system is planned for installation. In order to achieve the goal
of net zero operating energy by 2030, additional on-site renewables and the purchase of some renewable energy credits
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
(RECs) may be required over time.
REALIZE THE INTENDED DESIGN

Finally, to ensure full realization of the intended design, the team should create a plan to implement measurement and
verification (M&V) and continuous commissioning. Savings from energy efficiency projects cannot be measured directly,
since they represent the absence of energy use. M&V is the practice of measuring, computing and reporting savings for
energy-savings projects. M&V activities include site surveys, metering of energy flows, monitoring of independent variables,
calculations, and reporting. As part of the integrative design process, the team should set collaborative performance targets
and ensure that each service provider understands their obligations and contributions to meeting them.
The optimal performance of the project post-construction should be achieved initially through commissioning and
M&V-related activities. An ongoing commissioning process should be developed and outlined as part of the M&V Plan. To
take this effort to the next level, an automated fault-detection and diagnosis (FDD) system could be considered to work in
conjunction with the BAS. The FDD can be programmed to function as a “smart” filter which will alert the building
operator and commissioning agent to the most critical operational concerns that impact overall performance and comfort.
Byron Rogers Example: Measurement and Verification

Working with the building owner and its independent testing inspectors, the Byron Rogers team will develop and
implement a plan for enhanced building commissioning. In addition, the team will provide management and coordination
for commissioning, including commissioning procedures, pre-functional test procedures, functional performance test (FPT)
procedures, documentation forms, and checklists, all in accordance with IPMVP requirements.
CONCLUSION
Existing commercial buildings represent a major opportunity in the US to reduce building energy consumption. For
many, however, there remains a technical question as to whether we can cost-effectively achieve deep energy savings in
existing buildings given legitimate concerns such as existing orientation and massing constraints, security requirements, and
unknown future building plug loads and uses. The Byron Rogers design team used the right steps in the right order to re-
frame these design concerns into major efficiency opportunities. By identifying the appropriate inquiries to make
throughout design and post-occupancy, the design team undertook a rigorous approach to identify, test, select, and verify
efficiency measures. The process encouraged the team to use old technology and concepts in innovative ways to capture
synergies between measures, and produce multiple benefits from single expenditures, in this case, resulting in an anticipated
50% or greater reduction from the pre-retrofit energy use baseline.
ACKNOWLEDGMENTS
The design team would like to thank the General Services Administration (project owner) and the building tenant
agencies for their cooperation in this process.
18
ASHRAE Transactions
6$&
Current Performance of Ground Source

Heat Pumps for Space Conditioning and for
Water Heating under Simulated Occupancy
Conditions
Moonis R. Ally, PhD Jeffrey D. Munk Van D. Baxter

Member, ASHRAE ASHRAE Fellow
Anthony C. Gehl
ABSTRACT
In this paper we present measured performance and efficiency metrics of Ground Source Heat Pumps (GSHPs) for space conditioning and for
water heating connected to a horizontal ground heat exchanger (GHX) loop. The units were installed in a 345 m2 (3700 ft2) high-efficiency test
house built with structural insulated panels (SIPs), operated under simulated occupancy conditions, and located in Oak Ridge, Tennessee (USA)
in US Climate Zone 4 . The paper describes distinctive features of the building envelope, ground loop, and equipment, and provides detailed
monthly performance of the GSHP system. Space conditioning needs of the house were satisfied by a nominal 2-ton (7.0 kW) water-to-air
GSHP (WA-GSHP) unit with almost no auxiliary (resistance) heat usage. Recommendations for further improvement through engineering
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
design changes are identified. The comprehensive set of data and analyses demonstrate the feasibility and practicality of GSHPs in residential
applications and their potential to help achieve source energy and greenhouse gas emission reduction targets set under the IECC 2012 Standard.
INTRODUCTION
Buildings account for about 40% of primary energy use in the U.S. (EIA 2009). Greenhouse-gas (GHG) emissions
from the building sector exceed those of both the industrial and transportation sectors in the U.S (Metz, Davidson et al.
2007). Buildings also represent approximately 40% of the European Union’s (EU’s) energy consumption and carbon
dioxide production (Baden, Waide et al. 2006). Clearly, building energy efficiency plays a vital part in addressing the
environmental challenges confronting industrialized and developing nations of the world. The 2012 International Energy
Conservation Code (IECC) offers an opportunity to achieve 50% increase in energy savings in new U.S. buildings by 2015
(DOE 2010). To achieve these objectives, data from buildings with cost-effective novel designs, and efficient methods to
utilize renewable energy resources are needed to elevate existing technology to higher levels of performance.
Although Ground Source Heat Pumps (GSHPs) are recognized (IGSHPA 2010) as the most efficient technology to

1RWLFH7KLVPDQXVFULSWKDVEHHQDXWKRUHGE\87%DWWHOOH//&XQGHU&RQWUDFW1R'($&25ZLWKWKH86'HSDUWPHQWRI
(QHUJ\7KH8QLWHG6WDWHV*RYHUQPHQWUHWDLQVDQGWKHSXEOLVKHUE\DFFHSWLQJWKHDUWLFOHIRUSXEOLFDWLRQDFNQRZOHGJHVWKDWWKH8QLWHG
6WDWHV*RYHUQPHQWUHWDLQVDQRQH[FOXVLYHSDLGXSLUUHYRFDEOHZRUOGZLGHOLFHQVHWRSXEOLVKRUUHSURGXFHWKHSXEOLVKHGIRUPRIWKLV
PDQXVFULSWRUDOORZRWKHUVWRGRVRIRU8QLWHG6WDWHV*RYHUQPHQWSXUSRVHV

0RRQLV5$OO\-HIIUH\'0XQN9DQ'%D[WHUDQG$QWKRQ\&*HKODUHUHVHDUFKHUVDWWKH2DN5LGJH1DWLRQDO/DERUDWRU\2DN5LGJH
7HQQHVVHH
$6+5$(

harvest thermal energy from the ground during winter, or to use the ground as a thermal sink during the summer, market
penetration has lagged its technical potential (Ally 2006). Thus, the goal of this 11-month (January-November 2010) field
test was to measure the performance of two GSHPs; one for space conditioning, and the other exclusively for water heating
in a low-energy residential test house built with structural insulated panel (SIP) technology. Since ground loops are a major
cost component of GSHPs, we utilized the house foundation and utility trenches for as much of the ground loop as
possible to reduce the amount of additional trenching required. Quantifying efficiency and energy consumption under
simulated occupancy conditions as per Building America Benchmark (Hendron and Engebrecht 2010) provides useful data
to assess the feasibility and practicality of GSHPs in residential applications and their potential to achieve source energy and
greenhouse gases reduction targets set under the IECC 2012 Standard.
Since subsurface ground temperatures are generally warmer in winter and cooler in summer than the ambient air,
GSHPs should theoretically operate more efficiently than air source heat pumps (ASHPs), especially in colder climates.
Although deep vertical bores offer the advantage of a fairly constant ground temperature, the same is not true of shallow
horizontal trenches where the sub-surface ground temperature varies with the surface air temperature. However, shallow
horizontal trenches may be easier to install and are immune from the vagaries of karstic geology experienced in vertical
wells. Our field study focuses on the performance achieved with a shallow horizontal ground loop utilizing building utility
trenches, and high efficiency water-source heat pumps currently on the market.
Building Envelope, Ground loop, and GSHP systems
The test house is a two-story building with a walk-out basement and a total conditioned space of 345 m2 (3700 ft2), a
SIP-based thermal envelope design, and weather resistive barriers to limit moisture infiltration. The house roof is a
standing seam metal with an infrared reflective (IRR) paint yielding a solar reflectance of 0.30 and a thermal emittance of
0.85. The roof assembly has an overall thermal resistance of about RSI-8.8 (RUS-50). The walls are 0.14 m (5.5-in.) thick
and have a thermal resistance of RSI-3.7 (RUS-21). Windows are triple pane with a U-factor of 1.64 W/m2·K (0.29
Btu/h·ft2·oF) and a solar heat gain coefficient (SHGC) of 0.25. One feature of SIP structures is that they can be built to
have low air leakage rates - blower door tests yielded 0.74 air changes per hour (ACH) at 50 Pa for the house. The building
has a Home Energy Rating System (HERS) Index of 44. In comparison, an EPA Energy Star home would have a HERS
score of 85, and a Net Zero Energy Home would have a HERS Index of 0 (where lower HERS Index values indicate better
energy efficiency). The thermostat set point temperatures are 21.7°C (71°F) for heating and 24.4°C (76°F) for cooling and
were maintained within 1°C (1.8°F) in each of four zones (top floor, master bedroom, living space, and basement). The
monthly relative humidity (RH) (averaged for all 4 zones) was maintained around 47% with the lowest and highest values of
36% and 59% in January and July, respectively. The temperatures and RH are well within comfort levels as defined by
(ASHRAE 2001).
The ground heat exchanger (GHX) loop consisted of three parallel circuits (6 pipes, 1.9 cm (3/4 in.) ID) made of high
density polyethylene (HDPE) with a total length of 559 m (1834 ft). A mixture of 20% polypropylene glycol and 80%
water, by volume was used as the GHX loop fluid (brine). A nominal 2-ton (7 kW), two-capacity water-to-air heat pump
was used for the space conditioning system. It had a rated high-stage heating COP of 4.0 and a rated high-stage cooling
COP of 5.4 (ground loop rating conditions per ANSI/AHRI/ISO 13256-1). A separate 1.5-ton (5.3 kW) water-to-water
GSHP (WW-GSHP) connected to the same GHX loop serviced a 303 L (80-gallon) water tank providing approximately
220 L/day (~58 gallons/day) of hot water at 49°C (120°F).
Equipment, Sensors and Data Acquisition Hardware

Compressor discharge and suction pressure transducer accuracy were ±0.25% FS. Brine flow through the space
conditioning heat pump was determined by measuring pressure drop across the brine/water-refrigerant HX (with a
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`
20
ASHRAE Transactions
differential pressure transducer with ±0.25% FS accuracy) and a detailed 'P vs. flow factory calibration. Brine and domestic
hot water flow through the water heating heat pump were measured with flowmeters having ±1.5% accuracy full-scale and
±0.5% repeatability. Compressor, fan+controls/pump, and total heat pump power were measured with watt/watt-hour
transducers with split-core current transformers (CTs) having a minimum accuracy of ±0.45% of reading plus 0.05% of FS.
Pressure drops in connecting lines were estimated at 1-1.5% of the upstream pressure, based on manufacturer’s data. The
pressure drop across the expansion valve included the pressure drop in the connecting lines between the indoor and
refrigerant-to-water/brine heat exchanger coils.
A total of 20 thermistors, 4 refrigerant line pressure transducers, 2 differential pressure transducers measuring brine
flow (not shown), 2 flow meters and 8 watt transducers were used for the measurements on the two heat pumps. A data
logger was programmed to collect data every 15 seconds for the WW-GSHP and every 30 seconds for the GSHP. The data
was then averaged over a 15 minute interval, and dispatched to a server via an internet connection for subsequent analysis.
A schematic of the GSHP indicating state points is shown in Figure 1. Data are reported for 11 continuous months from
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
January-November, 2010. The time-averaged values of key state variables are tabulated for each month, and several
measures of performance are reported to highlight areas of potential efficiency improvement.
Figure 1. Schematic of GSHP and WW-GSHP for space conditioning and for water heating connected to a common ground loop with
sensor placement.
Some practical trade-offs were necessary in taking our measurements. Although in situ measurements of refrigerant
temperatures would have a quicker response time and be more accurate, the potential for leaks would not be worth taking.
Thus surface-mounted thermistors with conductive paste to enhance contact were used for refrigerant temperatures. But
pressure measurements were made with in situ pressure transducers. Steady-state conditions were assumed in our mass and
energy balances and this assumption is valid when run times exceed five minutes. State points used in this paper refer to the
refrigerant side only. The expansion valve was assumed to be perfectly adiabatic. Pressure drops across each segment of the
GSHPs were according to the manufacturer-supplied data, except for the compressor whose suction and discharge side
pressures were directly measured. State points were calculated using the National Institute of Standards and Technology
(NIST) Reference Fluid Thermodynamic and Transport Properties Database (REFPROP) Version 8.0 (NIST)
Theoretical Analysis
$6+5$(

For purposes of our analyses, the heat pump cycle is assumed to operate under steady-state, steady flow conditions
(SSSF) with negligible kinetic and potential energy effects. The steady state, steady flow assumption holds true except
during compressor start-up or shut down or when cycling times are less than 5 minutes. The expansion valve is assumed to
be adiabatic; however the compressor is not. Its heat loss is assumed to be with the surrounding temperature in the
basement set at 24.4°C (76°F) during the summer and 21.7°C (71°F) during the winter by the thermostat whose accuracy is
±1°C (1.8°F).
Space Conditioning with WA-GSHP

The general energy balances for the WA-GSHP (space conditioning equipment) for any given month is obtained by
integrating the SSSF equation (Wylen, Sonntag et al. 1994; Seader, Henley et al. 2011) over the duration of the respective
month to yield
¦ (Q W ) ¦ (Q W )
i o
0 (1)
In component form, Eq (1) applied to the space conditioning GSHP becomes
[QBrineHX WComp. W fanControl WBrinePump We, Aux. ] [QIndoor HX QComp. Qe, Aux. W fancontrols ] 0 (2)
The Heating Coefficient of Performance is given by
COPh
Q
Brine HX WComp. W fan controls WBrine Pump We, Aux.
(3)
(WComp. W fan controls WBrine Pump We, Aux. )
The Cooling Coefficient of Performance is given by
(QIndoor HX W fan controls QComp. )
COPc (4)
(WComp W fan controls WBrine Pump )
The monthly energy balances for space conditioning are shown in Tables 1 through 3. The accuracy of accounting of
energy inflows to energy outflows is characterized by Δ% (Table 3).
Table 1. Measured Energy Input to WA-GSHP

Months WComp. Wfan+controls WBrine Pump QBrine HX We, Aux. Energy In, Ei
(yr. 2010) (kWh) (kWh) (kWh) (kWh) (kWh) (kWh)
Jan 647.7 102.2 95.14 2200.2 9.8 3055.0
Feb 633.9 102.0 88.04 2026.8 0.0 2850.8
Mar 431.7 72.4 61.27 1436.8 0.0 2002.2
Apr (heat) 47.1 7.6 6.653 197.2 0.0 258.6
Apr (cool) 24.3 6.4 5.587 -248.3 0.0 -212.0
May (heat) 0.353 0.059 0.053 1.7 0.0 2.2
May (cool) 111.3 24.4 22.65 -988.4 0.0 -830.1
Jun 282.0 49.1 47.14 -2047.5 0.0 -1669.3
Jul 413.4 63.6 62.366 -2647.9 0.0 -2108.6
Aug 504.2 67.5 70.154 -2955.5 0.0 -2313.6
Sept 301.3 43.0 43.437 -1844.5 0.0 -1456.7
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Oct (heat) 2.223 0.4 0.306 11.6 0.0 14.5

Oct (cool) 34.9 6.1 5.896 -252.0 0.0 -205.1
Nov 99.0 11.6 14.564 418.6 0.0 543.7
Comparison of the last column in Tables 1 and 2 is a measure of the degree to which the energy balance Eq. (2) holds
true. The minor discrepancy in the energy balance, due to errors in measurement, the heating and cooling COPs and the
22
ASHRAE Transactions
entering water temperatures averaged for the month, are all summarized in Table (3). Details of the WA-GSHP are
discussed in (Ally, Munk et al. 2012).
Table 2. Measured Energy Output of WA-GSHP

Months Q Indoor Wfan+controls QComp. Qe, Aux. Energy Out, Eo
(Yr. 2010) HX (kWh) (kWh) (kWh) (kWh)
(kWh)
Jan -2685.9 -102.2 -105.8 -9.78 -2903.7
Feb -2501.5 -102.0 -101.6 -0.009 -2705.1
Mar -1731.0 -72.4 -102.7 -0.005 -1906.0
Apr (heat) -227.6 -7.6 -14.1 0.0 -249.3
Apr (cool) 228.8 -6.4 -4.8 0.0 217.6
May (heat) -1.9 -0.1 -0.1 0.0 -2.1
May(cool) 898.7 -24.4 -21.2 0.0 853.2
Jun 1816.2 -49.1 -46.6 0.0 1720.5
Jul 2307.3 -63.6 -65.8 0.0 2177.9
Aug 2531.8 -67.5 -73.1 0.0 2391.2
Sept 1594.1 -43.0 -47.0 0.0 1504.0
Oct (heat) -13.2 -0.4 -0.5 0.0 -14.1
Oct (cool) 224.8 -6.1 -7.5 0.0 211.2
Nov -486.3 -11.6 -17.1 0.0 -515.0
Table 3. Satisfaction of Energy Balance Equation and

Performance Metrics of Space Conditioning WA-GSHP
Months EWT Energy Energy Δ%=(Ein- COPh COPc
(Yr. 2010) (°C/ °F) In Out, Eo)100/
Ei Eo Ei
(kWh) (kWh)
Jan 4.6/40.3 3055.0 -2903.7 5.0 3.57
Feb 2.8/37 2850.8 -2705.1 5.1 3.46
Mar 3.7/38.7 2002.2 -1906.0 4.8 3.54
Apr (heat) 10.1/50.2 258.6 -249.3 3.6 4.21
Apr (cool) 11.4/52.6 -212.0 217.6 -2.6 5.99
May (heat) 17.6/63.7 2.2 -2.1 3.3 4.73
May (cool) 14.7/58.4 -830.1 853.2 -2.8 5.39
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Jun 24.3/75.8 -1669.3 1720.5 -3.1 4.55
Jul 28.8/83.8 -2108.6 2177.9 -3.3 4.04
Aug 31.7/89.0 -2313.6 2391.2 -3.4 3.73
Sept 30.1/86.2 -1456.7 1504.0 -3.2 3.88
Oct (heat) 19.7/67.5 14.5 -14.1 2.9 5.02
Oct (cool) 24.7/76.5 -205.1 211.2 -3.0 4.51
Nov 16.1/60.9 543.7 -515.0 5.3 4.34
Water Heating with WW-GSHP

The same ground loop that services the space conditioning GSHP also serves the WW-GSHP for water heating. The
general energy balance equation Eq. (1) also applies to the WW-GSHP. In component form, the application of Eq (1) to
the WW-GSHP yields,
QBrineHX WBrinePump WDHW Pump WComp. (QHW Tank QComp. ) 0 (5)
$6+5$(

The COP for the WW-GSHP is
QHW Tank
COPWW GSHP (6)
(WBrine Pump WDHW Pump WComp. )
The monthly energy balances for water heating are shown in Tables 4 and 5. The accuracy of accounting of energy
inflows to energy outflows is characterized by Δ% (Table 6).
Table 4. Measured Energy Input to WW-GSHP

Month WBrine Pump WDHW Pump QBrine HX WComp. Energy
(Yr. 2010) (kWh) (kWh) (kWh) (kWh) Input,
Ei
(kWh)
Jan 12.2 2.9 289.0 132.9 437.0
Feb 10.2 2.5 254.8 114.4 381.9
Mar 11.0 2.6 282.2 122.9 418.7
April 7.6 1.8 235.1 89.2 333.8
May 7.0 1.7 250.0 87.3 345.9
Jun 5.8 1.5 244.1 77.3 328.7
Jul 4.6 1.3 216.6 67.5 289.9
Aug 5.9 2.9 264.2 84.9 357.9
Sept 5.9 2.9 292.1 85.3 386.2
Oct 7.3 3.2 306.6 94.8 412.0
Nov 8.0 3.3 291.3 95.6 398.1
Table 5. Measured Energy Output from WW-GSHP

Month QHW Tank QComp. Energy
(Yr. 2010) (kWh) (kWh) Output,
Eo
(kWh)
Jan -374.0 -37.9 -411.9
Feb -335.8 -33.3 -369.1
Mar -361.1 -44.0 -405.0
April -288.4 -36.0 -324.3
May -298.8 -38.4 -337.2
Jun -286.3 -35.1 -321.4
Jul -251.2 -32.8 -284.0
Aug -307.4 -36.9 -344.3
Sept -335.4 -40.4 -375.8
Oct -356.4 -42.2 -398.6
Nov -344.5 -40.3 -384.8
24
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---

ASHRAE Transactions
Table 6. Energy Balance and Performance Metrics of WW-
GSHP
Months EWT Energy Energy Out, Δ%=(Ein- COPWW-
(Yr. 2010) (°C/ °F) In, Eo Eout)100 GSHP
Ei (kWh) /Ei
(kWh)
Jan 4.5/40.2 437.0 -411.9 5.7 2.5
Feb 2.7/36.9 381.9 -369.1 3.3 2.6
Mar 4.0/39.2 418.7 -405.0 3.3 2.6
Apr 10.651.0 333.8 -324.3 2.8 2.9
May 16.6/61.8 345.9 -337.2 2.5 3.1
Jun 23.2/73.8 328.7 -321.4 2.2 3.4
Jul 28.1/82.5 289.9 -284.0 2.0 3.4
Aug 30.9/87.6 357.9 -344.3 3.8 3.3
Sept 28.9/84.0 386.2 -375.8 2.7 3.6
Oct 23.1/73.6 412.0 -398.6 3.3 3.4
Nov 17.3/63.1 398.1 -384.8 3.3 3.2
Comparison of the last columns in Tables 4 and 5 is a measure of the degree to which the energy balance for the WW-
GSHP, Eq. (5) holds true. The minor discrepancy in the energy balance, due to errors in measurement, the coefficient of
performance of the WW-GSHP and the entering water temperatures averaged for the month, are all summarized in Table
(6).
CONCLUSION
The highest and lowest monthly average space heating COP achieved was 5.02 (October) and 3.46 (February),
respectively. The highest and lowest monthly average space cooling COP achieved was 5.99 (April) and 3.73 (August). The
highest and lowest monthly average water heating COP was 3.6 (September) and 2.5 (January).COPs were calculated on the
basis of total power input (including duct, pumps, and control board power consumptions). Building insulation and air-
tightness significantly reduce loads. A judicious and cost-effective step towards energy conservation is to increase the level
of insulation to reduce thermal loads. In this study, winter auxiliary heating was virtually eliminated allowing a more
efficient use of electricity than would otherwise have been the case.. Brine pumps should be sized such that it operates in
the optimum range for a majority of the time, thus reducing its energy input and increasing the COP. Variable speed drive
technology in which the compressor speed can be modulated, further improves compressor performance. Ground loops
with phase change materials(Ally, Tomlinson et al. 2010) are another novelty for reducing the size of the loop and
improving its heat transfer characteristics.
ACKNOWLEDGMENTS
The authors acknowledge sponsorship of this work by the Tennessee Valley Authority and by the DOE Building
Technologies Program under contract DE-AC05-00OR22725 with UT-Battelle, LLC.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
NOMENCLATURE
' = per cent difference

COP = coefficient of performance
$6+5$(

EWT = entering water temperature
Q = heat transfer
RUS = thermal insulation in English units
RSI = thermal insulation in SI units
T = temperature
W = work
Subscripts
Aux. = auxiliary heat, same as resistance heat
Brine HX = Brine heat exchanger
c = cooling cycle
Comp. = compressor
DHW = domestic hot water
e =electrical
HW = hot water
h = heating
i = input
Indoor HX = indoor heat exchanger
o = output
WW-GSHP = water-to-water ground source heat pump
REFERENCES
Ally, M. R. (2006). Ground Source Heat Pumps in the USA. International Energy Agency (IEA) Heat Pump Meeting: Global
Advances in Heat Pump Technology, Applications, and Markets, Linz, Austria.
Ally, M. R., J. D. Munk, et al. (2012). "Exergy Analysis and Operational Efficiency of a Horizontal Ground Source Heat Pump
System Operated in a Low-Energy Test House under Simulated Occupancy Conditions." Int. J. Refrig.(In Press).
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Ally, M. R., J. J. Tomlinson, et al. (2010). Super Energy Saver Heat Pump with Daynamic Hybrid Phase Change Material. U.S.A,
UT-Battelle, LLC, Oak Ridge, TN. Patent No. US 7,757, 508 B2.
ASHRAE (2001). ASHRAE Fundamentals Handbook I-P Edition, Amer. Soc. Heating Refrigeration and Air-conditioning: 8.12.
Baden, S., P. Waide, et al. (2006). Hurdling Financial Barriers to Low Energy Buildings: Experiences from the USA and Europe
on Financial Incentives and Monetizing Building Energy Savings in Private Investment Decisions. Proceedings of 2006
ACEEE Summer Study on Energy Efficiency in Buildings, Washington, D. C.
DOE, U. S. (2010, November 15, 2010). "DOE Announces Historic Strides in Energy Efficiency for Residential and Commercial
Building Codes." from http://apps1.eere.energy.gov/news/progress_alerts.cfm/pa_id=437.
EIA (2009). "Buildings Energy Data Book, Table 1.1."
Hendron, R. and C. Engebrecht (2010). Building America Research Benchmark Definition., National Renewable Energy
Laboratory.
IGSHPA. (2010). "International Ground Source Heat Pump Association." from
http://www.igshpa.okstate.edu/geothermal/faq.htm.
Metz, B., O. R. Davidson, et al. (2007). Summary for policymakers. In Climate change 2007: Mitigation. Contribution of Working
Group III to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change Intergovernmental
Panel on Climate Change.
NIST. "NIST Reference Fluid Thermodynamic and Transport Properties Database (REFPROP): Version 8.0." 2009, from
http://www.nist.gov/srd/nist23.cfm.
Seader, J. D., E. J. Henley, et al. (2011). Separation Process Principles. Chemical and Biochemical Operations, John Wiley & Sons,
Inc.
Wylen, G. V., R. Sonntag, et al. (1994). Fundamentals of Classical Thermodynamics, McGraw-Hill.
26
ASHRAE Transactions
6$&
Measured Impact on Space Conditioning

Energy Use in a Residence Due to
Operating a Heat Pump Water Heater inside
the Conditioned Space
Jeffrey D. Munk Moonis R. Ally, Ph.D. Van D. Baxter

Member ASHRAE Fellow ASHRAE
ABSTRACT
The impact on space conditioning energy use due to operating a heat pump water heater (HPWH) inside the conditioned space is analyzed based
on 2010-2011 data from a research house with simulated occupancy and hot water use controls. The 2700 ft2 ( 251 m2) house is located in
Oak Ridge, TN (mixed-humid climate) and is equipped with a 50 gallon (189 L) HPWH that provided approximately 55 gallons/d (208
L/d) of hot water at 120°F (48.9°C) to the house during the test period. The HPWH has been operated every other week from December
2010 through November 2011 in two modes; a heat pump only mode, and a standard mode that utilizes 15355 Btu/h (4500 W) resistance
heating elements. The energy consumption of the air-source heat pump (ASHP) that provides space conditioning for the house is compared for the
two HPWH operating modes with weather effects taken into account. Impacts during the heating and cooling seasons are compared.
INTRODUCTION
Heat pump water heaters (HPWHs) with energy factors (EFs) from 2.0-2.5 provide a significant performance increase
over standard electric water heaters with EFs around 0.9 (AHRI 2011). The HPWHs achieve these performance levels by
using a vapor compression cycle to transfer heat from the surrounding air to the water that is being heated. If the HPWH
unit is installed inside the conditioned space, it will cool the air and increase the space heating load in the winter and
decrease the space cooling load in the summer. The net impact of this effect has been evaluated in a research home that
has been operated for a year (December 2010 – November 2011) under simulated occupancy. The house has a clothes
washer, dishwasher, and shower that have been operated on a schedule to draw between 50 and 60 gallons/d (189-227
L/d). The shower was adjusted throughout the year in order to maintain an average temperature of 105°F (40.5°C) at the
shower head. Adjustments were necessary due to seasonal changes in the temperature of the cold water supply, which
varied between 46.8°F (8.2°C) and 73.7°F (23.2°C) as seen in Figure 1. The HPWH was installed in a 3 ft by 10 ft (0.9 m
by 3.0 m) utility closet with two louvered doors connecting to the laundry room. It was set to provide 120°F (48.9°C) water
and was switched between standard electric mode and heat pump only mode every other week. Space conditioning was
All authors are research staff at Oak Ridge National Laboratory in Oak Ridge, TN
1RWLFH7KLVPDQXVFULSWKDVEHHQDXWKRUHGE\87%DWWHOOH//&XQGHU&RQWUDFW1R'($&25ZLWKWKH86'HSDUWPHQWRI
(QHUJ\7KH8QLWHG6WDWHV*RYHUQPHQWUHWDLQVDQGWKHSXEOLVKHUE\DFFHSWLQJWKHDUWLFOHIRUSXEOLFDWLRQDFNQRZOHGJHVWKDWWKH8QLWHG
6WDWHV*RYHUQPHQWUHWDLQVDQRQH[FOXVLYHSDLGXSLUUHYRFDEOHZRUOGZLGHOLFHQVHWRSXEOLVKRUUHSURGXFHWKHSXEOLVKHGIRUPRIWKLV
PDQXVFULSWRUDOORZRWKHUVWRGRVRIRU8QLWHG6WDWHV*RYHUQPHQWSXUSRVHV
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

provided by an air-source heat pump (ASHP) that maintained a temperature of 71°F (21.7°C) in the heating season and
76°F (24.4°C) in the cooling season. The fan coil of the ASHP was installed in the same utility closet as the HPWH.
In the following sections, two different approaches are used to estimate the HPWH impact on space conditioning
energy use for the test period, both for heating and cooling seasons and the net annual impact. First an energy balance is
performed on the water heater (WH) to determine the net impact (increase or decrease) to the space conditioning load on
the ASHP. These load impacts are then used with the average measured ASHP seasonal performance factors to estimate
the impact on HVAC energy use. Second, the measured energy use of the ASHP is analyzed directly to determine the net
change in its energy use due to HPWH operation.
Cold Water Supply Temperature

75
70
65
Temperature (°F)
60
55
50
45
Figure 1: Cold water supply temperature
ENERGY BALANCE ON WATER HEATER

Using a simple energy balance on the WH itself, shown in Figure 2 and Equation 1, the net heat transferred between
the water heater and its surroundings can be calculated for both the HP and standard electric modes. A positive value for
Q indicates heat transfer into the WH, and a negative value indicates heat transfer out of the WH.
ܳ ൌ ݉௪ ܿ௣ǡ௪ ሺܶ௢௨௧ െ ܶ௜௡ ሻ െ ܹ௘ (1)
28 ASHRAE Transactions
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---

݉௪ ǡ ܶ௢௨௧ ݉௪ ǡ ܶ௜௡
‫ݑݐܤ‬
ܿ‫݌‬ǡ‫ ݓ‬ൌ ͳǤͲͲ
݈ܾ ή ܴ
ܳ ܹ௘
Figure 2: Water heater energy balance assuming mass entering the tank is equal to mass leaving the tank
Table 1 provides average daily heat transfer to/from the WH, WH energy use, and hot water use for each month of
the test period. Annual as well as heating and cooling season averages are also given. In the heating season, the additional
space heating load due to HPWH operation (difference between the average WH heat transfer in the HP mode and the
standard mode) is 6685 W∙h/d. For the cooling season the average daily net space cooling provided by the HPWH is 5302
W∙h/d.
Table 1. Heat Transfer from Surroundings to Water Heater
HPWH Mode Standard Mode
Heat Transfer Energy Use Water Use Heat Transfer Energy Use Water Use
Month kBtu/d kW∙he/d gal/d (L/d) kBtu/d kW∙he/d gal/d (L/d)
(kW∙ht/d) (kW∙ht/d)
Dec 2010 18.42 (5.40) 3.14 53.7 (203.3) -3.61 (-1.06) 9.65 54.8 (207.4)
Jan 2011 20.81 6.10 3.47 56.2 (212.7) -4.01 (-1.17) 10.59 55.9 (211.6)
Feb 2011 20.70 (6.07) 3.51 57.0 (215.8) -3.86 (-1.13) 10.74 58.1 (220.0)
March 2011 20.74 (6.08) 3.53 60.8 (230.2) -3.71 (-1.09) 10.20 58.2 (220.3)
April 2011 18.85 (5.53) 3.37 59.6 (225.6) -3.78 (-1.11) 9.49 59.0 (223.3)
May 2011 15.77 (4.62) 2.93 57.1 (216.1) -4.01 (-1.18) 8.84 59.2 (224.1)
June 2011 13.52 (3.96) 2.64 56.3 (213.1) -4.57 (-1.34) 7.90 56.8 (215.0)
July 2011 12.18 (3.57) 2.47 52.5 (198.7) -4.36 (-1.28) 7.20 52.6 (199.1)
Aug 2011 12.13 (3.55) 2.53 53.8 (203.7) -4.49 (-1.32) 7.42 54.0 (204.4)
Sept 2011 13.15 (3.86) 2.48 53.8 (203.7) -4.43 (-1.30) 8.15 54.1 (204.8)
Oct 2011 14.13 (4.14) 2.74 54.2 (205.2) -4.11 (-1.21) 7.80 51.9 (196.5)
Nov 2011 15.85 (4.65) 2.99 53.5 (202.5) -4.29 (-1.26) 8.76 53.1 (201.0)
Heating
Season 18.91 (5.54) 3.29 56.4 (213.5) -3.90 (-1.14) 9.79 56.0 (212.0)
Average
Cooling
Season 13.79 (4.04) 2.67 55.0 (208.2) -4.31 (-1.26) 7.99 55.2 (209.0)
Average
Annual
Average 16.33 (4.79) 2.98 55.7 (210.8) -4.10 (-1.20) 8.88 55.6 (210.5)
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

These values can be used in conjunction with the measured heating and cooling season average coefficient of
performance (COP) values for the ASHP to determine the average impact on space conditioning energy use as seen in
Equations 2 and 3. In the heating season, the average COP for the ASHP was 3.11, which means that in order to provide
the additional heat that is used by the HPWH, the ASHP must use an additional 2.15 kW∙h/d. In the cooling season, the
average COP of the ASHP was 4.54, which would result in 1.17 kW∙h/d less space conditioning energy use due to the
HPWH operation in the HP mode.
൫ொಹ೐ೌ೟ǡಹುೈಹ ିொಹ೐ೌ೟ǡೄ೟ೌ೙೏ೌೝ೏ ൯
οܹ௘ǡு௘௔௧ǡௌ஼ ൌ ஼ை௉ಹ೐ೌ೟
(2)
൫ொ಴೚೚೗ǡೄ೟ೌ೙೏ೌೝ೏ ିொ಴೚೚೗ǡಹುೈಹ ൯
οܹ௘ǡ஼௢௢௟ǡௌ஼ ൌ ஼ை௉಴೚೚೗
(3)
Over the course of the year, this averages out to an additional 0.48 kW∙h/d of space conditioning energy use in this
house at the test location as shown in Equation 4. For added perspective, the average annual ASHP energy for this study
was 12.48 kW∙h/d. While these numbers will vary based on the performance of each home’s space conditioning equipment
and water use, it will still typically be nearly an order of one magnitude less than the associated water heating energy use
savings between a HPWH and a standard electric water heater, which in this study was 5.90 kW∙h/d.
൫οௐ೐ǡಹ೐ೌ೟ǡೄ಴ ௗಹ೐ೌ೟ ାοௐ೐ǡ಴೚೚೗ǡೄ಴ ௗ಴೚೚೗ ൯

οܹ௘ǡ௔௡௡௨௔௟ǡௌ஼ ൌ ሺଷ଺ହሻ
(4)
MEASURED SPACE CONDITIONING ENERGY USE

As noted earlier, the actual ASHP energy use was recorded throughout the test period as well. This data was
summed over each day and tabulated along with the average outdoor air temperature and the operating mode of the
HPWH. Prior to the analysis, the data was filtered by removing days when there was no energy use by the ASHP for space
conditioning, as well as any days when there were known issues with the data. Days in which the operating mode of the
HPWH was switched were also removed.
Heating Season Data

The heating season data covered the periods of 12/1/2010 to 4/20/2010 and 10/20/1011 to 11/30/2011 and is
plotted in Figure 3. Also shown in Figure 2 are curve fits for the space heating energy use when the WH operated either in
heat pump mode or standard mode, as a function of daily outdoor air temperature.
As seen in Figure 3, there is significant variance in the ASHP energy use that is not accounted for solely by the average
outdoor air temperature. This variation appears to be much larger than the difference in energy use indicated by operating
the HPWH in the HP mode. Applying the curve fits over the entire heating season suggests that on average the ASHP in
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
this test house required 1.00 kW∙h/d more energy with the HPWH operating in the HP mode when compared to the
standard mode. For outdoor temperatures below about 35°F (1.7°C) the additional space heating energy use exceeded 2.00
kW∙h/d.
30
ASHRAE Transactions
Figure 3: Daily space heating energy use vs. daily average outdoor air temperature
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Cooling Season Data
The cooling season data, 4/21/2011 to 10/19/2011, was evaluated in the same fashion as the heating season data.
Unfortunately, there was a zone damper in the central air distribution system that was malfunctioning during a large portion
of test period. This caused higher than expected space cooling energy use and this data was removed before analysis.
Figure 4 shows the resulting space cooling energy data which was available for the cooling season analysis. The curve fits
shown in Figure 3 indicate that the ASHP used less energy when the WH was operating in the HPWH mode when
compared to the standard mode at outdoor air temperatures above about 77°F (25°C). Applying these curve fits over the
entire cooling season suggests that the ASHP in this house required about 0.21 kW∙h/d less energy on average with the
HPWH operating in the conditioned space.
$6+5$(

Figure 4: Daily space cooling energy use vs. daily average outdoor air temperature
Annual HVAC Energy Impact

Applying the heating and cooling seasonal curve fits to the entire 2010-2011 test period, the annual net impact on
HVAC energy use of the HPWH operating indoors was estimated to be 0.39 kW∙h/d. This compares reasonably well to
the 0.48 kW∙h/d impact estimated via the energy balance on the WH. In both cases the HVAC energy use impact is
substantially less than the WH energy savings obtained using the HPWH (5.90 kW∙h/d in this study).
CONCLUSIONS
Both methods of analysis indicated a small net increase in space conditioning energy use over an entire year due to
operating a HPWH within the conditioned space for the test house in its mixed-humid climate location. The energy
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
balance analysis estimated an additional 0.48 kW∙h/d over the entire year, while the measured ASHP data estimated an
additional 0.39 kW∙h/d. While the energy balance indicated a larger magnitude effect in both the heating and cooling
seasons, when averaged over an entire year, the results were reasonably close to those obtained via analysis of the HVAC
system measured data. Both methods indicate that the impact on space conditioning energy use of operating a HPWH
inside the conditioned space of a residence in a mixed-humid climate is quite small in comparison with the water heating
efficiency gains achieved by HPWHs (~0.4-0.5 kWh-d vs. 5.90 kW∙h/d in this case).
32
ASHRAE Transactions
This study has presented the results of a single case study in one location – Oak Ridge, TN. HVAC energy use
impacts due to indoor location of a HPWH will vary for other homes based on climactic location, hot water use pattern,
entering cold water temperature, HPWH efficiency, HVAC system efficiency, etc. However, in many cases, additional
HVAC energy use will be substantially less than the associated water heating energy use savings between a HPWH and a
standard electric water heater.
ACKNOWLEDGMENTS
The authors acknowledge sponsorship of this work by the Tennessee Valley Authority and by the DOE Building
Technologies Program under contract DE-AC05-00OR22725 with UT-Battelle, LLC.
NOMENCLATURE
c = specific heat capacity
d = day
݉ = mass
Q = heat transfer
7 = temperature
W = work
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Subscripts
annual = annual average
e = electric
in = in
out = out
p = constant pressure
t = thermal
w = water
Cool = cooling season
Heat = heating season
SC = space conditioning
REFERENCES
Moran, M. J., H. N. Shapiro. 2000. Fundamentals of Engineering Thermodynamics, John Wiley & Sons, Inc.
Air-Conditioning, Heating, and Refrigeration Institute. 2011. “Directory of Certified Product Performance”, from
http://www.ahridirectory.org/ahriDirectory/pages/home.aspx, AHRI
$6+5$(

SA-12-C005
Thermal Storage Properties of a Hybrid

Ground Source Heat Pump
Amanda Pertzborn Greg Nellis, PhD Sanford Klein, PhD

Student Member ASHRAE Member ASHRAE Fellow ASHRAE
ABSTRACT
+\EULGJURXQGVRXUFHKHDWSXPS+\*6+3V\VWHPVDUHJDLQLQJSRSXODULW\DVDPHDQVRIGHFUHDVLQJORQJWHUPHQHUJ\DQGPDLQWHQDQFH
FRVWVZKLOHPDLQWDLQLQJPDQDJHDEOHILUVWFRVWV,QWKHVHV\VWHPVWKHUHLVDSRWHQWLDOWRXVHWKHJURXQGQRWRQO\DVDQLPPHGLDWHKHDWVRXUFHRUVLQN
EXWDOVRDVDQHQHUJ\VWRUH2QDVFDOHRIPRQWKVKHDWUHMHFWHGWRWKHJURXQGGXULQJWKHFRROLQJVHDVRQFRXOGSRWHQWLDOO\EHUHFRYHUHGGXULQJWKH
KHDWLQJVHDVRQ2QDVFDOHRIKRXUVWKHJURXQGVWRUDJHFRXOGEHFRROHGE\DK\EULGFRPSRQHQWVXFKDVDFRROLQJWRZHUGXULQJWKHQLJKWLQRUGHUWR
LPSURYHWKHKHDWUHMHFWLRQHIILFLHQF\LQFRROLQJGXULQJWKHIROORZLQJGD\7KLVSDSHUZLOOSUHVHQWWKHUHVXOWVRIDGHWDLOHGVWXG\RIWKHJURXQGDVD
WKHUPDOHQHUJ\VWRUDJHPHGLXPRQDVKRUWWHUPEDVLVLQDFRROLQJGRPLQDWHGHQYLURQPHQWDOVRUHIHUUHGWRDVQLJKWSUHFRROLQJ$VLPSOLILHGPRGHO
RIDERUHILHOGLVGHYHORSHGWRFDOFXODWHVWRUDJHHIILFLHQF\DQGWKHQDPRUHGHWDLOHGPRGHOLVXVHGWRHYDOXDWHWKHSRZHUFRQVXPSWLRQDQGRSHUDWLRQDO
FRVWRIDV\VWHPXVLQJWKHJURXQGIRUGLXUQDOWKHUPDOVWRUDJH
INTRODUCTION
,QD+\EULG*URXQG6RXUFH+HDW3XPS+\*6+3WKHUHLVWKHSRWHQWLDOWRXVHWKHJURXQGKHDWH[FKDQJHU*+;
QRWRQO\DVDKHDWVRXUFHRUVLQNEXWDOVRIRUWKHUPDOVWRUDJHRQDVHDVRQDORUGLXUQDOEDVLVRUDFRPELQDWLRQRIWKHWZR,Q
WKHFDVHRIVHDVRQDOWKHUPDOVWRUDJHIRUH[DPSOHHQHUJ\IURPDVRODUWKHUPDOSODQWFRXOGEHVWRUHGLQWKHJURXQGGXULQJ
WKHVXPPHUDQGWKHQUHPRYHGIURPWKHJURXQGGXULQJWKHZLQWHUWRPHHWVRPHRIWKHZLQWHUKHDWLQJGHPDQG0F'RZHOO
7KRUQWRQ3DKXG:RQJ6QLMGHUV 0F&OXQJ$QH[DPSOHRIGLXUQDOWKHUPDOVWRUDJHLVRSHUDWLQJD
FRROLQJWRZHUDWQLJKW²HQHUJ\UDWHVDUHORZHURSHUDWLRQLVJHQHUDOO\PRUHHIILFLHQWGXHWRORZHUDPELHQWWHPSHUDWXUHDQG
FRROLQJ ORDGV DUH ORZ ² LQ RUGHU WR SUHFRRO WKH JURXQG IRU WKH IROORZLQJ GD\ 7KH FRROHU JURXQG VWRUHUHVXOWV LQ ORZHU
HQWHULQJZDWHUWHPSHUDWXUH(:7WRWKHKHDWSXPSUHVXOWLQJLQORZHUKHDWSXPSSRZHUFRQVXPSWLRQ)DQ-LDQJ<DR
0D-LQJJDQJ;LDR[LD=KHQMLDQJ )DQJ3DUGR0RQWHUR0DUWRV 8UFKXHJXtD<DYX]WXUN 6SLWOHU

7KHUHDUHVHYHUDOSRWHQWLDOEHQHILWVWRXVLQJWKH*+;IRUWKHUPDOVWRUDJH)LUVWWKHVWUDWHJ\PRUHIXOO\XWLOL]HVWKH
*+; ZKLFK ZRXOG RWKHUZLVH EH SUHVHQW EXW XQXVHG GXULQJ WKHUPDO VWRUDJH SHULRGV 6HFRQG WKHUH LV D SRWHQWLDO WR
GHFUHDVHWKHSRZHUFRQVXPSWLRQRIWKHKHDWSXPSVDYLQJERWKHQHUJ\DQGPRQH\7KLUGE\XVLQJWKHUPDOVWRUDJHGXULQJ
RIISHDNSHULRGVWKHUHFDQEHIXUWKHUFRVWVDYLQJV
7KHSRWHQWLDOSUREOHPZLWKWKLVV\VWHPLVWKDWPRUHHQHUJ\PLJKWEHXVHGDQGRUPRQH\VSHQWRSHUDWLQJWKHSXPS
DQGVXSSOHPHQWDOHTXLSPHQWHJFRROLQJWRZHUGXULQJWKHUPDOVWRUDJHWKDQLVVDYHGLQKHDWSXPSRSHUDWLRQ7KHVWXG\
Pertzborn is a PhD candidate in the Department of Mechanical Engineering, University of Wisconsin, Madison, WI. Nellis is a professor in the
Department of Mechanical Engineering, University of Wisconsin, Madison, WI. Klein is a professor in the Department of Mechanical
Engineering, University of Wisconsin, Madison, WI.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
34
ASHRAE Transactions
GHWDLOHGLQWKLVSDSHUZLOOIRFXVRQHYDOXDWLQJWKH*+;DVDWKHUPDOVWRUDJHGHYLFHLQDFRROLQJWRZHUK\EULGRQDGLXUQDO
EDVLV:KHQWKH*+;LVXVHGDVWKHUPDOVWRUDJHWKHSURFHVVLVUHIHUUHGWRDVSUHFRROLQJEHFDXVHWKH*+;LVEHLQJFRROHG
SULRUWREHLQJXVHGIRUFRROLQJDEXLOGLQJ
NUMERICAL MODEL
7KH ILUVW VWHS LQ HYDOXDWLQJ WKH *+; DV D WKHUPDO VWRUH LV WR VLPSOLI\ WKH SUREOHP IURP D IXOO V\VWHP PRGHO WKDW
FRQWDLQVPXOWLSOHERUHKROHVKHDWSXPSVDQGDFRROLQJWRZHUWRDUHODWLYHO\VLPSOHPRGHOWKDWFRQVLVWVRIDVLQJOHERUHKROH
DQGLVHYDOXDWHGRQDSHUXQLWOHQJWKEDVLV8VLQJWKLVPRGHOWKHWKHUPDOVWRUDJHHIILFLHQF\FDQEHFDOFXODWHGLQRUGHUWR
HYDOXDWH KRZ GLIIHUHQW SUHFRROLQJ VWUDWHJLHV SHUIRUP$ UHDO ERUHKROH FRQVLVWV RI WZR SLSHV FRXSOHG WR WKH VXUURXQGLQJ
JURXQG YLD D WKHUPDOO\ FRQGXFWLYH JURXW KRZHYHU LQ WKH VLPSOLILHG PRGHO XVHG LQ WKHVH FDOFXODWLRQV WKHVH GHWDLOV DUH
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
QHJOHFWHG 7KH RXWHU GLDPHWHU RI WKH ERUHKROH LV WKH LQQHU ERXQGDU\ RI WKH V\VWHP 7KH KHDW WUDQVIHU IURP WKLV LQQHU
ERXQGDU\LQWRWKHVXUURXQGLQJJURXQGLVFDOFXODWHGE\WKHPRGHO7KHLQQHUERXQGDU\FRQGLWLRQLVDWLPHYDU\LQJVSHFLILHG
WHPSHUDWXUHZKLOHWKHRXWHUERXQGDU\LVDFRQVWDQWWHPSHUDWXUHWKDWLVHTXDOWRWKHXQGLVWXUEHGJURXQGWHPSHUDWXUH7KH
KHDWWUDQVIHULVPRGHOHGDV'IURPWKHERUHKROHLQWRWKHJURXQG
$ ' QXPHULFDO PRGHO ZDV GHYHORSHG XVLQJ PHWKRGV GHVFULEHG LQ 1HOOLV .OHLQ DQGXVHG WR SUHGLFW WKH
WHPSHUDWXUHSURILOHLQWKHJURXQGDVDIXQFWLRQRIWLPHIRUWKHSUHFRROLQJFDVHVVKRZQLQ)LJXUHZKHUHQLJKWLVEHIRUH
DQGGD\LVDIWHU7KHEDVHOLQHFDVHKDVDΔ&Δ)VSHFLILHGLQQHUERXQGDU\WHPSHUDWXUHGXULQJWKHQLJKW
DQGDVSHFLILHGLQQHUERXQGDU\WHPSHUDWXUHRI&)DERYHWKHXQGLVWXUEHGJURXQGWHPSHUDWXUHWJGXULQJWKHGD\
3UHFRROLQJ&DVH$KDVDVSHFLILHGLQQHUERXQGDU\WHPSHUDWXUHGXULQJWKHQLJKWRI&)EHORZWJDQG&)
DERYHWJGXULQJWKHGD\3UHFRROLQJ&DVH%KDVDVSHFLILHGLQQHUERXQGDU\WHPSHUDWXUHGXULQJWKHILUVWVL[KRXUVRIWKH
QLJKWRI&)EHORZWJΔ&Δ)IRUWKHUHPDLQLQJVL[KRXUVRIWKHQLJKWDQG&)DERYHWJGXULQJWKHGD\
3UHFRROLQJ&DVH&KDVDVSHFLILHGLQQHUERXQGDU\WHPSHUDWXUHGXULQJWKHILUVWVL[KRXUVRIWKHQLJKWRIΔ&Δ)&
)EHORZWJIRUWKHUHPDLQLQJVL[KRXUVRIWKHQLJKWDQG&)DERYHWJGXULQJWKHGD\
)LJXUHVKRZVWKHJURXQGWHPSHUDWXUHWKHLQLWLDOWHPSHUDWXUHLV&RU)SURILOHDIWHURQHGD\RIRSHUDWLRQ
DQGWKHKHDWWUDQVIHUSHUXQLWOHQJWKDWWKHLQQHU ERXQGDU\IRUWKHILUVWKRXUVIRUWKH EDVHOLQHDQGSUHFRROLQJFDVHV
$IWHUGD\WKHWKHUPDOZDYHIRUDOOFDVHVKDVSHQHWUDWHGDSSUR[LPDWHO\WKHVDPHGLVWDQFHLQWRWKHJURXQG+RZHYHUWKH
EDVHOLQHFDVHKDVRQO\DGGHGKHDWWRWKHJURXQGDQGWKHUHIRUHWKHJURXQGWHPSHUDWXUHKDVLQFUHDVHGHYHU\ZKHUH)RUWKH
SUHFRROLQJFDVHVWKHUHLVDGHFUHDVHLQJURXQGWHPSHUDWXUHIURPWKHLQLWLDOFRQGLWLRQEHWZHHQDQGPDQG
IW&DVH$VKRZVWKHJUHDWHVWLQFUHDVHLQKHDWWUDQVIHUWRWKHJURXQGDIWHUSUHFRROLQJIROORZHGE\&DVH&DQG&DVH%
7KHVHILJXUHVLOOXVWUDWHKRZSUHFRROLQJRSHUDWHV
7KH VWRUDJH HIILFLHQF\ LV WKH GLIIHUHQFH EHWZHHQ WKH WRWDO KHDW UHMHFWHG WR WKH JURXQG GXULQJ WKH GD\ IRU WKH SUH
FRROLQJFDVHDQGWKHEDVHOLQHFDVHGLYLGHGE\WKHKHDWUHPRYHGIURPWKHJURXQGGXULQJWKHQLJKWIRUWKHSUHFRROLQJFDVH
7KHVWRUDJHHIILFLHQF\ZDVFDOFXODWHGIRUDORQJGD\PRGHOLQZKLFKWKHYDOXHVRIWKHWLPHSHULRGVτDQGτIURP)LJXUH
DUHODUJHDQGIRUDPRGHORIDVLQJOHGD\UHSHDWHGRYHUDQGRYHULQZKLFKτDQGτDUH,QERWKFDVHVWKHPRGHOZDV
UXQ XQWLO WKH HIILFLHQF\ UHDFKHG DQ DV\PSWRWH )RU WKH ORQJ GD\ PRGHO WKLV UHTXLUHG GD\V DQG IRU WKH UHSHDWHG GD\
PRGHOWKLVUHTXLUHGUHSHDWHGGD\V7KHUHVXOWLQJHIILFLHQFLHVDUHJLYHQLQ7DEOH7KHORQJGD\PRGHOLVLQWHQGHGWR
HYDOXDWHWKHPD[LPXPSRVVLEOHVWRUDJHHIILFLHQF\ZKLFKIRUWKHWKUHHSUHFRROLQJPRGHOVVWXGLHGKHUHLV7KLVPHDQV
WKDWRIWKHHQHUJ\UHPRYHGGXULQJWKHQLJKWFDQODWHUEHUHFRYHUHGLQWKHIRUPRIDGGLWLRQDOWKHUPDOHQHUJ\UHMHFWLRQ
7KHUHPDLQLQJHQHUJ\OHDYHVWKHV\VWHPWKURXJKWKHRXWHUERXQGDU\
7KHVWRUDJHHIILFLHQF\LVVKRZQIRUWKHUHSHDWHGGD\PRGHOLQ)LJXUH,QLWLDOO\WKHHIILFLHQF\LQFUHDVHVUDSLGO\EXW
DIWHUDSSUR[LPDWHO\GD\VWKHLQFUHDVHVORZVGRZQDQGHYHQWXDOO\DSSURDFKHVDQDV\PSWRWH

$6+5$(

ĂƐĞůŝŶĞĂƐĞ WƌĞͲŽŽůŝŶŐĂƐĞ
& &
нϯΣΔ нϯΣΔ
;нϱ͘ϰΣΔ&Ϳ ;нϱ͘ϰΣΔ&Ϳ
τϭсϬ τϭсϬ
θ θ
τϮсϭϮ͗ϬϬ τϯ τϯ
ͲϯΣΔ ͲϯΣΔ
;Ͳϱ͘ϰΣΔ&Ϳ ;Ͳϱ͘ϰΣΔ&Ϳ τϮсϭϮ͗ϬϬ

WƌĞͲŽŽůŝŶŐĂƐĞ WƌĞͲŽŽůŝŶŐĂƐĞ
& &
нϯΣΔ нϯΣΔ
;нϱ͘ϰΣΔ&Ϳ ;нϱ͘ϰΣΔ&Ϳ τ сϬ
τϭсϬ
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
ϭ
θ θ
τϯсϭϮ͗ϬϬ τϯсϭϮ͗ϬϬ τϰ
ͲϯΣΔ τϰ ͲϯΣΔ
;Ͳϱ͘ϰΣΔ&Ϳ τϮсϲ͗ϬϬ ;Ͳϱ͘ϰΣΔ&Ϳ τϮсϲ͗ϬϬ
)LJXUH3URILOHVRIWKHEDVHOLQHDQGSUHFRROLQJIRUFLQJIXQFWLRQV
r (ft)
22.50 0.98 1.97 2.95 3.94 4.92
72.5 150 156
&DVH$
&DVH&
21.8 71.2 100 &DVH% 104
q (Btu/hr-ft)
50 52
q (W/m)
21.1 70.0 &DVH$

t (C)
t (F)

0 0
20.4 68.7 %DVHOLQH
%DVHOLQH
&DVH& -50 -52
19.7 67.5
&DVH% &DVH&
&DVH$ -100 -104
&DVH% 0 0.2 0.4 0.6 0.8 1
19 66.2
0 0.3 0.6 0.9 1.2 1.5 time (day)
r (m)
)LJXUH/HIW&RPSDULVRQRIJURXQGWHPSHUDWXUHIRUWKHGLIIHUHQWPRGHOVDWKRXUV7KHLQLWLDOJURXQGWHPSHUDWXUHLV&5LJKW
&RPSDULVRQRIKHDWWUDQVIHUDWWKHLQQHUERXQGDU\IRUWKHGLIIHUHQWPRGHOVIRUWKHILUVWKRXUV
36
ASHRAE Transactions
Table 1. Storage efficiency for different pre-cooling cases.
Pre-Cooling Case Long Day 1000 Repeated Days
$
%
&

0.7
&DVH&
0.6
Cumulative Efficiency
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
&DVH%
0.5
&DVH$
0.4

0.3
0.2
0 200 400 600 800 1000
Time (day)
)LJXUH&XPXODWLYHVWRUDJHHIILFLHQF\IRUWKHWKUHHSUHFRROLQJFDVHV&DVHV$DQG%DUHYHU\VLPLODU
,Q RUGHU WR HYDOXDWH ZKHQ SUHFRROLQJ LV PRVW HIIHFWLYH WKH VWRUDJH HIILFLHQF\ ZDV FDOFXODWHG DV D IXQFWLRQ RI WKH
GXUDWLRQ RI SUHFRROLQJ ZKHUH SUHFRROLQJ RFFXUV LPPHGLDWHO\ EHIRUH WKH GD\ EHJLQV )RU H[DPSOH LI WKH GXUDWLRQ RI
SUHFRROLQJZDVKRXUVDQGSUHFRROLQJHQGHGDWWKHQSUHFRROLQJVWDUWHGDW7KHUHVXOWVDUHVKRZQLQWKHOHIW
VLGHRI)LJXUHDVWKHSUHFRROLQJGXUDWLRQGHFUHDVHVWKHVWRUDJHHIILFLHQF\LQFUHDVHV3UHFRROLQJFDQDOVRRFFXUDWWKH
EHJLQQLQJRIWKHQLJKWZLWKDEUHDNEHIRUHGD\RSHUDWLRQEHJLQV)RUH[DPSOHLIWKHSUHFRROLQJGXUDWLRQLVKRXUVWKHQ
SUHFRROLQJHQGVDWDQGWKHUHLVQRSUHFRROLQJIURPWRWKHWZRKRXUVEHIRUHGD\WLPHRSHUDWLRQVWDUWV
7KHUHVXOWVRIWKLVFDVHDUHVKRZQLQWKHULJKWRIWKHILJXUH%RWKVHWVRIUHVXOWVLQGLFDWHWKDWSUHFRROLQJLVPRVWHIIHFWLYH
ZKHQLWLVRIVKRUWHUGXUDWLRQ
$VHQVLWLYLW\VWXG\ZDVFRQGXFWHGWRHYDOXDWHWKHHIIHFWRIWKHVRLOSURSHUWLHVRQWKHVWRUDJHHIILFLHQF\7KHUHDUHWZR
LPSRUWDQWSDUDPHWHUVWKHWKHUPDOFRQGXFWLYLW\NDQGWKHYROXPHWULFKHDWFDSDFLW\&,INLVGRXEOHGDQG&UHPDLQVWKH
VDPHWKHQWKHVWRUDJHHIILFLHQF\GHFUHDVHVE\EHFDXVHHQHUJ\PRYHVDZD\IURPWKHERUHKROHPRUHTXLFNO\DQGOHVVLV
VWRUHG,INLVKDOYHGDQG&UHPDLQVWKHVDPHWKHQWKHVWRUDJHHIILFLHQF\LQFUHDVHVE\EHFDXVHHQHUJ\PRYHVDZD\
IURP WKH ERUHKROH PRUH VORZO\ ,IN LV KHOG FRQVWDQWDQG& LVGRXEOHGWKH VWRUDJH HIILFLHQF\ LQFUHDVHV E\ EHFDXVH
PRUHHQHUJ\LVVWRUHGLQWKHJURXQGDQGOHVVOHDYHVRUHQWHUVWKURXJKWKHRXWHUERXQGDU\RIWKHV\VWHP,QWKLVVWXG\WKH
EDVHOLQHYDOXHVIRUNDQG&DUH:P.%WXKUIW)DQGN-P.%WXIW)UHVSHFWLYHO\
7KH HIILFLHQF\ GRHV QRW FKDQJH ZKHQ WKH VSHFLILHG LQQHU ERXQGDU\ WHPSHUDWXUH FKDQJHV IURP IRU H[DPSOH Δ&
Δ)WRΔ&Δ)DVORQJDVWKHQLJKWDQGGD\WHPSHUDWXUHRIIVHWLVHTXDODQGRSSRVLWH,IWKHQLJKWWHPSHUDWXUH
RIIVHWLVWZLFHWKHGD\RIIVHWHJΔ&YVΔ&WKHUHLVDGHFUHDVHLQVWRUDJHHIILFLHQF\EXWWKHFKDQJHLVVRVOLJKWWKDW
LWFDQEHQHJOHFWHG7KHHIILFLHQF\LVDOVRLQGHSHQGHQWRIWKHORFDWLRQRIWKHRXWHUERXQGDU\6WRUDJHHIILFLHQF\WKHUHIRUHLV
QRWVLJQLILFDQWO\LPSDFWHGE\WKHPDJQLWXGHRIWKHERXQGDU\FRQGLWLRQVEXWFDQEHLPSDFWHGE\WKHVSHFLILFFKDUDFWHULVWLFV
RI WKH VRLO SURSHUWLHV 7KH LGHDO VRLO IRU WKHUPDO VWRUDJH KDV KLJK KHDW FDSDFLW\ DQG ORZ WKHUPDO FRQGXFWLYLW\ 1RWH
KRZHYHUWKDWORZWKHUPDOFRQGXFWLYLW\LVJHQHUDOO\QRWLGHDOIRUVWDQGDUG+\*6+3V\VWHPV
$6+5$(

0.82 0.9
0.8
0.85
0.78
Efficiency
Efficiency
0.76 0.8
0.74
0.75
0.72
0.7
0.7
0.68 0.65
0 5 10 15 0 5 10 15
Duration of pre-cooling Duration of pre-cooling
)LJXUH6WRUDJHHIILFLHQF\DVDIXQFWLRQRIWKHGXUDWLRQRISUHFRROLQJ/HIW3UHFRROLQJLVSULRUWRWKHVWDUWRIWKHGD\5LJKW3UHFRROLQJ

RFFXUVWKHQWKHUHLVDEUHDNWKHQWKHGD\VWDUWV
SYSTEM MODEL
7KH VWRUDJH HIILFLHQF\ LV D XVHIXO PHWULF IRU XQGHUVWDQGLQJ WKH EHKDYLRU RI WKH JURXQG KHDW H[FKDQJHU ZKHQ XVHG
ZLWKLQDSUHFRROLQJVWUDWHJ\+RZHYHUSUHFRROLQJPXVWXOWLPDWHO\EHHYDOXDWHGE\H[DPLQLQJLWVLPSDFWRQWKHHQHUJ\
XVDJHDQGHFRQRPLFVRID+\*6+3V\VWHP)LJXUHLVDVFKHPDWLFRIWKHV\VWHPWKDWLVDQDO\]HGLQWKLVVWXG\WKHFRROLQJ
WRZHU &7 LV XVHG RQO\ GXULQJ SUHFRROLQJ SHULRGV DQG WKH JURXQG KHDW H[FKDQJHU *+; LV XVHG ZKHQHYHU WKHUH LV D
FRROLQJORDG7KH&7IORZUDWHGXULQJSUHFRROLQJLVVHWE\WKHSUHFRROLQJVFKHPHVHH7DEOH7KHKHDWSXPS+3GRHV
QRWRSHUDWHGXULQJSUHFRROLQJ
d

,W ',y
)LJXUH6FKHPDWLFRIWKHV\VWHPPRGHOZLWKSUHFRROLQJ
7KLVV\VWHPZDVHYDOXDWHGXVLQJDV\QWKHWLFFRROLQJORDGVKRZQLQ)LJXUHIRUHDFKGD\RIWKH\HDUVLPXODWLRQ
7KHORDGSURILOHZDVGHVLJQHGWRPLPLFDW\SLFDOFRROLQJGD\DJUDGXDOLQFUHDVHDQGGHFUHDVHDWWKHVWDUWDQGHQGRIWKHGD\
ZLWKDSHDNLQPLGDIWHUQRRQ7KHGU\DQGZHWEXOEWHPSHUDWXUHVIRUWKLVGD\DUHVKRZQDWWKHULJKWRI)LJXUHWKHVHDUH
WKHWHPSHUDWXUHVIRU-XO\LQ/DV9HJDV1915(/7KHXQGLVWXUEHGJURXQGWHPSHUDWXUHW\SLFDOIRUWKLVORFDWLRQ
LVDOVRVKRZQLQWKHILJXUH7KHORZZHWEXOEWHPSHUDWXUHLVEHQHILFLDOIRUFRROLQJWRZHURSHUDWLRQEHFDXVHLWLVDOZD\VOHVV
WKDQWKHXQGLVWXUEHGJURXQGWHPSHUDWXUHVRWKDWWKHJURXQGZLOOEHDOZD\VEHFRROHGE\WKHFRROLQJWRZHU/DV9HJDVZDV
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
38
ASHRAE Transactions
FKRVHQIRUWKLVVWXG\EHFDXVHWKHVHWHPSHUDWXUHVDUHIDYRUDEOHIRUVXFFHVVIXOSUHFRROLQJ
7KLVPRGHOZDVLPSOHPHQWHGLQ)2575$1EXWVRPHRIWKHFRPSRQHQWVZHUHEDVHG RQPRGHOVGHYHORSHGLQWKH
7516<6HQYLURQPHQW6$.OHLQHWDO7KH*+;ZDVPRGHOHGXVLQJWKH'67'XFW6WRUDJHPRGHO+HOOVWU|P
ZKLFKFRPELQHVERWKDQDO\WLFDODQGQXPHULFDOVROXWLRQVWRFUHDWHD'PRGHORIWKHERUHILHOGZKLFKDFFRXQWVIRU
WKH LQWHUDFWLRQ EHWZHHQ ERUHKROHV 7KH '67 PRGHO ZDV UXQ ZLWK D VLQJOH ERUHKROH DQG FKDUDFWHULVWLFV PDWFKLQJ WKH
QXPHULFDO PRGHO GLVFXVVHG LQ WKH SULRU VHFWLRQ )RU &DVH $ WKH ORQJ WHUP VWRUDJH HIILFLHQF\ ZDV IRU WKH '67
PRGHODQGIRUWKHQXPHULFDOPRGHO7KLVUHVXOWVKRZVWKDWDWWKHPRVWIXQGDPHQWDOOHYHOWKH'67PRGHOSURGXFHV
WKHVDPHUHVXOWVDVWKHVLPSOHQXPHULFDOPRGHO3HUW]ERUQ+DFNHO1HOOLV .OHLQ
7KH FRROLQJ WRZHU VL]H 6L]H&7 DQG *+; OHQJWK ZHUH RSWLPL]HG IRU HDFK FDVH WR PLQLPL]H OLIH F\FOH FRVW /&&
UHVXOWLQJLQWKHYDOXHVVKRZQLQ7DEOH/&&ZDVFDOFXODWHGRYHUD\HDUSHULRGE\XVHRIWKH33PHWKRG-$'XIILH
%HFNPDQ
6RPHRIWKHNH\UHVXOWVRIWKHVHVLPXODWLRQVDUHVKRZQLQ
7DEOH 7KH HOHFWULFDO FRVWV ZHUH FDOFXODWHG XVLQJ D WLPH RI GD\ UDWH VWUXFWXUH 7KH RSHUDWLRQ FRVWV LQFOXGH ERWK
HOHFWULFDODQGZDWHUFRVWV

200 682 40 104
160 546 34 93.2

Cooling load (kBtu/hr)
Temperature (°C)
Temperature (°F)
Cooling load (kW)
120 409 28 82.4
Dry Bulb Temperature

80 273 22 71.6
Wet Bulb Temperature
Ground Temperature
40 136 16 60.8
0 0 10 50
0 5 10 15 20 25 0 5 10 15 20 25
Time (hr) Time (hr)
)LJXUH/HIW&RROLQJORDGSURILOH5LJKW'U\DQGZHWEXOEWHPSHUDWXUHV
Table 2. Optimal design parameters for all cases.

Baseline Case A Case B Case C
*+;OHQJWKPIW
6L]H&7N:WRQV
6WDUWRISUHFRROLQJ
(QGRISUHFRROLQJ
&7IORZUDWHNJKUOEPKU HH HH HH
$OORIWKHSUHFRROLQJFDVHVVKRZDQLQFUHDVHLQSRZHUFRQVXPSWLRQHOHFWULFDOFRVWVDQGRSHUDWLRQDOFRVWV$OWKRXJKWKHUH
LVDVOLJKWUHGXFWLRQLQKHDWSXPSSRZHU FRQVXPSWLRQLW LVRIIVHWE\WKH LQFUHDVHLQSXPSSRZHUFRQVXPSWLRQDQG&7
SRZHU FRQVXPSWLRQ 7KH VLJQLILFDQW LQFUHDVH LQ RSHUDWLRQDO FRVWV LV GXH WR WKH VXEVWDQWLDO ZDWHU FRQVXPSWLRQ RI SUH
FRROLQJ %DVHG RQ WKH RSHUDWLRQDO SHUIRUPDQFH SUHFRROLQJ KDV QR EHQHILW KRZHYHU WKHUH LV DQ HFRQRPLF EHQHILW 7KH
DGGLWLRQRISUHFRROLQJOHDGVWRDUHGXFWLRQLQWKHOHQJWKRIWKH*+;ZKLFKFRUUHVSRQGVWRDVLJQLILFDQWUHGXFWLRQLQWKH
ILUVWFRVWRIWKHV\VWHP$VFDQEHVHHQLQWKHILQDOURZVRI
7DEOHWKHLQFUHDVHLQDQQXDORSHUDWLRQDOFRVWLVPRUHWKDQRIIVHWE\WKHUHGXFWLRQLQWKHILUVWFRVWRIWKHV\VWHP
--`,`,``,,`,`,```,`,,``,`,``,,,-`
$6+5$(

Table 3. Key results for the optimal designs.
&RROLQJ7RZHU3RZHUN:K0%WX
3XPSLQJ3RZHUN:K0%WX
+HDW3XPS3RZHUN:K0%WX
,QFUHDVHLQ7RWDO(OHFWULFDO3RZHU
,QFUHDVHLQ7RWDO(OHFWULFDO&RVWV
,QFUHDVHLQ2SHUDWLRQ&RVWV
5HGXFWLRQLQ)LUVW&RVW
,QFUHDVHLQ$QQXDO2SHUDWLRQ&RVW
5HGXFWLRQLQ/&&
1HW6DYLQJV
,QPRVW+\*6+3V\VWHPVWKH*+;OHQJWKLVEDVHGRQWKHKHDWLQJORDGDQGWKH&7LVVL]HGWRPHHWZKDWHYHUFRROLQJORDG
WKH*+;LVXQDEOHWRPHHW7KHUHIRUHLWLVXQOLNHO\WKDWWKH*+;OHQJWKFRXOGEHUHGXFHGWRWKHH[WHQWVKRZQKHUH,Q
RUGHU WR DVVHVV WKLV VLWXDWLRQ D N: FRROLQJ WRZHU ZDVXVHG LQ SUHFRROLQJ ZLWK WKH EDVHOLQH *+; OHQJWK RI P
IW7KHNH\UHVXOWVIRUWKLVFDVHDUHVKRZQLQ7DEOH
Table 4. Key results for the optimal design with the baseline GHX length.
&RROLQJ7RZHU3RZHUN:K0%WX
3XPSLQJ3RZHUN:K0%WX
+HDW3XPS3RZHUN:K0%WX
,QFUHDVHLQ7RWDO(OHFWULFDO3RZHU
,QFUHDVHLQ7RWDO(OHFWULFDO&RVWV
,QFUHDVHLQ2SHUDWLRQ&RVWV
,QFUHDVHLQ)LUVW&RVW
,QFUHDVHLQ$QQXDO2SHUDWLRQ&RVW
,QFUHDVHLQ/&&
1HW,QFUHDVH
/&& LQFUHDVHV IRU HDFK RI WKH SUHFRROLQJ FDVHV EHFDXVH RI WKH DGGLWLRQDO ILUVW FRVW RI WKH FRROLQJ WRZHU DQG ODUJHU
RSHUDWLRQDOFRVWV7KHLQFUHDVHLQRSHUDWLRQDOFRVWLVOHVVWKDQLQWKHRSWLPDOGHVLJQVEHFDXVHWKHZDWHUFRQVXPSWLRQDQG
SRZHUUHTXLUHPHQWVDUHOHVVIRUDVPDOOHUFRROLQJWRZHU7KHKHDWSXPSSRZHUFRQVXPSWLRQGHFUHDVHVPRUHWKDQLQWKH
RSWLPDOGHVLJQVEHFDXVHWKHVL]HRIWKH*+;LVODUJHU,QWKHFDVHVZLWKDN:FRROLQJWRZHUWKHODUJHEDVHOLQH*+;LV
XVHGVRKHDWLVUHMHFWHGWRDODUJHUYROXPHDVFRPSDUHGWRWKHRSWLPDOGHVLJQVDQGWKHJURXQGWHPSHUDWXUHLQFUHDVHVPRUH
VORZO\GXULQJWKHGD\OHDGLQJWRDORZHURYHUDOOHQWHULQJZDWHUWHPSHUDWXUHDQGPRUHHIILFLHQWKHDWSXPS
CONCLUSIONS
%DVHGRQWKHVHLQLWLDOUHVXOWVWKH*+;LVQRWDJRRGWKHUPDOVWRUDJHGHYLFHIRUVKRUWWHUPFRROLQJ,QWKHEHVWFDVH
VLJQLILFDQWVDYLQJVZHUHDFKLHYHGQRWGXHWRWKHEHQHILWRIWKHUPDOVWRUDJHEXWGXHWRWKHHIIHFWRIKDYLQJDVHFRQGKHDW
UHMHFWLRQGHYLFHZKLFKEDODQFHGWKHORDGRQWKHJURXQGDQGOHGWRDUHGXFWLRQLQ*+;OHQJWK:KHQWKH*+;OHQJWKZDV
KHOG FRQVWDQW WKH DGGLWLRQ RI D FRROLQJ WRZHU OHG WR D VOLJKW LQFUHDVH LQ /&& DQG RSHUDWLRQDO FRVW ,Q DGGLWLRQ WKH VRLO
SURSHUWLHV ZKLFK IDYRU SUHFRROLQJ DUH ORZ WKHUPDO FRQGXFWLYLW\ DQG KLJK KHDW FDSDFLW\ ZKLOH *6+3 V\VWHPV JHQHUDOO\
SHUIRUP EHVW ZLWK KLJK WKHUPDO FRQGXFWLYLW\ DQG ORZ KHDW FDSDFLW\ VR WKDW HQHUJ\ FDQ PRYH DZD\ IURP WKH ERUHKROH
TXLFNO\)XUWKHUZRUNLVEHLQJFRQGXFWHGWREHWWHUXQGHUVWDQGSUHFRROLQJLQPRUHUHDOLVWLFV\VWHPV
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
40
ASHRAE Transactions
ACKNOWLEDGEMENTS
7KHDXWKRUVZRXOGOLNHWRUHFRJQL]HWKHDVVLVWDQFHRI6FRWW+DFNHODWWKH(QHUJ\&HQWHURI:LVFRQVLQLQREWDLQLQJ
VRPHRIWKHLQSXWVWRWKHPRGHOVXVHGLQWKLVVWXG\
NOMENCLATURE
α WKHUPDOGLIIXVLYLW\
& KHDWFDSDFLW\
&7 FRROLQJWRZHU
'7 GLIIHUHQWLDOWHPSHUDWXUHVHWSRLQWWRGHWHUPLQHZKHQWKHFRROLQJWRZHULVXVHG
(:7 HQWHULQJZDWHUWHPSHUDWXUHWRWKHKHDWSXPS
*+; JURXQGKHDWH[FKDQJHU
+3 KHDWSXPS
+\*6+3 K\EULGJURXQGVRXUFHKHDWSXPS
N WKHUPDOFRQGXFWLYLW\
6L]H&7 VL]HRIWKHFRROLQJWRZHU
W WHPSHUDWXUH
τ WLPHDWZKLFKDVWHSFKDQJHRFFXUVLQWKHERXQGDU\FRQGLWLRQ
7& WHPSHUDWXUHVHWSRLQWWRGHWHUPLQHZKHQWKH*+;LVXVHGLQFRROLQJ
Subscripts
J JURXQG
REFERENCES
'XIILH-$ %HFNPDQ:$6RODU(QJLQHHULQJRI7KHUPDO3URFHVVHVUGHG:LOH\DQG6RQV,QF
)DQ5-LDQJ<<DR< 0D=7KHRUHWLFDOVWXG\RQWKHSHUIRUPDQFHRIDQLQWHJUDWHGJURXQGVRXUFHKHDWSXPS
V\VWHPLQDZKROH\HDU(QHUJ\
+HOOVWU|P**URXQG+HDW6WRUDJH0RGHO0DQXDOIRU&RPSXWHU&RGH'HSDUWPHQWRI0DWKHPDWLFDO3K\VLFV8QLYHUVLW\RI
/XQG6ZHGHQ
-LQJJDQJ:;LDR[LD*=KHQMLDQJ< )DQJ/6WXG\RI2SHUDWLQJ&RQWURO6WUDWHJLHVIRU+\EULG*URXQG6RXUFH
+HDW3XPS6\VWHPZLWK6XSSOHPHQWDO&RROLQJ7RZHU,,7$,QWHUQDWLRQDO&RQIHUHQFHRQ6HUYLFHV6FLHQFH0DQDJHPHQWDQG
(QJLQHHULQJSS,HHH
.OHLQ6$%HFNPDQ:$0LWFKHOO-:'XIILH-$7KRUQWRQ-:0LWFKHOO-&%UDXQ-(HWDO7516<6$
7UDQVLHQW6\VWHP6LPXODWLRQ3URJUDP8VHU·V0DQXDO9HUVLRQ6RODU(QHUJ\/DERUDWRU\8QLYHUVLW\RI:LVFRQVLQ
0F'RZHOO73 7KRUQWRQ-:6,08/$7,21$1'02'(/&$/,%5$7,212)$/$5*(6&$/(62/$5
6($621$/6725$*(6<67(06LP%XLOG7KLUG1DWLRQDO&RQIHUHQFHRI,%36$86$SS%HUNHOH\&$
15(/1DWLRQDO6RODU5DGLDWLRQ'DWD%DVH7\SLFDO0HWHRURORJLFDO<HDU5HWULHYHGIURP
KWWSUUHGFQUHOJRYVRODUROGBGDWDQVUGEWP\
1HOOLV*) .OHLQ6$+HDW7UDQVIHUVWHGS&DPEULGJH8QLYHUVLW\3UHVV
3DKXG'&HQWUDOVRODUKHDWLQJSODQWVZLWKVHDVRQDOGXFWVWRUDJHDQGVKRUWWHUPZDWHUVWRUDJHGHVLJQJXLGHOLQHVREWDLQHG
E\G\QDPLFV\VWHPVLPXODWLRQV6RODU(QHUJ\
3DUGR10RQWHURÉ0DUWRV- 8UFKXHJXtD-)2SWLPL]DWLRQRIK\EULG²JURXQGFRXSOHGDQGDLUVRXUFH²KHDW
SXPSV\VWHPVLQFRPELQDWLRQZLWKWKHUPDOVWRUDJH$SSOLHG7KHUPDO(QJLQHHULQJ(OVHYLHU/WG
3HUW]ERUQ$+DFNHO61HOOLV* .OHLQ6([SHULPHQWDOYDOLGDWLRQRIDJURXQGKHDWH[FKDQJHUPRGHOLQDK\EULG
JURXQGVRXUFHKHDWSXPS+9$& 55HVHDUFK
:RQJ%6QLMGHUV$ 0F&OXQJ/5HFHQW,QWHUVHDVRQDO8QGHUJURXQG7KHUPDO(QHUJ\6WRUDJH$SSOLFDWLRQVLQ
&DQDGD,((((,&&OLPDWH&KDQJH&RQIHUHQFH,HHH
<DYX]WXUN& 6SLWOHU-'&RPSDUDWLYH6WXG\WR,QYHVWLJDWH2SHUDWLQJDQG&RQWURO6WUDWHJLHVIRU+\EULG*URXQG
6RXUFH+HDW3XPS6\VWHPV8VLQJD6KRUW7LPHVWHS6LPXODWLRQ0RGHO$6+5$(7UDQVDFWLRQV
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

SA-12-C006
Field Trials of a Waterless Home Heating

and Humidification Technology
Dexin Wang, PhD Shawn Scott Ainan Bao, PhD William Liss
ABSTRACT
It is generally accepted, and has been confirmed by studies, that humidification of dry indoor air to raise relative humidity (RH) during the heating
season is beneficial to the comfort and health of building occupants. Humidification also prevents adverse effects on wood floors and furniture and
reduces static electricity buildup which can harm electronic equipment.
Currently, the most widely used residential humidification technologies are forced air furnace mounted bypass wetted media, spray mist, and steam
humidifiers. These use city water as a water source and require additional furnace heat or electricity to evaporate the water, which consumes 4% or
more of the furnace fuel input. Mineral deposition, white dust and microbial growth problems are associated with these humidifiers. For
commercial building humidification, demineralized water is typically used for humidification equipment such as steam heat exchangers, electric and
ultrasonic humidifiers, compressed air atomizers, and high pressure cold water foggers. In addition to the energy consumption for the water
evaporation, energy is also needed to produce high-quality demineralized water through a reverse osmosis process.
A Transport Membrane Humidifier (TMH) technology was developed by using nanoporous membrane capillary condensation separation
mechanism to transport water vapor only from furnace combustion flue gas to humidify building air. After proving the technology in a laboratory
environment for an equivalent 4-year operation, two TMH units were installed for two home furnaces with AFUE ratings of 80%. The two
furnaces are from two different manufacturers with different ductwork configurations and different heating capacities, so two separate designs were
made to accommodate the difference. Both TMH units had been in operation through the 2010-2011 and 2011-2012 heating seasons in
Chicago area homes, and provided satisfactory whole house humidification with both occupied homes maintained at 40 to 60% RH. At the same
time, they boosted the two furnace’s efficiency from around 80% to more than 95%, providing significant energy savings. Compared with
conventional whole house humidification technology, the TMH humidification benefit comes with no water connection, no need to change
filters/wetting pads/drums, no white dust to rooms, and no bacteria growth concerns from standing water. In addition no maintenance is required
for the TMH units.
INTRODUCTION
It is generally accepted, and has been confirmed by studies, that humidification of dry indoor air to raise relative
humidity (RH) during the heating season is beneficial to the comfort and health of building occupants. There are also
significant energy savings possible due to the “apparent temperature” phenomenon that allows people to feel more
comfortable (i.e., warmer) at higher RH. Humidification also prevents adverse effects on wood floors and furniture and
reduces static electricity buildup which can harm electronic equipment. ASHRAE Standard 62-1989, states, “relative
humidity in habitable space preferable should be maintained between 30% and 60%... to minimize growth of allergenic and
pathogenic organisms”. Notably, the lack of proper space humidification enhances the rate of influenza virus, resulting in a
significant number of illnesses and deaths each year. Humidity control is important in commercial buildings – including
hospitals – as well as many industrial processes, such as electronic and semiconductor manufacturing, medical supply,
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
42
ASHRAE Transactions
printing application, woodworking and storage, and textile industries.
Currently, the most widely used residential humidification technologies are forced air furnace-mounted bypass wetted
media, spray mist, and steam humidifiers. These use city water as a water source and require additional furnace heat or
electricity to evaporate the water. Mineral deposition, white dust and microbial growth problems are associated with most
of these humidifiers. For commercial building humidification, demineralized water is typically used for humidification
equipment like steam heat exchangers, electric and ultrasonic humidifiers, compressed air atomizers, and high pressure cold
water foggers. In addition to the energy consumption for the water evaporation, energy is also needed to produce high-
quality demineralized water through a reverse osmosis process.
The Transport Membrane Humidifier (TMH) technology was developed by using a nanoporous membrane that
facilitates a capillary condensation separation mechanism which transports water vapor only from furnace combustion flue
gas to humidify building air. The capillary condensation action enables high water transport rates while also blocking non-
condensible gases from transporting across the membrane.
There are other research efforts aimed at using membranes to separate and transport water vapor for gas stream
dehydration, humidity control, and energy recovery in commercial HVAC systems. None of these applications, however,
has attempted to extract water vapor from a flue gas stream to humidify air. For all these and similar applications, only very
small trans-membrane total pressure is available. The driving force for water vapor to transport from one side of the
membrane to the other side relies mainly on the water vapor partial pressure difference between the two gas streams. For all
these reported applications, they are dealing with transporting moisture from a high humidity air stream to a low humidity
air stream, the water partial pressure difference is relatively small, less than 0.4 psi (2,760 pascal).
A flue gas stream typically has a dew point of 120 to 136°F (49 to 58°C) . This high temperature high humidity level
can create a greater than 2 psi (13,800 pascal) water vapor partial pressure difference with the circulating room air, which
usually has a dew point of 50ÜF (10°C) or lower. Using flue gas moisture to humidify the room air can provide five times
larger driving force across the membrane, therefore substantially less membrane surface area is needed. The reduced surface
area greatly lowers the cost and improves the prospect for a cost effective commercial application using the TMH. In
addition, since the flue gas is typically at much higher temperature (over 250ÜF, or 121°C), the TMH functions as a heat
exchanger to preheat the air stream to save energy.
The combined energy saving and humidification function with no potable water consumption makes this technology
unique. The reduced membrane surface area and simple design make it promising for a commercial product. To our
knowledge, no practical technology has ever been developed for humidifying room air with flue gas moisture for residential
use. TMH technology can reduce fuel use, eliminate city water consumption, completely avoid mineral deposition and white
dust, and avoid microbial growth, improving both the physical and financial health of the homeowners.
The TMH technology has been developed from concept to laboratory prototype, and the laboratory prototype TMH
has been tested and proved working well in a wide operation range for a residential furnace to add moisture into the
circulation air and enhance the mid-efficiency residential furnaces (around 80% AFUE) by about 15%. A long term testing
has also been carried out for this laboratory TMH for about 5,000 hour furnace operating time, which is equivalent to about
4 year operation time of a typical furnace. At the end of the testing period, the furnace efficency still can be enhanced by
13% from its baseline condition, with 5.0 lb/hr (2.27 kg/hr) moisture transport rate to the air, enough for home
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
humidification.
This paper will mainly introduce two actual home TMH installations and the test results, to show the TMH
technology real world performance on both whole house humidification effect and furnace efficiency enhancement in the
two occupied homes.
FIELD TRIAL DESIGN AND HOME INSTALLATION
TMH Installation Arrangement And System Setup
$6+5$(

As shown in Figure 1 a), the TMH is installed in the furnace air inlet ductwork. Inside the TMH, flue gas flows from
the membrane feed side, while the room circulating air that requires heating and humidification flows on the permeate side.
The low-temperature, high-flow-rate room circulating air passing over the membrane surfaces provides adequate membrane
cooling to facilitate the high-performance capillary condensation water vapor separation mode. Water from the flue gas is
transported to the air side, simultaneously heating and humidifying the air.
Figure 1: a) TMH install arrangement for a residential mid-efficiency furnace, b) P&ID for the TMH field trial installation
Detailed P&ID of the TMH installations with all the measurement is shown in Figure 1 b). The furnace natural gas
flow rate was measured by a natural gas flow meter. The furnace flue gas temperature, TMH air inlet/outlet temperatures,
and furnace air delivery temperature were measured by thermocouples. The air inlet and outlet dew points were measured
by hygrometers. An ID fan was installed to overcome the flue gas pressure drop through the TMH, and its electrical usage
was measured by a power meter. All experimental data were collected by a data acquisition system for post-processing.
TMH Module Assembly And Field Installation

Two occupied single family homes were selected to demonstrate the whole house TMH heat recovery and
humidification technology for residential furnaces, to verify their real world performance on furnace efficiency
improvement and whole house humidification.
Based on the laboratory prototype TMH design and assembling experience, two TMH modules with even lower air
and flue gas pressure drops were designed, and the two TMH module overall dimensions were based on the corresponding
furnace air ductwork cross sections and their fuel input capacities. Figure 2 c) shows pictures of the two TMH modules
built for the two field trail installations.
Pictures for the two TMH home installations are shown in Figure 2 a) and b). Furnace in home 1 has a 110,000
BTU/hr (3.22 kW) fuel input, furnace in home 2 has a 90,000 BTU/hr (2.63 kW) fuel input, both are mid-efficiency
furnaces with AFUE rated 80%. The TMH modules were installed into the return air ductwork going into the furnaces, and
the flue gas heat and water were simultaneously recovered in the TMH and distributed into the homes after being further
heated by the furnaces. The flue gas side pressure drops through the TMH were measured as, 0.35-0.4 inches of water (87-
99 pascal) for home 1 TMH module, and 0.2-0.25 inches of water (50-62 pascal) for home 2.
44
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
ASHRAE Transactions
Figure 2: TMH installations for Home 1 a) (left) and Home 2 b) (middle), and the assembled TMH modules for Home 1 c)
(right, top) and Home 2 c)(right, bottom)
TMH FIELD TRIAL RESULTS
Overall Furnace Efficiency Enhancement And Whole House Humidification Effect

The furnace overall efficiency is calculated based on the fuel higher heating value (HHV) and the furnace exhaust flue
gas temperature and moisture content. For both mid-efficiency furnaces with the TMH installations, the flue gas exhaust
temperatures decrease significantly from around 350ºF (177ºC) to around 105ºF (41ºC) for Home 1, and to around
95ºF(35ºC) for Home 2. Flue gas outlet dew points decrease from around 125ºF (52ºC) to around 90ºF(32ºC) for both
cases, and the furnace overall efficiencies thus increase significantly based on these lower flue gas outlet temperatures and
dew points. Calculation results show that the home 1 furnace efficiency increases from 81.5% without the TMH to 95.5%
with the TMH, and the home 2 furnace efficiency increases from 80.6% to 96.9%. The average moisture transport rates are
2.7-6.2 gallon per day (10-23 L per day) for home 1, and 1.5-4.8 gallon per day (5.7-18 L per day) for home 2, depending on
different room air temperatures and dew point conditions. Humidity levels for both homes have been maintained in a
comfortable range of 40-55% relative humidity with the TMH in operations.
For both homes, we have selected some days to operate the furnaces under TMH bypass mode to check the baseline
conditions without the TMHs. The results proved a significant humidity increase with the TMH in operation. For example,
relative humidity for Home 1 was 33-38% under TMH bypass mode, and 40-50% under TMH mode. Figure 3 shows the --`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
humidification effect with and without the TMH operation in January, 2011 for Home 1. This figure shows the temperature
and relative humidity in the first and second floors for Home 1. Figure 4 shows similar conditions for Home 2, which is a
one-story single family home, with temperature and humidity loggers placed in its family room (FR) and living room (LR).
Detailed Furnace Performance With The TMH

Different furnaces have different operating characteristics, which is related to the furnace capacity, the heating area
size, and the customized thermostat programming. The mid-efficiency furnace for Home 1 has shorter heating cycles; and
the furnace for Home 2 has longer heating cycles.
$6+5$(

80 TMHHome1(Jan,2011)
70
Temperature(F)/RH(%) 60
50
40
1stFloorTemp
30 2ndFloorTemp
Vacation 1stFloorRH
2ndFloorRH Bypassmode
20
1/1 1/4 1/7 1/10 1/13 1/16 1/19 1/22 1/25 1/28 1/31
Date/Time
Figure 3: Home 1 room temperature and humidity with and without TMH in operation in January 2011
80 TMHHome2(Jan,2011)
70
Temperature(F)/RH(%)
60
FRTemp
50 LRTemp
FRRH
LRRH
40
30
Vacation Bypassmode Vacation

20
1/1 1/4 1/7 1/10 1/13 1/16 1/19 1/22 1/25 1/28 1/31
Date/Time
Figure 4: Home 2 room temperature and humidity with and without TMH in operation in January 2011
Figures 5 and 6 show the furnace characteristic temperatures, such as flue gas outlet temperature and dew point,
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
TMH air inlet and outlet temperatures and dew points, and the furnace final air delivery temperature, in a typical furnace
operation cycle for both homes, under the TMH mode and TMH bypass mode. From Figure 5, we can see the circulating
air dew point (Td) increases about 3ÜF(1.7ÜC) after it passes through the TMH module in the TMH mode in one heating
cycle, but has no change when the TMH was bypassed. For Home 2 as shown in Figure 6, the heating cycle is much
longer, and there is no obvious difference between the TMH inlet and outlet air dew points, but at the end of the heating
cycle, we can see the air dew point increased about 15ÜF (8.3ÜC) with the TMH, but only increased about 11ÜF(6.1ÜC) and
stayed at a lower level when the TMH is bypassed.
Figure 7 shows the instantaneous efficiency of one typical heating cycle for the two home furnaces at TMH bypass
mode and TMH mode. Averaged efficiency increases for these two typical heating cycles are from about 82% to 96% for
Home 1 furnace and 81% to 96% for Home 2 furnace.
46
ASHRAE Transactions
160 Home1Bypassmode 160 Home1TMHmode
AirinTemp
150 AirinTemp 150 AiroutTemp
140 AiroutTemp 140 AirinTd
130 AirinTd 130 AiroutTd
120 AiroutTd 120 FlueoutTemp
Temperature(°F)
Temperature(°F)
110 DeliverAirTemp 110 DeliverAirTemp

100 100
90 90
80 80
70 70
60 60
50 50
40 40
30 30
20 20
0 1 2 3 4 5 6 7 8 9 10 11 12 0 1 2 3 4 5 6 7 8 9 10 11 12
Time(minute) Time(minute)
Figure 5: Home 1 Furnace and TMH characteristic temperatures for a short heating cycle
170 Home2Bypassmode 170 Home2TMHmode

160 160
150 150
140 AirinTemp
140 AiroutTemp
130 AirinTemp 130 AirinTd
Temperature(°F)
Temperature(°F)
120 AiroutTemp 120 AiroutTd

AirinTd FlueoutTemp
110 110 DeliverAirTemp
100 AiroutTd 100
DeliverAirTemp
90 90
80 80
70 70
60 60
50 50
40 40
30 30
0 20 40 60 80 100 120
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
0 20 40 60 80 100 120
Figure 6: Home 2 Furnace and TMH characteristic temperatures for a long heating cycle
Home1FurnaceEfficiency Home2FurnaceEfficiency
100 100
90 90
80 80
70 70
Efficiency(%)
Efficiency(%)
60 60
50 50
40 40
30 30
20 TMHmode 20 TMHmode
10 Bypassmode 10 Bypassmode
0 0
0 1 2 3 4 5 6 0 2 4 6 8 10 12 14 16 18 20 22
Figure 7: Home 1 (left) and Home 2 (right) furnace instantaneous efficiency under TMH bypass mode and TMH mode for
one heating cycle
$6+5$(

Economic Analysis and Potential Markets
There are about 35 million gas furnaces currently operating in U.S. homes. In 1998, 12% of furnaces available in the
market are considered high efficiency furnaces, and by 2010 high efficiency furnaces represented about 30% of national
furnace shipments. So it is estimated now more than 70% of furnaces in use are still mid efficiency furnaces (Federal
requirement for minimum 78% AFUE, most of them are around 80% AFUE). This TMH technology is first targeted for
this huge retrofit market to significantly boost the mid-efficiency furnace efficiency while at the same time providing whole
house humidification without external water consumption and other health benefits. The high efficiency furnace shipment
percentage is not expected to increase significantly in the near future considering various federal and state high efficiency
rebates are winding down, and the payback period is less attractive to customers for the much higher equipment cost of the
high efficiency furnaces, which are typically more than doubled of the mid efficiency furnace price. Although have not been
demonstrated yet, the TMH technology has already been further developed and proved in our laboratory to have the
potential to be used for high efficiency furnaces too. Many of the high efficiency furnaces have lower than 92% AFUE,
only a small amount of flue gas water vapor in these furnaces are condensed therefore the remaining water vapor is still
enough to humidify the homes, though the efficiency gain by the TMH installation will be lower for these furnaces. For
much higher efficiency furnaces, the TMH modules can be built into the furnaces to replace their condensing heat transfer
modules, so all the flue gas water vapor is available for the home humidification. Table 1 summarizes how the TMH stacks
up against main conventional humidifier types. Besides the energy and health benefits listed in the table, there is no wetting
medium needed to be replaced regularly during a heating season compared with conventional humidifiers, which typically
costs about $30/year. The payback period of the TMH installation for a mid efficiency furnace is estimated at less than 4
years.
Table 1. Comparison of Current Furnace Humidifier Types with Proposed TMH

Commercial types Proposed
Type Bypass Steam Spray mist TMH
humidifier humidifier humidifier
Additional furnace fuel 4% 0 4% 0
consumption
Electricity usage 12 watts 1,400 watts Negligible 20 watts
Mineral deposition Yes Yes Yes a No
"White dust" in home Medium Zero High Zero
Microbial growth High None Very low Very low
potential
Water consumption b 15 gal/day 15 gal/day 12 gal/day Zero
Equipment cost $150-$225 $525-$850 $160-$200 $400 c
a potential clogging of spray nozzle; also requires water filter.
b assumes average 3 gal/day additional water throughput to control mineral deposits.
c preliminary cost target.
CONCLUSION
Two field trail TMH units were designed and tested for two typical mid-efficiency residential furnaces in two occupied
single family homes. The real world operating results showed the TMH units are capable of transferring enough water
vapor from furnace flue gas to circulating room air for humidification, and enhancing furnace efficiency by about 15%. The
home room temeprature and humidity continous monitoring data indicates both homes have been humidified to a
comfortable humidity level (40 to 60% RH) with the benefits of no external water consumption, no white dust and no
baterial growth concerns. For the two heating season operation of the two TMH units, the technology was proved can
provide comfortable and healthy humidification for the home owners and also greatly increase their furnace efficiencies.
48
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
ASHRAE Transactions
The TMH technology will be first targeted for the existing furnace retrofit market, and further development is for emerging
high efficiency furnace market for both retrofit and new installations.
ACKNOWLEDGMENTS
This work was sponsored by the Utilization Technology Development NFP.
NOMENCLATURE
RH: relative humidity
HHV: Higher Heating Value
Td: dew point
REFERENCES
Sterling, E.M., Arundel, A., and Sterling, T.D. 1985. Criteria for Human Exposure to Humidity in Occupied Buildings.
ASHRAE Transactions 91:611-622.
ASHRAE. 1989. ANSI/ASHRAE Standard 62.1989,Ventilation for Acceptable Indoor Air Quality. Atlanta: American
Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc.
ASHRAE. 1992. ASHRAE Standard 55.1992,Thermal Environmental Conditions for Human Occupacy. Atlanta: American
Asaeda, M., L. Du and M. Ushijima. 1985. Feasibility study on dehumidification of air by thin porous alumina gel
membrane p472-478; Drying’ 85. R. Toei and A.S. Mujumkar, ed. Washington D.C.: Hemisphere Publishing Corp.
Ray, R., D. D. Newbold, S.B. McCray, and D. T. Frlesen. 1992. A novel membrane device for the removal of water vapor
and water droplets from air. 22nd international conf. on environmental system, Seattle, Washington, SAE 921322.
Scovazzo, P., A. Hoehn, and P. Todd. 2000. Membrane Porosity and Hydrophilic Membrane-based Dehumidification
Performance. J. Membrane Science 167:217-225.
Scovazzo, P., J. Burgos, A. Hoehn, P. Todd. 1998. Hydrophilic Membrane-based Humidity Control. J. Membrane Science
149:69-81.
Zhang, L. and Y. Jiang. 1999. Heat and Mass Transfer in a Membrane-based Energy Recovery Ventilator. J. Membrane
Science 163:29-38.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

SA-12-C007
Environmental Noise Case Studies:

Air-Cooled Refrigeration Chiller
Installations near Residential Structures
Jack B. Evans, PE Chad N. Himmel, PE Joshua D. Leasure, PE

Member ASHRAE Member ASHRAE Member ASHRAE
ABSTRACT
The air conditioning design process includes decisions to use outdoor air-cooled refrigeration chillers or indoor chiller with outdoor cooling tower. Many air-
cooled chillers utilize screw or rotary helical compressors, which emit very tonal noise and vibration in the outdoor environment. Complaints arise from
residential properties in the vicinity, even though investigations show conformance with building code, ordinance or regulation permitted sound levels. Sound
quality, in terms of tonality, unbalanced spectrum and on/off cycles, appears to increase perception and annoyance. Simple conformance to a noise
ordinance may not be adequate to prevent neighborhood disturbances and complaints.
This paper presents acoustical issues that should be considered in site selection and orientation, equipment selection and details of air-cooled chiller
installations. Environmental noise criteria are presented along with case studies that include on-site noise spectrum measurement results. Acoustical
characteristics and site parameters are explored with respect to overall spectrum and level, relative contributions of compressors and condenser fans, other
equipment noise radiation issues, sound paths to sensitive receivers, noise build-up between large structures, distances from source to receiver and effects of
reflectors and barriers. Noise mitigation solutions are presented with photos, illustrations and discussions of several installations and the resulting
improvements and performance limitations determined in the case studies.
INTRODUCTION
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
The case studies presented herein are air-cooled chiller installations at various schools subject to the same city noise
ordinance, which is based on Uniform Building Code (UBC) Ch. 35 which restricts exterior noise levels received at a
residential property to 63 dBA. In addition, 7 dB corrections are to be applied for tonality and for nighttime, so that net
permissible level is 49 dBA (63-7-7) at residential properties adjacent to the schools.1 Ordinance conformation was not
enough, however. The owners (school districts) and design engineers wished to avoid complaints from the public, even if in
some cases the levels would have to be lower than ordinance allowables.
CASE STUDY #1: SCHOOL A

School A is an elementary school receiving a replacement for an aging, single chiller in a three-sided, brick enclosure.
The fourth wall of the enclosure is chain-link fence, causing the enclosure to be open to single family residences across the
street from the school, the closest of which is 135 feet (41 m) away. There is a history of complaints about chiller noise at
those residences.
Jack B. Evans is Principal Engineer of JEAcoustics / Engineered Vibration Acoustic & Noise Solutions, an acoustical consulting firm in Austin,
TX. Chad N. Himmel and Joshua D. Leasure are Associate and Sr. Consulting Engineers, respectively, at JEAcoustics, Austin, TX.
50
ASHRAE Transactions
The location is subject to the municipal noise ordinance, which has a limit of 63 dBA measured at a residential
property line under separate ownership. There is a nighttime penalty of 7 dBA. A further penalty of 7 dBA is applied to
noises that are impulsive in nature (changing more than 10 dB per second) or having strong pure-tone components. Based
on the lack of a strong pure-tone in the measured data (see below), and the possibility for nighttime operation of the chiller,
the unmodified nighttime limit of 56 dBA was chosen as the applicable criterion for conformance with the ordinance.
Daytime sound level measurements were taken to characterize noise emissions and propagation for the existing chiller
and ambient conditions using a handheld, ANSI Type 1, 1/3-octave sound spectrum analyzer. Octave band equivalent
sound levels (Leq, integrated average over measurement time span) were examined for ambient conditions near the existing
chiller (with the chiller off) and at the nearest residential property line with the chiller running at 25% capacity. At the time
of the measurement (early weekday afternoon), ambient noise levels were 49 dBA, and the chiller running at 25% capacity
measured 65 dBA at the nearest residential property line, exceeding the noise ordinance for both daytime and nighttime.
The existing chiller showed strong sound energy in the 250-Hz octave band, measured as approximately 64 dB. The
ambient noise spectrum showed a dip in the 250-Hz band. The high sound energy from the chiller in this band compared
to the ambient spectrum could result in higher audibility of the existing chiller when compared to other chiller models with
similar broadband noise generation.
Figure 1 Location of existing and new chiller with respect to neighboring residences.
Figure 2 Measured chiller and ambient levels compared to predicted levels for new chiller.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

The proposed replacement chiller is available in three noise reduction packages: Quiet Fan, Ultra Quiet Fan, and Low
Noise. The Quiet Fan and Ultra Quiet Fan packages include alternate fan blades and speeds, while the Low Noise package
includes a compressor covering in addition to the Ultra Quiet fans. The manufacturer provided sound pressure data at 30
feet (9 m) for all three packages at 25%, 50%, 75%, and 100% loads. For comparison to the measured levels of the existing
chiller, the data for 25% load were normalized for distance (-2 dB) and for placement in a reflecting enclosure (estimated at
+3 dB).
Figure 2 compares the measured levels of the existing chiller to measured ambient levels and the normalized 25%
load manufacturer’s data for all three packages. Other than its smoother spectrum, the Quiet Fan package shows no
improvement over the existing chiller, with an estimated property line noise level of 65 dBA. The Ultra Quiet Fan package
shows a courser spectrum with potential audible tone issues but would be 4 dBA less at the property line, while the Low
Noise package shows a smoother spectrum 7 dBA less at the property line. It does not appear that any package will allow the
new chiller to be a drop-in replacement that can meet the 56 dBA nighttime noise limit. Selection of the Low Noise package
was recommended based on its smoother spectrum and lowest overall broadband noise emission.
The 100% load data for the Low Noise package was normalized to the existing conditions for comparison to the noise
ordinance. The manufacturer’s data reports 68 dBA at 30 feet (9 m). Normalized to the applicable distance and placement
in a reflective enclosure, the predicted property line level is 69 dBA, 6 dBA above the nighttime limit and 13 dBA above the
daytime limit. To achieve the additional 13 dBA of attenuation needed to meet the criterion, recommendations were made
for modification of the chiller enclosure, the addition of sound absorbing surfaces to the chiller enclosure, and the
installation of flexible refrigerant pipe connections to the new chiller.
The addition of a fourth, solid wall to the chiller enclosure was recommended to provide shielding between the chiller
and the residences across the street. Additionally, an increase in the wall height of the enclosure was recommended for
better containment of chiller noise in all directions. A study of the chiller’s ventilation requirements was recommended to
determine if the additional wall would reduce air flow to the chiller unacceptably. If the need for additional ventilation is
determined, the use of acoustic louvers in the new wall is recommended and/or the addition of acoustical louvers in one of
the existing walls.
Installations of spray-on absorptive material or mounted sound absorbing panels were recommended to reduce noise
build-up inside the enclosure. The recommendation was for 50% coverage with a material having NRC 0.65 or higher.
The manufacturer indicated that no internal vibration isolation is installed on the chiller beyond isolation mounts for
compressors. To further isolate the compressor bodies from the chiller framework, recommendations include the
installation of flexible couplers in refrigerant pipes between the compressors and the chiller frame, and the use of resilient
mounts where refrigerant piping is attached to the compressor frame. The intent of these recommendations is to reduce
vibration in the chiller frame that can be re-radiated as noise by metal structural members or casing cover panels.
CASE STUDY #2: SCHOOL B

Cooling capacity requirements were increased at a middle school, due to a facility expansion. In addition, to reduce
noise intrusion into classrooms, the mechanical yard was to be relocated. Two larger air-cooled refrigeration chillers were
specified to replace the original chillers. Original and new chillers incorporate rotary helical or “screw” compressors. The
new mechanical yard was designed with walls around three sides, with north and northwest open (chain link fence), whereas
the former mechanical yard was open only to the northeast. The new chiller installation was near building walls that could
reflect sound in the northwest to northeast directions. When chillers were initially operated during commissioning, a noise
complaint was received from a neighborhood residence approximately 1000 feet (305 m) to the northwest. The municipal
noise ordinance permitted up to 49 dBA at residential property boundaries when tonality and nighttime corrections were
applied, as described in Introduction above. A noise study was requested to determine chiller noise conditions due to the
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
chiller installation and to develop and recommend noise mitigation, if necessary to conform to noise ordinance and/or
resolve the residential noise complaint. Post-installation performance validation measurements were to be conducted also.
52
ASHRAE Transactions
a) b)
Figure 3 (a) Plan view of chiller installation and (b) aerial view of school and nearby neighborhood(s).
Existing Conditions: On-Site Observations and Measurement Results

Acoustical measurements were conducted on-site to determine chiller noise level and spectrum for normal ambient
(without chiller contributions) and for part- and full-load chiller operations using a handheld, ANSI Type 1, 2-channel
spectrum analyzer and two Type 1, 1/3-octave, 1-channel analyzers. Equivalent levels (Leq) and statistical percentile levels
(Ln) were obtained 5 feet (1.5 m) above ground at locations approximately 60 feet (18 m) from the chillers, at the school
property boundary approximately 300 feet (92 m) to the northwest and approximately 1000 feet (305 m) away, in the
neighborhood near complaint origin, as shown in Figure 3(b). Other measurements were made within and immediately
adjacent to the mechanical yard enclosure to determine and compare spectra of the chillers with some indications of relative
contributions from compressors and condenser fans.
a) b)
Figure 4 (a) Noise levels near chillers and in neighborhood and (b) view of mechanical yard open end (chain link).
Findings and Recommendations

Measurements indicated that chiller noise levels below 49 dBA in the neighborhood conformed to municipal
ordinance requirements. Although it is not easily seen in the 1/3 octave spectrum, narrowband measurements showed some
tonal peaks at various frequencies, which appeared to represent the new chillers and possibly other air conditioning
equipment in the neighborhood. The spectrum near the chillers was very tonal with most significant prominent frequencies
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`
$6+5$(

in 315-Hz and 125-Hz 1/3 octaves. Projected to the school property boundary, an approximate 49-51 dBA level only
marginally conformed to the ordinance. Observations and measurements on and near the chillers indicated broadband
noise from condenser fans and tonal noise and vibration from chillers. Concrete retaining walls around three sides of the
chiller mechanical yard were contributing to a reverberant build up around the chillers.
Noise containment jacket or enclosure was recommended for the compressors. (The equipment had been ordered
with noise jackets, but only temporary enclosures formed out of duct board, not noise barrier material, had been installed.)
Pulsation damper or diffuser for compressor discharge was recommended. Resilient connections between refrigerant piping
and chiller chassis/frame was recommended, and to reduce radiated sound, damping was recommended on interior side of
sheet metal side panels of chiller. Acoustical absorption treatment (suitable for outdoors) was recommended for the
mechanical yard retaining wall enclosure. A barrier wall with height greater than the chillers was recommended to replace
the chain link fence at the open end of the yard.
Results and Conclusions

With the noise barrier enclosure and compressor noise jacket recommendations implemented, the chiller noise was
significantly reduced, particularly in the northwest direction, so that chiller noise conformed to ordinance limits at the
property boundary and was only 43 dBA in the neighborhood. The noise spectrum outside the mechanical yard noise
barrier wall showed reduced tonality. The original complaint was not adequately resolved, however, indicating incremental
development of receiver sensitivity or receiver hearing a different source and attributing to the chiller installation. The
owner and engineer have proceeded with supplemental implementation of the chiller panel damping and minor
modifications of refrigerant piping attachment mountings to reduce radiated noise.
This installation is substantially improved, with reductions in noise intrusion into classrooms, reduced noise and
tonality at the neighborhood, and it conforms to municipal ordinance limits.
a) b)
Figure 5 (a) Noise Levels near Chillers and in Neighborhood after Enclosure Modifications (b) View of Mechanical
Yard Noise Barrier (note that inside surfaces are acoustically absorptive spray plaster and slotted masonry units).
CASE STUDY #3: SCHOOL C

School C is an elementary school with a new building addition and new air-cooled chiller installation. When
construction on the new addition was nearly complete, the chiller was turned on and began to draw noise complaints from
residential neighbors. The aerial photo shown in Figure 6 illustrates the approximate location and orientation for the new
chiller. The design included a new ~7-foot tall (2.1 m) masonry enclosure wall with security fence. The topography in the
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
54
ASHRAE Transactions
area is relatively flat. Neighboring residential houses are typically single-story. The new school addition would be three
stories tall once completed. Neighbors that were exposed to noise disturbance from the new chiller had already been
subjected to noise from two existing chillers, without any complaints. Therefore, during design and equipment selection,
adverse conditions were not anticipated from the additional chiller. The new chiller was specified to include a noise control
enclosure or lagging around the compressors to contain radiated compressor noise. However, shortly after the new chiller
was operational, and before the compressor enclosure was installed, neighbors began to complain about the new chiller’s
noise.
Figure 6 Site aerial view and illustration showing new and existing chillers, the new building under construction,
and residential neighbors.
Existing Conditions: On-Site Observations and Measurement Results

A site visit was arranged to conduct observations and noise measurements with the two existing chillers and one new
chiller at different compressor loading and staging conditions. Measurement results were reviewed to determine spectrum
and level of the noise from the new chiller, relative to background existing conditions and city ordinance limits. A particular
focus was to identify the chiller noise levels near locations of complaints for the purpose of determining mitigation required
to mitigate the disturbance complaints. Measurements included equivalent levels (Leq) with the microphone located
approximately 5 feet (1.5 m) above ground. During that first site visit, the following findings and observations were made:
1. The new chiller noise received at complainant’s property was about 64 dBA, approximately 15 dBA louder than
allowed by city ordinance during nighttime hours.
2. The new chiller’s noise was tonal.2 Results showed prominent tonality for the new chiller in the 315-Hz 1/3 octave
band, related to the compressors. The 63-Hz, 125-Hz, and 630-Hz 1/3 octave bands also indicated contributions
from various harmonics apparently related to condenser fans (63 & 125 Hz) and compressors (315 & 630 Hz).
3. The existing chillers’ noise was also tonal, but not as prominent at residential property lines as the new chiller.
4. The shape of the mechanical yard enclosure walls around the new chiller appeared to contribute reflections to
reinforce the chiller noise and direct it over the walls towards neighbors as shown in Figure 7(a).
5. The orientation and height of the building wall and roof soffit shown in Figure 7(b) also contribute reflections to
reinforce chiller noise, directing it over walls2 towards the neighbors as shown in Figure 7(a).
6. Surface radiation from parts of the chiller produce audible airborne noise that reinforces noise radiated directly
from the compressors.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

a) b)
Figure 7 (a) Site aerial view and illustration showing chiller noise reflections off new building and chiller walls and
(b) the new chiller (in the foreground) with new school building wall and soffit above and behind it.
Existing Conditions: Recommendations

Based on evaluations of existing conditions, the following noise mitigation measures were recommended and
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
coordinated for implementation:

1. Complete the installation of planned noise control enclosure or lagging around the compressors to contain greatest
radiated tonal noise.
2. Install and extend a taller solid barrier wall along north and west sides of the existing chiller enclosure. Add or
extend noise barrier wall around the north end of the chiller where the open chain-link fence exists now, as shown
in Figure 8.
3. Provide acoustical seals, overlapping barrier materials to seal joints at entry/access doors as shown in Figure 8b.
4. Apply acoustically absorptive surface finish to the existing masonry enclosure wall surfaces facing the chiller.
a) b)
Figure 8 Illustration of recommendation to (a) increase the west chiller wall height and (b) enclose the open
fencing on the north end for noise mitigation.
Post Construction: Observations and Measurement Results

After noise mitigation measures were implemented to extend barrier walls and add acoustically absorptive surfaces,
neighbors continued to complain, indicating they did notice a reduction in high-frequency “whine” or what was described
by neighbors as “machine” noise but did not notice much reduction in a low-frequency “rumble” or “hum” that was still
annoying to them. To evaluate the results for comparison with city ordinance requirements and complainants’ descriptions,
additional noise measurements were conducted on site. Results indicated the following:
56
ASHRAE Transactions
1. The new chiller noise received at complainant’s property was reduced by about 5 dBA, almost achieving the city’s
allowance for daytime noise, but still approximately 10 dBA louder than allowed by city ordinance at night.
2. As shown in Figure 9b, a significant noise reduction of 5 to 10 decibels was achieved in the frequency spectrum
from 80 Hz to 1000 Hz, reducing the tonality of the equipment noise received at neighbors’ properties.
3. Very little noise reduction was achieved for condenser fan tone in the 63-Hz 1/3 octave band. It was assumed the
63-Hz tone represented the prominent low-frequency “rumble” that neighbors continued to be annoyed by.
a) b)
Figure 9 Chiller noise received on neighboring property from (a) the existing chillers and (b) the new chiller before
and after the noise mitigation was installed.
Post Construction: Further Recommendations

For additional noise mitigation, the following additional measures were recommended, and are currently being
coordinated for implementation. To reduce fan noise, apply silencers supported separately above the condenser fans with a
tight clearance to the top surface of the chiller (but not touching). This measure would reduce low frequency tonal noise
from condenser fans but would not be likely to significantly reduce the overall A-weighted (dBA) chiller noise level with
respect to city noise requirements. However, since the primary complainants indicated their concern was with the lower-
frequency “rumbly” noise rather than the higher-pitched “whine,” it is hoped that this measure would improve the sound
spectrum, improving the sound quality with respect to the neighbors’ specific concerns, regardless of dBA city noise limits.
CONCLUSION
Air-cooled chiller noise is often tonal. Noise reduction that is designed for outdoor chiller equipment to achieve a
reduction in A-weighted (dBA) noise level will not necessarily reduce the perceived impact on sensitive receivers. Designers
should consider equipment tonality, noise radiating from equipment surfaces and pipes, reflecting building and wall
surfaces, topography and lines of sight to potential noise receivers, and existing ambient noise conditions when evaluating
the potential impacts of new outdoor chiller equipment on project sites and at neighboring properties.
REFERENCES
1 Code of Ordinances, Part II, Ch. 21, Article III, Division 1, Sec. 21-52. – Noise nuisance enumeration (a.9) and (b) Table
1, San Antonio, Texas, 2010
2 ASHRAE. 2011. ASHRAE Handbook-HVAC Applications, Ch. 48 – Noise and Vibration Control. Atlanta: American
Society of Heating Refrigeration and Air Conditioning Engineers, Inc.

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

SA-12-C008
Field-Test of the New

ASHRAE/CIBSE/USGBC Performance
Measurement Protocols: Intermediate and
Advanced Level Indoor Environmental
Quality Protocols
Hyojin Kim Jeff S. Haberl, PhD, PE

Student Member ASHRAE Fellow ASHRAE
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
ABSTRACT
This paper presents the preliminary results of an effort to develop and apply a field test to evaluate the
ASHRAE/CIBSE/USGBC Performance Measurement Protocols (PMP) Intermediate and Advanced Level Indoor
Environmental Quality (IEQ) performance protocols in a case-study building in Central Texas. In accordance with the
Intermediate and Advanced Level IEQ protocols, a comprehensive instrumentation cart was developed to collect continuous
data from several IEQ-related parameters, including: air and globe temperatures at four levels, humidity, air speed, CO2,
VOCs, horizontal and vertical illuminance level, as well as A-weighted and C-weighted sound pressure levels. Using the
developed instrumentation cart, continuous IEQ measurements were conducted in several office spaces of a case-study
building from July to September 2011. In addition to the instrumented measurements, an occupant ‘right-now’ survey of
thermal sensation and comfort was concurrently conducted. The data collected were then analyzed and compared with the
appropriate PMP recommended benchmarks.
INTRODUCTION
Research into building Indoor Environmental Quality (IEQ) performance measurements began at the beginning of 20th
century, and professional societies were launched at the same time for the respective fields of study (i.e., thermal comfort,
Indoor Air Quality (IAQ), lighting and acoustics). Since then a number of studies on building IEQ performance
measurements have been conducted, and guidelines, protocols and standards on measurement procedures have been
published in the different fields. However, current efforts toward standardizing the measurement procedures were made
separately for each field. Hence, there remained a need for an integrated protocol since an occupant’s assessment of comfort
is often influenced by their experience in a complex indoor environment including: thermal, indoor air quality (IAQ), lighting,
and acoustics. One recent effort to develop such protocols is the ASHRAE/CIBSE/USGBC Performance Measurement
Protocols (PMP) for Commercial Buildings (ASHRAE 2010a).
The purpose of the PMP is to provide a standardized set of protocols for measuring and comparing the operational
performance of occupied commercial buildings including: energy use, water use, and IEQ. The PMP has been developed at
three levels of cost/accuracy, Basic (Indicative), Intermediate (Diagnostic), and Advanced (Investigative), for the following
H. Kim is a Ph.D. candidate in the Department of Architecture and is a research engineering associate at the Energy Systems Laboratory,
Texas A&M University, College Station, TX. J.S. Haberl is a professor in the Department of Architecture and is an associate director at the
Energy Systems Laboratory, Texas A&M University, College Station, TX.
58
ASHRAE Transactions
six performance categories: energy use, water use, thermal comfort, indoor air quality, lighting and acoustics. The issues with
implementing the PMP basic level protocol identified from a field test were reported by Kim and Haberl (2012). This paper
presents the preliminary results to develop and apply a field test to evaluate the PMP Intermediate and Advanced Level
Indoor Environmental Quality (IEQ) performance protocols in a case-study building in Central Texas.
METHODOLOGY
Based on a review of the PMP, a field test was developed and applied to data taken from a large office case-study
building. Table 1 presents the performance metrics selected and modified in this field study, which were required at the
intermediate and advanced levels. The metric selections and modified approaches were made based on the results of the basic
level field test (Kim and Haberl 2012) while also considering significance and practicality of the measures. For example, a
diagnostic lighting satisfaction survey was not considered in this study since the case-study building’s lighting performance
was consistently well above the average benchmarking scores based on the basic level field test results. A spot measurement
of full-grid illuminance and luminance was not performed partly due to the practical applicability of these measures (i.e., low
availability and high cost requirements). On the other hand, to improve a limitation of spot measurement observed from the
basic level field test1, continuous measurements of horizontal and vertical illuminance as well as A-weighted and C-weighted
sound pressure levels were performed in this study.
To accomplish the selected measurements, a comprehensive instrumentation cart was developed to collect continuous
data from several IEQ-related parameters while logging the occupancy using an occupancy sensor 2 , including: four air
temperatures, four globe temperatures, humidity, air speed, CO2, volatile organic compounds (VOCs) 3 , horizontal and
vertical illuminance, and A-weighted and C-weighted sound pressure levels (SPL), as shown in Table 2 and Figure 1. To
construct the cart, previous studies were reviewed to identify the instruments that had been used to measure the
corresponding metrics. The survey of currently available equipment on the market, included sensor type, accuracy, resolution,
response time, and power consumption, along with calibration and cost information to determine the appropriate sensors4.
Using the portable instrumentation cart, continuous IEQ measurements were conducted in eleven office spaces of a
case-study building from July to September 2011. The eleven offices were selected based on the results of the basic level IEQ
Table 1. IEQ Performance Metrics Required at the PMP Intermediate and Advanced Levels
Type of Measures Area1) Required Performance Metrics Tested
Descriptive Specific IEQ-related building/system characteristics Yes
Intermediate
TC ‘Right-now’ thermal comfort survey Yes

Subjective
L Diagnostic lighting satisfaction survey No
Instrumented TC (C) Temp., humidity, MRT, and air speed Yes
(C) Continuous IAQ (C) CO2 concentration Yes
measurement L (S) Full grid illuminance and luminance No (Modified)
(S) Spot Meas. A (S) Background noise and reverberation time. No (Modified)
Descriptive Detailed IEQ-related building/system characteristics Yes
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Advanced
Subjective TC Local comfort survey for asymmetrical or transient environment No

Instrumented TC (C) Temp., humidity, MRT, and air speed w/ detailed spatial resolution Yes
(C) Continuous IAQ (C) CO2, PM2.5, and TVOCs Partially
measurement L (S) High dynamic range (HDR) photography Yes
(S) Spot Meas. A (S) Speech privacy, speech communication, and sound/vibration isolation No
1)
TC: Thermal Comfort; IAQ: Indoor Air Quality; L: Lighting; A: Acoustics

1
IEQ spot measurements have one major limitation: it is hard to catch dynamic responses of IEQ with spot measurements. Therefore, in light of rapidly
evolving field instruments and data loggers, the PMP should consider providing advice about how to collect and analyze dynamic measurements across the
six categories (Kim and Haberl 2012).
2
The occupancy was measured using a dual-technology occupancy sensor that employs both passive infrared and ultrasonic technologies.
3
The VOCs were measured using an instrument that produces a voltage signal that increases as the VOCs level increases. This instrument was selected to
quantify the relative amount of VOCs in the measurement space by detecting most solvent based VOCs, including acetone, benzene, diacetone alcohol,
formaldehyde, methylene chloride, methyl ethyl ketone, perchloroethylene, toluene, and trichloroethylene.
4
The results of market survey will be available in Kim (2012).
$6+5$(

Table 2. IEQ Continuous Measurement Parameters and Instrumentation Specifications
Parameter Meas. Height (m (in.)) Sensor Type Accuracy Range
Air temperature (C, F) T-type thermocouple
0.1, 0.6, 1.1, 1.7 T-type thermocouple inside ±0.5°C (0.9°F) or 0 to 350°C
Globe temperature (C, F) (4, 24, 43, 66) a 38 mm diameter table tennis 0.4% reading (0 to 662°F)
ball painted black or gray
Relative humidity (%) 1.1 (43) Capacity polymer sensor ±2.0% RH (0 to 90% RH) 0 to 100% RH
0.13 to 5.0 m/s
Air speed (m/s, fpm) 1.1 (43) Hot-film anemometer ±2.0% reading
(25 to 1,000 fpm)
Single beam absorption ±50 ppm or
CO2 (ppm) 1.0 (39) 0 to 5,000 ppm
infrared sensor ±5% reading
Heated metal oxide 0 to 140 ppm5
VOCs (mV) 1.0 (39) Not specified
semiconductor (HMOS) (0 to 1,300 mV)
Horizontal illuminance (lux, fc) 0.76 (30)1) 0 to 100 klux
Silicon photovoltaic detector CIE curve within ± 5%
Vertical illuminance (lux, fc) 1.1 (43) (0 to 10 kfc)
LAeq3) (dBA) 1.2 (47)2) Electret condenser microphone ±1.5 dB 30 to 130 dB
LCeq3) (dBC) 1.2 (40)2) Electret condenser microphone ±1.5 dB 35 to 90 dB
1)
At the height of work plane
2)
At the height of seated occupant’s ear
3)
A-weighted equivalent sound pressure level (dBA); and C-weighted equivalent sound pressure level (dBC)
assessment survey (Kim and Haberl 2012). Four offices were chosen where occupants were
dissatisfied with their IEQ environments and seven additional offices were chosen where
occupants were satisfied. The measurements were made over a one week period in each office with
a scan interval of 10 seconds and three different data logging intervals (1 minute, 5 minutes, and
15 minutes). The cart was placed as close to the occupant as possible while ensuring enough space
for occupants to minimize disturbances. In addition to the instrumental measurements, an occupant
‘right-now’ survey of thermal sensation, comfort, acceptability, and preference was conducted
concurrently. The occupants were asked to repeat the survey twice a day over the measurement
period. Once the data were collected, the performance metrics required in the PMP were calculated
and compared with the appropriate benchmarks.
Case-Study Building Description
The seven-story, case-study building is a typical office building in College Station, Texas.
The building was constructed in 1992 and currently serves as one of the Texas A&M University
facilities occupied by about 323 employees. The conditioned floor area of the building is 11,516
m2 (123,960 ft2). The windows are non-operable, double-pane, tinted glazing with a window-to-
wall ratio of approximately 40%. The space usage includes: offices, meeting rooms, and a gym.
Figure 1: The building is served by 17 single duct variable air volume (SDVAV) air handling units (AHUs)
Comprehensive IEQ with variable frequency drives (VFDs) and two 100% outside air AHUs that provide the SDVAV
Monitoring Cart
units with conditioned outside air. There are 230 VAV terminal boxes with hot water reheat coils.
RESULTS
Thermal Comfort
The evaluation of the thermal comfort performance metrics was performed by comparing the results to the ASHRAE
Standard 55-2010 (ASHRAE 2010b). Figure 2 presents the statistical distributions (max, 95th, median, 5th, min) of
measurement results during the occupied periods for the eleven offices (four dissatisfied and seven satisfied) with the
appropriate benchmarking criteria (if applicable), including operative temperature, predicted mean vote (PMV), predicted
percentage of dissatisfied (PPD), humidity ratio, relative humidity, air speed, vertical air temperature difference between head

5
From the partial calibration for perchloroethylene,specified in a manufacturer’s note (Eco Sensors, Inc. 2011).
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---

and ankles and operative temperature drifts/ramps6. The X-axis of the plots consists of a 7-point thermal satisfaction scale (3:
very satisfied, 0: neutral, -3: very dissatisfied) surveyed at the basic level field test for each office.
The median operative temperatures of the seven offices were lower than the lower boundary for ASHRAE 55-2010
Section 5.2.1.1 Graphic Comfort Zone Method’s summer (0.5 clo) comfort temperature. When the Computer Model Method
(i.e., PMV model) in Section 5.2.1.2 of ASHRAE 55-2010 is used, nine offices were out of the acceptable comfort range for
more than 50% of occupied hours during a measurement week. The median PMVs of nine of the eleven offices were lower
than í0.5, which is the lower boundary for Computer Model Method’s acceptable PMV range, and the median PPDs of the
same nine offices exceeded the 10% criterion, which means more than 10% of people are predicted to be dissatisfied due to
local thermal discomfort in these offices. Conversely, the humidity environments of all eleven offices were maintained well
below 0.012 of a humidity ratio, which is the high limit specified in the Section 5.2.2 of the ASHRAE 55-2010. Thus the
humidity environments of all eleven offices conform to ASHRAE 55-2010.
The thermal stratification was examined using the temperature difference between head (1.1 m, 43 in.) and ankles (0.1
m, 4in.). For all eleven offices, the vertical temperature differences were lower than the allowable differences in Section
5.2.4.3 of ASHRAE 55-2010, which is 3.0°C (5.4°F). Lastly, the temperature fluctuations with time were examined using the
change in operative temperature during a 15 minutes period. Although the two offices approached the limit, all eleven offices
maintained the operative temperature variations less than 1.1°C (2.0°F), which is the limit based on Section 5.2.5.
When comparing the thermal conditions of the offices expressing dissatisfaction (four offices) versus the offices
expressing satisfaction (seven offices), the differences are not noticeable. However, the median PMV of one dissatisfied
office is í0.3 PMV, which is the highest thermal sensation among the eleven offices. The occupant in this office also
expressed warm discomfort. For about 93% of occupied hours, the thermal conditions of this office were inside the comfort
range specified in the ASHRAE 55-2010 based on the PMV-PPD, but the occupant preferred lower temperatures in the
‘right-now’ surveys which were answered seven times during a measurement period. In a time-of-day analysis, increasing
temperatures due to solar radiation from the window were observed in the morning in this east-facing office, which also
caused temperature drifts near the maximum acceptable limit of the standard7.
The lowest median PMV (í1.2 PMV and 38% PPD) is found in one of the dissatisfied offices (third line from the left in
the PMV distribution plot), and the occupant complained about cool discomfort due to high air movement and drafts. The air
speed distribution in this office during a measurement week revealed higher air speeds compared to other offices, and in the
results of ‘right-now’ survey, the occupant preferred less air movement. However, since the current version of ASHRAE 55-
2010 revised the provisions on draft by excluding its de facto draft limit (0.2 m/s, 40 fpm) in the previous 2004 version and
adding a limit (0.15 m/s, 30 fpm) only for the conditions below 22.5°C (72.5°F) operative temperature, there are no
applicable air speed criteria for this office with an operative temperature higher than 22.5°C based on the ASHRAE 55-2010.
The thermal environments of two other dissatisfied offices were similar to those observed in satisfied offices. The
occupant who answered “very dissatisfied” in the basic level IEQ survey complained about the inaccessible thermostat which
was adjusted by other people in another office and preferred lower temperatures in the ‘right-now’ surveys. This may indicate
individual differences beyond two well-known personal variables, clothing and activity. The other occupant also complained
about the inaccessible thermostat and cold temperatures in winter which could not be verified in this measurement.
Indoor Air Quality
The evaluation of the IAQ performance metrics was performed by comparing the survey results to the ASHRAE
Standard 62.1-2010 (ASHRAE 2010c) as required in the PMP. Figure 3 presents the statistical distributions of measurement
results during the occupied periods for eleven offices (two dissatisfied, one neutral, and eight satisfied) with the appropriate
benchmarking criteria, including CO2, VOCs, and outdoor ventilation rate8. The X-axis of the plots consists of a 7- point IAQ
satisfaction scale surveyed at the basic level field test.

6
Air and globe temperatures measured at the 0.6 m were used to calculate operative temperature, PMV, and PPD as per Section 7.2.2 of ASHRAE 55-2010.
7
Additional discussions on the advantages of a time-of-day analysis will be available in Kim (2012).
8
Outdoor ventilation rate was estimated from the daily maximum CO2 concentrations using equilibrium CO2 analysis method per Section 10 of the ASTM
D6245-2007 (ASTM 2007).
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,
$6+5$(

60
Predicted Percentage of Dissatisfied

30 86.0 3
Dissatisfied Satisfied Dissatisfied Satisfied Dissatisfied Satisfied
Predicted Mean Vote (PMV)

Operative Temperature (F)
2 50
Operative Temperature (C)
28 82.4
ASHRAE55Ǧ2010Graphiccomfortzonemethod: 1 40
0.5 clocomforttemperatureat40%RH
Acceptable
(PPD %)
26 78.8
Acceptable
Range
ASHRAE55Ǧ2010Ǧ0.5<PMV<+0.5
Range
0 30
24 75.2
-1 20
22 71.6 10
Acceptable
-2
Range
ASHRAE55Ǧ2010PPD<10
20 68.0 -3 0
-4 -3 -2 -1 0 1 2 3 4 -4 -3 -2 -1 0 1 2 3 4 -4 -3 -2 -1 0 1 2 3 4
80 1.0 200
Humidity Ratio (kg H2O / kg dry air)
0.014 Dissatisfied Satisfied Dissatisfied Satisfied Dissatisfied Satisfied

0.9 180
0.012 70
Relative Humidity (%)

ASHRAE55Ǧ2010Graphiccomfortzonemethod: 0.8 160
below0.012humidityratio
Air Speed (fpm)

Air Speed (m/s)
0.010 0.7 140
60
0.008 0.6 120
Acceptable
50 0.5 100
Range
0.006 0.4 80
40 0.3 60
0.004
0.2 40
0.002 30
0.1 20
0.000 20 0.0 0
-4 -3 -2 -1 0 1 2 3 4 -4 -3 -2 -1 0 1 2 3 4 -4 -3 -2 -1 0 1 2 3 4
2.0 3.6
Vertical Temperature Difference (F)
Vertical Temperature Difference (C)
4 Dissatisfied Satisfied 7.2

1.8 Dissatisfied Satisfied 3.2
Operative Temperature
3 5.4
ASHRAE55Ǧ2010Vertical airtemperature 1.6 2.9
Drifts/Ramps (C)
2 difference<3C(5.4F) 3.6
Drifts/Ramps (F)
1.4 2.5
ASHRAE55Ǧ2010Operative temperaturechange
1 1.8 1.2 allowed<1.1C(2.0F)in 15minutes 2.2
Acceptable
Range
0 0.0 1.0 1.8

0.8 1.4
Acceptable
-1 -1.8
Range
0.6 1.1
-2 -3.6
0.4 0.7
-3 -5.4 0.2 0.4
-4 -7.2 0.0 0.0
-4 -3 -2 -1 0 1 2 3 4 -4 -3 -2 -1 0 1 2 3 4
Satisfaction: Thermal Comfort Satisfaction: Thermal Comfort
(-3: Very dissatisfied, 0: Neutral, 3: Very satisfied) (-3: Very dissatisfied, 0: Neutral, 3: Very satisfied)
Figure 2: Statistical Distributions (Max, 95th, Median, 5th, Min) of Thermal Comfort Measurement Results.
The following observations can be made regarding the results of the IAQ evaluation. The maximum CO2 concentrations
of all eleven offices were below 950 ppm, corresponding to 8.5 L/s (17 cfm) per person of outdoor air rate which is the
minimum ventilation rate for office spaces in the ASHRAE 62.1-20109. The VOCs levels of all eleven offices were far below
the cautionary threshold limit value (650 mV) specified in a manufacturer’s note (Eco Sensors, Inc. 2011). In several offices,
there were some events in which VOCs levels increased, but the increase was short lived10. The outdoor air flow rates are
displayed with the uncertainty bar, and the calculated rates ranged from 11 to 17 L/s (23 to 36 cfm) per person, which
exceeded the ASHRAE’s minimum ventilation rate for offices (8.5 L/s (17 cfm) per person of outdoor air rate).
No differences were identified between the measurement of CO2 and VOCs in offices where occupants expressed
dissatisfaction (two offices) versus satisfaction (eight offices). The two dissatisfied offices were ventilated adequately based
on the CO2 measurement results, however, the occupants complained about stuffy and stale air due to poor circulation within
a room. Figure 4 illustrates vertical temperature profiles of two dissatisfied (dotted lines), one neutral (dashed line), and eight
satisfied offices (solid lines) using the median temperatures at four heights during the occupied periods measured over one
week. For the two dissatisfied offices, hot air stratification was observed at 1.1 m (43 in.), possibly indicating poor air
circulation in these two offices that might prevent the delivery of fresh air in the breathing zones. In addition, the measured
air speeds at 1.1 m (43 in.) in these two offices remained constant and was lower when compared to other offices (two north
offices in a dissatisfied group in Figure 2), which also implies poor circulation.

9
950 ppm of indoor CO2 concentration threshold limit was estimated using the Informative Appendix C of the ASHRAE 62.1-2010 with the following two
assumptions: 350 ppm of outdoor CO2 concentration and 0.0051 L/s (0.011 cfm) of indoor CO2 generation rate.
10
These one-time events were likely to occur during the use of office stationery supplies or cleaning products that contain VOC gases and vapors, including
felt marker pens, hand sanitizers, and floor cleaners.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
62
ASHRAE Transactions
1,200 700 Dissatisfied Satisfied 24 51
Dissatisfied Satisfied Dissatisfied Satisfied
Daily Outdoor Air Flow Rate

Daily Outdoor Air Flow Rate
600
CO2 Concentration (ppm)
1,000 Manufacturer'srecommendations 20 42
Cautionaryat 650mV,Hazardousat1,040 mV
(cfm per person)

500
(L/s per person)
Acceptable
800 16 34
VOCs (mV)
Range
400
Acceptable
Range
600 12 25
Acceptable
Range
300
400 8 17
200 ASHRAE62.1Ǧ2010 VentilationRateProcedure
ASHRAE62.1Ǧ2010 InformativeAppendixC: MinimumOAratewithdefaultoccupantdensity
200 1.1met,8.5L/s(17cfm)perpersonofOA, 100 4 8
(20m2/person)=8.5 L/s(17cfm)perperson
350ppmOACO2 level
0 0 0 0
-4 -3 -2 -1 0 1 2 3 4 -4 -3 -2 -1 0 1 2 3 4 -4 -3 -2 -1 0 1 2 3 4
Satisfaction: IAQ Satisfaction: IAQ Satisfaction: IAQ
(-3: Very dissatisfied, 0: Neutral, 3: Very satisfied) (-3: Very dissatisfied, 0: Neutral, 3: Very satisfied) (-3: Very dissatisfied, 0: Neutral, 3: Very satisfied)
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 3: Statistical Distributions (Max, 95th, Median, 5th, Min) of CO2 and VOCs Measurement Results and Daily
Outdoor Air Flow Rate with Uncertainty Bars
Neutral Satisfied offices Dissatisfied offices
Lighting
Vertical Air Temperature Differentials
Relative to the 0.6 m Temperature (F)
The evaluation of the lighting performance metrics was performed by comparing
-1.4 -1.1 -0.7 -0.4 0.0 0.4 0.7 1.1 1.4
2.0 79
1.9 the survey results to the Table 3-9 in the PMP and the ISO Standard 8995:2002 (ISO
75
1.8 71
1.7 66
67
1.6 2002). Figure 5 presents the statistical distributions of measurement results during the
63
Height (in.)
Height (m)
1.5 59
1.4
1.3 occupied periods for the eleven offices (two dissatisfied and nine satisfied) with the
55
51
1.2 47
1.1
1.0
appropriate benchmarking criteria, including horizontal illuminance and vertical
43
43
39
0.9 35
0.8 illuminance. The X-axis of the plots consists of a 7-point lighting level satisfaction
31
0.7 28
0.6 24
scale surveyed at the basic level field test.
24
0.5 20
0.4 16
0.3
0.2
The measured median horizontal illuminance of four offices were higher than the
12
8
0.1 44
0.0 recommended value for task areas (500 lux, 50 fc) in both the PMP and ISO Standard
0
-0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8
8995:2002. In a second comparison to the value for the immediate surrounding areas
Vertical Air Temperature Differentials
(within 0.5 m (20 in.) from the task area) in the ISO Standard 8995:2002, ten offices
Relative to the 0.6 m Temperatue (C)
Figure 4: Vertical Air Temperature complied with the provision (i.e., the median illuminance higher than 300 lux, 30 fc).
Profiles during the Occupied Periods One office showed very low horizontal illuminance with the median illuminance
around 73 lux (6.8 fc). The ceiling light fixtures in this office remained turned off during the most of the measurement week,
but the occupant was satisfied with his lighting environment that was maintained by daylight only. Since the task areas were
closer to the window than the measurement location, the occupant was expected to be exposed to a brighter environment.
The median vertical illuminance of ten offices was higher than 50 lux (5 fc) which is recommended in the PMP. A large
variation in vertical illuminance was found in two offices where the blinds were partially open, and in one of these offices,
the computer monitors face a window, which indicates a high possibility of glare. In an analysis of this west-facing office
using HDR photography taken around noon on a typical summer weekday, the luminance ratio between the task and monitor
was approximately1:56 (2.5 cd/m2:140 cd/m2, 0.7 fL :41 fL) which far exceeded the 1:3 limit in the PMP11.
No differences were identified from the measurements of illuminance between the offices where occupants expressed
dissatisfaction (two offices) and satisfaction (nine offices). One of the dissatisfied occupants complained about darkness in
the task areas because of a hutch over this area. This problem could not be verified using the instrumentation cart since it was
located in the immediate surrounding areas. The cart can be improved by using detachable or remote illuminance sensors
located at the task equipped with wireless devices. Another dissatisfied occupant complained that the room was too bright
with overhead lights but not bright enough with just a desk lamp, which may be due to different preferences on lighting levels
since the measured lighting levels in this office were near the recommended values.
Acoustics
The evaluation of the acoustics performance metrics was performed by comparing the survey results to the Table 3-10 in the
PMP. Figure 6 presents the statistical distributions of measurement results for eleven offices (one dissatisfied, one neutral,
and nine satisfied) with the appropriate benchmarking criteria, including LAeq and difference between LCeq and LAeq12.
The X-axis of the plots consists of a 7-point noise level satisfaction scale surveyed at the basic level field test.

11
Additional discussions on the HDR photography analysis will be available in Kim (2012).
12
To identify low frequency noise components in the measured noise, the difference between C- and A-weighted SPL were calculated instead of octave band
measurements. The LCeq – LAeq difference has been regarded as a simple indicator to estimate low frequency noise annoyance (Berglund et al. 1999).
$6+5$(

1,800 Dissatisfied Satisfied 180 1,800 Dissatisfied Satisfied 180
1,600 160 1,600 160
Horizontal Illuminance (lux)
Horizontal Illuminance (fc)
Vertical Illuminance (lux)
Vertical Illuminance (fc)

1,400 140 1,400 140
1,200 120 1,200 120
1,000 100 1,000 100
800 80 800 80
PMPǦ2010
600 Table3Ǧ9:500lux(50fc) 60 600 60
400 ISO8995:2002 40 400 40
200 500lux,50fc(task) 20 200 PMPǦ2010 20
300lux,30fc (immediatesurroundings) Table3Ǧ9:50lux,5fc
0 0 0 0
-4 -3 -2 -1 0 1 2 3 4 -4 -3 -2 -1 0 1 2 3 4
Satisfaction: Lighting Level Satisfaction: Lighting Level
Figure 5: Statistical Distributions (Max, 95th, Median, 5th, Min) of Lighting Measurement Results
90 Satisfied
40 Dissatisfied Satisfied
Dissatisfied
A-Weighted SPL (Leq(dBA))
80 35
Leq(dBC) í Leq(dBA)
70 30
60 55dB(A)line 25dBdifferenceline
25
50
20 20dBdifferenceline
Acceptable
40
Range
15dBdifferenceline
15
30
PMPǦ2010Table3Ǧ10: 10
20 Idealat30dB(A),
Maximumat40dB(A)forprivateoffices 5
10
0 0
-4 -3 -2 -1 0 1 2 3 4 -4 -3 -2 -1 0 1 2 3 4
Satisfaction: Noise Level Satisfaction: Noise Level
Figure 6: Statistical Distributions (Max, 95th, Median, 5th, Min) of Acoustics Measurement Results
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Of eleven offices, three offices met the criteria for background noise specified in the PMP with a minimum LAeq lower
than 40 dBA which corresponds to the maximum allowable background noise level. The median LAeq of three offices were 5
dB or more above the background noise, which might be partly affected by the use of portable fans in these three offices. Ten
offices had a median LCeq – LAeq difference higher than 15 dB, and two of them had a median difference over 25 dB. Using
the method by the Swedish National Board of Health and Welfare (1996) and the Swedish Royal Board of Building (1992) 13,
the measured noise in ten offices can be considered as low frequency noise, which likely annoyed the occupants.
When comparing the acoustics environments of offices between satisfied and dissatisfied, the dissatisfied office had a
relatively high LCeq – LAeq difference. The LCeq of one dissatisfied office was higher than LAeq by 24 dB or more for
more than 50% of occupied hours during a measurement week, which indicates a possibility of annoyance related to low
frequency. In the subjective IEQ survey, the occupant in this office complained about noise made by people in neighboring
areas and rooms. The measured background sound level of this office was the third lowest level with a minimum LAeq of 39
dBA, which complies with the criteria for background noise levels in the PMP. However, due to individual differences, there
is still likelihood that the background sound was not enough to mask intruding noise for the occupant in this office.
SUMMARY AND RECOMMENDATIONS

This paper presents the preliminary results of an effort to develop and apply a field test to evaluate the ASHRAE PMP
intermediate and advanced level IEQ protocols in a case-study office building in central Texas. Overall, the PMP IEQ
protocols accomplished its goal of providing a standardized set of protocols for measuring and comparing the IEQ
performance of a building in one document, including thermal comfort, IAQ, lighting and acoustics. However, some of the
performance categories failed to provide specific measurement protocols (i.e., instrumentation, sampling methods,
measurement length and time of day, data intervals, etc.) as well as analysis methods.
Observations from the field test are as follows with the corresponding recommendations:
1) One of the main benchmarks to evaluate thermal comfort performance in the PMP is the ASHRAE Standard 55
compliance provisions. Unfortunately, ASHRAE Standard 55-2004 and 2010 do not have any criteria on lower humidity

13
Based on some Swedish recommendations (Swedish National Board of Health and Welfare 1996; Swedish Royal Board of Building 1992), the measured
noise can be regarded as low frequency if the LCeq – LAeq difference exceeds 15-20 dB, and if the difference is greater than 25 dB, there is a chance of
serious low frequency noise annoyance.
64
ASHRAE Transactions
limits although non-thermal comfort issues (i.e., skin drying, eye dryness) are recognized in these standards. Low humidity
has been reported as one of the contributors to sick building syndrome (SBS) symptoms (Mendell 1993 and Menzies and
Bourbeau 1997). Therefore, the PMP should consider providing reliable benchmarks for acceptable low humidity limit.
2) The vertical temperature profile is an effective method to evaluate room air circulation which is related to the stuffiness
and staleness of the room air. In this study, for two dissatisfied offices where the occupants complained about stale air, hot air
stratification was observed at the 1.1 m (43 in.) height, indicating poor air circulation in these two offices that might prevent
the delivery of fresh air in the breathing zones. Therefore, it is recommended that this method be added to the PMP.
3) There were some issues that could not be verified using the instrumentation cart. Several occupants expressed
dissatisfaction although their physical IEQ environments met the criteria specified in the PMP. This may indicate individual
differences in subjective IEQ preferences. Therefore, the PMP should provide advice to the users how to interpret the results
when different results arise between instrumented and subjective measurements.
4) The IEQ spot measurements are helpful to discover possible causes of problems if the measurements are conducted at the
right location and at the same time when discomfort arises, which is not easy to accomplish. Unfortunately, the PMP lighting
and acoustics protocols do not require a continuous measurement while it is required in thermal comfort and IAQ protocols.
In this study, time-of-day variations in measured lighting and acoustics performance metrics were observed. For example, the
measured illuminance level was highest in the afternoon in west-facing office, while the peak level was observed in the
morning in east-facing office. Therefore, the PMP should consider providing advice about how to collect and analyze
dynamic measurements, or should adequately explain the limitation of spot measurements (i.e., which is hard to catch
dynamic responses) with advice about how to interpret the results when continuous measurements are not available.
5) The LCeq – LAeq difference can be regarded as a simple indicator to estimate low frequency noise annoyance when the
frequency analysis is not available due to the practical applicability of octave band measurements, (i.e., low availability and
high cost requirements). Therefore, it is recommended that this method be discussed in the PMP.
6) The current version of the PMP intermediate and advanced level IEQ protocols is not a standalone document which relies
on lots of external standards and protocols. Thus for the PMP to become more than a combination of several standards for
each performance category, it needs to provide details that are sufficient enough for the users to perform the measurements
without referencing other external documents.
REFERENCES
ASHRAE. 2010a. ASHRAE/CIBSE/USGBC Performance Measurement Protocols for Commercial Buildings. Atlanta, GA:
American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc.
ASHRAE. 2010b. ANSI/ASHRAE Standard 55-2010, Thermal Environmental Conditions for Human Occupancy. Atlanta,
GA: American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc.
ASHRAE. 2010c. ANSI/ASHRAE Standard 62.1-2007, Ventilation for Acceptable Indoor Air Quality. Atlanta, GA:
American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc.
ASTM. 2007. ASTM Standard D6245-07, Standard Guide for Using Indoor Carbon Dioxide Concentrations to Evaluate
Indoor Air Quality and Ventilation. West Conshohocken, PA: American Society for Testing and Materials.
Berglund, B., T. Lindvall, and D. Schwela. (eds) 1999. Guidelines for Community Noise. Geneva: World Health Organization.
Eco Sensors, Inc. 2011. VOC Gas Sensor Model C-21 Instructions for Use. Santa Fe, NM: Eco Sensors, Inc.
ISO. 2002. ISO/CIE 8995:2002, Lighting of Work Places. Geneva: International Organization for Standardization.
Kim, H. 2012. Methodology for rating a building’s overall performance based on the ASHRAE/CIBSE/USGBC performance
measurement protocols for commercial buildings. Ph.D. dissertation in preparation, Department of Architecture, Texas
A&M University.
Kim, H. and J. Haberl. 2012. Field-test of the New ASHRAE/CIBSE/USGBC Performance Measurement Protocols for
Commercial Buildings: basic level. ASHRAE Transactions, Accepted for publication.
Mendell, M.J. 1993. Non-specific symptoms in office workers: a review and summary of the epidemiologic literature. Indoor
Air 3:227-236.
Menzies, D. and J. Bourbeau. 1997. Building related illness. New England Journal of Medicine 337:1524-1531.
SNBHW. 1996. SOSFS 1996: 7/E. Indoor Noise and High Sound-Levels. Stockholm: Swedish National Board of Health and
Welfare.
SRBOB. 1992. Low Frequency Noise from Ventilation Installations. Solna, Sweden: Swedish Royal Board of Building.
$6+5$(
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---

SA-12-C009
An Experimental and Simulation Study of

Lighting Performance in Offices with
Automated Roller Shades
Hui Shen Athanasios Tzempelikos, Ph.D.

Student Member ASHRAE Associate Member ASHRAE
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
ABSTRACT
7KLVSDSHULQYHVWLJDWHVWKHDQQXDOGD\OLJKWSHUIRUPDQFHRIDVPDOOSULYDWHRIILFHDQGLQDODUJHRSHQSODQHRIILFHHTXLSSHGZLWKDXWRPDWHG
VKDGHV$QQXDOGD\OLJKWDXWRQRP\DQGXVHIXOGD\OLJKWLOOXPLQDQFHDUHFRPSXWHGDVDIXQFWLRQRIWKHIROORZLQJIDoDGHGHVLJQSDUDPHWHUVZLQGRZ
WRZDOOUDWLRJODVVWUDQVPLWWDQFHVKDGLQJWUDQVPLWWDQFHDQGRULHQWDWLRQIRUWKUHHGLIIHUHQWORFDWLRQVLQWKH8QLWHG6WDWHV$XWRPDWHGRSHUDWLRQRI
UROOHU VKDGHV SUHYHQWV GLUHFW VXQOLJKW WR HQWHU WKH VSDFH LQGHSHQGHQWO\ IRU HDFK RULHQWDWLRQ LQ RUGHU WR DYRLG JODUH 2YHUDOO YDOXHV IRU GD\OLJKW
SHUIRUPDQFHDUHREWDLQHGE\DYHUDJLQJWKHYDOXHVRIDOOHYDOXDWHGSRLQWVLQVSDFHRUE\UHTXLULQJDOOSRLQWVWRPHHWWKHLOOXPLQDQFHUHTXLUHPHQWDW
WKHVDPHWLPH7KHUHVXOWVVKRZWKDWIRUSULYDWHRIILFHVQRUWKDQGHDVWIDFLQJIDFDGHVDOORZKLJKHUGD\OLJKWSURYLVLRQGXHWRVKDGLQJRSHUDWLRQ
GXULQJ ZRUNLQJ KRXUV +RZHYHU LOOXPLQDQFHV EHWZHHQ OX[ IF²DQ LPSRUWDQW LQGH[² UHDFK D PD[LPXP IRU
ZLQGRZWRZDOOUDWLRLQDOPRVWDOOVWXGLHGFDVHV)RURSHQSODQRIILFHVGD\OLJKWDXWRQRP\YDULHVZLWKFOLPDWHDQGVKDGLQJSURSHUWLHV([SHULPHQWV
ZHUHSHUIRUPHG LQD IXOOVFDOH WHVW RIILFH VSDFH WR YDOLGDWH WKHVLPXODWLRQ UHVXOWV ²WKH FRPSDULVRQEHWZHHQPHDVXUHG LOOXPLQDQFHDQG VLPXODWHG
YDOXHVVKRZHGJRRGDJUHHPHQW
INTRODUCTION
(OHFWULFOLJKWLQJHQHUJ\FRQVXPSWLRQDFFRXQWVIRUDQLPSRUWDQWSDUWRIWKHWRWDOHQHUJ\FRQVXPSWLRQRIFRPPHUFLDO
EXLOGLQJV /LJKWLQJHQHUJ\ VDYLQJ WHFKQLTXHV IURP OLWHUDWXUH LQFOXGH GD\OLJKW KDUYHVWLQJ DQG VKDGLQJ FRQWURO 'XERLV DQG
%ORPVWHUEHUJ &RODFR HW DO 2SWLPL]HG IHQHVWUDWLRQ GHVLJQ PD\ LPSURYH H[SORLWDWLRQ RI GD\OLJKW DQG UHVXOW
VLJQLILFDQWVDYLQJVLQHOHFWULFLW\FRQVXPSWLRQIRUOLJKWLQJ3URSHUHYDOXDWLRQRIWKHGD\OLJKWLQJSHUIRUPDQFHRIDVSDFHIRUD
VSHFLILFIHQHVWUDWLRQGHVLJQLVLPSRUWDQW6RPHUHVHDUFKHUVXVHGD\OLJKWLQJIDFWRUVWRHYDOXDWHWKHGD\OLJKWLQJSHUIRUPDQFH
*KLVLDQG7LQNHUZKLFKLVOLPLWHGLQGHVFULELQJWKHGD\OLJKWLQJSHUIRUPDQFHDWDSRLQWLQDVSDFHZLWKRXWVKDGLQJ
GHYLFHXQGHURYHUFDVWVN\/DWHUWZRDGYDQFHGGD\OLJKWPHWULFVZHUHEURXJKWLQWRXVHGD\OLJKWDXWRQRP\5HLQKDUW
7]HPSHOLNRV DQG $WKLHQLWLV DQG XVHIXO GD\OLJKW LOOXPLQDQFHV 1DELO DQG 0DUGDOMHYLF 6KHQ DQG
7]HPSHOLNRV$OOWKHVHGD\OLJKWPHWULFVJHQHUDOO\UHIHUWRRQHSRLQWLQWKHVWXGLHGVSDFH%XWXVXDOO\WKHUHZLOOEH
PRUH WKDQ RQH SRLQWV DUH RI LQWHUHVW VR KRZ WR JHW RQH YDOXH WR HYDOXDWH WKH RYHUDOO VSDFH GD\OLJKW SHUIRUPDQFH LV
Hui Shen is a Ph.D. student and Athanasios Tzempelikos is an Assistant Professor of Architectural Engineering in the School of Civil
Engineering, Purdue University, West Lafayette, Indiana.
66
ASHRAE Transactions
VLJQLILFDQW
7KLVSDSHULQYHVWLJDWHVWKHGD\OLJKWSHUIRUPDQFHRIRIILFHVDVDIXQFWLRQRIZLQGRZWRZDOOUDWLRJOD]LQJDQGVKDGLQJ
RSWLFDO SURSHUWLHV DQG IDoDGH RULHQWDWLRQ IRU WKUHH FOLPDWLF W\SHV LQ WKH 86 $ UHFHQWO\ GHYHORSHG G\QDPLF GD\OLJKWLQJ
VLPXODWLRQPRGHOLVXVHG6KHQDQG7]HPSHOLNRV7ZRVHSDUDWHFDVHVZHUHFRQVLGHUHGIRUDPRUHFRPSOHWHDQDO\VLV
D VPDOO SULYDWH RIILFH ZLWK RQH H[WHULRU IDoDGH DQG D ODUJH RSHQ SODQ RIILFH ZLWK WKUHH H[WHULRU IDFDGHV 7ZR GLIIHUHQW
PHWKRGV DUH XVHG WR REWDLQ WKH RYHUDOO VSDFH GD\OLJKW SHUIRUPDQFH IURP DOO VWXGLHG SRLQWV LQ VSDFH 7KH FRPSDULVRQ
EHWZHHQ WKH RYHUDOO VSDFH GD\OLJKW SHUIRUPDQFHV XVLQJ WKH WZR PHWKRGV \LHOGV D EHWWHU ZD\ WR REWDLQ WKH VSDFH GD\OLJKW
PHWULFVVWXGLHG)LQDOO\IXOOVFDOHH[SHULPHQWVZHUHSHUIRUPHGWRYDOLGDWHWKHGD\OLJKWLQJVLPXODWLRQPRGHOV
DAYLIGHTING CALCULATION MODEL AND PERFORMANCE METRICS

7KH GD\OLJKWLQJ PRGHO FRQVLVWV RI WKH FDOFXODWLRQ RI LQFLGHQW DQG WUDQVPLWWHG LOOXPLQDQFH XVLQJ WKH 3HUH] PRGHO
3HUH]HWDODQGFDOFXODWLRQRIZRUNSODQHLOOXPLQDQFHXVLQJWKHUDGLRVLW\PHWKRG$GHWDLOHGGHVFULSWLRQDERXWWKH
PRGHOLVSUHVHQWHGLQSUHYLRXVVWXGLHV6KHQDQG7]HPSHOLNRV7KLVVWXG\HYDOXDWHVWKHGD\OLJKWLQJSHUIRUPDQFHRI
RIILFHVZLWKDXWRPDWHGLQWHULRUUROOHUVKDGHV7KHVKDGHVRQHDFKIDoDGHDXWRPDWLFDOO\FORVHZKHQLQFLGHQWEHDPUDGLDWLRQ
RQWKHIDoDGHLVKLJKHUWKDQ :P:IW:KHQWKHVKDGHVDUHFORVHGWKHKRXUO\HIIHFWLYHWUDQVPLWWDQFHRIWKH
JOD]LQJVKDGHV\VWHPIRUGLUHFWƴHEDQGGLIIXVHƴHGLOOXPLQDQFHDUHFDOFXODWHGXVLQJ(TVDQGUHVSHFWLYHO\
τ sh
τ eb = τ gb ⋅
1 − ρ gd
b
⋅ ρ sh
τ sh
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
τ ed = τ gd ⋅
1 − ρ gd
b
⋅ ρ sh

ZKHUHƴJEDQGƴJEDUHWKHGLUHFWDQGGLIIXVHJOD]LQJLOOXPLQDQFHWUDQVPLWWDQFHƴVKLVWKHYLVLEOHWUDQVPLWWDQFHRILQWHULRUUROOHU
VKDGHƱVKLVWKHYLVLEOHUHIOHFWDQFHRILQWHULRUUROOHUVKDGHDQGƱJGELVWKHEDFNVLGHJOD]LQJGLIIXVHUHIOHFWDQFH
+RXUO\ZRUNSODQHLOOXPLQDQFHLVILUVWFDOFXODWHGIRUWKHHQWLUH\HDUIRUDOOVWXGLHGFDVHV+RZHYHUWKLVLVD´VWDWLFµ
LQGH[ WKDW FDQQRW EH XVHG WR GHVFULEH WKH RYHUDOO DQQXDO GD\OLJKW SHUIRUPDQFH RI D VSHFLILF GHVLJQ 'D\OLJKW DXWRQRP\
'$ DQ DQQXDO PHDVXUH RI KRZ RIWHQ D PLQLPXP ZRUN SODQH LOOXPLQDQFH UHTXLUHPHQW O[ IF LV D W\SLFDO VHW
SRLQWIRURIILFHVSDFHVFDQEHPHWE\GD\OLJKWDORQHLVFDOFXODWHGIRUWKHZRUNSODQHJULGIRUDOOZRUNLQJKRXUVLQWKH\HDU
DPSP7KLVLQGH[LQFOXGHVWKHLPSDFWRIDOOYDULDEOHVFOLPDWHZLQGRZVL]HDQGSURSHUWLHVURRPJHRPHWU\HWFDQG
FDQDOVREHXVHGWRSUHGLFWOLJKWLQJHQHUJ\VDYLQJLIRQRIIOLJKWLQJFRQWUROLVWREHXVHG7]HPSHOLNRVDQG$WKLHQLWLV
$OLPLWDWLRQRIWKHGD\OLJKWDXWRQRP\LQGH[LVWKDWLWH[FOXGHVLOOXPLQDQFHYDOXHVVOLJKWO\EHORZWKHWKUHVKROGDQG
WKDW LW GRHV QRW FRQVLGHU WKH SUREOHP RI JODUH GXH WR H[FHVVLYH GD\OLJKW +HQFH RWKHU PHWULFV FDOOHG XVHIXO GD\OLJKW
LOOXPLQDQFHV8',DUHDOVRFDOFXODWHG7KHVHGHILQHWKHLOOXPLQDQFHVWKDWIDOOZLWKLQWKHUDQJHRIO[IF
1DELODQG0DUGDOMHYLF,QWKLVVWXG\LQRUGHUWRV\VWHPDWLFDOO\LQYHVWLJDWHWKHZRUNSODQHLOOXPLQDQFHSURILOHV
WKH UDQJH RI XVHIXO GD\OLJKW LOOXPLQDQFH LV IXUWKHU VXEGLYLGHG LQWR WKUHH ELQV O[ IF O[
IFDQGO[IF
)RUHDFKFDOFXODWLRQWLPHVWHSWKHLOOXPLQDQFHIRUWKHSRLQWLVFRPSDUHGZLWKWKHSUHGHILQHGUHTXLUHPHQWVDQGWKHQD
FUHGLWLVJLYHQIRUWKDWWLPHVWHSLIWKHYDOXHLVVXIILFLHQW6XPPLQJXSDOOWKHFUHGLWVIRUDOOFDOFXODWHGWLPHVWHSVDQGWKHQ
GLYLGHLWE\WKHWRWDOFRQVLGHUHGSHULRG\LHOGVWKHGD\OLJKWPHWULFIRUDVSHFLILFSHULRG%XWXVXDOO\PRUHWKDQRQHSRLQWZLOO
EH RI LQWHUHVW DQG WKHQ KRZ WR XVH WKH GD\OLJKW PHWULFV RQ VHYHUDO SRLQWV LQ WKH VWXGLHG VSDFH WR HYDOXDWH WKH RYHUDOO
GD\OLJKWLQJSHUIRUPDQFHRIWKHVSDFH")RUH[DPSOHLIJULGUHVROXWLRQVRIïPïIWDQGïPïIW
DUHWREHXVHGIRUDVSDFHVKRXOGRQHJLYHFUHGLWWRDWLPHVWHSZKHQLOOXPLQDQFHUHTXLUHPHQWDUHPHWDWDOOSRLQWVRUJLYHD
UHVSHFWLYHVFRUHFUHGLWWRHDFKSRLQWDQGWKHQXVHWKHDYHUDJHYDOXHDWWKDWWLPHVWHS"1RFOHDUFODLPLVIRXQGLQOLWHUDWXUH
6KHQ DQG 7]HPSHOLNRV XVH WKH SRLQWDYHUDJHG GD\OLJKW PHWULFV WR SUHVHQW WKH RYHUDOO SHUIRUPDQFH IRU WKH HQWLUH
$6+5$(

VWXGLHGVSDFH
BUILDING DESCRIPTION AND DESIGN VARIABLES

7ZR GLIIHUHQW EXLOGLQJ FRQILJXUDWLRQV DUH FRQVLGHUHG LQ WKLV VWXG\ D W\SLFDO SULYDWH DQG DQ RSHQ RIILFH VSDFH 7KH
GLPHQVLRQVRIWKHSULYDWHRIILFHVSDFHDUHïïPïïIW)LJXUHD,WKDVRQO\RQHH[WHULRUIDoDGHZLWK
RQHZLQGRZ7KHGLPHQVLRQVRIWKHRSHQSODQHRIILFHVSDFHDUHïïPZLWKWKHORQJVLGHIDFLQJVRXWKDQGZLQGRZV
RQWKHVRXWKHDVWDQGZHVWIDFDGHV)LJXUHE)RUWKHWZRRIILFHPRGHOVWKHVXUIDFHUHIOHFWDQFHRIWKHIORRUFHLOLQJDQG
ZDOOVDUHDQGUHVSHFWLYHO\7KHZLQGRZVLOOKHLJKWLVVHWWREHPIWDERYHWKHIORRUVDPHDVWKH
ZRUNSODQHKHLJKW
)LJXUH 6SDFHJHRPHWU\IRUWKHLQYHVWLJDWHGDSULYDWHRIILFHDQGERSHQSODQRIILFH

7KH GHVLJQ YDULDEOHV FRQVLGHUHG LQ WKH PRGHO LQFOXGH EXLOGLQJ ORFDWLRQ ZLQGRZ VL]H H[SUHVVHG DV ZLQGRZWRZDOO
UDWLR ::5 JOD]LQJ YLVLEOH WUDQVPLWWDQFH VKDGLQJ WUDQVPLWWDQFH DQG ZLQGRZ RULHQWDWLRQ 0RUH VSHFLILFDOO\ WKLV SDSHU
FRQVLGHUVWKHLPSDFWVRI
• 7KUHHFLWLHVUHSUHVHQWLQJGLIIHUHQWFOLPDWLF]RQHVLQWKH86&KLFDJR3KLODGHOSKLDDQG/RV$QJHOHV
• )RXUZLQGRZVL]HVïPïIWOHQJWKïKHLJKWïPïIWïPïIWDQGïP
ïIWZLWKUHVSHFWLYH::5YDOXHVRIDQG
• 7KUHHW\SHVRIJODVVDGRXEOHJOD]HGFOHDUJOD]LQJ*,DGRXEOHJOD]HGORZHJOD]LQJZLWKDVHOHFWLYHFRDWLQJ*,,
DQGDGRXEOHJOD]HGORZHJOD]LQJZLWKDVHOHFWLYHFRDWLQJRIKLJKUHIOHFWLYLW\*,,,7KHLUGLUHFWYLVLEOHWUDQVPLWWDQFHV
DUHVKRZQLQ)LJXUHDVDIXQFWLRQRIVRODULQFLGHQFHDQJOH
• )RXUWUDQVPLWWDQFHYDOXHVIRUUROOHUVKDGHVDQGVKDGHDEVRUSWDQFH
• $OOIRXUPDLQRULHQWDWLRQVIRUWKHSULYDWHRIILFHVSDFHVRXWKZHVWHDVWDQGQRUWK
)LJXUH 9DULDWLRQRIGLUHFWYLVLEOHWUDQVPLWWDQFHZLWKVRODULQFLGHQFHDQJOHIRUWKHWKUHHVWXGLHGJOD]LQJW\SHV
RESULTS AND DISCUSSION

&RQVLGHULQJWKHFRPELQDWLRQVEHWZHHQGHVLJQYDULDEOHVFDVHVZHUHVLPXODWHGIRUWKHSULYDWHRIILFHDQGFDVHV
IRUWKHRSHQRIILFHIRUHDFKORFDWLRQ%DVLFUHVXOWVDUHSUHVHQWHGDVFRPSDUDWLYHJUDSKVLQWKHVHFWLRQVEHORZVKRZLQJWKH
LPSDFWRIORFDWLRQ::5JODVVWUDQVPLWWDQFHVKDGLQJWUDQVPLWWDQFHDQGRULHQWDWLRQ
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
68
ASHRAE Transactions
Shading operation schedule
,Q DOO FDVHV VKDGHV DXWRPDWLFDOO\ FORVH ZKHQ WKHUH LV GLUHFW VXQOLJKW RQ WKH UHVSHFWLYH IDoDGH %DVHG RQ WKH DQQXDO
FDOFXODWLRQV WKH VKDGLQJ VFKHGXOHV GXULQJ ZRUNLQJ KRXUV DPSP WKURXJKRXW WKH \HDU FDQ EH H[SUHVVHG DV WKH
SHUFHQWDJH RI WLPH GXULQJ ZKLFK VKDGHV UHPDLQ RSHQ DV VKRZQ LQ )LJXUH 6RXWKIDFLQJ ZLQGRZV UHVXOW LQ WKH ORZHVW
RSHQLQJ WLPHV IRU VKDGHV VLQFH VRXWK IDFDGHV UHFHLYH GD\OLJKW GXULQJ PRVW RI WKH GD\ (DVWIDFLQJ IDFDGHV KDYH WKH
DGYDQWDJH RI NHHSLQJ VKDGHV RSHQ IRU ORQJHU SHULRGV VLQFH WKH\ VHH WKH VXQ RQO\ IRU WKUHH RI WKH ZRUNLQJ KRXUV DP
SP ²FRPSDUHG WR ZHVW IDFDGHV ZKLFK PD\ UHFHLYH GLUHFW VXQOLJKW IRU ILYH KRXUV LQ WKH DIWHUQRRQ /RV $QJHOHV KDV
EULJKWHUVNLHVWKHUHIRUHOHVVWLPHIRURSHQVKDGHVFRPSDUHGWR&KLFDJRDQG3KLODGHOSKLD
)LJXUH 3HUFHQWDJHRIDQQXDOZRUNLQJWLPHGXULQJZKLFKVKDGHVUHPDLQRSHQIRUDOOORFDWLRQVDQGRULHQWDWLRQV
Daylighting performance results

3ULYDWHRIILFHFDVH7KHIROORZLQJVHFWLRQVVKRZUHSUHVHQWDWLYHUHVXOWVIRUWKHVSDFHRYHUDOOYDOXHVRI'$DQG8',
IRU GLIIHUHQWZLQGRZ VL]HVDQG RULHQWDWLRQV LQHDFK ORFDWLRQ 7R JHW WKH RYHUDOO GD\OLJKW SHUIRUPDQFHPHWULFV IURP ZRUN
SODQHSRLQWLOOXPLQDQFHVWZRPHWKRGVDUHFRPSDUHGLXVHWKHDYHUDJHGD\OLJKWPHWULFVRIDOOSRLQWVDWHDFKWLPHVWHSLL
DWHDFKWLPHVWHSDOOSRLQWVPXVWPHHWWKHLOOXPLQDQFHUHTXLUHPHQWWRJHWRQHFUHGLWIRUWKDWWLPHVWHS)LJXUHVVKRZ
'$DQG8',LQWKHSULYDWHRIILFHIRU*,,DQGVKDGHWUDQVPLWWDQFHDWIRUHDFKRULHQWDWLRQUHVSHFWLYHO\
)LJXUH 'D\OLJKWDXWRQRP\DVDIXQFWLRQRI::5*,,VKDGLQJWUDQVPLWWDQFHƴV IRUSULYDWHRIILFHVIDFLQJ

WKHIRXUPDLQRULHQWDWLRQVLQ&KLFDJR3KLODGHOSKLDDQG/RV$QJHOHVXVLQJPHWKRGLWRSDQGPHWKRGLLERWWRP

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

)URPDILUVWORRNLWDSSHDUVVWUDQJHWKDWPD[LPXP'$LVUHDFKHGZKHQWKHZLQGRZLVIDFLQJQRUWK7KLVLVKRZHYHU
H[SODLQHGE\WKHORQJHVWVKDGHRSHQWLPHIRUQRUWKZLQGRZV)LJXUH,QHDFKORFDWLRQGD\OLJKWDXWRQRP\LQFUHDVHVZLWK
WKHLQFUHDVHRI::5IRUERWKWKHWZRRYHUDOOSHUIRUPDQFHHYDOXDWLRQPHWKRGV7KHYDOXHVIRU'$DUHORZHUZLWKPHWKRG
LL WKDQ ZLWK PHWKRG L 7KLV LV DSSDUHQW VLQFH PHWKRG LL XVHV D VWULFWHU VWDQGDUG UHTXLULQJ DOO SRLQWV WR PHHW WKH
LOOXPLQDQFHVHWSRLQWDWWKHVDPHWLPH$OVRGD\OLJKWDXWRQRP\LQFUHDVHVQRQOLQHDUO\ZLWKLQFUHDVLQJJODVVWUDQVPLWWDQFH
DQGVKDGLQJWUDQVPLWWDQFHDOWKRXJKWKHYDULDWLRQVDUHQRWVKRZQLQWKHDERYHJUDSKV'D\OLJKWDXWRQRP\GHILQHVWKHWLPH
IUDFWLRQZKHQWKHZRUNSODQHLOOXPLQDQFHVHWSRLQWPHWE\GD\OLJKWDORQHVRLWLVPRUHUHDVRQDEOHWRXVHPHWKRGLLWR
HYDOXDWHWKHRYHUDOOVSDFHGD\OLJKWSHUIRUPDQFHIURPWKHVLPXOWDQHRXVLOOXPLQDQFHVRIPDQ\HYDOXDWHGSRLQWVLQWKHVSDFH
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
)LJXUH 8VHIXOGD\OLJKWLOOXPLQDQFHDVDIXQFWLRQRI::5*,,VKDGLQJWUDQVPLWWDQFHƴV IRUSULYDWHRIILFHV
IDFLQJWKHIRXUPDLQRULHQWDWLRQVLQ&KLFDJR3KLODGHOSKLDDQG/RV$QJHOHVXVLQJPHWKRGLWRSDQGPHWKRGLLERWWRP
7KH8',UHVXOWVSUHVHQWPRUHFRPSOH[DQGLQWHUHVWLQJWUHQGV7KHWZRHYDOXDWLRQPHWKRGVVKRZGLIIHUHQWYDULDWLRQV
RI8',DVIXQFWLRQRIZLQGRZVL]H$VWKHZLQGRZVL]HLQFUHDVHV8',ILUVWLQFUHDVHVWRPD[LPXPDWDERXW::5
DQG WKHQ GHFUHDVHV ZKHQ ZLQGRZ VL]H EHFRPHV ODUJHU IRU PRVW RI WKH FDVHV ZLWK HYDOXDWLRQ PHWKRG L +RZHYHU ZLWK
HYDOXDWLRQPHWKRGLLWKH8',IRUVRXWKDQGZHVWIDFLQJIDFDGHVLQFUHDVHZLWKWKHLQFUHDVHRIZLQGRZVL]HZKLOH8',IRU
HDVW DQG QRUWK IDFLQJ IDFDGHV GHFUHDVH ZKHQ ZLQGRZ VL]H LQFUHDVH )URP WKH JUDSKV LW LV KDUG WR VD\ ZKLFK HYDOXDWLRQ
PHWKRGLVEHWWHUVLQFH8',GHILQHVDQLOOXPLQDQFHUDQJHEHWZHHQO[IFDQGO[IFLQFOXGLQJWKHYDOXHV
EHORZDQGKLJKHUWKDQWKHVHWSRLQWO[IFIRURIILFHVSDFHV8',RYHUFRPHVWKHOLPLWDWLRQRI'$WKDWLJQRUHV
WKHFRQWULEXWLRQRIGD\OLJKWVOLJKWO\EHORZWKHVHWSRLQWDQGWKHULVNRIJODUHEXWLWFDQQRWWHOOZKLFKIDoDGHGHVLJQRSWLRQLV
EHWWHUEDVHGRQO\RQWKHYDOXHRI8',)RUH[DPSOHDKLJK8',YDOXHRIZLWK8',O[IFLVQRW
EHWWHUWKDQDORZHU8',YDOXHRIZLWK8',O[IFVLQFHO[IFDUHXVXDOO\
ORZ HQRXJK WR SUHYHQW JODUH DQG KLJK HQRXJK WR UHGXFH OLJKWLQJ RSHUDWLRQ )XUWKHU LQYHVWLJDWLRQ RI WKLV ELQ 8',
O[IFDOORZVXVHIXOFRQFOXVLRQV)LJXUHVKRZVWKH8',O[IFLQWKHSULYDWHRIILFHIRU
*,,DQGVKDGHWUDQVPLWWDQFHDWIRUHDFKRULHQWDWLRQ(YDOXDWLRQPHWKRGLLVXVHGWRREWDLQWKHUHVXOWVLQ)LJXUH
VLQFHDOOYDOXHVEHFRPH]HURZKHQXVLQJPHWKRGLL7KH8',O[IFILUVWLQFUHDVHVWRPD[LPXPDWDERXW
::5DQGWKHQGHFUHDVHVIRUODUJHUZLQGRZV$XVHIXOREVHUYDWLRQRIWKLVDQDO\VLVLVWKDWPHWKRGLVKRXOGEH
XVHGIRUHYDOXDWLRQRI8',O[IFZKLOHPHWKRGLLVKRXOGEHXVHGIRUHYDOXDWLRQRIVSDFHRYHUDOO'$
+RZHYHULIOLJKWIL[WXUHVFDQEHFRQWUROOHGLQGHSHQGHQWO\WKHQFRQWLQXRXVDXWRQRP\PHWULFVZLWKOLJKWGLPPLQJVKRXOG
EHXVHGIRUHVWLPDWLRQRIHQHUJ\VDYLQJV%DVHGRQWKHWUHQGVLOOXVWUDWHGLQ)LJXUHVDQGLWLVUHDVRQDEOHWRVD\WKDWIRU
70
ASHRAE Transactions
VXFK D W\SLFDO SULYDWH RIILFH ZLQGRZ WR ZDOO UDWLR LV HQRXJK WR H[SORLW GD\OLJKW VLQFH WKHUH LV QR VLJQLILFDQW
LQFUHDVHLQGD\OLJKWDYDLODELOLW\IRUODUJHUZLQGRZVL]HV1RWHWKDWWKHWKHUPDOSHUIRUPDQFHQRWFRQVLGHUHGLQWKLVSDSHULV
DOVRDNH\IDFWRULQIDoDGHGHVLJQRSWLRQVDQGODUJHUZLQGRZVZRXOGOLNHO\LQFUHDVHWKHKHDWLQJDQGFRROLQJHQHUJ\GHPDQG
$OVR D JRRG GHVLJQ VKRXOG XWLOL]H PD[LPXP GD\OLJKW ZLWKRXW FDXVLQJ JODUH DQG GLVFRPIRUW 7KH VSDWLDO GLVWULEXWLRQV RI
GD\OLJKWDXWRQRP\DQGXVHIXOGD\OLJKWLOOXPLQDQFHRQWKHZRUNSODQHVKRXOGEHXVHGWRDQDO\]HWKHDQQXDOUHVSRQVHV
)LJXUH 8', O[ IF DV D IXQFWLRQ RI ::5 *,, VKDGLQJ WUDQVPLWWDQFH ƴV IRU SULYDWH
RIILFHVIDFLQJWKHIRXUPDLQRULHQWDWLRQVLQ&KLFDJR3KLODGHOSKLDDQG/RV$QJHOHVXVLQJPHWKRGL
2SHQSODQRIILFHFDVH)LJXUHVVKRZWKHYDULDWLRQRI'$DQG8',O[IFIRUWKHRSHQSODQ

RIILFHIRU*,,DVDIXQFWLRQRI::5DQGVKDGLQJWUDQVPLWWDQFH)RUWKLVFDVH'$DQG8',O[IFKDYH
DPRUHOLQHDUGHSHQGHQFHRQ::5ZKHUHDVWKHLPSDFWRIVKDGLQJWUDQVPLWWDQFHLVLQWHUGHSHQGHQWRQZLQGRZVL]H1R
´VDWXUDWLRQµRIGD\OLJKWDYDLODELOLW\LVQRWLFHGLQWKLVFDVHVLQFHWKHVKDGHVRQHDFKRIWKHWKUHHIDoDGHVDXWRPDWLFDOO\FORVH
GXULQJGLIIHUHQWSDUWVRIWKHGD\DQGWKHGHSWKRIWKHVSDFHLVVLJQLILFDQW)RUVPDOO::5VWKHGLIIHUHQFHVLQ'$DUHVPDOO
DOPRVW WKH VDPH IRU::5 EXW IRU ODUJHUZLQGRZV WKH LPSDFW RI VKDGHWUDQVPLWWDQFH EHFRPHVKLJKHU)RU8',
O[WKHGLIIHUHQFHVDUHKLJKHUEHWZHHQDQGVKDGHWUDQVPLWWDQFHDWODUJHZLQGRZVL]H::5 EXW
IXUWKHULQFUHDVHRIVKDGHWUDQVPLWWDQFHWRGRHVQRWJUHDWO\LQFUHDVHWKHYDOXHVRI8',O[IF
Figure 7 Daylight autonomy as a function of WWR (G-II) and shading transmittance for the open-plan office in
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Chicago, Philadelphia and Los Angeles.
)LJXUH 8', O[ DV D IXQFWLRQ RI ::5 *,, DQG VKDGLQJ WUDQVPLWWDQFH IRU WKH RSHQSODQ RIILFH LQ
&KLFDJR3KLODGHOSKLDDQG/RV$QJHOHV
$6+5$(

7KHVSDWLDOGLVWULEXWLRQRI'$DQG8',O[IFRQWKHZRUNSODQHVXUIDFHRIWKHRSHQSODQRIILFHLQ
&KLFDJRZLWK::5DQGVKDGHWUDQVPLWWDQFHHTXDOWRLVVKRZQLQ)LJXUHVDDQGE1RWHWKDWWKHQRUWKVLGH
WRSD[LVKDVQRZLQGRZVXVHGIRUHQWUDQFHHOHYDWRUVHWFDQGWKHUHIRUHWKHUHLVQROLJKWFRPLQJIURPWKDWVLGHLQWRWKH
VSDFH
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$XWRPDWHGFORVLQJRIVKDGHVXQGHUGLUHFWVXQOLJKWUHVXOWVLQORZHU'$QHDUWKHVRXWKIDoDGHDQGKLJKHUYDOXHVQHDU
WKHHDVWIDoDGHWKHVXQGRHVQRWVHHWKLVVLGHLQWKHDIWHUQRRQVRVKDGHVUHPDLQRSHQ7KH8',UHVXOWVIROORZDVLPLODU
WUHQGZLWKORZHUYDOXHVVLQFHRQO\WKHWDUJHWJURXSEHWZHHQOX[IFLVFRQVLGHUHG7KHVHUHVXOWVFKDQJH
VLJQLILFDQWO\ZLWK::5LIWKHUHZHUHZLQGRZVRQWKHQRUWKVLGHWKHGD\OLJKWOHYHOVUHDFKKLJKYDOXHVDQGWKHSHUIRUPDQFH
PHWULFVYDOXHVFRXOGEHGRXEOHGZLWKSURIRXQGHIIHFWVRQOLJKWLQJHQHUJ\VDYLQJV

)LJXUH D 6SDWLDO GLVWULEXWLRQ RI '$ IRU WKH RSHQ SODQ RIILFH LQ &KLFDJR ::5 DQG E 6SDWLDO
GLVWULEXWLRQRI8',O[IFIRUWKHRSHQSODQRIILFHLQ&KLFDJR::5
Experimental validation
([SHULPHQWDOPHDVXUHPHQWVZHUHSHUIRUPHGLQDVRXWKIDFLQJIXOOVFDOHRIILFHVSDFHWRYDOLGDWHWKHVLPXODWLRQUHVXOWV
7KHWHVWURRPKDVGLPHQVLRQVRIïïPïïIW7KHVRXWKZLQGRZGLPHQVLRQVDUHïPïIW
VWDUWLQJIURPPDERYHWKHIORRU6HYHQSKRWRPHWHUVORFDWHGLQWKHPLGGOHOLQHSHUSHQGLFXODUWRWKHZLQGRZDUHXVHGWR
PHDVXUHWKHLOOXPLQDQFHRQWKHZRUNSODQH7KHGLVWDQFHLQWHUYDOEHWZHHQWKHVHQVRUVLVPZLWKWKHILUVWRQHORFDWHGDW
P IW IURP WKH ZLQGRZ )LJXUH VKRZV WKH FRPSDULVRQ EHWZHHQ WKH PHDVXUHG LOOXPLQDQFH DQG WKH VLPXODWHG
YDOXHV DW GLIIHUHQW WLPHV IRU FORVHG UROOHU VKDGHV a WUDQVPLWWDQFH DQG RSHQ UROOHU VKDGHV GXULQJ D VXQQ\ GD\ 7KH
PRGHOJHQHUDOO\JLYHVJRRGHVWLPDWLRQIRUWKHIURQWSRLQWVDQGEDFNSRLQWVEXWWHQWVWRVOLJKWO\XQGHUHVWLPDWHWKHPLGGOH
SRLQWV7KHPD[LPXPFRPSDUDWLYHHUURULV
)LJXUH &RPSDULVRQEHWZHHQWKHPHDVXUHGDQGVLPXODWHGZRUNSODQHLOOXPLQDQFHVIRUDFORVHGVKDGHFRQGLWLRQ

DQGERSHQVKDGHFRQGLWLRQ
CONCLUSION
7KLV SDSHU LQYHVWLJDWHV WKH DQQXDO GD\OLJKW SHUIRUPDQFH LQ D VPDOO SULYDWH RIILFH DQG LQ D ODUJH RSHQ SODQH RIILFH
HTXLSSHGZLWKDXWRPDWHGVKDGLQJ7KHVSDFHRYHUDOOGD\OLJKWSHUIRUPDQFHZDVHYDOXDWHGEDVHGRQWZRPHWKRGVDQGWKH
72
ASHRAE Transactions
UHVXOWVVXJJHVWWKDWIRUFDOFXODWLRQRI'$DOOSRLQWVVKRXOGPHHWWKHLOOXPLQDQFHUHTXLUHPHQWDWHDFKWLPHVWHSZKLOH8',
LVEHWWHUHYDOXDWHGXVLQJWKHDYHUDJHYDOXHRIDOOVHOHFWHGSRLQWVLQWKHVSDFH7KHLPSDFWVRIZLQGRZWRZDOOUDWLRJOD]LQJ
DQG VKDGLQJ RSWLFDOSURSHUWLHV DQG IDoDGH RULHQWDWLRQ DUHHYDOXDWHG IRU WKUHHFOLPDWLF W\SHV LQ WKH 86 ,QDOO FDVHVUROOHU
VKDGHVZHUHFRQWUROOHGEDVHGRQLQFLGHQWGLUHFWVRODUUDGLDWLRQIRUHDFKIDoDGHVRDVWRSUHYHQWGLUHFWVXQOLJKWIURPHQWHULQJ
WKHVSDFH)RUWKHSULYDWHRIILFHWKHUHVXOWVVKRZHGWKDWQRUWKZLQGRZVKDYHWKHDELOLW\WRDOORZHQRXJKGD\OLJKWLQWRWKH
VSDFHIRUDOOORFDWLRQVIRUZLQGRZWRZDOOUDWLRVKLJKHUWKDQ)RUWKHRWKHURULHQWDWLRQVHDVWSHUIRUPVEHWWHUVLQFHWKH
VKDGHVKDYHWRFORVHIRUDVPDOOHUSHULRGRIWLPHLQWKHPRUQLQJFRPSDUHGWRVRXWKDQGZHVWIDFDGHV'D\OLJKWDXWRQRP\
LQFUHDVHV ZLWKZLQGRZ VL]HJOD]LQJWUDQVPLWWDQFHDQGVKDGLQJ WUDQVPLWWDQFH UHDFKLQJD PD[LPXP RI IRUVRXWK
IDFLQJIDFDGHVGHSHQGLQJRQWKHORFDWLRQ8VHIXOGD\OLJKWLOOXPLQDQFHVEHWZHHQOX[IFZHUHFRQVLGHUHG
LQ GHWDLO LW UHDFKHG D PD[LPXP IRU ZLQGRZWRZDOO UDWLR IRU PRVW FDVHV LQGLFDWLQJ WKH ERXQGDULHV IRU VL]LQJ
ZLQGRZVZLWKUHVSHFWWRGD\OLJKWSURYLVLRQ7KHRSHQSODQRIILFHFDVHQRZLQGRZVRQQRUWKVLGHFRQWUROOHGVKDGHVRQDOO
RWKHU RULHQWDWLRQV VKRZHG WKDW '$ DQG 8', KDYH D PRUH OLQHDU UHVSRQVH WR FKDQJLQJ ZLQGRZ VL]H DQG WKH HIIHFW RI
VKDGLQJ WUDQVPLWWDQFH LV KLJK RQO\ IRU KLJKHU ZLQGRZ VL]HV 7KH PD[LPXP GD\OLJKW DXWRQRP\ YDULHG EHWZHHQ DQG
GHSHQGLQJ RQFOLPDWHDQG VKDGLQJWUDQVPLWWDQFH 8VHIXO GD\OLJKW LOOXPLQDQFHV IROORZHG WKHVDPHWUHQG LQWKLVFDVH
)RUWKHRSHQSODQRIILFHGD\OLJKWDXWRQRP\LVKLJKHUQHDUWKHIDFDGHVEXWWKHHDVWVLGHDOORZVPRUHGD\OLJKWWKURXJKRXWWKH
\HDU EHFDXVH RI WKH LQGHSHQGHQW FRQWUROOHG VKDGLQJ RSHUDWLRQ RQ HDFK IDoDGH )LQDOO\ JRRG DJUHHPHQW ZDV REVHUYHG
EHWZHHQVLPXODWHGDQGPHDVXUHGLOOXPLQDQFHVLQDIXOOVFDOHRIILFH
ACKNOWLEDGMENTS
7KHDXWKRUVZRXOGOLNHWRWKDQNWKH$PHULFDQ6RFLHW\RI+HDWLQJ5HIULJHUDWLQJDQG$LU&RQGLWLRQLQJ(QJLQHHUVWKH
86'HSDUWPHQWRI(QHUJ\WKH3XUGXH5HVHDUFK)RXQGDWLRQ/XWURQ(OHFWURQLFV,QF9LUDFRQ,QFDQG.DZQHHU$OFRD
IRUWKHLUNLQGVXSSRUW
REFERENCES
$WKLHQLWLV$.7]HPSHOLNRV$$PHWKRGRORJ\IRUVLPXODWLRQRIGD\OLJKWURRPLOOXPLQDQFHGLVWULEXWLRQDQGOLJKW
GLPPLQJIRUDURRPZLWKDFRQWUROOHGVKDGLQJGHYLFH6RODU(QHUJ\
'XERLV0&%ORPVWHUEHUJ$(QHUJ\VDYLQJSRWHQWLDODQGVWUDWHJLHVIRUHOHFWULFOLJKWLQJLQIXWXUH1RUWK(XURSHDQ
ORZHQHUJ\RIILFHEXLOGLQJV$OLWHUDWXUHUHYLHZ(QHUJ\DQG%XLOGLQJV
*KLVL(7LQNHU-$$QLGHDOZLQGRZDUHDFRQFHSWIRUHQHUJ\HIILFLHQWLQWHJUDWLRQRIGD\OLJKWDQGDUWLILFLDOOLJKWLQ
EXLOGLQJV%XLOGLQJDQG(QYLURQPHQW
1DELO $ 0DUGDOMHYLF - 8VHIXO GD\OLJKW LOOXPLQDQFHV D QHZ SDUDGLJP IRU DVVHVVLQJ GD\OLJKW LQ EXLOGLQJV /LJKWLQJ
5HVHDUFKDQG7HFKQRORJ\
1DELO$0DUGDOMHYLF-8VHIXOGD\OLJKWLOOXPLQDQFHVDUHSODFHPHQWIRUGD\OLJKWIDFWRUV(QHUJ\DQG%XLOGLQJV

3HUH] 5 ,QHLFKHQ 3 6HDOV 5 0RGHOLQJ GD\OLJKW DYDLODELOLW\ DQG LUUDGLDQFH FRPSRQHQWV IURP GLUHFW DQG JOREDO
LUUDGLDQFH6RODU(QHUJ\
5HLQKDUW&7KHDQQXDOGD\OLJKWDYDLODELOLW\LQSHULSKHUDORIILFHVLQ&DQDGD²DVLPXODWLRQVWXG\5HSRUW%/
IRUWKH1DWLRQDO5HVHDUFK&RXQFLORI&DQDGD
6KHQ +7]HPSHOLNRV$ 'D\OLJKWLQJ DQGHQHUJ\DQDO\VLVRI SULYDWH RIILFHVZLWK DXWRPDWHG LQWHULRU UROOHU VKDGHV
6RODU(QHUJ\
6KHQ + 7]HPSHOLNRV $ $ SDUDPHWULF DQDO\VLV IRU WKH LPSDFW RI IDFDGH GHVLJQ RSWLRQV RQ WKH GD\OLJKWLQJ
SHUIRUPDQFHRIRIILFHVSDFHV3URFHHGLQJVRIVW+LJK3HUIRUPDQFH%XLOGLQJV&RQIHUHQFHDW3XUGXH:HVW/DID\HWWH,186$
7]HPSHOLNRV$$WKLHQLWLV$.7KHLPSDFWRIVKDGLQJGHVLJQDQGFRQWURORQEXLOGLQJFRROLQJDQGOLJKWLQJGHPDQG
6RODU(QHUJ\
7]HPSHOLNRV $ $WKLHQLWLV $. ,QWHJUDWHG GD\OLJKWLQJ DQG WKHUPDO DQDO\VLV RI RIILFH EXLOGLQJV $6+5$(
7UDQVDFWLRQV
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

SA-12-C010
,PSDFWVRI$LU)LOWHUVRQ(QHUJ\
&RQVXPSWLRQLQ7\SLFDO+9$&6\VWHPV

1DELO1DVVLI3K'3(
$6+5$(0HPEHU
$%675$&7
)LOWHUVDUHXVHGLQ+9$&V\VWHPVIRUERWKFRPPHUFLDODQGUHVLGHQWLDOEXLOGLQJVWRSURWHFWWKHHTXLSPHQWDQGLPSURYH
LQGRRU DLU TXDOLW\ LQ FRQGLWLRQHG VSDFHV $OWKRXJK WKHUH DUH PDQ\ EHQHILWV RI XVLQJ WKH DLU ILOWHU LQ DQ DLU FRQGLWLRQLQJ
V\VWHP WKH UHVLVWDQFH DVVRFLDWHG ZLWK LW FDQ LQFUHDVH IDQ HQHUJ\ XVH DQG PD\ DGYHUVHO\ DIIHFW DLU FRQGLWLRQLQJ V\VWHP
SHUIRUPDQFH DQG HIILFLHQF\ 7KH SDSHU H[SORUHV WKH LPSDFW RI DLU ILOWUDWLRQ RQ HQHUJ\ FRQVXPSWLRQ IRU D W\SLFDO DLU
FRQGLWLRQLQJ $& V\VWHP ZLWK FRQVWDQW RU YDULDEOHVSHHG IDQ $ ZKROH EXLOGLQJ VLPXODWLRQ PRGHO LV XVHG WR VLPXODWH WKH
DQQXDOHQHUJ\FRQVXPSWLRQIRUYDULRXVDLUFRQGLWLRQLQJV\VWHPFDSDFLWLHVGLIIHUHQWOHYHOVRIILOWHUFOHDQOLQHVVDQGYDULRXV
ILOWHUPLQLPXPHIILFLHQF\UHSRUWLQJYDOXHV0(597KHUHVXOWVLQGLFDWHVWKDWZLWKDFRQVWDQWVSHHGIDQWKHFRROLQJHQHUJ\
XVHLQFUHDVHVDVWKHILOWHUJHWVGLUW\RYHUWLPHDQGWKHHQHUJ\XVHLQWKHIDQPD\LQFUHDVHEXWWKLVGHSHQGVKHDYLO\RQWKH
LQYHVWLJDWHG IDQ SHUIRUPDQFH FXUYH :LWK D YDULDEOHVSHHG IDQ WKH IDQ HQHUJ\ XVHV LQFUHDVHV ZLWK D GLUW\ ILOWHU EXW WKH
FRROLQJDQGKHDWLQJHQHUJ\XVHVDUHVOLJKWO\DIIHFWHG7KHIDQHQHUJ\XVHULVHGXHWRWKHGLUW\ILOWHUGHSHQGVPDLQO\RQDLU
V\VWHPFDSDFLWLHVILOWHU0(59UDWLQJVDQGWKHGHJUHHRIWKHILOWHUFOHDQOLQHVV
,1752'8&7,21
)LOWHUVDUHW\SLFDOO\XVHGLQERWKFRPPHUFLDODQGUHVLGHQWLDOEXLOGLQJV\VWHPV7KH\DUHORFDWHGLQWKHPDLQDLUVWUHDPWR
SURWHFWWKHHTXLSPHQWDQGLPSURYHLQGRRUDLUTXDOLW\LQFRQGLWLRQHGVSDFHV$6+5$(:LWKRXWILOWHUVSDUWLFOHVPD\
DFFXPXODWHRQIDQVDQGKHDWH[FKDQJHUFRLOVDGYHUVHO\DIIHFWLQJKHDWWUDQVIHU<DQJHWDOD<DQJHWDOE6LHJHODQG
1D]DURII $OWKRXJK WKHUH DUH PDQ\ EHQHILWV RI XVLQJ WKH DLU ILOWHU LQ DQ DLU FRQGLWLRQLQJ V\VWHP WKH UHVLVWDQFH
DVVRFLDWHG ZLWK LW FDQ DGG H[WUD HQHUJ\ SRZHU LQ WKH IDQ RSHUDWLRQ DQG PD\ DIIHFW DGYHUVHO\ RQ DLU FRQGLWLRQLQJ V\VWHP
SHUIRUPDQFH6WHSKHQVHWDO$VWKHILOWHUJHWVGLUW\RYHUWLPHWKHSUHVVXUHGURSDFURVVLWUDLVHVDQGFDXVHVQRWRQO\DQ
LQFUHDVHLQIDQSRZHUEXWDUHGXFWLRQRIDLUFRQGLWLRQLQJV\VWHPSHUIRUPDQFHDQGDLUGLVWULEXWLRQV\VWHPHIILFLHQF\:LWKD
W\SLFDOFRQVWDQWVSHHGIDQPRWRUDVILOWHUJHWVGLUW\WKHVWDWLFSUHVVXUHLQFUHDVHVDQGDLUIORZUDWHGURSV7KHUHGXFHGDLUIORZ
UDWHOHDGVWRUHGXFHWKHV\VWHPFDSDFLW\DQGVHQVLEOHKHDWUDWLR-DPHVHWDO3DODQLHWDO3DUNHUHWDO:LWK
YDULDEOHVSHHGIDQDVWKHUHVLVWDQFHLQFUHDVHVGXHWRDGLUW\ILOWHUWKHIDQVSHHGLQFUHDVHVDQGWKHUHE\WKHIDQSRZHULQRUGHU
WRPDLQWDLQWKHVDPHDLUIORZUDWHDQGPHHWVSDFHVHQVLEOHORDGUHTXLUHPHQWV7KXVWKHVWXG\ZLOOLQYHVWLJDWHWKHVHHIIHFWVRQ
ZKROHV\VWHPSHUIRUPDQFHWKDWLQFOXGHVIDQDQGDLUFRQGLWLRQLQJV\VWHP7KHLPSDFWVRIDLUILOWHUVRQHQHUJ\FRQVXPSWLRQ
IRUW\SLFDODLUFRQGLWLRQLQJ$&V\VWHPVHTXLSSHGZLWKFRQVWDQWRUYDULDEOHVSHHGIDQDUHLQYHVWLJDWHG7KHVWXG\FRQVLGHUV
DVPDOOFRPPHUFLDOEXLOGLQJVPDOOUHWDLOZLWKYDULRXVV\VWHPFDSDFLWLHVDQGORFDWLRQV

1DELO1DVVLILVDDVVLVWDQWSURIHVVRULQWKH'HSDUWPHQWRI&LYLODQG$UFKLWHFWXUDO(QJLQHHULQJ1RUWK&DUROLQD$ 76WDWH8QLYHUVLW\
*UHHQVERUR1&
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
74
ASHRAE Transactions
0(7+2'2/2*<
$16,$6+5$(6WDQGDUG$16,$6+5$(XVHVWKHPLQLPXPHIILFLHQF\UHSRUWLQJYDOXH0(59WR
UDWHWKHHIIHFWLYHQHVVRIDLUILOWHUV7KHVFDOHLVGHVLJQHGWRUHSUHVHQWWKHZRUVWFDVHSHUIRUPDQFHRIDILOWHUZKHQGHDOLQJZLWK
SDUWLFOHVLQWKHUDQJHRIWRPLFURPHWHUVîWRîLQFKHV7KH0(59UDWLQJLVIURPWR+LJKHU
0(59UDWLQJVFRUUHVSRQGWRDJUHDWHUSHUFHQWDJHRISDUWLFOHVFDSWXUHGRQHDFKSDVVZLWKD0(59ILOWHUFDSWXULQJPRUH
WKDQRISDUWLFOHVRYHUWKHIXOOUDQJH7ZR0(59UDWLQJILOWHUV0(59DQG0(59DUHFRQVLGHUHGPRVWO\XVHGLQ
UHVLGHQWLDODQGVPDOOFRPPHUFLDOV\VWHPV7DEOH(RI$6+5$(6WDQGDUG,Q7DEOHRIWKH6WDQGDUG
WKH6WDQGDUGUHFRPPHQGVDPLQLPDOILQDOUHVLVWDQFHWKDWVKRXOGEHDWOHDVWWZLFHWKHLQLWLDOUHVLVWDQFHWKURXJKILOWHUV
7KHILOWHULVDVVXPHGWREHFOHDQZKHQWKHILOWHUUHVLVWDQFHLVHTXDOWRWKHLQLWLDOYDOXHLQ7DEOHRI6WDQGDUGDQGWREH
YHU\GLUW\ZKHQWKHILOWHUUHVLVWDQFHLVHTXDOWRWKHILQDOUHVLVWDQFHYDOXH7KHILOWHULVWKHQDVVXPHGWREHGLUW\LQWKHPLGGOH
RIWKDWUDQJH7DEOHDQG7DEOHVKRZWKHILOWHUUHVLVWDQFHSUHVVXUHGURSDFURVVWKHILOWHUWRWDOSUHVVXUHDQGDLUIORZUDWH
XVHG LQ WKLV VWXG\ 6HYHUDO ILHOG WHVWLQJ KDV EHHQ SHUIRUPHG E\ DLU ILOWHU PDQXIDFWXUHV IRU VHYHUDO W\SHV DQG VL]H RI ILOWHUV
7KHVHWHVWVVKRZHGWKDWWKHSUHVVXUHGURSWKURXJKFOHDQILOWHUVZLWKGLIIHUHQWDLUYHORFLWLHVDQGILOWHUGHSWKVLVLQWKHUDQJHRI
LQ:*3D7KHVHUHVXOWVDUHFRQVLVWHQWZLWKWKHGDWDIRUWKHFOHDQILOWHUVSURYLGHGLQ7DEOHDQG7DEOH
)RUDFRQVWDQWVSHHGIDQRQO\WRQDQGWRQ$&V\VWHPVDUHFRQVLGHUHGZLWK0(59ZKHUHDVWKHWRQWRQDQG
WRQV\VWHPVDUHFRQVLGHUHGIRUWKHYDULDEOHVSHHGIDQ ZLWK 0(59:LWKDFRQVWDQWVSHHGIDQ ZKHQWKHVWDWLFSUHVVXUH
LQFUHDVHVWKHDLUIORZUDWHGURSVDQGWKHDPRXQWRIIORZUHGXFWLRQGHSHQGVRQIDQSHUIRUPDQFHGDWD,QDGGLWLRQWKHIORZDQG
SUHVVXUH YDULDWLRQV FKDQJH WKH IDQ HIILFLHQF\ 7KH DLUIORZ UDWHV LQ 7DEOH DQG HIILFLHQF\ FRQVLGHUHG LQ WKLV VWXG\ DUH IRU
FHUWDLQIDQFKDUDFWHULVWLFVFROOHFWHGIURPSDUWLFXODUPDQXIDFWXUHU¶VGDWD$VH[DPSOHIRUWKHWRQ$&V\VWHPZKHQWKHWRWDO
SUHVVXUHLQFUHDVHVIURPWRLQ:*WR3DWKHIORZUDWHGURSVIURPFIPWRFIPWR
/V7KHIORZUDWHIRUFOHDQILOWHULVEDVHGRQFIPSHUWRQ/VSHUWRQ
7KH VWXG\ GLVFXVVHV D W\SLFDO VLQJOH VWRUHIURQW EXLOGLQJ ORFDWHG LQ WKH IROORZLQJ FOLPDWH ]RQHV *UHHQVERUR 1&
2UODQGR1HZ<RUNDQG6DQ)UDQFLVFR7KHEXLOGLQJDUHDYDULHVWRJHWYDULRXVV\VWHPFDSDFLWLHVUDQJLQJIURPWRQWR
WRQ,QWKLVSDSHUIRUVPDOOV\VWHPVDQGWRQWKHFRQGLWLRQHGDLULVSURYLGHGE\DVSOLWV\VWHPVLQJOH]RQHZLWKIXUQDFH
DQG 0(59 ILOWHU )RU UHODWLYHO\ ODUJHU V\VWHPV WRQ DQG WRQ WKH FRQGLWLRQHG DLU LV SURYLGHG E\ D SDFNDJHG
YDULDEOHVSHHGYDULDEOHWHPSHUDWXUHV\VWHPZLWKIXUQDFHDQG0(59ILOWHU7RHYDOXDWHWKHDQQXDOHQHUJ\FRQVXPSWLRQIRU
DLUFRQGLWLRQLQJV\VWHPVDZKROHEXLOGLQJHQHUJ\VLPXODWLRQPRGHOH48(67LVXVHG7KHVRIWZDUHXVHV'2((KRXUO\
EXLOGLQJHQHUJ\VLPXODWLRQVRIWZDUHWRVLPXODWHHQHUJ\XVH,QDGGLWLRQWRLQYHVWLJDWHWKHHIIHFWRIUHGXFHGDLUIORZUDWHRQ
FRROLQJ SHUIRUPDQFH WKH GLUHFW H[SDQVLRQ $& PRGHO $&'; LV XVHG %UDQGHPXHKO HW DO 7KH $&'; PRGHO LV
UHOLDEOHIRUDLUIORZUDWHVEHWZHHQDQGFIPSHUWRQHJFIPDQGFIPIRUDWRQV\VWHP

7DEOH)ORZDQGSUHVVXUHGDWDIRUFRQVWDQWVSHHGIDQ$&V\VWHPZLWKGLIIHUHQW0(59ILOWHUFOHDQOLQHVV
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
3-Ton system 5-Ton system
Flow Filter Total Flow Filter Total
CFM (L/s) In WG (Pa) In WG (Pa) CFM (L/s) In WG (Pa) In WG (Pa)
Clean 1260 (595) 0.1 (25) 0.63 (157) 2100 (991) 0.1 (25) 0.63 (157)
Dirty 1120 (529) 0.2 (50) 0.73 (182) 1850 (873) 0.2 (50) 0.73 (182)
Very dirty 1000 (472) 0.3 (75) 0.83 (207) 1550 (779) 0.3 (75) 0.83 (207)

$6+5$(


7DEOH)ORZDQGSUHVVXUHGDWDIRUYDULDEOHVSHHGIDQ$&V\VWHPZLWKGLIIHUHQW0(59ILOWHUFOHDQOLQHVV
5-Ton system 7.5-Ton system 10-Ton system
Filter Flow Total Flow Total Flow Total
In WG (Pa) CFM (L/s) In WG (Pa) CFM (L/s) In WG (Pa) CFM (L/s) In WG (Pa)
Clean 0.2 (50) 2100 (991) 0.73 (182) 3150 (1487) 0.73 (182) 4200 (1982) 0.73 (182)
Dirty 0.3 (75) 2100 (991) 0.83 (207) 3150 (1487) 0.83 (207) 4200 (1982) 0.83 (207)
Very dirty 0.4 (100) 2100 (991) 0.93 (232) 3150 (1487) 0.93 (232) 4200 (1982) 0.93 (232)
5(68/76$1'',6&866,21
7KHDQDO\VLVFRQVLGHUVGLIIHUHQWVL]HVRIDVPDOOFRPPHUFLDOEXLOGLQJVPDOOUHWDLOVWRUHIURQWZLWKGLIIHUHQW$&V\VWHP
FDSDFLWLHVDQGILOWHUSUHVVXUHGURSV,WLVDVVXPHGWKDWWKHVWRUHRSHQVDW$0DQGFORVHVDW307KHVWRUHDUHDLVYDULHG
IURPIWWRIWP WRPGHSHQGLQJRQWKHORFDWLRQVWRYDU\WKH$&V\VWHPFDSDFLWLHVUDQJLQJIURPWRQ
WRWRQ)RUDSDUWLFXODUORFDWLRQDQGV\VWHPDQGEXLOGLQJVL]HWKHLQSXWYDULDEOHVRIH48(67PRGHOVXFKDVDLUIORZUDWH
WRWDOVWDWLFSUHVVXUHDQGIDQHIILFLHQF\DUHYDULHGWRVLPXODWHWKHDQQXDOHQHUJ\FRQVXPSWLRQV
,Q WKLV VHFWLRQ D YDULDEOH VSHHG IDQ PRWRU $& V\VWHP SDFNDJHG YDULDEOHVSHHG YDULDEOHWHPSHUDWXUH V\VWHP ZLWK
0(59LVFRQVLGHUHGZLWKWKUHHGLIIHUHQWFDSDFLWLHVWRQWRQDQGWRQ$VWKHUHVLVWDQFHLQFUHDVHVGXHWRDGLUW\
ILOWHUWKHIDQVSHHGLQFUHDVHVWRPDLQWDLQWKHVDPHDLUIORZUDWHDQGPHHWWKHVSDFHVHQVLEOHORDGUHTXLUHPHQWV$VWKHUHLVQR
FKDQJHLQWKHDLUIORZUDWHWKHUHLVQRFKDQJHRUVOLJKWO\FKDQJHLQWKHFRROLQJDQGKHDWLQJHQHUJ\XVH7KHFRROLQJHQHUJ\
XVHVOLJKWO\ LQFUHDVHVDQGKHDWLQJHQHUJ\XVHVOLJKWO\GHFUHDVHVGXHWRLQFUHDVHGIDQSRZHUDQGKHDWUHOHDVHGIURPWKHIDQ
PRWRUWRVXSSO\DLUVWUHDP,QDGGLWLRQWKHHOHYDWHGSUHVVXUHLQWKHDLUGXFWLQFUHDVHVWKHDLUOHDNDJHDQGHIIHFWVQHJDWLYHO\RQ
FRROLQJDQGKHDWLQJV\VWHPSHUIRUPDQFH,QWKLVSDSHUWKHDLUOHDNDJHLVQRWFRQVLGHUHG7KHSUHVVXUHGURSWKURXJKWKHILOWHU
GHSHQGVPDLQO\RQFHUWDLQ IDFWRUVVXFKDVDLUVWUHDP YHORFLW\ILOWHUW\SH0(59UDWLQJ DUHDDQGFOHDQOLQHVV$VVKRZQ LQ
7DEOHIRUDW\SLFDOWRQV\VWHPWKHSUHVVXUHGURSDFURVVWKHILOWHUYDULHVIURPLQ:*WRLQ:*3DWR3D
DQGWKHWRWDOVWDWLFSUHVVXUHYDULHVIURPLQ:*WRLQ:*3DWR3D7KHSUHVVXUHGURSDVVRFLDWHGZLWKWKH
DLUILOWHUOHDGVWRKLJKHUWRWDOVWDWLFSUHVVXUHDQGKLJKHUIDQSRZHU)LJXUHVKRZVWKHDQQXDOIDQHQHUJ\FRQVXPSWLRQIRUWKH
SDFNDJHG YDULDEOHVSHHG $& V\VWHP ZLWK YDULRXV VWDWLF SUHVVXUHV *UHHQVERUR 1& )RU WKH WRQ V\VWHP ORFDWHG LQ
*UHHQVERURWKHDQQXDOIDQHQHUJ\XVHLQFUHDVHVIURPN:KWRN:KZKHQWKHWRWDOVWDWLFSUHVVXUHLQFUHDVHVIURP
LQ:*WRLQ:*3DWR3D'XHWRWKHDVVXPSWLRQVSUHVHQWHGLQ7DEOHWKHSUHVVXUHGURSWKURXJKWKHFOHDQ
ILOWHULVLQ:*3DDQGEHFRPHVLQ:*3DZKHQWKHILOWHUJHWVYHU\GLUW\7DEOHVKRZVWKHDQQXDOIDQ
HQHUJ\ FRQVXPSWLRQ IRU WKH YDULDEOH VSHHG $& V\VWHP ZLWK FOHDQ GLUW\ DQG YHU\ GLUW\ ILOWHUV 7KH FOHDQ GLUW\ YHU\ GLUW\
ILOWHUV DUH FRUUHVSRQGLQJ WR WKH WRWDO SUHVVXUH LQ :* LQ :* DQG LQ :* 3D 3D DQG 3D DV
SUHVHQWHGLQ7DEOH7KHIDQHQHUJ\FRQVXPSWLRQLQFUHDVHVDVWKHILOWHUJHWVGLUW\$VH[DPSOHIRUWKHWRQV\VWHPORFDWHG
LQ6DQ)UDQFLVFRWKHDQQXDOIDQHQHUJ\FRQVXPSWLRQVDUHN:KN:KDQGN:KIRUFOHDQGLUW\DQGYHU\GLUW\
ILOWHUVUHVSHFWLYHO\7KLVUHVXOWLQGLFDWHVWKDWWKHDQQXDOIDQHQHUJ\FRQVXPSWLRQLQFUHDVHVN:KZLWKWKHGLUW\ILOWHUDQG
N:KZLWKWKHYHU\GLUW\ILOWHUDVFRPSDUHGWRWKDWIRUWKHFOHDQILOWHUDQLQFUHDVHRIDQGUHVSHFWLYHO\7KH
DQQXDO IDQ HQHUJ\ XVH ULVH FRXOG EH XS WR DERXW GHSHQGLQJ RQ ORFDWLRQV FRQVLGHUHG DQG WKH GHJUHH RI WKH ILOWHU
FOHDQOLQHVV

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
76
ASHRAE Transactions
^ƚĂƚŝĐƉƌĞƐƐƵƌĞWĂ
ϭϬϳ͘Ϭϳ ϭϯϭ͘ϵϳ ϭϱϲ͘ϴϳ ϭϴϭ͘ϳϳ ϮϬϲ͘ϲϳ Ϯϯϭ͘ϱϳ
ϴϬϬ
ϳϬϬ
ϲϬϬ
&ĂŶĞŶĞƌŐǇƵƐĞŬtŚ
ϱϬϬ
ϱƚŽŶ
ϳ͘ϱƚŽŶ
ϭϬƚŽŶ
ϰϬϬ
ϯϬϬ
ϮϬϬ
ϭϬϬ
Ϭ͘ϰϯ Ϭ͘ϱϯ Ϭ͘ϲϯ Ϭ͘ϳϯ Ϭ͘ϴϯ Ϭ͘ϵϯ
^ƚĂƚŝĐƉƌĞƐƐƵƌĞŝŶt'
)LJXUH $QQXDO IDQ HQHUJ\ FRQVXPSWLRQ IRU WKH SDFNDJHG YDULDEOH VSHHG $& V\VWHP ZLWK YDULRXV VWDWLF SUHVVXUHV
*UHHQVERUR1&

7DEOH$QQXDOIDQHQHUJ\FRQVXPSWLRQIRUWKHYDULDEOHVSHHG$&V\VWHP
5 ton system 7.5 ton system 10 ton system
Energy Increase Energy Increase Energy Increase
kWh % kWh % kWh %
Clean 285 0.0 435 0.0 570 0.0
Greensboro Dirty 330 15.8 495 13.8 660 15.8
Very dirty 360 26.3 555 27.6 735 28.9
Clean 270 0.0 405 0.0 540 0.0
NY Dirty 300 11.1 465 14.8 615 13.9
Very dirty 345 27.8 525 29.6 690 27.8
Clean 255 0.0 390 0.0 510 0.0
San Francisco Dirty 285 11.8 435 11.5 585 14.7
Very dirty 330 29.4 495 26.9 660 29.4
Clean 270 0.0 495 0.0 570 0.0
Orlando Dirty 315 16.7 570 15.2 645 13.2
Very dirty 345 27.8 630 27.3 720 26.3

,Q WKH IROORZLQJ VHFWLRQV D FRQVWDQWVSHHG IDQ PRWRU $& V\VWHP VSOLW V\VWHP VLQJOH ]RQH ZLWK 0(59 ILOWHU LV
FRQVLGHUHGZLWKWZRGLIIHUHQWFDSDFLWLHVWRQDQGWRQV\VWHPV$VWKHILOWHUJHWVGLUW\WKHVWDWLFSUHVVXUHLQFUHDVHVDQG
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

DLUIORZUDWHGURSV7KLVFDXVHVQRWRQO\DQLQFUHDVHLQWKHIDQSRZHUEXWDUHGXFWLRQRIDLUFRQGLWLRQLQJV\VWHPSHUIRUPDQFH
DQGDLUGLVWULEXWLRQV\VWHPHIILFLHQF\)RUVSHFLILFIDQSHUIRUPDQFHGDWDDVSUHVHQWHGLQ7DEOHWKHDLUIORZUDWHGURSVIURP
FIPWRFIP/VWR/VZKHQWKHWRWDOVWDWLFSUHVVXUHLQFUHDVHVIURPLQ:*WRLQ:*3DWR
3DGXHWRDQLQFUHDVHRIWKHSUHVVXUHGURSWKURXJKWKHILOWHUIURPLQ:*WRLQ:*3DWR3D7KHIDQ
SRZHULVLQIOXHQFHGE\ERWKDLUIORZUDWHDQGVWDWLFSUHVVXUHYDULDEOHVZKHUHDVWKH$&FRROLQJSHUIRUPDQFHLVLQIOXHQFHGE\
RQO\WKHYDOXHRIDLUIORZUDWH7KHGLUHFWH[SDQVLRQ$&PRGHO$&';LVXVHG%UDQGHPXHKOHWDOWRVWXG\WKHHIIHFW
RI UHGXFHG DLUIORZ UDWH RQ FRROLQJ SHUIRUPDQFH 7KH VLPXODWLRQV ZHUH GRQH RQ WKH WRQ $& V\VWHP ZLWK YDULRXV DLUIORZ
UDWHV DQG WKH IROORZLQJ FRQGLWLRQV HQWHULQJ GU\ WHPSHUDWXUH RI R) R& KXPLGLW\ UDWLRQ RI DQG RXWGRRU DLU
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
WHPSHUDWXUH RI R) R& 7KH UHVXOWHG FRROLQJ FDSDFLW\ DQG DLUIORZ UDWHV DUH GHSLFWHG LQ )LJXUH 7KH UHVXOWV DUH
UHSUHVHQWHGDVDSHUFHQWDJHRIWKHVHOHFWHGGHVLJQYDOXHHJFIPWRQ/VSHUWRQWRWDODLUIORZUDWHRIFIP
/V WRWDO FRROLQJ FDSDFLW\ RI : DQG VHQVLEOH FRROLQJ FDSDFLW\ RI : 7KH WRWDO FRROLQJ FDSDFLW\ DQG
VHQVLEOHFRROLQJFDSDFLW\ERWKGURSZLWKDORZHUDLUIORZUDWHZKHUHDVWKHODWHQWFDSDFLW\LQFUHDVHV)RULQVWDQWZKHQWKHIORZ
UDWHGURSVWRRIGHVLJQ YDOXH WKH WRWDOFDSDFLW\GURSVWRDQGWKHVHQVLEOH FDSDFLW\GURSVWRRIGHVLJQ
YDOXH LH : 7KH UHGXFHG DLUIORZ UDWH PDNHV WKH IDQ DQG DLU FRQGLWLRQLQJ V\VWHP UXQ ORQJHU WR PHHW WKH VDPH
FRROLQJ ORDG 7KHVH HIIHFWV RI UHGXFHG DLU IORZ UDWH RQ FRROLQJ SHUIRUPDQFH ZHUH VLPLODUO\ LQYHVWLJDWHG DQG GLVFXVVHG LQ
GHWDLOVHOVHZKHUH3DUNHUHWDO

6HQVLEOH
7RWDO
/DWHQW

'HVLJQFRQGLWLRQVFIPWRQ/VWRQ

&RROLQJ&DSDFLW\

)ORZUDWH
)LJXUH6\VWHPFDSDFLW\DVDIXQFWLRQRIDLUIORZUDWH

7KH HQHUJ\ VLPXODWLRQV UXQ IRU WKH FRQVWDQWVSHHG WRQ DQG WRQ $& V\VWHPV ZLWK 0(59 DQG ZLWK YDULRXV
FRPELQDWLRQV RI IORZ UDWH DQG VWDWLF SUHVVXUH WR HVWLPDWH WKH DQQXDO HQHUJ\ XVH )LJXUH DQG )LJXUH VKRZ WKH DQQXDO
FRROLQJDQGIDQHQHUJ\FRQVXPSWLRQVIRUWKHV\VWHPORFDWHGLQ*UHHQVERUR,IWKHIORZUDWHGURSVIURPFIPWRFIP
/VWR/VIRUWRQV\VWHPWKHFRROLQJHQHUJ\FRQVXPSWLRQLQFUHDVHVIURPN:KWRN:K7KHLQFUHDVH
LQFRROLQJHQHUJ\FRQVXPSWLRQLVEHFDXVHRIDORZHUDLUIORZUDWHDQGDVVRFLDWHGORZHUFRROLQJFDSDFLW\DVVKRZQLQ)LJXUH
2WKHUHIIHFWEXWQRWFRQVLGHUHGKHUHLVWKHLQFUHDVHGDLUOHDNDJHLQWKHUHWXUQVLGHGXFWRUSOHQXPGXHWRWKHKLJKSUHVVXUH
ZLWKDGLUW\ILOWHU'XULQJWKHKHDWLQJVHDVRQWKHUHLVDVOLJKWFKDQJHLQWKHKHDWLQJHQHUJ\FRQVXPSWLRQGXHWRWKHUHGXFHGDLU
IORZWKHLQFUHDVHGUHWXUQGXFWOHDNDJHDQGWKHLQFUHDVHGKHDWJHQHUDWHGE\HOHYDWHGIDQSRZHU7KHFKDQJHLQWKHFRROLQJ
FRQVXPSWLRQRIDFRQVWDQWVSHHGIDQYDULHVRQO\ZLWKWKHDLUIORZUDWH+RZHYHUZKHQWKHVWDWLFSUHVVXUHLQFUHDVHVDQGDWWKH
78
ASHRAE Transactions
VDPHWLPHWKHDLUIORZUDWHGURSVWKHIDQHQHUJ\YDULHVGHSHQGLQJRQERWKYDOXHVDQGIDQHIILFLHQF\DVZHOO)RUH[DPSOHIRU
WKH WRQ V\VWHP ORFDWHG LQ *UHHQVERUR WKH IDQ HQHUJ\ LQFUHDVHV IURP N:K WR N:K ZKHQ WKH SUHVVXUH LQFUHDVHV
IURPLQ:*WRLQ:*3DWR3DDQGWKHIORZUDWHGURSVIURPFIPWRFIP/VWR/V
7KLVFKDQJHLQWKHIDQHQHUJ\FRQVXPSWLRQGHSHQGVRQWKHLQYHVWLJDWHGIDQSHUIRUPDQFHFXUYHWKDWUHSUHVHQWVWKHUHODWLRQVKLS
EHWZHHQWKHSUHVVXUHDQGIORZUDWHLQFOXGLQJWKHIDQHIILFLHQF\

&ůŽǁƌĂƚĞ>ͬƐ
ϯϳϳ͘ϱϮ ϰϳϭ͘ϵ ϱϲϲ͘Ϯϴ ϲϲϬ͘ϲϲ ϳϱϱ͘Ϭϰ ϴϰϵ͘ϰϮ ϵϰϯ͘ϴ ϭϬϯϴ͘ϭϴ ϭϭϯϮ͘ϱϲ ϭϮϮϲ͘ϵϰ
ϱϱϬϬ
ϱϬϬϬ
ϰϱϬϬ
ŽŽůŝŶŐĞŶĞƌŐǇƵƐĞŬtŚ
ϯƚŽŶ
ϰϬϬϬ
ϱƚŽŶ
ϯϱϬϬ
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
ϯϬϬϬ
ϮϱϬϬ
ϴϬϬ ϭϬϬϬ ϭϮϬϬ ϭϰϬϬ ϭϲϬϬ ϭϴϬϬ ϮϬϬϬ ϮϮϬϬ ϮϰϬϬ ϮϲϬϬ
&ůŽǁƌĂƚĞ&D
)LJXUH $QQXDO FRROLQJ HQHUJ\ FRQVXPSWLRQ IRU WKH FRQVWDQW VSHHG VSOLW V\VWHP ZLWK YDULRXV DLUIORZ UDWHV
*UHHQVERUR1&

7DEOHVXPPDUL]HVWKHUHVXOWVIRUWKHFOHDQGLUW\DQGYHU\GLUW\0(59ILOWHUV7KHFOHDQGLUW\YHU\GLUW\ILOWHUVDUH
FRUUHVSRQGLQJWRWKHWRWDOSUHVVXUHRILQ:*LQ:*DQGLQ:*3D3DDQG3DDVSUHVHQWHGLQ
7DEOH:KHQWKHFOHDQ0(59ILOWHUEHFRPHVYHU\GLUW\WKHWRWDOVWDWLFSUHVVXUHLVDVVXPHGWRULVHIURPLQ:*WR
LQ:*3DWR3DDQGWKHIORZFRQVHTXHQWO\GURSVIURPFIPWRFIP/VWR/VIRUWKH
WRQ$&V\VWHPDQGIURPFIPWRFIP/VWR/VIRUWKHWRQ$&V\VWHPGXHWRIDQSHUIRUPDQFHFXUYH
REWDLQHGIRUWKRVHVSHFLILF XQLWV7KLVFKDQJHFDXVHVDQLQFUHDVHLQWKH IDQHQHUJ\FRQVXPSWLRQIURPN:KWR
N:KDQLQFUHDVHRIIRUWKHWRQ$&V\VWHPDQGIURPN:KWRN:KDQLQFUHDVHRIIRUWKHWRQH
$&V\VWHPORFDWHGLQ*UHHQVERURDQGDQLQFUHDVHLQWKHFRROLQJHQHUJ\FRQVXPSWLRQIURPN:KWRN:KDQ
LQFUHDVHRIIRUWKHWRQ$&V\VWHPDQGIRUPN:KWRN:KDQLQFUHDVHRIIRUWKHWRQ$&V\VWHP
,QGU\FOLPDWHVXFKLQ6DQ)UDQFLVFRGXHWRWKHUHODWLYHO\ORZODWHQWORDGWKHFRROLQJHQHUJ\XVHLVVOLJKWO\DIIHFWHGE\WKH
DLUIORZUDWHYDULDWLRQV

$6+5$(

^ƚĂƚŝĐƉƌĞƐƐƵƌĞWĂ
ϭϬϳ͘Ϭϳ ϭϯϭ͘ϵϳ ϭϱϲ͘ϴϳ ϭϴϭ͘ϳϳ ϮϬϲ͘ϲϳ Ϯϯϭ͘ϱϳ
ϭϱϬϬ
ϭϰϬϬ
ϭϯϬϬ
ϭϮϬϬ
&ĂŶĞŶĞƌŐǇƵƐĞŬtŚ
ϭϭϬϬ
ϭϬϬϬ
ϯƚŽŶ
ϱƚŽŶ
ϵϬϬ
ϴϬϬ
ϳϬϬ
ϲϬϬ
ϱϬϬ
Ϭ͘ϰϯ Ϭ͘ϱϯ Ϭ͘ϲϯ Ϭ͘ϳϯ Ϭ͘ϴϯ Ϭ͘ϵϯ
^ƚĂƚŝĐƉƌĞƐƐƵƌĞŝŶt'
)LJXUH $QQXDO IDQHQHUJ\FRQVXPSWLRQ IRUWKHFRQVWDQWVSHHGVSOLWV\VWHP ZLWKGLIIHUHQWFRPELQDWLRQVRIDLUIORZ
UDWHDQGVWDWLFSUHVVXUHDVSUHVHQWHGLQ7DEOH*UHHQVERUR1&

7DEOH$QQXDOIDQDQGFRROLQJHQHUJ\FRQVXPSWLRQIRUWKHFRQVWDQWVSHHG$&V\VWHP
3 ton AC system 5 ton AC system
Fan Cooling Fan Cooling
Energy Increase Energy Increase Energy Increase Energy Increase
kWh % kWh % kWh % kWh %
Clean 790 0.0 3210 0.0 1310 0.0 4540 0.0
Greensboro Dirty 810 2.5 3360 4.7 1340 2.3 4750 4.6
Very dirty 820 3.8 3460 7.8 1360 3.8 4870 7.3
Clean 820 0.0 2180 0.0 1370 0.0 3090 0.0
NY Dirty 850 3.7 2290 5.0 1400 2.2 3240 4.9
Very dirty 860 4.9 2350 7.8 1410 2.9 3310 7.1
Clean 821 0.0 1094 0.0 1310 0.0 1550 0.0
San Francisco Dirty 845 2.9 1111 1.6 1340 2.3 1570 1.3
Very dirty 858 4.5 1120 2.4 1343 2.5 1584.0 2.2
Clean 820 0.0 5430 0.0 1310 0.0 6720 0.0
Orlando Dirty 850 3.7 5540 2.0 1340 2.3 6940 3.3
Very dirty 860 4.9 5599 3.1 1345 2.7 7120 6.0

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---

$VDVXPPDU\IRUDFRQVWDQWVSHHG$&V\VWHPWKHLQFUHDVHLQWKHFRROLQJHQHUJ\XVHLVLQWKHUDQJHRIN:KSHU
WRQH[FHSW6DQ)UDQFLVFRDQGWKHLQFUHDVHLQWKHIDQHQHUJ\XVHLVLQWKHUDQJHRIN:KSHUWRQ,WVKRXOGEHQRWHGWKDW
WKHFKDQJHLQWKHFRROLQJHQHUJ\XVHGHSHQGVRQO\RQWKHYDULDWLRQRIDLUIORZUDWHEXWWKHFKDQJHLQWKHIDQHQHUJ\XVHGXHWR
WKHGLUW\ILOWHUVWURQJO\GHSHQGVRQWKHIDQFKDUDFWHULVWLFVFXUYHXVHG)RUDYDULDEOHVSHHGIDQQRVLJQLILFDQWFKDQJHLQWKH
FRROLQJDQGKHDWLQJHQHUJ\XVHLVRFFXUUHG+RZHYHUWKHLQFUHDVHLQWKHIDQHQHUJ\LVLQWKHUDQJHRIN:KSHUWRQ7KH
UHVXOWVLQGLFDWHWKDWWKHGLUW\ILOWHUFDQFDXVHKLJKHUHQHUJ\XVHDQGVKRZWKHLPSRUWDQWRIUHSODFLQJWKHILOWHUIUHTXHQWO\DQG
LQRSWLPDOPDQQHU
&21&/86,21
7KHZKROHEXLOGLQJHQHUJ\VLPXODWLRQ PRGHOH48(67LVXVHGWRLQYHVWLJDWHWKHLPSDFWVRIDLUILOWHURQIDQHQHUJ\
FRQVXPSWLRQLQDLUFRQGLWLRQLQJV\VWHPVZLWKFRQVWDQWRUYDULDEOHVSHHGIDQ7KHVWXG\WDNHVLQWRFRQVLGHUDWLRQWZR0(59
ILOWHUVDQGV\VWHPFDSDFLWLHVUDQJLQJIURPWRWRQV7KHVLPXODWLRQUHVXOWVVKRZHGWKDWWKHDQQXDOHQHUJ\FRQVXPSWLRQ
LQFUHDVHVDVWKHILOWHUVJHWGLUW\RYHUWLPH)RUDYDULDEOHVSHHGIDQ$&V\VWHPWKHDQQXDOIDQHQHUJ\FRQVXPSWLRQLQFUHDVHV
ZLWK YHU\ GLUW\ ILOWHU XS WR GHSHQGLQJ RQ WKH ORFDWLRQV DQG WKH GHJUHH RI WKH ILOWHU FOHDQOLQHVV 7KH IDQ HQHUJ\ XVH
LQFUHDVHFRXOGEHLQWKHUDQJHRIN:KSHUWRQ7KHUHLVQRFKDQJHRUVOLJKWO\FKDQJHLQWKHFRROLQJDQGKHDWLQJHQHUJ\
XVHVGXHWRWKHFKDQJHLQIDQSRZHUWKHKHDWUHOHDVHGWRWKHVXSSO\DLUDQGDLUGXFWOHDNDJH)RUDFRQVWDQWVSHHGIDQ$&
V\VWHPWKHDQQXDOIDQHQHUJ\FRQVXPSWLRQDOVRLQFUHDVHVEXWOHVVWKDQZLWKWKHYDULDEOHVSHHGIDQDQGLWFRXOGUHDFKXSWR
GHSHQGLQJ VWURQJO\ RQ WKH IDQ SHUIRUPDQFH GDWD FRQVLGHUHG 7KH DQQXDO FRROLQJ HQHUJ\ XVH LQFUHDVHV ZLWK YHU\ GLUW\
ILOWHU XS WR GHSHQGLQJ RQ ORFDWLRQV ,Q UHODWLYHO\ GU\ FOLPDWH DV LQ 6DQ )UDQFLVFR WKH FRROLQJ HQHUJ\ XVH LQFUHDVHV
UHODWLYHO\OHVVWKDQRWKHUDUHDVGXHWRWKHORZHUODWHQWORDG7KHUHLVQRFKDQJHRUVOLJKWO\FKDQJHLQWKHKHDWLQJHQHUJ\XVH
GXHWRWKHLQFUHDVHGIDQSRZHUDQGDLUGXFWOHDNDJH7KHLQFUHDVHLQWKHFRROLQJHQHUJ\XVHFRXOGEHLQWKHUDQJHRI
N:KSHUWRQH[FHSW6DQ)UDQFLVFRDQGWKHLQFUHDVHLQWKHIDQHQHUJ\XVHFRXOGEHLQWKHUDQJHRIN:KSHUWRQ7KH
UHVXOWVLQGLFDWHGWKDWWKHDLUILOWHUFDQEHLPSRUWDQWVRXUFHRIHQHUJ\ZDVWHLIWKHGLUW\ILOWHULVQRWUHSODFHGLQDQRSWLPDOWLPH
PDQQHU
5()(5(1&(6
$16,$6+5$($16,$6+5$(6WDQGDUG0HWKRGRIWHVWLQJJHQHUDOYHQWLODWLRQDLUFOHDQLQJGHYLFHVIRUUHPRYDO
HIILFLHQF\E\SDUWLFOHVL]H$PHULFDQ6RFLHW\RI+HDWLQJ5HIULJHUDWLQJDQG$LU&RQGLWLRQLQJ(QJLQHHUV$WODQWD*$
$6+5$( $6+5$( +DQGERRN+9$& 6\VWHPV DQG (TXLSPHQW &KDSWHU DQG $PHULFDQ 6RFLHW\ RI +HDWLQJ
5HIULJHUDWLQJDQG$LU&RQGLWLRQLQJ(QJLQHHUV,QF
%UDQGHPXHKO0-*DEHO6DQG$QGHUVHQ,$7RRONLWIRU6HFRQGDU\+9$&6\VWHP(QHUJ\&DOFXODWLRQ$PHULFDQ
6RFLHW\RI+HDWLQJ5HIULJHUDWLQJDQG$LU&RQGLWLRQLQJ(QJLQHHUV$6+5$(
-DPHV 3 - &XPPLQJV - 6RQQH 5 9LHLUD DQG - .ORQJHUER 7KH (IIHFW RI 5HVLGHQWLDO (TXLSPHQW &DSDFLW\ RQ
(QHUJ\8VH'HPDQGDQG5XQ7LPH$6+5$(7UDQVDFWLRQV
3DODQL0'2¶1HDODQG-+DEHUO7KH(IIHFWRI5HGXFHG(YDSRUDWRU$LU)ORZRQWKH3HUIRUPDQFHRID5HVLGHQWLDO
&HQWUDO$LU&RQGLWLRQHU,QWKH3URFHHGLQJVRIWKH6\PSRVLXPRQ%XLOGLQJ6\VWHPVLQ+RW+XPLG&OLPDWHV
3DUNHU ' - 6KHUZLQ 5 5DXVWDG DQG ' 6KLUH\ ,,, ,PSDFW RI (YDSRUDWRU &RLO $LUIORZ LQ 5HVLGHQWLDO $LU
&RQGLWLRQLQJ6\VWHPV$6+5$(7UDQVDFWLRQV
<DQJ/-(%UDXQDQG:$*UROOD7KHLPSDFWRIHYDSRUDWRUIRXOLQJDQGILOWUDWLRQRQWKHSHUIRUPDQFHRISDFNDJHG
DLUFRQGLWLRQHUV,QW-5HIULJHUDWLRQ
<DQJ/-(%UDXQDQG:$*UROOE7KHLPSDFWRIIRXOLQJRQWKHSHUIRUPDQFHRIILOWHUHYDSRUDWRUFRPELQDWLRQV
,QW-5HIULJHUDWLRQ
6LHJHO -$ DQG : 1D]DURII 3UHGLFWLQJ SDUWLFOH GHSRVLWLRQ RQ +9$& KHDW H[FKDQJHUV $WPRV (QYLURQ

6WHSKHQV%$1RYRVHODFDQG$-6LHJHO,PSDFWVRI+9$&)LOWUDWLRQRQ$LU&RQGLWLRQHU(QHUJ\&RQVXPSWLRQLQ
5HVLGHQFHV3URFHHGLQJVRI+HDOWK\%XLOGLQJV3DSHU
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

SA-12-C011
Design and Analysis of an Integrated Heat

and Energy Recovery Ventilation System
with Economizer Control for Net-Zero
Energy Solar Houses
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Carl Yu Chen Jiten Mistry Alan S. Fung, PhD, Peng Wey H. Leong, PhD, PEng Sumeet Jhingan EIT, LEED® AP
Student ASHRAE Member Student ASHRAE Member ASHRAE Member ASHRAE Member ASHRAE Member
ABSTRACT
There is a requirement now within residential and commercial building sectors to be as energy efficient as possible. Government
organizations are issuing more stringent energy efficiency regulations, including the mandatory energy efficiency requirements in the National
Building Code of Canada (NBC); supplementary SB-10 code requirements in the Ontario Building Code (OBC); and National Resources
Canada (NRCan) administered standards, including R-2000, the EnerGuide Rating System, and the ENERGY STAR for New Homes.
R-2000, which is a voluntary standard for residential houses, includes requirements such as energy efficiency, indoor air quality and the use of
environmentally friendly products and materials. As these standards and regulations are becoming more stringent, the focus has shifted towards
building airtight homes to provide significant energy savings. While an airtight house does offer significant savings, it also causes indoor air quality
(IAQ) issues due to a lack of adequate fresh-air delivery through the cracks. Therefore, mechanical ventilation becomes vital as evident by its
requirement in all these standards. Ventilation systems that are currently available in the industry are heat recovery ventilation (HRV) and
energy recovery ventilation (ERV) systems. HRV’s can be equipped in homes to recover sensible heat from the exhaust air onto the incoming
supply air, whereas ERV’s can recover both moisture and sensible heat from exhaust air to be used in the fresh incoming air in order to maintain
the indoor air quality and reduce the energy costs. As well as an economizer mode for free cooling when suitable conditions are conducive. A
combination of the individual HRV and ERV systems would be beneficial, which can lead to further energy conservation. Hence the purpose of
this project is to design and develop a prototype which will be utilized in net-zero energy solar houses, and it will be tested in the Ryerson
University campus lab and at the TRCA Archetype House in Toronto. The control system will be combined with the HRV, ERV, and
Economizer with the Air Handling Unit (AHU) which can control the zones separately and have features like recirculation between zones to
recycle the heat and reduce the energy costs. In order to do this project, a conceptual design of the combined system will be established followed by
schematics, controls systems and then finally the specifics like fan rating, the AHU capacity etc. A model will then be prepared in 2D and 3D,
which will later be used in manufacturing of the prototype. The model will then be tested to obtain data to determine its benefits and efficiency
under different weather conditions. An Excel based program, called TRVE, for conducting preliminary analysis and potential benefits of the
proposed HRV/ERV/Economizer ventilation system was developed. TRVE was used to determine when, where, and how such proposed
system could be beneficial in different regions of North America.
Carl Y. Chen and Jiten Mistry are BEng students in the Department of Mechanical Engineering, Ryerson University, Toronto, Ontario. Alan S.
Fung and Wey H. Leong are associate professors in the Department of Mechanical Engineering, Ryerson University, Toronto, Ontario.
Sumeet Jhingan is a Building Energy & Sustainability Analyst with Morrison Hershfield Limited.
82
ASHRAE Transactions
INTRODUCTION
The growing energy crisis and the growth of research and finding in both medical and technological fields have
increased the awareness on energy recovery and also the need of proper ventilation in closed environments. With improper
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
ventilation in homes the built up of excess moisture causes toxins, pollutants and gases, nurturing the growth of mold and
dust mites. Therefore, mechanical ventilation systems such as HRV and ERV have been introduced to recover latent and
sensible energy loss within homes. An HRV is normally designed for cooler climates to heat the cool fresh air using indoor
warm air through a sensible heat recovery core. The ERV, on the other hand, is beneficial for dry and polar climates where
moisture retention is crucial and tropical climates, where moisture removal is vital to reduce latent cooling load. As well,
certain weather conditions (like spring or fall where minimal heating/cooling is required) may only require an economizer
mode, which enables fresh air to be pumped into the home without any conditioning.
As North America is one of the largest energy consumers, they have taken steps to moderate this issue in various
ways (Natural Resources Canada, 2000). Due to Canada’s cold and dry climate, one third of the country’s energy is
consumed in the building sectors (Marrone, 2007). Therefore, residential heating and cooling is a major part of this energy
usage. In order to reduce the energy demand, modern homes are increasingly constructed to be air tight and mechanical
systems such as HRV and ERV are used for providing the required ventilation to maintain an acceptable level of indoor air
quality. Hence, the purpose of HERV is an integration of both HRV and ERV with an economizer feature.
OBJECTIVE
The main objective of this project is to design and build a prototype of a system which includes; HRV, ERV, and
Economizer, for the North America market. This system is called the HERV, which will integrate the benefits of each of
the discussed systems as well as providing a smart automated control system. The human control interface will enable the
user to control the level of fresh air input into the system and other parameters of the AHU. From the outside temperature
and relative humidity inputs, the HERV system will determine whether the system should run on the economizer, HRV,
ERV or both the HRV and ERV mode. Based on which mode is required, the control system will open and close the
appropriate dampers and turn on and off the appropriate fans.
Design Requirements:
Proposed Markets: Regions in North America where there are ranges of temperature from hot to cold, and
conditions from humid to dry
Efficiency: Maximize energy efficiency by reducing energy consumption and frictional losses
Size: The Heat and Energy Recovery Ventilation System (HERV) must be relatively small and compact similar to
most HRV and ERV systems on the market
Simplicity: The system has to be user friendly and operable by homeowners
METHODOLOGY
This project aims to use two methodologies to test the HRV/ERV/Economizer device - one an experimental
approach, the other - TRVE (A customized Excel based program for the evaluation of the potential benefits of the
proposed HERV system in different climatic conditions). The experimental method will be used once the device is
manufactured during the winter of 2012. In order to do the experimental analysis the device will be installed in a super
energy efficient TRCA Archetype house (Zhang et al., 2011) in Toronto.
The TRVE program methodology, however, is an analytical approach based on general HVAC formulas and
climatic condition of the region considered for the analysis. This program is of a particular benefit because based on the
energy savings obtained using the TRVE program for a specific region the user may or may not opt in for installing the
combined HRV/ERV/HERV device. Figure 1 illustrates the logic of and input and output of the TRVE program.
$6+5$(

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 1: Process flow chart for the TRVE Program
84
ASHRAE Transactions
The program is a Microsoft Excel Spreadsheet based tool that determines the heat loss in the house based on the
hourly weather data of the climatic location; building footprint; the extent of airtightness; and exterior envelope
information, including exterior walls, roof, windows, basement floor and underground walls and their U-values. The tool
uses the Bin Method principles for load calculations instead of the hourly load analysis used by softwares such as HAP,
eQuest, EE4, and HOT2000. Further, the tool takes the performance of the building’s furnace and air conditioning into
account for load calculations. It should be noted that the internal heat gains have also been accounted for; including the
heat load from occupants and the electrical appliances, which are based on NRCan’s R-2000, EnerGuide standards and are
included in the HOT2000 analysis tool as well. The tool also accounts for the ventilation requirements, based on the
ASHRAE Standard 62.2-2010: Ventilation and Acceptable Indoor Air Quality in Low-Rise Residential Buildings, and
compares the performance of the HERV system against a traditional ERV or an HRV system.

After conducting the preliminary analysis for various zones across North America using the developed TRVE tool,
the annualized average efficiencies of the proposed HERV system were obtained. It is clear from the Table 1 that the
combined HRV/ERV (HERV) is almost always as efficient as individual HRV or ERV alone except for the Tropical/
Megathermal Climates. The reasoning is due to the HERV’s functionality, where it has the capability to switch between
different modes; sensible recovery (when there is a minor difference in the relative humidities between indoors and
outdoors), latent recovery (when there is a minor difference in the temperatures between indoors and outdoors); Sensible &
Latent (when there is a significant difference in both temperatures and relative humidities between indoors and outdoors);
and the economizer control (when both temperature and relative humidity variations are minor). Due to this control system
that evaluates the difference between the indoor and outdoor setpoints (temperature and humidity), higher savings are
achieved compared to a traditional ERV or an HRV. Thus, the HERV system offers ease of control and the capability to
operate under different conditions and at the same time maintain required thermal comfort within the house.
Table 1: Weighted annual efficiencies for each system in different climatic zones
Zones Type of Climates HERV System ERV System HRV System

Zone A Tropical/Megathermal 38% 31% 77%
Climates
Zone B Dry Climates 61% 58% 73%
Zone C Temperate/Mesothermal 53% 20% 72%
Climates
Zone D Continental/Microthermal 59% 57% 69%
Climates
Zone E Polar Climates 62% 63% 63%
Note: The efficiencies shown for the HRV system represent sensible efficiency only, while the HERV and ERV systems show a combined Sensible & Latent
efficiency.
Furthermore, the annual operating savings, shown in Table 2, clearly indicates that a HERV system is always
beneficial in any type of climate. Table 3 indicates that the polar climates have the shortest payback period of 13 years.
Table 2: Annual savings for each system in different climatic zones

Zone A Tropical/Megathermal $7 -$156 -$43
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Climates
Zone B Dry Climates $171 $45 $30
$6+5$(

Zone C Temperate/Mesothermal $75 $59 $34
Climates
Zone D Continental/Microthermal $193 $71 $76
Climates
Zone E Polar Climates $282 $203 $125
Table 3: Payback periods for each system in different climatic zones

Zone A Tropical/Megathermal 487 No Payback No Payback
Climates
Zone B Dry Climates 21 32 130
Zone C Temperate/Mesothermal 54 No Payback 44
Climates
Zone D Continental/Microthermal 20 223 17
Climates
Zone E Polar Climates 13 8 9
It should be noted that the determined efficiencies of the HERV system are based on the individual hourly
performance data of the ERV and HRV in the TRCA Archetype house in Toronto. This, however, will be revised based
on the testing and final experimental results for the device. These will be obtained once the device is manufactured and
proper tests are performed, which will tentatively be performed before May 2012.
CONCLUSION AND ONGOING WORK

The HERV system is a new concept and hopefully this will be a breakthrough in the HVAC industry, which will
lead to greater and more efficient energy conservation. North America is one of the highest energy consuming countries in
the world, and the benefit from this concept could lead to further research and development in energy efficient homes. The
HERV system combines the benefits of both ERV and HRV systems, allowing it to maximize sensible and latent energy
recovery.
A virtual prototype of the HERV was created and evaluated using the developed spreadsheet program called
TRVE. From the preliminary analysis, it was recommended that a HERV be used for dry climates (Zone B),
continental/microthermal climates (Zone D) and possibly polar climates (Zone E) due to their expected annual energy
savings. HERV would be beneficial in these climates due to its energy savings and short payback period. However, its
potential high incremental cost may be a disadvantage in certain cities resulting in unacceptably high payback periods.
Further research and development will be conducted for the proposed HERV system. Experimental results for the
proposed HERV system will be obtained once the final design and prototype are created and properly tested at the Ryerson
Campus Labs as well as the TRCA Archetype House in Toronto.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
86
ASHRAE Transactions
Benefits of the proposed HERV System
The following is the schematic/ physical layout of the proposed HERV system—
Figure 2: Proposed HERV System
Most of the heat recovery systems available on market are either HRV or ERV or ERV without economizer controls.
However it has been found that if HRV used along with ERV has benefits and can be beneficial for a particular weather
conditions. When the temperature difference between the indoor and the outdoor is greater than 2 oC and the relative
humidity difference is greater than 5% there is a requirement for both sensible and latent heat transfer. ERV by itself can
accomplish this particular task, however, ERV cores are less efficient then the HRV cores. Hence it is useful if the indoor
air is allowed to exchange heat with the outside air and then the indoor air is allowed to exchange both heat and humidity
with the outside air as show in the schematic above. This way the efficiency for heat and moisture transferred is increased
because of the use of both the HRV and ERV in series.
Furthermore, the sensible efficiency of the HRV core is given by (Zhang, L-Z ., 2008)
------------------------------------Eq. 1
The latent efficiency of the ERV core is given by—
------------------------------------Eq. 2
And the combined efficiency is given by—
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

--------------------------------Eq. 3
Where
T1: Temperature of outside air before the heat exchanger (oC)
T2: Temperature of outside air after the heat exchanger (oC)
T3: Temperature of inside air before the heat exchanger (oC)d
ω1: Humidity ratio of outside air before the heat exchanger (kg/kg)
ω2: Humidity ratio of outside air after the heat exchanger (kg/kg)
ω3: Humidity ratio of inside air before the heat exchanger (kg/kg)
mf: Mass flow rate of fresh air
me: Mass flow rate of indoor air
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
H*= 2501 Дω/Дt
Using this above formulation and test data for separate efficiency of the HRV and ERV cores combined efficiency for the
above system can be calculated. Based on this efficiency payback periods and annual savings were calculated as seen in
Table 2 and 3 for the combined HERV system. As seen the combined HERV system has benefits over individual HRV or
ERV.
ACKNOWLEDGMENTS
The authors would like to thank ASHRAE for providing the funding for this student capstone project through its
Senior Undergraduate Project Grant Program.
NOMENCLATURE
ACH = Air Changes per Hour
AHU = Air Handling Unit
ERV = Energy Recovery Ventilator
HVAC = Heating, Ventilating and Air Conditioning
HERV = Heat and Energy Recovery Ventilator
HRV = Heat Recovery Ventilator
TRCA = Toronto and Region Conservation Authority
REFERENCES
American Society of Heating, Refrigerating and Air-Conditioning Enginners. (2009). 2009 ASHRAE Handbook -
Fundamentals. American Society of Heating, Refrigerating and Air-Conditioning Engineers
Barua, Rupayan, Zhang, Dahai, and Fung, Alan S. (2010). Analysis of Energy Performance of the Sustainable Archetype
House at Kortright Centre. 1st International High Performance Buildings Conference, West Lafayette, Indiana, USA,
July 12-15, 2010.
88
ASHRAE Transactions
Dembo, Aya, Ng, Ka Long Ringo, Pyrka, Agatha, and Fung, Alan. (2010) The Archetype Sustainable House: Investigating
its potentials to achieving the net-zero energy status based on the results of detailed energy audit. 1st International
High Performance Buildings Conference, West Lafayette, Indiana, USA, July 12-15, 2010.
Marrone, J. (2007). Getting to Zero: Defining the Path to Net Zero Energy Home Construction. Natural Resources
Canada.
Natural Resources Canada. Office of Energy Efficiency. (2000). Demand Policy and Analysis. Ottawa.
Rupayan Barua, D. Z. (2010). Implementation of TRCA Archetype Sustainable House Monitoring System. Department of
Mechanical and Industrial Engineering, Ryerson University.
Toronto Region Conservation Authority. (2009, Feburary 7). Archetype Sustainable House. Retrieved December 19 2011,
from TRCA Kortright Center: http://www.sustainablehouse.ca/rentals/vanEE. (n.d.). How VanEE Air Exchanger's
Work. Retrieved December 20 2011, from VanEE: http://www.vanee.ca/hw- vanee- airexchanger.php
Zhang, Dahai, Barua, Rupayan, and Fung, Alan S. (2010). Development of Monitoring System for the Sustainable
Archetype House at Kortright Centre. 1st International High Performance Buildings Conference, West Lafayette,
Indiana, USA, July 12-15, 2010.
Zhang, Dahai, Barua, Rupayan, and Fung, Alan S. (2011). TRCA-BILD Archetype Sustainable House--Overview of
Monitoring System and Preliminary Results for Mechanical Systems. ASHRAE Transactions, 117(2), 2011, pp. 597-
612.
Zhang, L.-Z. (2008). Total Heat Recovery: Heat and Moisture Recovery from Ventialtion Air. Nova Science Publisher, Inc.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

SA-12-C012
Validation Methodology to Allow Simulated

Peak Reduction and Energy Performance
Analysis of Residential Building Envelope
with Phase Change Materials
Paulo Cesar Tabares-Velasco, PhD Craig Christensen Marcus V. A. Bianchi, PhD, PE

Associate Member ASHRAE Member ASHRAE
ABSTRACT
Phase change materials (PCM) represent a potential technology to reduce peak loads and HVAC energy consumption in residential
buildings. This paper summarizes NREL efforts to obtain accurate energy simulations when PCMs are modeled in residential buildings: the
overall methodology to verify and validate Conduction Finite Difference (CondFD) and PCM algorithms in EnergyPlus is presented in this study.
It also shows preliminary results of three residential building enclosure technologies containing PCM: PCM-enhanced insulation, PCM
impregnated drywall and thin PCM layers. The results are compared based on predicted peak reduction and energy savings using two algorithms
in EnergyPlus: the PCM and Conduction Finite Difference (CondFD) algorithms.
INTRODUCTION
Energy storage in buildings can be accomplished by sensible, latent, or chemical means. From these three types, latent
storage using phase change materials (PCMs) has been the focus of multiple studies and considerations due to its potential
thermal energy storage per volume compared to sensible storage. Currently, there are multiple PCMs commercially available
that vary in type (salts, paraffins, fatty acids), encapsulation technology (micro and macro encapsulation), and melting
temperatures (covering a range useful for building wallboard and enclosure applications, 64-104oF (18-40oC)).
Previous PCM studies have shown that there are important benefits related to thermal comfort, energy savings, and
the potential for HVAC downsizing when thermal storage is added into buildings (Zhu et al. 2009). These benefits are
achieved through the appropriate design and selection of PCMs, because PCMs can reduce temperature fluctuations, and
shift heating and cooling demand (Khudhair and Farid 2004). Early building energy simulations efforts in this area focused
on wallboards impregnated with PCMs (Solomon 1979; Drake 1987; Tomlinson and Heberle 1990; Peippo et al. 1991;
Stovall and Tomlinson 1995). Later building simulation studies focused on different PCM applications: PCM impregnated
in fiber insulation (Kosny et al. 2010a; Shrestha et al. 2011), macro encapsulated PCMs in walls (Kosny et al. 2010b), and
floor heating systems that utilize shape-stabilized PCMs (Lin et al. 2004; Zhang et al. 2006). PCM energy modeling is a
complex task that requires sophisticated building energy simulation programs able to accurately simulate PCMs for different
applications in homes: attic, walls, and floor. There have also been efforts to benchmark PCM models with whole building
field data using a house with PCM distributed in the attic and wall insulation (Shrestha et al. 2011). However, the ability to
Paulo Cesar Tabares-Velasco is a Postdoctoral Researcher, Residential Buildings Research Group, National Renewable Energy Laboratory
(NREL). Craig Christensen is Principal Engineer, Residential Buildings Research Group, NREL. Marcus Bianchi was a Senior Engineer,
Residential Buildings Research Group, NREL. He presently is Building Science Program Leader at Owens Corning.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
90
ASHRAE Transactions
reproduce the experimental results in EnergyPlus was limited due to issues with the software code in multi-zone buildings.
The objective of this study is to validate the EnergyPlus Conduction Finite Difference (CondFD) algorithm for multi-
zone buildings and to investigate the potential benefits when PCMs are installed in homes. This study also summarizes
NREL validation efforts to conduct accurate energy simulations with PCMs, as this is part of an ongoing investigation with
the overall goal of optimizing PCMs in buildings.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
VERIFICATION AND VALIDATION METHODOLOGY
EnergyPlus can simulate PCMs only with the conduction finite difference (CondFD) solution algorithm. CondFD
discretizes walls, floors, and ceilings into several nodes and uses an implicit finite difference scheme to numerically solve the
appropriate heat transfer equations (EnergyPlus 2011). To model PCMs, the CondFD method is coupled with an enthalpy-
temperature function that users input to account for enthalpy changes during phase change. This function is used to
develop an equivalent specific heat at each time step that inputs the conduction finite difference equations (EnergyPlus
2011). Thus, accurate modeling of PCMs requires validation of the PCM and CondFD algorithms.
This validation work was completed at NREL following ASHRAE Standard 140 and NREL validation
methodologies, which consist of analytical verification, comparative testing, and empirical validation (Judkoff and Neymark
2006). Overall, the process is divided in two levels (wall or building) and the two algorithms (CondFD and PCM), as shown
in Table 1. The wall level tests were detailed tests that focused on a single wall subjected to particular boundary conditions
on both sides for a relatively short period of time. Thus, the wall level validation only tested the ability of the specific
CondFD and PCM algorithms to accurately model building envelope applications. The wall level tests for CondFD and PCM
algorithms were presented in previous studies (Tabares-Velasco and Griffith 2012; Tabares-Velasco et al. 2012). This was an
important step, but it did not guarantee that the algorithms would work with more realistic situations and different HVAC
systems.
Table 1. Verification and Validation
Verification/
CondFD PCM
Validation Level
Analytical:
x Variable k Analytical:
x Composite wall x Stefan Problem
x Const heat flux (2)
Comparative:
Wall x Periodic BC
x Heating 7.3
x Symmetry
Comparative: Empirical:
x H7.3 transient variable k x DuPont Hotbox
x H7.3 transient multilayer wall
Comparative: Evaluation:
x ASHRAE 140 Case 600 x ASHRAE 140 Case 600
Building 1
x BEopt retrofit house x BEopt retrofit house
x BEopt new house x BEopt new house
In contrast, the whole house tests of this study focus on an entire building, considering interactions between the
weather, building envelope, HVAC, and internal loads for periods of time that vary from few days to a year. Building-level
validation tests single-zone or multiple-zone houses and is mainly a comparison between Conduction Transfer Functions
(CTFs) and CondFD. Building level verification compares hourly surface temperatures, average zone temperatures, hourly

1
Building Energy Optimization (BEopt) is residential building energy optimization software. See http://beopt.nrel.gov/
$6+5$(

heating and cooling energy, and total heating and cooling energy between CondFD and CTFs. A building-level evaluation
of the PCM algorithm is presented in another study (Tabares-Velasco et al. 2012). However, PCM results could not be
compared to the CTF2 results because PCMs cannot be modeled using the CTF algorithm. Thus, this work was called
“evaluation” instead of “verification” (see Table 1). Nevertheless, the evaluation compared the run time and number of
iterations required for CondFD to converge for different PCM applications as a measure to confirm models were
converging without previous stability problems detected. All simulations had similar run times and required similar numbers
of iterations, which was consistent with the wall level verification for PCMs (Tabares-Velasco et al. 2012). Overall the PCM
model is now validated after wall level validation and building level evaluation. Thus this paper presents the comparative
verification for CondFD using the whole-building test (lower-left quadrant of Table 1).
WHOLE-BUILDING COMPARATIVE VERIFICATION OF CONDFD

The comparative building level verification for CondFD without PCMs consisted of:
1. ASHRAE Standard 140 Case 600: a simple one-zone building with lightweight construction (ASHRAE 2004).
2. BEopt retrofit house: a one-story 1,280 ft2(120m2), 1960s-era house with three bedrooms, two bathrooms, and an
unconditioned attic modeled in six out of the eight original U.S. cities considered in previous retrofit analysis
efforts (Polly, Gestwick et al. 2011)
3. BEopt new house: a two-story, 2500 ft2 (231 m2) house with three bedrooms, two bathrooms, a garage and an
unconditioned attic. The new homes were developed in BEopt according to the 2010 Building America Simulation
Protocols, which represents a house built according to the 2009 International Energy Conservation Code (IECC)
and federal appliance standards in effect as of January 1, 2010 (Hendron and Engebrecht 2010). More details about
the house can be found in the literature (Casey and Booten 2011).
BEopt houses were created in BEopt version 1.1 and the simulation files (idf) were later modified outside of BEopt to
allow comparison between CTF and CondFD. The foundation type of the homes varied according to location, as shown in
Figure 1. The verification was done using 1-minute time step (smallest allowed in EnergyPlus; default time step for CTFs is
10 minutes) for CondFD as well as for CTFs and other time steps. Preliminary, we have found that CondFD users should
use time steps less than or equal to 3 minutes. However, future research is necessary to address time step sensitivity more
fully.
The retrofit and new house validation consisted of three-day summer and winter tests (instead of full annual tests) to
reduce the run time. Initial findings from this procedure helped identify multiple problems in EnergyPlus that resulted in
average energy use differences of up to 30%. The problems were related to the modeling of 1) walls between zones (e.g.
walls between living space and garage/attic) and 2) adiabatic walls and massless walls. In both cases, CondFD was not
correctly distinguished between external and internal boundary conditions when adiabatic walls were presented (among
other minor bugs). Working with the EnergyPlus development team, the issues were resolved and the next release
EnergyPlus in 2012 (version 7.1) will have a more reliable model. The “corrected” EnergyPlus model was used to conduct
all simulations presented in this study. Figure 1 shows the cooling energy calculated using the CTF and CondFD algorithms
for the BEopt new houses. The green line represents the percent difference (on the right axis) between CTF and CondFD;
all were less than 0.4%. Additional annual simulations for Chicago and Phoenix houses showed that the annual percent
energy differences were even less than those calculated by the 3-day tests. All tests showed very close agreement between
CTF and CondFD.

2
The default conduction transfer function (CTF) solution algorithm in EnergyPlus cannot simulate materials with variable
properties such as PCMs (EnergyPlus 2011).
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
92
ASHRAE Transactions
3ͲDAYCooling EnergyUse
150 CTF CondFD %Diff 0.4%
0.3%
CoolingEnergy(kWh)
Difference(%)
100 Basement Crawlspace Slab
0.2%
0.1%
50
0.0%
0 Ͳ0.1%
Chicago Seattle Atlanta LA Houston Phoenix
Location
Figure 1. 3-Day Cooling Energy use calculated with EnergyPlus using the CTF and CondFD algorithms for BEopt
new houses without PCMs. The green line represents differences between the algorithms, with values displayed on the
right-Y axis.
DETAILED DIAGNOSTICS AVAILABLE WITH WHOLE BUILDING ANALYSIS FOR PCM SYSTEMS WITH
VALIDATED PCM MODEL
Once the CondFD and PCM algorithms were validated, a similar house to the BEopt new house was analyzed in
more detail in EnergyPlus for the Phoenix location and weather data using 2010 Building America Simulation Protocols
(Hendron and Engebrecht 2010). Thus, the house has natural ventilation as long as it can meet the cooling load and the
indoor temperature stays above the heating set point. Cooling and heating set points were 76oF (24.4oC) and 71oF (21.7 oC)
respectively with a COP of 3.9. The intend of this analysis is to show how the whole building analysis of a PCM system can
produce information useful in selecting the type, amount, and location of PCMs for improved energy performance.
However, it is not the intention of this study to perform a parametric analysis of PCMs or show the potential energy savings
from PCMs. The original building model generated in BEopt was modified to accommodate PCMs: walls and attic were
assumed to be stud-less and truss-less. This assumption was necessary as BEopt cannot model PCMs at this time. BEopt
calculates “equivalent” material layers for stud and cavity using the parallel path method and therefore there is currently no
easy way to correctly model PCM materials in parallel with other materials using BEopt. These assumptions resulted in R-
values for the stud-less wall of RSI-2.7 (RIP-15.3) and RSI-6.7 (RIP-40) for the truss-less attic. Table 2 shows the
properties of the PCM aggregate materials analyzed in this study: PCM distributed in insulation, PCM distributed in drywall,
and concentrated PCM layer. These three cases represent products that are already on the market or that have been
previously investigated (Kedl 1990; Kosny et al. 2010a; Kosny et al. 2010b). From these three materials, concentrated PCM
has more thermal storage capacity (in Btu/ft2 or kJ/m2) as it is the only concentrated application analyzed in this study.
Figure 2 shows predicted monthly cooling electric energy reduction for the analyzed PCM applications. All monthly
savings were divided by the annual energy use for the building without PCMs. The PCM applications analyzed are: PCM
distributed in drywall on the ceiling (DW_Ceil) and walls (DW_Walls), PCM distributed in wall insulation (Walls), PCM
concentrated at the middle of the wall cavity insulation (RCRWall) and a combination of RCRWall with PCM distributed in
Drywall (RCRWall-DW). The last application was selected because the two different PCM applications showed the best
performance during cooling season. Thus, it is added here to show the potential savings when multiple PCM applications
are used within a single building. For this particular case, the combination of RCRWall with PCM distributed in Drywall
(RCRWall-DW) obtained the highest annual percentage savings, which were around 4% or approximately 716 kBtu (210
kWh). However, more in depth parametric analysis is needed before any conclusions can be made as this study did not
analyze several melting temperatures or storage capacities in order to optimize energy and/or peak load savings.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

Table 2. Properties of PCM Strategies Analyzed
PCM Distributed PCM Distributed Concentrated PCM
Properties
Insulation Drywall Layer
% Weight of PCM in Wall Layer 40% 30% 100%
Equivalent Latent Heat 9.5 Btu/lb (68 kJ/kg) 14 Btu/lb (33 kJ/kg) 56 Btu/lb (130 kJ/kg)
Melting Temperature Range 83-87oF (28.5-30.5oC) 72-76 oF (22.5-24.5oC) 83-87oF (28.5-30.5oC)
0.2337 Btu in/(h ft R) 1.595Btu in/(h ft R) 1.109 Btu in/(h ft R)
Thermal Conductivity
(0.0337 W/m K) (0.23 W/m K) (0.16 W/m K)
Density 4.4 lb/ft (70 kg/m )
3 3 62.4 lb/ft3 (1000 kg/m3) 53.1 lb/ft3 (850 kg/m3)
0.229Btu/lb R 0.334 Btu/lb R 0.597 Btu/lb R
Specific Heat
(960 J/kg K) (1400 J/kg K) (2500 J/kg K)
Thickness Wall 3.5in (0.089m) 0.5in (0.0127m) 0.2in (0.005m)
Walls Latent Storage 37.2 Btu/ft2 (423 kJ/m2) 36.8 Btu/ft2 (420 kJ/m2) 48.6 Btu/ft2 (550 kJ/m2)
Amount of PCM 0.5 lb/ft (2.5 kg/m )
2 2 0.8 lb/ft2 (3.8kg/m2) 0.9 lb/ft2 (4.2kg/m2)
DW_Ceil Walls RCRWall Drywall DW_Walls RCRWall_DW

0.8% 45
(Monthsavings/yearenergyuse)
0.7% 40
EnergySavings(kWh)
0.6% 35
30
EnergySavings
0.5%
25
0.4%
20
0.3%
15
0.2% 10
0.1% 5
0.0% 0
Mar Apr May Jun Jul Aug Sep Oct Nov
Figure 2. Predicted cooling energy savings for different PCM strategies.
It is also interesting to see the seasonal performance variability of the different PCM applications: all PCM-drywall
applications performed the best during the swing seasons, while the system with the concentrated application offers more
continuous benefits. Moreover, the results show only small differences between installing the PCM drywall in the walls and
the ceiling (DW_Walls vs. DW_Ceil). Figure 3 shows the peak cooling reduction benefits for the same house and PCM
applications. Similar to energy savings, the PCM application with the highest reduction was the combination of RCRWall
with PCM distributed in Drywall. However, the highest peak reduction was obtained during April and May, months when
electric peak demand typically is not a concern as July and August are typically the months when local utilities struggle to
meet the cooling demand.
To help understand the behavior of the different PCM applications studied, Figure 4 shows the predicted hourly
cooling energy savings from four of the six PCM applications analyzed for four sunny days in April with diurnal
temperature variation from 59 to 95 oF (15 to 35oC). Energy savings from installing PCMs are shown in the left Y-axis. The
cooling energy use of the house without PCM (CoolingEnerg) is shown in the right Y-axis as a reference. During these
days, PCM savings peak at the same time when the cooling energy use peaks. This is due to the large temperature variation
that went below and above the melting temperature range of the PCMs. Thus, allowing them to melt and solidify daily.
Day-to-day variation is due to weather change: outdoor air temperatures were slightly warmer for the last two days. Overall
PCM location have an important role that needs to be consider when designing PCMs.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
94
ASHRAE Transactions
DW_Ceil Walls RCRWall Drywall DW_Walls RCRWall_DW
8% 0.30
PeakReduction(Monthenergy) 6% 0.25
PeakReduction(kW)
4% 0.20
0.15
2%
0.10
0%
Mar Apr May Jun Jul Aug Sep Oct Nov 0.05
Ͳ2% 0.00
Ͳ4% Ͳ0.05
Figure 3. Predicted peak cooling reduction from different PCM strategies.
CoolingEnergySavings andUse:4/05Ͳ4/08
RCRWall Walls DW_walls RCRWall_DW CoolingEner(kWh)
1.0 PeakDemand 2
CoolingEnergySavings(kWh)
CoolingEnergyUse(kWh)
0.8 1.6
0.6 1.2
0.4 0.8
0.2 0.4
0.0 0
0 12 24 36 48 60 72 84 96
Ͳ0.2 Ͳ0.4
Time(hrs)
Figure 4. Predicted hourly energy savings from different PCMs wall applications in April in Phoenix. Right Y-axis shows
hourly energy use for house without PCMs (CoolingEner).
Figure 5 shows the predicted hourly cooling energy savings for the same PCM applications shown in Figure 4, but
for four days in July with diurnal temperature variation from 86 to 113 oF (30 to 45oC). During these days, PCM energy
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
savings peak earlier than the peak cooling demand and overall energy savings are smaller than in April. This is due to the
mismatch between the thermal properties for the selected PCMs and the outdoor temperatures during the warm summer
weather; the PCMs selected melt between 83-87oF (28-30oC), so the outdoor temperature is above the melting range for
most of the four-day period. Thus, the chosen PCMs were not able to melt and solidify daily, which limited their ability to
store additional heat. Higher savings could be obtained if a PCM with higher melting temperature and/or storage capacity is
selected. Overall these results show strong seasonal variation for both energy savings and peak load reduction, which
suggests that the characteristics of optimal PCM solutions depend strongly on the user’s energy goals.
Figure 6 shows predicted monthly heating energy savings for the analyzed PCM applications. For the heating season,
the PCM application with drywall in the interior of the house obtained the best performance, with 25% predicted annual
savings. These savings could potentially be improved if PCM properties are optimized in future studies. Similar to the
cooling case, there were only small changes between PCM drywall installed in the ceiling and walls. For this particular
climate and building for these selected PCMs, percent heating energy savings were higher than percent cooling energy
savings. Despite the fact that annual onsite heating energy use represented about 1/2 of the total cooling onsite energy, the
$6+5$(

predicted absolute heating energy savings were almost 3 times higher than cooling energy savings for this particular house,
climate and PCM selection.
CoolingEnergySavings andUse:7/16Ͳ7/19
RCRWall Walls DW_walls RCRWall_DW CoolingEner(kWh)
0.30 PeakDemand 4
CoolingEnergySavings(kWh)
0.25
CoolingEnergyUse(kWh)
3.5
0.20
3
0.15
0.10 2.5
0.05 2
0.00 1.5
Ͳ0.05 0 12 24 36 48 60 72 84 96
1
Ͳ0.10
Ͳ0.15 0.5
Ͳ0.20 Time(hrs) 0
Figure 5. Predicted hourly energy savings from different PCM wall applications in July in Phoenix. Right Y-axis shows
hourly energy use for house without PCMs (CoolingEner).
DW_Ceil Walls RCRWall Drywall_All DW_Walls RCRWallͲDW a
8% 180
(Monthsavings/yearenergyuse)
7% 160
140
EnergySavings(kWh)
6%
EnenrgySavings
120
5%
100
4%
80
3%
60
2% 40
1% 20
0% 0
Jan Feb Mar Apr Nov Dec
Figure 6. Predicted heating energy savings from different PCM strategies.
CONCLUSION
This paper summarizes the approach to validate the CondFD and PCM algorithms in EnergyPlus. This study also
shows how the whole building analysis of a PCM system can produce information useful in selecting the type, amount, and
location of PCMs for improved energy performance. This application of whole building analysis is demonstrated for three
residential building envelope technologies containing specific PCMs: PCM distributed in insulation, PCM distributed in
drywall and concentrated PCM layer between insulation in the walls. These technologies are compared based on peak
reduction and energy savings using the CondFD algorithm in EnergyPlus. Preliminary results suggest that considerable
annual energy savings could be achieved using PCMs in residential buildings, but that careful design is needed to optimize
PCM solutions according to the specific user goals for peak demand and energy use reductions. Future research will include
more detailed parametric analyses to optimize PCM wall strategies as well as analyzing cooling and heating control
strategies.
ACKNOWLEDGMENTS
This work was supported by the U.S. Department of Energy under Contract No. DE-AC36-08-GO28308 with the
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
96
ASHRAE Transactions
National Renewable Energy Laboratory. The authors would like to thank Marc LaFrance, for supporting research on phase
change materials. The authors would like to thank Ben Polly of NREL for his critical reviews and Brent Griffith and Linda
Lawrie from the EnergyPlus developing team for their technical support.
REFERENCES
ASHRAE. 2004. ANSI/ASHRAE Standard 140-2004. Standard Method of Test for the Evaluation of Building Energy
Analysis Computer Programs. Atlanta, GA, ASHRAE: 151.
Casey, S. and C. Booten. 2011. Energy Savings Measure Packages: New Homes. Golden, National Renewable Energy
Laboratory: 205.
Drake, J. B. 1987. A study of the optima; transition tempreature of PCM wallboard for solar energy storage, Oak Ridge
National Laboratory. ORNL\TM -10210.
EnergyPlus. 2011. EnergyPlus Engineering Reference: The Reference to EnergyPlus Calculations, Ernest Orlando
Lawrence Berkeley National Laboratory: 1130.
Hendron, R. and C. Engebrecht. 2010. Building America House Simulation Protocols Golden, CO, National Renewable
Energy Laboratory: 80.
Judkoff, R. and J. Neymark. 2006. Model Validation and Testing: The Methodological Foundation of ASHRAE Standard
140; Preprint.
Kedl, R. J. 1990. Conventional wallboard with latent heat storage for passive solar applications.
Khudhair, A. M. and M. M. Farid. 2004. A review on energy conservation in building applications with thermal storage by
latent heat using phase change materials. Energy Conversion and Management 45(2): 263-275.
Kosny, J., S. Shrestha, D. Yarbrough, P. Childs, W. Miller, M. Bianchi, J. Smith, T. Fellinger, E. Kossecka and E. Lee.
2010a. Theoretical and Experimental Thermal Performance Analysis of Building Shell Components Containing Blown
Fiberglass Insulation Enhanced with Phase-Change Materials (PCM). Thermal Performance of the Exterior Envelopes
of Whole Buidings XI International Conference. Clearwater Beach, Florida: 13.
Kosny, J., S. Shrestha, T. Stovall and D. Yarbrough. 2010b. Theoretical and Experimental Thermal Performance Analysis of
Complex Thermal Storage Membrane Containing Bio-base Phase Change Material (PCM). Thermal Performance of
The Exterior Envelopes of Whole Buidings XI International Conference. Clearwater Beach, Florida 11.
Lin, K., Y. Zhang, X. Xu, H Di, R. Yang, and P Qin. 2004. Modeling and simulation of under-floor electric heating system
with shape-stabilized PCM plates. Building and Environment 39(12): 1427-1434.
Peippo, K., P. Kauranen, and P.D. Lund. 1991. Multicomponent PCM wall optimized for passive solar heating. Energy and
Buildings 17(4): 259-270.
Polly, B., M. Gestwick, M. Bianchi, R. Anderson, S. Horowitz, C. Christensen, and R Judkoff. 2011. A Method for
Determining Optimal Residential Energy Efficiency Retrofit Packages NREL Technical Report DOE/GO-102011-
3261. Golden, CO, National Renewable Energy Laboratory: 79.
Polly, B., N. Kruis, and D. Roberts. 2011. Assessing and Improving the Accuracy of Energy Analysis for Residential
Buildings. Golden, CO, National Renewable Energy Laboratory: 41.
Shrestha, S., W. Miller, T. Stovall, A. Desjarlais, K. Childs, W. Porter, M. Bhandari and S Coley. 2011. Modeling PCM-
Enhanced Insulation System and Benchmarking EnergyPlus against Controlled Data Field Building Simulation 2011:
12th Conference of International Building Performance Simulation Association Sidney.
Solomon, A. D. 1979. Design criteria in PCM wall thermal storage. Energy 4(4): 701-709.
Stovall, T. K. and J. J. Tomlinson. 1995. What are the potential benefits of including latent storage in common wallboard?
Journal of Solar Energy Engineering, Transactions of the ASME 117(4): 318-325.
Tabares-Velasco, P. C., C. Christensen, M. Bianchi. 2012. Verification and Validation of EnergyPlus Phase Change Material
Model for Opaque Wall Assemblies. Building and Environment DOI: 10.1016/j.buildenv.2012.02.019.
Tabares-Velasco, P. C. and B. Griffith Diagnostic test cases for verifying surface heat transfer algorithms and boundary
conditions in building energy simulation programs. Journal of Building Performance Simulation: DOI:
10.1080/19401493.2011.595501.
Tomlinson, J. J. and D. D. Heberle. 1990. Analysis of wallboard containing a phase change material, Reno, NV, USA, Publ
by IEEE.
Zhang, Y., K. Lin, Q. Zhang and H Di. 2006. Preparation, thermal performance and application of shape-stabilized PCM in
energy efficient buildings. Energy and Buildings 38(10): 1262-1269.
Zhu, N., Z. Ma, S Wang. 2009. Dynamic characteristics and energy performance of buildings using phase change materials:
A review. Energy Conversion and Management 50(12): 3169-3181.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

6$&
A Comparison of Thermal Comfort between

the Archetype Sustainable Homes at the
Kortright Centre for Conservation
Agatha Pyrka Alan S. Fung, PhD, PEng

Member ASHRAE
ABSTRACT
Thermal comfort plays a vital role in the well being of occupants and is therefore considered very important to control. This paper compares
thermal comfort of the semi-detached Archetype Sustainable Homes at the Kortright Centre in Vaughn, Ontario. This was done by following the
requirements of ASHRAE Standard 55 - 2010 Thermal Environmental Conditions for Human Occupancy. These semi-detached homes have
similar building envelope components however differ in HVAC technologies. The intent of House A is to demonstrate practices and technologies
currently in the residential building market. The intent of House B is to demonstrate advanced technologies such as thermal mass, zoned space
conditioning and in-floor heating that could be used by mass markets in the future. Results indicate House B is comfortable to a greater percent of
the population. House A did not meet the draft criteria under the local thermal discomfort requirement. This was a result of the combination of
drafts and lower air temperature during heating season measurements which reduced comfort levels. It is important to understand what makes a
home comfortable so that improvements to energy efficiency and demand management can be realized.
INTRODUCTION
Thermal comfort is defined as the condition of mind which expresses satisfaction with the surrounding environment
(ASHRAE, 2010). Since each occupant is unique, the sense of thermal satisfaction may be different from one individual to
another. To accommodate this, laboratory and field tests have been conducted to determine what thermal environmental
conditions are required to be accepted by a certain percentage of building occupants. The majority of this data pertains to
sedentary or near sedentary activity levels such as working at a desk or sitting on a couch. These results are used (or should
be) to design thermally comfortable buildings.
Similar houses with different heating and cooling systems will also have varying degrees of comfort. This paper
assesses comfort levels of two similar houses with different heating and cooling systems.
The Archetype houses at the Kortright Center for Conservation located in Vaughan, Ontario were used for this
research. These houses were developed by the Toronto and Region Conservation Authority (TRCA) in partnership with
Building Industry and Land Development (BILD) Association, both of who play an important role in research, education
and promotion of green buildings (Dahai, Barua & Fung, 2011). The houses were built with advanced materials and
mechanical equipment to ensure high energy efficiency and have been LEED Platinum certified (Dembo et al., 2010).
These semi-detached houses are comparable and a good fit for this comparison because they experience the same climatic
Agatha Pyrka has recently completed her Masters of Building Science degree from Ryerson University, Toronto, Ontario. Alan Fung is an
Associate Professor in the Department of Mechanical and Industrial Engineering at Ryerson University, Toronto, Ontario.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
98
ASHRAE Transactions
conditions, have similar layouts, square footage and building envelope characteristics.
ARCHETYPE SUSTAINABLE HOUSE DETAILS

The intent of House A is to demonstrate best practices and technologies currently in the residential building market.
The intent of House B is to demonstrate advanced technologies such as zoned space conditioning and in-floor heating that
could be used by mass markets in the future. Figure 1 shows the south elevation of the houses where House A is located
on the left side and B on the right.

Figure 1 South elevation of the Archetype Sustainable Houses
Various features and characteristics for each of the houses are summarized in Table 1.
Table 1. Summary of characteristics for both houses (Barua et al., 2010 and Zhang et al, 2011)
Characteristics House A House B
# of stories (not including basement) 3 3
Area (excluding basement) 261 m2 (2808 ft2) 317 m2 (3412 ft2)
Area of north facing bedroom 12.73 m2 (136.98 ft2) 12.38 m2 (133.29 ft2)
Insulation: above grade walls RSI-5.31 (R-32) RSI-5.31 (R-32)
Overall insulation: basement walls RSI- 3.54 (R-20) RSI-3.54 (R-20)
Overall insulation: roof RSI-7 (R-40) RSI-7 (R-40)
triple glazed; U-value = 1.59
Windows double glazed; U-value =2.19 W/m2 K
W/m2 K
Ground source heat pump
Heating & cooling season: Central variable
(GSHP) for both:
capacity two-stage air-to-air source heat
pump (ASHP) with multi-speed air handling
Heating season:-radiant in-floor
Heating / cooling system unit (AHU), supplying forced air to above
heating on all floors
basement zones. Supplemented during
Cooling season: GSHP supplies
heating season by mini boiler which heats
chilled water to the multi-zone
water and delivers to coils in AHU.
AHU.
Ventilations system Heat Recovery Ventilator Energy Recovery Ventilator

In addition to the characteristics presented in Table 1, it is also important to understand the specifics of the in-floor
heating and cooling system in House B since it is one of the main components of the HVAC system being analyzed. A thin
slab over a frame floor is used and composed of 2 cm (0.79 in) plywood subfloor below the 6.5 cm (2.56 in) concrete. The
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
tubing within the concrete is 1.25 cm (0.5 in) in diameter and is set 23 cm (9.05 in) off center from one another. The
finished wood floor on top of the concrete is 1.5 cm (0.59 in) thick with an RSI value of approximately 0.1 (R value of
0.57). For heating season, hot water is brought to a particular zone when there is a demand from the thermostat. Logically,
analysis has confirmed space heating increases with increased interior-exterior temperature differential. The thermal
resistance of the floor was optimal since a higher or lower value would results in less comfort. (Barua, Zhang, & Fung,
$6+5$(

2010)
For the purpose of this research, the most distinguishing factor between the two houses is the space conditioning
systems. Table 2 summarizes the space conditioning source for each house, where it goes, how it gets there and what
equipment is used to minimize cycling. For heating season, comparisons are made between a central two-stage variable
capacity air source heat pump (ASHP) by means of a single zone AHU (House A) and in-floor hydronic radiant heating
system (House B). For cooling season, comparisons are made between the central two-stage variable capacity ASHP by
means of a single zone AHU (House A) and multi zoned AHU fed by the ground source heat pump (GSHP), both of
which are forced air systems. (Barua et al., 2010 ; Barua Zhang, & Fung, 2009)
Table 2. Space conditioning summary for both houses (Barua et al., 2010 and Zhang et al., 2011)
Heating Season Cooling Season
House A House B House A House B
Space conditioning source Variable Capacity ASHP GSHP Variable Capacity ASHP GSHP
3 zones (one for 3 zones (one for
Zoning 1 zone (above basement) 1 zone (above basement)
each floor) each floor)
Heating delivery Two-Stage AHU, Two-Stage AHU, Multi-zone AHU,
In-floor piping
mechanism Forced Air Forced Air Forced Air
Equipment used to Variable capacity ASHP Variable capacity ASHP
Buffer tank Buffer tank
decrease cycling? with Two-Stage AHU with Two-Stage AHU
OBJECTIVE AND INTENT OF RESEARCH

The objective is to determine which house is more comfortable by testing the requirements set out in ASHRAE
Standard 55 - 2010 Thermal Environmental Conditions for Human Occupancy (ASHRAE, 2010). This analysis allows the
heating and cooling systems of the two houses to be compared to one another. This analysis also determines if the houses
meet requirements of the standard.
Occupant comfort is important because it addresses energy use and conservation. In the residential sector, many
existing and a vast majority of new houses use forced air furnaces and air conditioners. These systems work by ensuring the
whole house is conditioned. Rather than looking to heat or cool an entire building, perhaps the priority should be set on
ensuring specific areas occupied by occupants are comfortable.
METHODOLOGY
Indoor thermal comfort levels in the north bedroom on the second floor of each house were compared for heating
and cooling season. This location was chosen because they would be least likely disrupted by walk through tours as well as
not affected by incoming solar radiation from the larger south, east and west windows. The comparison was done by
measuring environmental parameters and assigning personal factors. Environmental parameters include dry bulb and
radiant temperatures, relative humidity and air velocity. Personal environmental factors include metabolic rate and clothing
insulation where values were chosen that reflect typical occupant activity and clothing levels for each season. ASHRAE
Standard 55 specifies metabolic rate needs to be between 1.0-1.3 and clothing levels between 0.5-1.0. A value of 1.1 met
was chosen for metabolic rate for simplicity since this is the value required when interpreting the Graphic Comfort Zone
Method (ASHRAE, 2010). Clothing levels were chosen to be 1.0 for heating season and 0.5 for cooling season also for
simplicity to comply with graphs in the standard.
These environmental and personal factors are assessed using the Predicted Mean Vote and Predicted Percent
Dissatisfied (PMV-PPD) index to determine if a space is thermally comfortable by measuring the following parameters:
Humidity Limits
Elevated Air Speed
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
100
ASHRAE Transactions
Local Thermal Discomfort (draft, vertical air temperature differences, floor surface temperature, radiant
asymmetry)
Temperature Variations with Time
Each of these parameters have well defined limits where each must be met to achieve compliance with the standard. The
standard also specifies how and when measurements should be taken. It should be noted that heating season
measurements could not be started until May and it was therefore not possible to achieve the temperature differential
required in ASHRAE Standard 55. Indoor air temperatures were raised to 24.0 °C (75.2 °F) to help meet this differential
however the lowest recorded exterior temperature during the night time low was 5.5 °C (41.9 °F) while the differential
required it to be 3.5 °C (6.3 °F). Interior temperatures during heating season were set to 24.0 °C (75.2 °F) for both houses.
Cooling season temperatures were set to 24.0 °C (75.2 °F) for house A and 23.0 °C (73.4 °F) for House B.
Two sets of measuring equipment were procured for this research. Each set includes an air velocity / thermometer
combo meter (accuracy of ±0.3 °C (0.54 °F) and 0.04 m/s (0.13 f/s)), a black globe ball with a diameter of 15.0 cm (5.9
inches) (accuracy of ±0.3 °C (0.54 °F)) to measure radiant temperature, and a relative humidity/thermometer combo meter
(accuracy of ±0.3 °C (0.54 °F)). Floor, ceiling and wall temperatures were taken using calibrator thermometer (accuracy of
±0.5 °C (0.9 °F)). Sensors were placed away from windows, air vents, lamps and the computer. Once the sensors made
contact with the surface, a reading was taken after the temperature stabilized which occurred after approximately 5 - 10
seconds.
The house with the highest percent of satisfied occupants is considered to be more thermally comfortable.
RESULTS
The subsections below will discuss the results for both heating and cooling seasons for each required parameter.
Measurements were taken as close as possible to the centre of each bedroom. Operative Temperature was calculated
by summing the dry bulb air and radiant temperature and then dividing by two. Air and radiant temperature measurements
were taken every five seconds at a height of 0.6 m (24 in.) to mimic a seated occupant. Averages over a 15 minute period
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
were used to limit fluctuations causes by students and tour groups entering and leaving the houses. Each house met the
operative temperature requirement for both heating and cooling season. Graphic Comfort Zone Method was used and the
intersecting points for operative temperature and percent relative humidity were plotted. For heating season, both houses
had operative temperature of 24.4 °C (75.9 °F) which corresponds with clo levels on the border of the 1.0 and 0.5 comfort
zones. Occupants typically wear clothing levels around 1.0 during heating season which would allow the operative
temperature to slightly decrease while maintaining within the comfort zone of 1.0 clo. If this clothing level is maintained,
the lowest operative temperature could be 21.5 °C (70.7 °F) which would conserve energy.
For cooling season, both houses were within the 1.0 clo comfort zone with corresponding operative temperatures of
23.4 °C (74.1 °F) and 22.6 °C (72.7 °F) for House A and B respectively. To decrease energy use, both houses could increase
temperatures to move into the 0.5 clo comfort zone which would more typically represent clothing worn during cooling
season.
Humidity Limits
Each house met the humidity limit requirement for both heating and cooling season. The measured relative humidity
limits were 28.9 % / 28.8 % during heating season and 51.1 % / 54.5 % during cooling season for House A and B
respectively. This was found by using the acceptable limit of the Graphic Comfort Zone Method of 0.012 kg water / kg dry air
(0.026 lb water / lb dry air).
$6+5$(

Elevated Air Speed
Increased air speeds allow the upper temperature limit to increase for occupant acceptability. Such elevated air speed
techniques have not consciously been employed for either house. In addition, it makes logical sense that the benefit of
elevated air speeds to increase upper temperature limits is useful for cooling season applications only. If the houses were
assessed based on current temperatures and air speeds, both would have insufficient air velocity to increase the upper
temperature limit. For the houses to be able to consider using elevated air speeds, the temperature would have to increase
and additional air flow would need to be introduced such as higher velocity mechanical equipment, opening windows or
installing fans. This parameter is therefore not applicable.
Draft
House A failed to meet this criteria for heating season when temperatures are set below 22.5 °C (72.5 °F) due to a
maximum cyclic air speed of 0.24 m/s (0.79 f/s) as shown in Figure 2. Air speed should not exceed 0.15 m/s (0.49 f/s)
unless it is under the occupant’s control which they are not. Due to this, House B was considered to be more thermally
comfortable.
Air Velocity (m / s)
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Time of Day (A.M)
Figure 2 Two hour air velocity profile of House A considering spatial and temporal average for May 10 from 4-6AM
Vertical Air Temperature Difference

Vertical air temperatures during heating season were not possible to take as specified in the standard for both houses.
Floor and ceiling temperatures were however taken and the difference was 1.8 °C (3.24 °F) and 0.5 °C (0.9 °F) for House A
and B respectively. Neither the ceiling nor floor were adjacent to an exterior envelope component and therefore with some
reservation, it can be assumed heating season measurements are below the maximum 3 °C (5.4 °F) limit.
Cooling season measurements for both houses met the criteria. The associated PD is 4 % and 0 % with a
corresponding ankle to head temperature difference of 2.0 °C (3.6 °F) and 0.4 °C (0.72 °F) for House A and B respectively.
House B was therefore more thermally comfortable for this criteria. The distance of the cooling vents from the sensors
may have affected this result where it was closer for House A versus House B.
Floor Surface Temperature

Both houses were well within the acceptable floor temperature range for both heating and cooling season. Floor
temperatures were found to be 24.2 °C (75.6 °F) and 24.6 °C (76.3 °F) for House A and B respectively during heating
season. For cooling season, House A floor temperature readings was 22.3 °C (72.1 °F) while House B was 23.7 °C (74.7
102
ASHRAE Transactions
°F). For heating season, both houses had a PD of 6% while for cooling season, House A had a PD of 7 % and House B, 6
%. House B was more comfortable to approximately 1 % of occupants and is therefore considered to be more thermally
comfortable.
Radiant Asymmetry
Both houses were well within the acceptable radiant asymmetry range for both heating and cooling season. For
heating season, House A’s warm ceiling results in a PD of approximately 1 %. The ceiling temperature for House B and
the walls of both houses represent less than 1 % PD. For cooling season, both houses had similar results and produce the
same PD of less than 1 %.
Temperature Variations with Time

The maximum allowable temperature difference in a 3 hour period is 3.3 °C (5.94 °F) and 1.1 °C (1.98 °F) for a 15
minute period (ASHRAE, 2010). Table 3 shows the maximum measured temperature difference for 15 minute and 8 hour
periods for heating and cooling season for each house. The 8 hour temperature profile was used to show greater
temperature variation. Both houses were well within the acceptable range for heating and cooling season.
Table 3. Maximum measured temperature difference for 15 minute and 8 hour periods for both
heating and cooling seasons
Heating Cooling
House A House B House A House B
8 Hour 1.1 °C (1.98 °F) 0.7 °C (1.26 °F) 1.6 °C (2.88 °F) 1.1 °C (1.98°F)
15 Minute 0.2 °C (0.36 °F) 0.1 °C (0.18 °F) 0.1 °C (0.18 °F) 0.1 °C (0.18 °F)
Figure 3 shows the temperature variations for House A and B during heating season for an 8 hour period. Both
houses maintained internal temperatures for a long period of time. For House A, the low temperature variance is a result of
its variable capacity two-stage air source heat pump which runs at part-load most of time preventing frequent on-off cycles.
It may only cycle on-off once or twice in a typical day. For House B, the low temperature variance suggests the thermal
mass in the room helps maintain temperature over a long period of time.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 3 Eight hour operative and dry bulb temperature profile for House A (left) and House B (right)
For cooling season, House B had lower temperature variation in all scenarios with the exception of the 15 minute
$6+5$(

profile where both houses had the same temperature variation of 0.1 °C (0.18 °F). House B was therefore more
comfortable when considering temperature variations.
CONCLUSION
The results section suggests House B met all the requirements of ASHRAE Standard 55 while House A did not meet
draft requirement during heating season. This is based on clothing levels between 0.5-0.7 which is inconsistent with
ASHRAE Standard 55 and is discussed further in the Contributions section. Regardless of whether House A met or did
not meet this particular requirement, House B was comfortable to a greater percent of the population when all parameters
are considered as shown in Table 4.
Table 4. Summary of PPD results and if houses met requirements
Requirement met or PPD
Parameter House A House B House A House B
Heating Season Cooling Season
Yes Yes Yes Yes
Humidity Limits
Elevated Air Speed n/a n/a
Draft (assuming T < 22.5°C/72.5°F) No Yes Yes Yes
Vertical Air Temperature Difference n/a 4% 0%
Floor Surface Temperature 6% 6% 7% 6%
Radiant Asymmetry 1% 1% <1% <1%
Temperature Variations with Time Yes Yes Yes Yes
The temperature variations with time analysis revealed both houses had the ability to maintain internal temperatures
for a long period of time before the threshold temperature is reached and the on-cycle is activated. This has positive
implications for both comfort and energy demand savings. In terms of energy demand, drifts could be timed to align with
high electricity use such as during hot summer afternoons. House A for example experienced a drift from 1 PM to past 5
PM during cooling season which with high energy demand periods in Ontario. A combination of mechanical equipment
and thermal mass to maintain internal temperatures could be used to off-set energy use during high demand times. Further
analysis should be conducted to determine the feasibility of this and associated energy and demand savings.
As shown in the results analysis section of this paper, PPD is dependent on clothing and activity levels, air speed and
temperature. Understanding how residential occupants behave in terms of setting and changing their thermostat, what they
wear, how often they open windows and doors and their activity level is very important for HVAC and house design. If
houses are designed to be within a certain PPD range, behaviors that influence these factors need to be researched first.
Rather than trying to change the behaviors of occupants, it may be worthwhile to build houses that compliment their
lifestyle. In other words, a PPD range should not be set until there is a better understanding of our behaviors not only on a
societal level but more specifically on a cultural basis as well.
CONTRIBUTIONS
There are a number of items the authors wish to address related to ASHRAE Standard 55 for the residential sector.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
First, ASHRAE Standard 55 2010 made changes to the draft local thermal comfort criteria. The standard explicitly
indicates draft could feel pleasant if occupants have a thermal sensation on the warm side. However, it is unclear how
drafts would make occupants feel when temperatures are above 22.5 °C (72.5 °F) and PMV is on the cool side of thermal
sensation scale. It is suggested the provision for increased draft needs to be discussed further when PMV is on the cool
104
ASHRAE Transactions
side to determine a limit for PPD.
Second, local comfort analysis is based on clothing levels between 0.5-0.7. This is inconsistent with ASHRAE
Standard 55 which expects individuals to wear clothing levels of 1.0 during heating season. Higher clothing levels decrease
the perception of local discomfort due to less exposed skin and increased insulation layers (ASHRAE, 2010). It is
suggested heating and cooling seasons have separate sections to address each. This is important because if an occupied
zone is being designed using the current criteria, the design temperatures may be unnecessarily high.
Third, the local comfort criteria for floor temperature is based on occupants wearing lightweight indoor shoes. This
standard can be applied to residential spaces however, occupants often wear less insulation on their feet such as socks or
even go bare. If bare feet are considered, the temperature range may decrease due to less insulation.
Unrelated to the standard, there is a lack of quality research specific to thermal comfort of forced air and in-floor
space conditioning systems for residential buildings. This may be the current situation simply because residential occupants
have control over their thermal environment. Larger buildings do not tend to offer individual control. For these larger
buildings, designing one that is comfortable for a high percent of individuals is extremely important or else complaints
would overwhelm building operators. Regardless of occupant controllability, this gap in research may hinder the
advancement of occupant comfort which is correlated to energy efficiency and demand savings.
ACKNOWLEDGMENTS
The authors would like to thank Regional Municipality of Peel, Regional Municipality of York, City of Toronto,
BILD, TRCA, MITACS/Accelerate Ontario, Reliance Home Comfort, and Union Gas for their financial support.
NOMENCLATURE
AHU Air handling unit
ASHP Air source heat pump
BILD Building Industry and Land Development
clo Clothing level (W/m2)
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
GSHP Ground source heat pump

met Metabolic rate (W/m2)
PMV Predicted Mean Vote
PPD Predicted Percent Dissatisfied
TRCA Toronto and Region Conservation Authority
REFERENCES
ASHRAE. 2009. 2009 ASHRAE Handbook-Fundamentals. Atlanta: American Society of Heating Refrigeration and Air
Conditioning Engineers, Inc.
ASHRAE. 2010. ASHRAE Standard 55 - 2010. Thermal Environmental Conditions for Human Occupancy. Atlanta: American
Society of Heating Refrigeration and Air Conditioning Engineers, Inc.
Barua, R., Zhang, D., & Fung, A.S. 2009. Implementation of TRCA Archetype House Monitoring Systems. Report
prepared for TRCA, Reliance House Comfort and Union Gas.
Barua, R., Zhang, D., & Fung, A.S. 2010. Analysis of Energy Performance of the Sustainable Archetype House At the
Kortright Centre. 1st International High Performance Buildings Conference, July 12-15, 2010. Purdue University, West
Lafayette, USA.
Dembo, A., K.L.R. Ng, A. Pyrka, and A.S. Fung. 2010. The Archetype Sustainable House: Investigating its potentials to
achieving the Net-zero energy based on the results of detailed energy audit. 1st International High Performance
Buildings Conference, West Lafayette, Indiana.
Zhang, D., Barua, R., & Fung, A.S. 2011. TRCA-BILD Archetype Sustainable House - Overview of Monitoring System
and Preliminary Results for Mechanical Systems. ASHRAE Transactions 98 (2)
$6+5$(

SA-12-C014
Long-term Evaluation of Internal Thermal

Comfort with Different Kinds of Glazing
Systems and Window Sizes: from Energetic
Considerations to Users’ Comfort
A. Gasparella, PhD F.Cappelletti, PhD G. Pernigotto
P. Romagnoni, PhD
ASHRAE Member
ABSTRACT
Comfort limits are generally expressed in relation to the thermal balance of human body as a whole by two indexes proposed by Fanger in 1970:
the Predicted Mean Vote, PMV, and the Predicted Percentage of Dissatisfied, PPD, which is correlated to the PMV. Inside a building three of
the ambient parameters determining these values, i.e. the air temperature, the air humidity and the air velocity can be controlled by the system,
while the mean radiant temperature depends on the surfaces temperature that means on the envelope characteristics. Glazed surfaces generally less
performing in terms of thermal resistance than opaque components can affect the general human sensation in terms of thermal comfort. From energy
point of view the presence of large window surfaces is preferable during the heating season because solar gains through glazing can overcome heating
losses from glazed envelope. In contrast during the cooling season more attention has to be paid in order to limit the inlet of solar radiation which
causes the increment of cooling load. As already underlined by the authors in a previous paper, the use of particular glass t ypes, with low thermal
transmittance and high solar factor which is the best choice for heating energy savings, can considerably worsens the cooling performance. Moreover
considering the user point of view, the choice of glazing type and the design of windows on a façade may depend on comfort consideration besides
energetic evaluations. In this paper a study of internal environment conditions in a simple case study is presented considering 192 configurations for
an office application in three European climates. Energetic evaluations are represented with comfort ones.
INTRODUCTION
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
The contribute of the windows physical properties on the building energy need is, without doubt, fundamental in
determining buildings energy need especially when the building opaque envelope is well insulated. Many studies have
analyzed the importance of the glazing system in the optimization of energy need both during the heating and the cooling
season. In previous researches (Gasparella et al., 2011) the authors underlined the fact that energetic optimization is not
only dependent on the use of insulating glasses, but also on the quantity of solar radiation entering through the windows. In
particular in European localities with latitude between 40°N and 49°N the use of large glazings enhances buildings winter
performance, especially for the south orientation. Moreover the use of windows with low thermal transmittance is useful
F. Cappelletti is assistant professor in the Department of the Research, Università IUAV, Venice, Italy. A. Gasparella is associate professor in
the Faculty of Science and Technology, Free University of Bozen-Bolzano, Bolzano, Italy. G. Pernigotto is PhD Student in the Department of
Management and Engineering, Università di Padova, Padova, Italy. P. Romagnoni is full professor in the Department of the Research,
Università IUAV, Venice, Italy.
106
ASHRAE Transactions
only if accompanied by high solar transmittance (or SHGC), with some drawbacks in summer. However the choice of
energy optimized glazing system can affect thermal comfort of occupants, which is the topic of the present paper.
The long-term evaluation of internal thermal comfort especially in buildings with large glazed façade and in relation to
the reduction of energy need has already been performed by many authors. Hwang et al. (2011) developed a similar
investigation on the effect of different envelope parameters on thermal comfort. Using the PMV as the comfort index they
carried out a parametric study considering the lobby of a tourist service and changing glazing types, depths of overhang and
glazing areas. Their analysis compared the cooling energy need of a space controlled by a thermostat with the same space
with a PMV-based control evaluating the standard Mean Radiant Temperature and the PMV with and without the effect of
the solar radiation through the glazed areas directly reaching the occupants (La Gennusa et al., 2005, Singh et al., 2008).
Recently some authors have underlined the importance of taking into account of adaptation possibilities in order to
reduce energy needs (Nicol et al., 2002, Ferrari et al., 2011), but the Technical Standard EN ISO 15251:2007considers this
approach only for buildings without mechanical cooling system.
In this paper a parametrical study of indoor conditions in a simple case study is presented considering 192
configurations for an office application in three European climates (Rome, Milan and Paris-Trappes). The long-term
analysis has been carried out considering two seasons of six months each (summer from 1 st April till 30th September and
winter from 1st October till 31st March) and using the traditional Predicted Mean Vote index as a comfort index because the
indoor air temperature is supposed to be controlled by mechanical system during the whole year. The Mean Radiant
Temperature has not been corrected to consider the effect of the diffuse and beam solar radiation through the windows
directly reaching the occupants. This simplification doesn’t affect the conclusions in the comparison of the different glazing
systems and it follows the Technical Standard methodology (EN ISO 7730:2006 and EN ISO 15251:2007).
MODEL AND METHOD

A simple three-dimensional office model has been considered and a parametric simulation plan has been developed.
The base module is an open office box shaped (fig. 1 a), with a square floor of 100 m2 / 1076.391 ft2 and an internal height
of 3 m / 9.843 ft. The roof is horizontal and the floor is considered adiabatic, as could be for the top floor of an office
building. The four vertical walls are oriented towards the four cardinal directions.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Both horizontal and vertical walls are composed by a two-layer structure (table 1): the internal layer is a clay block of
0.200 m / 0.656 ft of thickness and the external one is an insulation layer, whose thickness is dependent on the simulation
plan. The solar absorptance coefficients have been set to 0.6 for the floor (internal side) and 0.3 for the vertical walls and
the roof (both sides). The wall emissivity is 0.9, both for the internal and the external side. Different kind of glazings have
been considered, following the simulation plan. A wooden frame with a thermal transmittance of 1.2 W/(m 2 K) / 0.211
Btu/(h ft2 °F) and a class-1 edge correlation have been chosen. 192 configurations to analyze were obtained by changing
insulation level, the kind of glazings, the glazing size, the windows distribution and the climatic conditions (table 2).
Hourly climatic data were calculated from average monthly values from the Italian Standard UNI 10349:1994 for
Milan and Rome and from TRY weather files (CEC, 1985) for Paris using the TRNSYS subroutine Type 54 Weather Data
Generator.
The office is occupied from 8:00 am to the 6:00 pm, Monday to Friday. An ideal air system with two setpoints and a
dead band has been implemented, all year long. The heating system has a setpoint equal to 20°C / 68 °F from 7:00 am to
6:00 pm and equal to 15 °C / 59 °F for the rest of the time (i.e., when the office is unoccupied). The cooling system is on
only when the office is occupied and has a setpoint of 26 °C / 78.8 °F.
Table 1. Properties of the opaque envelope layers
Thermal
Parameter Density Specific Heat
Conductivity
Clay layer 0.250 W/(m K) / 0.144 Btu/(h ft °F) 850 kg/m3/ 53.016 lb/ft3 840 J/(kg K) / 0.201 Btu/(lb °F)
Insulation layer 0.040 W/(m K) / 0.023 Btu/(h ft °F) 40 kg/m3/ 2.495 lb/ft3 1470 J/(kg K)/0.351 Btu/(lb °F)
$6+5$(

The ventilation rate is set equal to 1.58 ach/h during the occupation, in accordance to the Italian technical standard
UNI 10339:1995, and 0.3 ach/h, as infiltration, for the rest of the time. The internal gains in accordance with the ISO
13790:2008 are 20 W/m2 / 6.337 Btu/(h ft2) during occupation, and 2 W/m2 / 0.634 Btu/(h ft2) otherwise, a half
convective and a half radiant.
The occupants are assumed with a sedentary activity corresponding to 1.2 met (ISO 15251:2007) and a clothing level
of 1 clo in winter and of 0.5 clo in summer. To describe the PMV results simple statistical indicators, aimed at emphasizing
the distribution of the hourly results, were used such as the minimum and the maximum for each configuration, the median
and the interquartile range, referring only to the occupancy period. In particular, the median represents the value which
separates the upper half of the distribution from the lower half and the interquartile range the distributions of the points -
the larger is the IQR the more dispersed are the data. The PMV have been evaluated for 9 positions in the room, as in
figure 1 b - each one at 1.1 m/3.6 ft from the floor level. The hourly standard deviation between the 9 points considered
for the PMV calculation is very low (less than 0.1) and so it was chosen to represent a mean value instead of analyzing all
the 9 points separately for each configuration.
The calculation were performed in TRNSYS 17.0 which enables 3D modeling of the ambient radiation exchanges.
Table 2. Parameters values varied in the considered configurations and symbols
Parameter Values
Insulation 1. Poor insulation: 0.050 m / 0.164 ft of polystyrene
2. High insulation: 0.100 m / 0.328 ft of polystyrene
Glazings 1. Double Glazings with high SHGC (H):Ugl= 1.140 W/(m2 K)/ 0.201 Btu/(h ft2 °F); SHGC = 0.608
2. Double Glazings with low SHGC (L): Ugl= 1.099 W/(m2 K)/0.193 Btu/(h ft2 °F); SHGC = 0.352
3. Triple Glazings with high SHGC (H): Ugl= 0.613 W/(m2 K)/0.108 Btu/(h ft2 °F); SHGC = 0.575
4. Triple Glazings with low SHGC (L): Ugl= 0.602 W/(m2 K)/0.106 Btu/(h ft2 °F); SHGC = 0.343
Glazing Size 1. Size1 (S1): Agl = 11.670 m2 / 125.615 ft2 (in each façade with windows)
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
2. Size2 (S2): Agl = 23.340 m2 / 251.230 ft2 (in each façade with windows)
Windows 1. East (E)
distribution 2. East + West (E+W)
3. South (S)
4. South + North (S+N)
Location 1. Rome (Latitude N 42° 54’ 39’’; Heating Degree-Days (HDD18): 1420 K d/2556 °F d; Cooling Degree-
Days (CDD18): 827 K d/1488 °F d)
2. Milan (Latitude N 45° 27’ 51’’; HDD18: 2249 K d/4048 °F d – CDD18: 686 K d/1235 °F d)
3. Paris (Trappes; Latitude N 48° 46’ 0’’; HDD18: 3015 K d/5427 °F d – CDD18: 52 K d/94 °F d)
7 (2,8,1.1) 8 (5,8,1.1) 9 (8,8,1.1)
4 (2,5, 1.1) 5 (5,5, 1.1)6 (8,5, 1.1)
1 (2,2, 1.1) 2 (5,2, 1.1)3 (8,2, 1.1)
(a) (b)
Fig. 1. (a) Office module (large size windows(S2) up and small size down (S2)) (b) Coordinates of the positions for the PMV calculations
108
ASHRAE Transactions
RESULTS
Rome, Milan and Paris poor insulated envelope - Winter

As regards Rome, for a given kind of glazing increasing the size of the windows generally reduces the heating energy
needs, but less evidently for the East orientation, for which enhanced solar gains are probably balanced by enhanced
dispersions. Adding windows on the opposite side when the increased area is north exposed increases the heating energy
needs. Winter cooling needs become significant when increasing the window area only for the wider size windows (S2) and
for the orientation East+West, South and South+North, but only for the high SHGC solutions.
In parallel increasing window area and/or adding windows on the opposite side improve thermal comfort moving the
PMV distribution to the positive side. The PMV interquartile ranges, which generally are within ±0.5, are also widened
towards the higher values by larger window area and in East+West, South and South+North orientations.
As regards the type of glazing, there are small differences between double and triple glazings energy needs, when
equivalent SHGC glazings are compared. Triple glazings are slightly better performing than equivalent SHGC double in
terms of heating energy needs especially for larger window areas and for East and West orientation. Their behavior tends
worsen somewhat when considering cooling needs, mainly for the reason that in this climate the heating needs depend
more on solar gains than on thermal losses (Gasparella et al. 2011).Negligible differences appear when comparing the PMV
distributions. In that case, although during the 24 hours a higher thermal resistance of triple glazing should provide a higher
surface temperature with respect to the double glazings, during the occupation time this difference is reduced by the effect
of solar radiation and in part also by the internal radiation gains (Singh et al., 2008). A wider PMV interquartile range, in
particular with higher values for high SHGC glazings and larger window areas, is in accordance with this.
Also in Milan, for each type of glazing increasing the size of the windows generally reduces the heating energy needs,
except for the East orientation, but less evidently than in Rome. Adding windows on the opposite side is ineffective or
slightly penalizing when the increased area is north exposed. Winter cooling needs are almost always null.
Increasing window area and/or adding windows on the opposite side improve thermal comfort moving the PMV
distribution toward the positive side. The interquartile ranges are narrower than in the case of Rome and generally overlap
the value -0.5. The distribution of PMV values are also widened towards the higher values by larger window area in South
and South+North orientations with high SHGC.
Again there are small differences between double and triple glazings with similar SHGC values. Triple glazings are
slightly better performing than equivalent SHGC double in terms of heating energy needs not only for larger window areas
and for East and West orientation. As in this climate the dependence of heating needs on thermal losses tends to increase,
even triple glazing with low SHGC performs generally as well as double with high SHGC, confirming the results of some
previously cited works. The differences are negligible when comparing the PMV distributions of different kind of glazings.
Again a wider PMV interquartile range is shown, in particular with higher values for each configuration with high SHGC
glazings and for larger window areas.
In the climate of Paris increasing the size of the windows for each type of glazing does not affect so much the heating
energy needs. For orientation East the winter energy needs increases as also happens when adding windows on the opposite
side in particular for South orientation. Winter cooling needs are always null.
This time increasing window area and/or adding windows on the opposite side is quite ineffective on thermal comfort
moving only the maximum values toward the positive side. South and South+North orientations have the wider
dispersions. The interquartile ranges are narrower than in the case of Rome and Milan and generally between -0.75 and -0.5.
As regards the type of glazing, there are some differences between double and triple glazing seven with similar SHGC.
Triple glazings are performing better than the double with equivalent SHGC in terms of heating energy needs in all the
considered orientations. Again the triple glazing with low SHGC performs generally as well as double with high SHGC,
confirming the results of some previously cited works. The differences are negligible when comparing the PMV
distributions of double and triple glazings. Higher maximum values are connected to South and South+North orientations.
$6+5$( --`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---

Rome, Milan and Paris poor insulated envelope - Summer
As regards Rome, no heating needs are present in summer. For each type of glazing and for each orientation
increasing the size of the windows strongly increases the cooling energy needs. The largest cooling needs are found for
orientation East which is also the orientation with the most important sensitivity to the window size and to the addition of
windows on the opposite side (East+West).
The distribution of PMV values is tighter than in the winter case and interquartile ranges generally overlap or are
pretty close to the value +0.5, although there is a larger difference between maximum and minimum values. Increasing
window area and/or adding windows on the opposite side decrease thermal comfort moving the PMV distribution to the
larger values. As for the cooling needs, the largest PMV values are found for orientation East and East+West.
Concerning the impact of the type of glazing on energy needs, in summer the differences between double and triple
glazings become negligible, when comparing glazings with similar SHGC. This appears to be the only relevant parameter in
this climate in which the thermal losses have modest impact on the building heat balance, that is in accordance with the
findings of some previously cited works. The differences keeps even more limited comparing the PMV distributions of
glazings with similar SHGC. A wider dispersion of the PMV and a larger sensitivity to the window size are present in
particular, with higher values for each configuration with high SHGC glazings and for larger window areas.
As regards Milan, no heating needs are present. Again increasing the size of the windows strongly increases the
cooling energy needs. This time the cooling needs for orientation East do not prevail over the South orientation but still
have the most important sensitivity to the addition of windows on the opposite side (East+West).
The distribution of PMV values and the interquartile ranges tend to be wider than in the case of Rome and in the
winter case. They generally keep overlapping or are close to the value +0.5, maintaining the large difference between
maximum and minimum values. As in Rome, increasing window area and/or adding windows on the opposite side decrease
thermal comfort moving the PMV distribution to the larger values. As for the cooling needs, the largest PMV values are
found for orientation East+West.
As regards the type of glazing, the differences between double and triple glazings with similar SHGCs are negligible,
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
either considering the energy needs or the PMV values. The cooling energy needs are lower for low thermal transmittance
glazings. The dispersion of PMV values is really uniform both for high and for low SHGC glazings although the latter have
slightly lower values.
As regards Paris, limited heating needs are present. Again increasing the size of the windows strongly increases the
cooling energy needs. The cooling needs for orientation East are very similar to the South ones but have a huge sensitivity
to the addition of windows on the opposite side (East+West).
The distribution of PMV values and the interquartile ranges is more than doubled with respect to the case of Milan
and. They generally are comprised between -1.0 and +0.5, maintaining the large difference between maximum and
minimum values. Increasing window area and/or adding windows on the opposite side increase thermal comfort moving
the PMV distribution to larger values. As in Milan, the largest PMV values are found for orientation East+West with large
window size.
As in Milan the differences between double and triple glazings are always negligible, when comparing glazings with
similar SHGCs, either considering the energy needs or the PMV values. The cooling energy needs are very low for low
thermal transmittance glazings. Again the dispersion of PMV values is really uniform both for high and for low SHGC
glazings although the latter have slightly lower values.
Rome, Milan and Paris high insulated envelope

The results and trends are so similar to the ones of poor insulated envelope that only the differences need to be
described, avoiding to report a complete graphical representation. In the case of high insulated envelope winter heating
energy needs are obviously reduced (generally by around 2 GJ/1.90 million Btu in Rome, 4 GJ/3.80 million Btu in Milan
110
ASHRAE Transactions
and 5 GJ/4.74 million Btu in Paris).Winter cooling needs increase in Rome but do not appears in Milan and Paris. They
become pretty important for East+West, South and South+North orientations. In Rome winter PMV distributions are very
similar to the poor insulated cases although the values rises by some decimal point. In Milan the interquartile ranges
increase for the large size and South or South+North orientation. In Paris PMV values are slightly larger while the
interquartile ranges do not increase significantly for the large size and South or South+North orientation as in the cases of
Milan.
In summer only the increase of the cooling energy needs (by approximately 1.5-2 GJ/1.42-1.90 million Btu in all the
considered climates) appears evident although it is a little larger for the East+West configuration in Rome. The PMV
interquartile ranges are slightly reduced in Milan and in Paris, where they are generally between -1.0 and +0.5 also for low
SHGC glazings.
CONCLUSIONS
In the present paper the long-term thermal comfort conditions and the seasonal energy performance for an office
application have been analyzed. The effects of different kinds of glazing systems have been evaluated in order to identify
the importance of the thermal transmittance and of the SHGC in relation to different orientations, window size and
envelope insulation for three European climatic conditions. The results confirmed the strong influence of the SHGC also
on the winter energy needs as already found in some previous works. In particular double glazings with high SHGC tend to
have at least as good performance as triple with low SHGC. Low SHGC double or triple glazings are preferable in summer
in particular for East orientation. A reasonable trade-off between winter and summer energy performance probably should
be pursued with the use of suitable shading devices.
Concerning the comfort conditions, the SHGC again is the controlling parameter, in particular in winter, while the
glazing thermal transmittance appears to maintain a limited effect. This is probably attributable to the fact that during the
occupancy period in which the comfort conditions are evaluated, the internal surface temperature depends on the radiation
gains and on solar radiation in particular. High SHGC triple glazings appear to be slightly preferable to double glazings in
Rome, Milan and Paris, providing suitable shading devices for the summer season.
REFERENCES
CEC (Commission of the European Communities). 1985. Test Reference Years TRY, weather data sets for computer
simulations of solar energy systems and energy consumption in buildings. Brussels, Belgium.
S. Ferrari and V. Zanotto. 2012. Adaptive comfort: Analysis and application of the main indices. Building and Environment
49: 25-32.
A. Gasparella, G. Pernigotto, F. Cappelletti, P. Romagnoni, P. Baggio. 2011. Analysis and modeling of window and glazing
systems energy performance for a well-insulated residential building. Energy and Buildings 43 (4): 1030-1037.
R.L. Hwang, S.Y. Shu. 2011. Building envelope regulations on thermal comfort in glass facade buildings and energy-saving
potential for PMV-based comfort control. Building and Environment 46: 824-834.
M. La Gennusa, A. Nucara, M. Pietrafesa and G. Rizzo. 2007. A model for managing and evaluating solar radiation for
indoor thermal comfort. Solar Energy 81: 594–606.
J.F. Nicol and M.A. Humphreys. 2002. Adaptive thermal comfort and sustainable thermal standards for buildings. Energy
and Buildings 34: 563–572.
M.C. Singh, S.N. Garg and R. Jha. 2008. Different glazing systems and their impact on human thermal comfort - Indian
scenario. Building and Environment 43: 1596–1602.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

Double H Double L Triple H Triple L
E E+W S S+N E E+W S S+N E E+W S S+N E E+W S S+N
S1 S2 S1 S2 S1 S2 S1 S2 S1 S2 S1 S2 S1 S2 S1 S2 S1 S2 S1 S2 S1 S2 S1 S2 S1 S2 S1 S2 S1 S2 S1 S2
3
1
PMV - Rome
-1
-2 Double H Double L Triple H Triple L

-3 E E+W S S+N E E+W S S+N E E+W S S+N E E+W S S+N
3
1
PMV - Milan
-1

3
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
1
PMV - Paris
-1
-2
-3
4
2
0
-2
-4
Energy needs [GJ]
-6
-8
-10
-12
-14
-16
-18
-20
Rome cooling Milan cooling Paris cooling
-22 Rome heating Milan heating Paris heating
Fig. 2. Rome, Milan and Paris, poor insulated envelope –Winter PMV Distributions (the upper dot represents the maximum, the
intermediate the median, the under dot the minimum and the box the interquartile range) and winter Energy Needs.
112
ASHRAE Transactions
E E+W S S+N E E+W S S+N E E+W S S+N E E+W S S+N
3
1
PMV - Rome
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
0
-1

3
1
PMV - Milan
-1
-2
3
1
PMV - Paris
-1
-2
-3
26 Rome cooling Milan cooling Paris cooling
24
Rome heating Milan heating Paris heating
22
20
18
Energy needs [GJ]
16
14
12
10
8
6
4
2
0
-2
Fig. 3. Rome, Milan and Paris, poor insulated envelope – Summer PMV Distributions (the upper dot represents the maximum, the
intermediate the median, the under dot the minimum and the box the interquartile range) and summer Energy Needs.
$6+5$(

SA-12-C015
Occupant Perceptions of an Indoor Thermal

Environment in a Naturally Ventilated
Building
Salman Ilyas, PE Ashley Emery, PhD Judith Heerwagen, PhD

Member ASHRAE Fellow ASHRAE
Dean Heerwagen
Life Member ASHRAE
ABSTRACT
$VWURQJHPSKDVLVLVFXUUHQWO\EHLQJSODFHGRQWKHXVHRIQDWXUDOYHQWLODWLRQDVDPHDQVIRUSURYLGLQJDVDIHDQGKHDOWKIXOLQGRRUHQYLURQPHQWDV
SDUWRIJUHHQEXLOGLQJSURJUDPV7KHUHKDVEHHQDQLQFUHDVLQJLQWHUHVWLQGHYHORSLQJQDWXUDOYHQWLODWLRQGHVLJQVWUDWHJLHVWKDWFDQIXUQLVKDGHTXDWH
IUHVKDLUWRWKHEXLOGLQJLQWHULRUDQGVLPXOWDQHRXVO\FRQWUROWKHLQGRRUDLUTXDOLW\HIIHFWLYHO\ZKLOHSURYLGLQJVLJQLILFDQWHQHUJ\VDYLQJV,QQDWXUDOO\
YHQWLODWHGVSDFHVIXUQLVKLQJDVXLWDEOHDLUH[FKDQJHUDWHEHWZHHQWKHEXLOGLQJH[WHULRUDQGLQWHULRUFDQFUHDWHDWKHUPDOO\FRPIRUWDEOHDQGKHDOWK\
LQGRRU HQYLURQPHQW +RZHYHU WKH DLUH[FKDQJHPXVWRFFXUVXFK WKDW WKH LQGRRUDLU TXDOLW\ RI WKH EXLOGLQJ LVQRWFRPSURPLVHG DQG WKHUPDOO\
FRPIRUWDEOHFRQGLWLRQVIRUWKHRFFXSDQWVFDQEHPDLQWDLQHG
$UFKLWHFWXUH+DOOLVDUHFHQWO\UHQRYDWHGQDWXUDOO\YHQWLODWHGEXLOGLQJORFDWHGRQWKH8QLYHUVLW\RI:DVKLQJWRQFDPSXVLQ6HDWWOH7KHQDWXUDO
YHQWLODWLRQLQWKLVEXLOGLQJZDVHYDOXDWHGXVLQJDYDULHW\RIH[SHULPHQWDOWHFKQLTXHVZKLFKLQFOXGHGPHDVXUHPHQWRIFDUERQGLR[LGH&2OHYHOV
DLUH[FKDQJHUDWHVDQGDLUYHORFLWLHV+LJK&2 FRQFHQWUDWLRQVDUHDJRRGLQGLFDWRURILQDGHTXDWHYHQWLODWLRQUDWHVDQGSRRUDLUPRYHPHQWLQD
VSDFH +HQFH D QXPEHU RI VWDQGDUGV DQG FHUWLILFDWLRQ SURJUDPV VSHFLI\ WKH XVH RI RXWGRRU DLU PRQLWRULQJ EDVHG RQ &2 FRQFHQWUDWLRQV LQ DQ
RFFXSLHGVSDFH2FFXSDQWVXUYH\VEDVHGRQD86(3$VWXG\ZHUHDOVRDGPLQLVWHUHGWRXQGHUVWDQGDQGDQDO\]HRFFXSDQWSHUFHSWLRQVDERXWWKH
LQGRRUWKHUPDOHQYLURQPHQWDQGWRLGHQWLI\WKHSUHYDOHQFHRIDQ\EXLOGLQJUHODWHGLOOQHVVV\PSWRPV
7KHGLVFXVVLRQLQWKLVSDSHUZLOOIRFXVRQWKHILQGLQJVRIWKHRFFXSDQWVXUYH\VDQGKRZWKH\UHODWHWRWKHPHDVXUHG&2 OHYHOVDLUH[FKDQJHUDWHV
DQGDLUYHORFLWLHVLQWKHQDWXUDOO\YHQWLODWHGVSDFHV7KHQDWXUDOYHQWLODWLRQIXQFWLRQLQ$UFKLWHFWXUH+DOOLVODUJHO\FOLPDWHGULYHQ)RUWKHSHULRG
RI1RYHPEHUWKURXJK0DUFKSDUWLFXODUO\RXWVLGHWHPSHUDWXUHVDUHTXLWHORZDQGZLQGRZVDUHVHOGRPRSHQHGE\WKHRFFXSDQWVLQVSLWHRIDODUJH
QXPEHU RI RFFXSDQWV EHLQJ GLVVDWLVILHG ZLWK WKH LQGRRU HQYLURQPHQWDO TXDOLW\ &RQVHTXHQWO\ &2 FRQFHQWUDWLRQV FRQVLVWHQWO\ H[FHHG DFFHSWDEOH
OHYHOVDQGYHU\OLWWOHDLUPRYHPHQWLVUHFRUGHG
INTRODUCTION
$LUPRYHPHQWLQDQDWXUDOO\YHQWLODWHGVSDFHLVDIIHFWHGE\DQXPEHURIEXLOGLQJDQGHQYLURQPHQWDOSDUDPHWHUV7KH
SULQFLSDO FDXVHV RI DLU PRWLRQ DUH SUHVVXUH JUDGLHQWV UHVXOWLQJ IURP ZLQG DQG EXR\DQF\ HIIHFWV ZKLFK UHVXOW IURP
Salman Ilyas is a Mechanical Engineer at Arup, Los Angeles. Ashley Emery is a professor in the Department of Mechanical Engineering,
University of Washington, Seattle, WA. Dean Heerwagen is a professor in the Department of Architecture, University of Washington, Seattle,
WA. Judith Heerwagen is a Program Expert at US General Services Administration.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
114
ASHRAE Transactions
WHPSHUDWXUH GLIIHUHQFHV 2WKHU SDUDPHWHUV WKDW PD\ DIIHFW WKH DLU PRWLRQ LQ EXLOGLQJV LQFOXGH WKH FRPSOH[LW\ RI WKH
SDWKZD\V WKDW WKH DLU PXVW WUDYHUVH ZKLOH HQWHULQJ SDVVLQJ WKURXJK DQG OHDYLQJ WKH RFFXSDQW XVDJH EHKDYLRU LQ WKH
EXLOGLQJ VSDFHV DQG DQ\ DFWLYH V\VWHP RSHUDWLRQ $OO WKHVH IDFWRUV FRPELQH WR SURGXFH ZLGHUDQJLQJ PDJQLWXGHV RI
LQIOXHQFHV XQGHU GLIIHUHQW FRQGLWLRQV $V D UHVXOW DLU PRWLRQ EHKDYLRU LQ QDWXUDOO\ YHQWLODWHG VSDFHV YDULHV FRQVLGHUDEO\
:KLOHWKHIORZSDWWHUQVIRUIRUFHGDLUV\VWHPVDUHEHWWHUXQGHUVWRRGWKHSDWWHUQVDVVRFLDWHGZLWKQDWXUDOYHQWLODWLRQDUHIDU
PRUHFRPSOH[+HQFHLQRUGHUWRLGHQWLI\SRVVLEOHRUFRPPRQDLUPRWLRQEHKDYLRUVLWLVJHQHUDOO\QHFHVVDU\WRPHDVXUH
WKHEHKDYLRUVZKLOHDOVRPHDVXULQJWKHLQIOXHQFLQJSDUDPHWHUV
,QDQLQWHULRUVSDFHKHDWLVJHQHUDWHGE\WKHRFFXSDQWVOLJKWLQJDQGSURFHVVHTXLSPHQW,IWKHDFFXPXODWHGKHDWLV
QRW GLOXWHG E\ WKH DGGLWLRQ RI FRROHU DLU IURP WKH H[WHULRU WKH DIIHFWHG VSDFH ZLOO EHFRPH RYHUKHDWHG DQG UHVXOW LQ D
WKHUPDOO\XQFRPIRUWDEOHHQYLURQPHQWIRUWKHRFFXSDQWV0RUHRYHUEXLOGLQJRFFXSDQWVSURFHVVHTXLSPHQWPRGHOPDNLQJ
PDWHULDOVDQGDUWVXSSOLHVHWFDUHDOOVRXUFHVRIFRQWDPLQDQWV7KHLQWURGXFWLRQRIIUHVKDLUFDQGLOXWHWKHFRQFHQWUDWLRQVRI
WKHVHFRQWDPLQDQWVDQGH[SHOWKHPIURPWKHDIIHFWHGVSDFHWKHUHE\PDLQWDLQLQJDQDFFHSWDEOHOHYHORILQGRRUDLUTXDOLW\DV
ZHOO
$ QXPEHU RI VWXGLHV KDYH DVVRFLDWHG SRRU YHQWLODWLRQ UDWHV LQ D VSDFH WR WKH SUHYDOHQFH RI EXLOGLQJ UHODWHG LOOQHVV
V\PSWRPV7KHPRVWFRPSUHKHQVLYHLQYHVWLJDWLRQRIDLUTXDOLW\LQ86RIILFHEXLOGLQJVEHJDQZLWKWKH%XLOGLQJ$VVHVVPHQW
6XUYH\ DQG (YDOXDWLRQ %$6( VWXG\ ZKLFK ZDV XQGHUWDNHQ E\ WKH 86 (QYLURQPHQWDO 3URWHFWLRQ $JHQF\ (3$
)ROORZLQJ WKH FRPSOHWLRQ RI GDWD JDWKHULQJ LQ RIILFH EXLOGLQJV UHVHDUFKHUV DW WKH /DZUHQFH %HUNHOH\ 1DWLRQDO
/DERUDWRU\ /%1/XQGHUWRRN H[WHQVLYH DQDO\VLV RIWKHVHGDWD ,QRQH VHW RIDQDO\VLV LQYROYLQJ D VXEVHW RI EXLOGLQJV
FRUUHODWLRQV ZHUH IRXQG EHWZHHQ HOHYDWHG LQGRRU FDUERQ GLR[LGH FRQFHQWUDWLRQV DQG KHDOWK PHDVXUHV WKDW DUH DVVRFLDWHG
ZLWK VLFN EXLOGLQJ V\QGURPHV\PSWRPV $SWH )LVN DQG'DLVH\ $QRWKHU VWXG\ LQYHVWLJDWHG WKHHIIHFWV RI YDU\LQJ
YHQWLODWLRQUDWHVLQDQRUPDORIILFHEXLOGLQJZKLOHPDLQWDLQLQJFRQVWDQWDQGQHXWUDOWKHUPDODFRXVWLFDQGYLVXDOFRQGLWLRQV
:DUJRFNL HW DO 7KH UHVXOWV VKRZHG WKDW WKH SUHYDOHQFH RI VLFN EXLOGLQJ V\QGURPH V\PSWRPV DPRQJVW WKH
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
RFFXSDQWVGHFUHDVHGDQGWKHLUSURGXFWLYLW\LQFUHDVHGIRUKLJKHUYHQWLODWLRQUDWHV
7KHUH DUH IHZHU VXFK GDWD DYDLODEOH IRU QDWXUDOO\ YHQWLODWHG EXLOGLQJV 2QO\ D IHZ VWXGLHV KDYH DVVHVVHG WKH
SHUIRUPDQFH RI WKHVH EXLOGLQJV LQ WHUPV RI WKHLU WKHUPDO HQYLURQPHQW DQG RFFXSDQW WKHUPDO FRPIRUW 2QH VXFK VWXG\
LQYHVWLJDWHGKRZSHUVRQDOFRQWURORIRSHUDEOHZLQGRZVLQRIILFHEXLOGLQJVDIIHFWVWKHWKHUPDOHQYLURQPHQWDOFRQGLWLRQVDQG
RFFXSDQW FRPIRUW %UDJHU 3DOLDJD DQG 5LFKDUG &RQGXFWLQJ RFFXSDQW VXUYH\V DQG SK\VLFDOO\ PHDVXULQJ WKHUPDO
FRQGLWLRQV DW WKH RFFXSDQW ZRUNVWDWLRQV DFFRPSOLVKHG WKLV ,Q UHFHQW \HDUV VHYHUDO UHVHDUFK VWXGLHV KDYH XVHG FRPSXWHU
EDVHG&RPSXWDWLRQDO)OXLG'\QDPLFV&)'WHFKQLTXHVWRHVWLPDWHDQGYLVXDOL]HDLUIORZLQQDWXUDOO\YHQWLODWHGEXLOGLQJV
+RUDQDQG)LQQ(YRODDQG3RSRY9LVDJDYHODQG6ULQLYDVDQ6RPHVWXGLHVKDYHDOVRXVHGVPDOOVFDOH
SK\VLFDOPRGHOVLQZLQGWXQQHOV/DUVHQDQG+HLVHOEHUJ-LDQJHWDO+RZHYHUWKHUHLVOHVVH[WHQVLYHUHVHDUFK
WKDWLQYROYHVPHDVXUHPHQWRIDLUIORZSDWWHUQVDQGSURSHUWLHVLQIXOOVFDOHQDWXUDOO\YHQWLODWHGEXLOGLQJV

METHODS OF EVALUATION
7KHXWLOL]DWLRQRIQDWXUDOYHQWLODWLRQLQ$UFKLWHFWXUH+DOOSUHVHQWHGDQRSSRUWXQLW\WRPHDVXUHDLUIORZEHKDYLRUVLQD
IXOOVFDOHRFFXSLHGEXLOGLQJ2YHUDSHULRGRIWZR\HDUV&2OHYHOVZHUHPHDVXUHGLQRFFXSLHGGHVLJQVWXGLRVZKLOHDLU
YHORFLWLHVDQGYHQWLODWLRQUDWHVZHUHDOVRPHDVXUHGXQGHUYDU\LQJFRQGLWLRQV2FFXSDQWVXUYH\VEDVHGRQD86(3$VWXG\
ZHUHDOVRDGPLQLVWHUHGLQDOOQDWXUDOO\YHQWLODWHGVSDFHV7KHPHDVXUHGGDWDZDVWKHQFRPSDUHGWRRFFXSDQWSHUFHSWLRQVRI
WKHLUWKHUPDOHQYLURQPHQWVDQGWKHSUHYDOHQFHRIDQ\EXLOGLQJUHODWHGLOOQHVVV\PSWRPV
7KH IROORZLQJ H[SHULPHQWDO PHWKRGV ZHUH GHYHORSHG GXULQJ WKH FRXUVH RI WKH VWXG\ WR V\VWHPDWLFDOO\ HYDOXDWH WKH
SHUIRUPDQFHRIWKHQDWXUDOO\YHQWLODWHGVSDFHVLQ$UFKLWHFWXUH+DOO
&2FRQFHQWUDWLRQVZHUHPHDVXUHGXVLQJWUDQVPLWWHUVPRXQWHGRQWZRLQVWUXPHQWHGFDUWVLQRFFXSLHGVSDFHV
$LU H[FKDQJH UDWHV ZHUH PHDVXUHG XVLQJ WUDFHU JDV FRQFHQWUDWLRQ GHFD\ DQDO\VLV LQ XQRFFXSLHG VSDFHV XVLQJ D
$6+5$(

SRUWDEOHLQIUDUHGDPELHQWDQDO\]HU
$LUYHORFLWLHVZHUHPHDVXUHGXVLQJKRWVSKHUHDQHPRPHWHUVPRXQWHGRQWZRLQVWUXPHQWHGFDUWVLQERWKRFFXSLHG
DQGXQRFFXSLHGVSDFHV
7KH LQVWUXPHQWHG FDUWV DOVR PHDVXUHG DQG UHFRUGHG LQVWDQWDQHRXV GU\EXOE DLU WHPSHUDWXUH JOREH WHPSHUDWXUH DQG
UHODWLYHKXPLGLW\7KHJOREHWKHUPRPHWHUKDGDGLDPHWHURILQFKHVP
CO2 Concentrations
&2LVSURGXFHGDVDUHVXOWRIWKHKXPDQUHVSLUDWLRQSURFHVV,IWKHYHQWLODWLRQUDWHVLQDVSDFHDUHLQDGHTXDWHOHYHOV
RIRGRUFDXVLQJELRHIIOXHQWVDQG&2ZLOOLQFUHDVH6LQFH&2FDQUHDGLO\EHPHDVXUHGLWFDQVHUYHDVDJRRGLQGLFDWRURI
YHQWLODWLRQUDWHVDQGDFFHSWDELOLW\RIDVSDFHLQWHUPVRIKXPDQSURGXFHGELRHIIOXHQWV*RGLVK7KHUHIRUHWKH&2
FRQFHQWUDWLRQVLQDVSDFHSULPDULO\GHSHQGRQWKHQXPEHURIRFFXSDQWVDQGWKHDLUH[FKDQJHUDWHV
7KH QDWXUDO YHQWLODWLRQ IXQFWLRQ LQ $UFKLWHFWXUH +DOO LV FOLPDWH GULYHQ 'XULQJ PXFK RI WKH VFKRRO \HDU IURP
1RYHPEHU WKURXJK 0DUFK RXWVLGH WHPSHUDWXUHV DUH RIWHQ TXLWH ORZ DQG YHU\ IHZ RU QR ZLQGRZV DUH RSHQHG E\ WKH
RFFXSDQWV)LJXUHVKRZVDURXJKFRUUHODWLRQEHWZHHQWKHRXWVLGHGU\EXOEDLUWHPSHUDWXUHV7RXWDQGPLQLPXPDPRXQW
RIZLQGRZRSHQDUHDV$UHDPLQUHFRUGHGIRUHDFKH[SHULPHQW$VDUHVXOW&2FRQFHQWUDWLRQVLQFUHDVHZKHQWKHVWXGLRV
DUH RFFXSLHG DQG FRQVLVWHQWO\ UHDFK VWHDG\ VWDWH OHYHOV WKDW VXEVWDQWLDOO\ H[FHHG WKH PD[LPXP DFFHSWDEOH LQ WKH GHVLJQ
VWXGLRV,O\DV(PHU\DQG'+HHUZDJHQ$&2FULWHULRQIRURXWGRRUDLUPRQLWRULQJIRUWKHVWXGLRVZDVHVWDEOLVKHG
XVLQJDPHWKRGRORJ\SUHVHQWHGLQDQ$6+5$(-RXUQDODUWLFOH/DZUHQFH%DVHGRQWKLVFULWHULRQLWZDVGHWHUPLQHG
WKDWWKH&2FRQFHQWUDWLRQVH[FHHGHGWKHVWHDG\VWDWHOHYHOVRQRIWKHWHVWGD\VLQGLFDWLQJSRRUYHQWLODWLRQUDWHV
)LJXUH 5HODWLRQVKLSEHWZHHQZLQGRZRSHQDUHDDQGRXWVLGHGU\EXOEWHPSHUDWXUHV
Air Exchange Rates

$SULQFLSDOWDVNRIDQ\YHQWLODWLRQV\VWHPLVWRFRQWUROWKHLQGRRUWKHUPDOHQYLURQPHQW2QHZD\WKLVLVDFKLHYHGLVE\
DGPLWWLQJ IUHVK RXWVLGHDLU WR FRROD VSDFHDQGWRGLOXWH LQGRRUFRQWDPLQDQW FRQFHQWUDWLRQV $ QDWXUDOO\ YHQWLODWHGVSDFH
SURYLGHV IUHVK DLU WKURXJK SDVVLYH PHDQV IRU H[DPSOH E\ RSHQLQJ ZLQGRZV 7KH DPRXQW RI IUHVK DLU UHTXLUHG FDQ EH
TXDQWLILHGLQWHUPVRIKRXUO\DLUH[FKDQJHUDWHV$&+
$&+·V ZHUH PHDVXUHG LQ XQRFFXSLHG GHVLJQ VWXGLRV XVLQJ WUDFHU JDV DQDO\VLV XQGHU YDU\LQJ FRQGLWLRQV 7KH UHVXOWV
LQGLFDWHGWKDWDLUH[FKDQJHEHWZHHQWKHLQWHULRUDQGH[WHULRUGRHVRFFXUSURYLGHGDGHTXDWHDPRXQWRIZLQGRZVZHUHRSHQ
:KHQ DOO ZLQGRZV ZHUHFORVHG WKH PHDVXUHG YHQWLODWLRQUDWHVZHUH LQ WKH UDQJH RI WR $&+ 7DEOH VKRZVD
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
116
ASHRAE Transactions
UDQJHRIPHDVXUHGYHQWLODWLRQUDWHVSHUSHUVRQLQWKHGHVLJQVWXGLRVIRUWKRVHH[SHULPHQWVZKHQDOOZLQGRZVZHUHFORVHG
DQG KRZ WKH\ FRPSDUH WR WKH PLQLPXP UHTXLUHG YHQWLODWLRQ UDWHV FDOFXODWHG XVLQJ $6+5$( 6WDQGDUG 7KH
UHVXOWVVXJJHVWWKDWWKHYHQWLODWLRQUDWHVLQWKHGHVLJQVWXGLRVZHUHFRQVLGHUDEO\EHORZLQGXVWU\VWDQGDUGVZKHQZLQGRZV
ZHUHFORVHGZKLFKZHUHWKHFDVHRQRIWKHWHVWGD\V
Table 1. Comparison of Measured and Calculated Ventilation Rates
* * *
6WXGLR &)0SHUVRQ
/VSHUVRQ
&DOFXODWHGSHU$6+5$(

0HDVXUHG0LQLPXP

0HDVXUHG0D[LPXP

6LQFHWKHWUDFHUJDVDQDO\VLVFRXOGRQO\EHFRQGXFWHGLQXQRFFXSLHGVSDFHVGXHWRWKHXVHRIVXOIXUKH[DIOXRULGH6)DVWKH

WUDFHU JDV DQG WKH URRPV ZHUH LQ XVH PRVW RI WKH DFDGHPLF \HDU RQO\ D VPDOO QXPEHU RI VXFK H[SHULPHQWV FRXOG EH
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
FRPSOHWHG+HQFHLWLVGLIILFXOWWRDVFHUWDLQZKDWWKHRSWLPXPDPRXQWRIZLQGRZRSHQDUHDLVLQRUGHUWRDFKLHYHDGHTXDWH
YHQWLODWLRQUDWHVXQGHUYDU\LQJLQIOXHQFLQJSDUDPHWHUV
Air Velocity Measurement

$LU PRYHPHQW LQ D EXLOGLQJ DIIHFWV RFFXSDQW WKHUPDO FRPIRUW DQG UDWHV RI KHDW JDLQ RU ORVV WKURXJK WKH EXLOGLQJ
HQYHORSHE\LQILOWUDWLRQEXWQRWE\WUDQVPLVVLRQ$OVRLWLVDPDMRUGHWHUPLQDQWRIZKHWKHUWKHLQGRRUDLUTXDOLW\ZLOOEH
VDWLVIDFWRU\'+HHUZDJHQ$QRFFXSDQWZRXOGIHHOWKHUPDOFRPIRUWRUGLVFRPIRUWGHSHQGLQJRQKRZUDSLGO\RU
VORZO\ WKH DLU PRYHV $ XQLIRUP GLVWULEXWLRQ RI DLU DW DFFHSWDEOH UDWHV ZRXOG HQVXUH D KHDOWKIXO LQGRRU HQYLURQPHQW E\
GLOXWLQJLQGRRUFRQWDPLQDQWVVXFKDVGXVWDOOHUJHQVDQGRWKHUWR[LFVXEVWDQFHV
$LU YHORFLW\ PHDVXUHPHQWV ZHUH GRQH LQ RFFXSLHG GHVLJQ VWXGLRV XVLQJ KRWVSKHUH DQHPRPHWHUV PRXQWHG RQ
LQVWUXPHQWHG FDUWV 7KH YHORFLW\ ZDV PHDVXUHG LQ D FHQWUDO ORFDWLRQ DW DSSUR[LPDWHO\ IW P DERYH IORRU OHYHO
7\SLFDOO\YHU\ORZDLUYHORFLWLHVZHUHUHFRUGHGLQWKHGHVLJQVWXGLRV7DEOHZKLFKFDQEHDWWULEXWHGWRWKHLQVXIILFLHQW
ZLQGRZRSHQDUHDV6WXGLRVZKHUHRFFXSDQWVKDGDWHQGHQF\WRRSHQUHODWLYHO\PRUHZLQGRZVUHFRUGHGKLJKHUYHORFLWLHV
EXWVWLOORQWKHORZHUHQG
Table 2. Statistical Summary of Air Velocity Measurements in Occupied Spaces
6WDWLVWLFDO0HDVXUH $LU9HORFLW\IWVHF $LU9HORFLW\PV
$YHUDJH
6WDQGDUG'HYLDWLRQ
0LQLPXP
0D[LPXP
0RGH
,QRWKHUH[SHULPHQWDWLRQYHORFLW\SUREHVZHUHSODFHGRQWKHZLQGRZVLOODQGDOVRGLUHFWO\LQWKHSDWKRILQFRPLQJDLULQ
RUGHUWRREVHUYHWKHEHKDYLRURILQFRPLQJDLU)LJXUHVKRZVDSORWRIRQHVXFKH[SHULPHQWZKHUHYHORFLW\ZDVPHDVXUHG
DWWKH RSHQ ZLQGRZ/RFDWLRQ $ DQG IW P GRZQVWUHDP/RFDWLRQ % 7KH JUDSK VKRZV PLQXWHV DYHUDJHG GDWD
RYHU D KRXU SHULRG $FFRUGLQJ WR ZLQG GDWD REWDLQHG IURP D QHDUE\ ZHDWKHU VWDWLRQ WKH ZLQGRZ ZDV UHFHLYLQJ GLUHFW
ZLQGZLWKDQDYHUDJHZLQGVSHHGRIIWVPV:HQRWHWKDWWKHPDJQLWXGHRIWKHDLUYHORFLW\GHFUHDVHGE\WR
WLPHVDIWHUWUDYHOLQJDGLVWDQFHRIRQO\IWP6LPLODUUHVXOWVZHUHREWDLQHGLQRWKHUVXFKH[SHULPHQWV
$6+5$(


)LJXUH 9HORFLW\RIDLUDWZLQGRZLQOHWDQGIWPGRZQVWUHDP
ENVIRONMENTAL QUALITY ASSESSMENT

,QDGGLWLRQWRSK\VLFDOPHDVXUHVRIDLUTXDOLW\LWLVLPSRUWDQWWRDVVHVVKRZRFFXSDQWVSHUFHLYHWKHDLUDQGWKHUPDO
FRQGLWLRQVDQGZKHWKHUWKH\H[SHULHQFHGLVFRPIRUWRULOOQHVVV\PSWRPV5HVXOWVIURPDGHFDGHRIUHVHDUFKLQ(XURSHDQG
WKH 86 VKRZ KLJK OHYHOV RI V\PSWRPV DVVRFLDWHG ZLWK DLU TXDOLW\ LQ W\SLFDO RIILFH EXLOGLQJV LH QRW LGHQWLILHG DV ´VLFN
EXLOGLQJVµ6\PSWRPVLQFOXGHKHDGDFKHH\HQRVHDQGWKURDWLUULWDWLRQXSSHUDQGORZHUUHVSLUDWRU\LUULWDWLRQDQGIDWLJXH
7KH HQYLURQPHQWDO FDXVHV KDYH EHHQ GLIILFXOW WR LGHQWLI\ LQ SDUW EHFDXVH RI WKH KLJK OHYHO RI SRWHQWLDO VRXUFHV WKDW DUH
SUHVHQWLQYHU\ORZLQGRRUFRQFHQWUDWLRQV6WROZLMN
'DWDIURPWKH%XLOGLQJ$VVHVVPHQWDQG6XUYH\(YDOXDWLRQ%$6(VWXG\86(3$DVZHOODVILQGLQJVIURP
(XURSHDQUHVHDUFKVKRZWKDWRUPRUHRIZRUNHUVH[SHULHQFHLOOQHVVV\PSWRPVDWZRUN$QDO\VLVRIEXLOGLQJVLQWKH
%$6(VDPSOHZLWKVXEMHFWVIRXQGWKDWH[SHULHQFHGLUULWDWLRQRIWKHH\HQRVHDQGWKURDWDQGWKDWUHSRUWHG
KHDGDFKHDQGIDWLJXH$SWH)LVNDQG'DLVH\*LYHQWKHLQGLFDWLRQVLQWKLVVWXG\WKDWDLUTXDOLW\LQ$UFKLWHFWXUH+DOO
ZDV FRPSURPLVHG IRU VLJQLILFDQW DPRXQWV RI WLPH ZH DGPLQLVWHUHG D VXUYH\ DVVHVVLQJ WKH QDWXUH DQG IUHTXHQF\ RI
V\PSWRPV
7KH SUHYDOHQFH RI LOOQHVV V\PSWRPV LQ WKH RFFXSDQWV RI $UFKLWHFWXUH +DOO ZDV FRPSDUHG WR WKH %$6( VWXG\ IRU
ZKLFK WKH 86 (3$ FRQGXFWHG D V\VWHPDWLF VXUYH\ RI 86 SXEOLF DQG SULYDWH RIILFH EXLOGLQJV IURP WR
%ULJKWPDQ HW DO 7KH EXLOGLQJV ZHUH UDQGRPO\ VHOHFWHG IURP FLWLHV ZLWK SRSXODWLRQV JUHDWHU WKDQ
SHRSOH $ SDSHUEDVHG VXUYH\ ZDV XVHG IRU WKLV VWXG\ DQG ZDV PRGLILHG IURP WKH %$6( VWXG\ TXHVWLRQQDLUH WR VXLW WKH
VSHFLILFUHTXLUHPHQWVRI$UFKLWHFWXUH+DOO5HVXOWVUHOHYDQWWRRFFXSDQWDLUPRYHPHQWSUHIHUHQFHVDQGLOOQHVVV\PSWRPVDUH
SUHVHQWHGDQGGLVFXVVHGLQWKLVSDSHU7KHVXUYH\ZDVDGPLQLVWHUHGGXULQJDWZRZHHNVWUHWFKLQODWH1RYHPEHUDQGHDUO\
'HFHPEHUZKHQPRVWRIWKHVWXGHQWVVSHQGDVLJQLILFDQWDPRXQWRIWLPHLQWKHGHVLJQVWXGLRV
Description of Questionnaire
7KH TXHVWLRQQDLUH XVHG IRU WKH VXUYH\ ZDV GLYLGHG LQWR WZR PDLQ SDUWV SHUFHSWLRQV DQG FRPIRUW DQG
HQYLURQPHQWDO H[SHULHQFHV 7KH ILUVW SDUW DVNHG TXHVWLRQV DERXW WKH RFFXSDQWV· LQWHUDFWLRQ ZLWK WKH VWXGLR WKHUPDO
HQYLURQPHQW 7KH VHFRQG SDUW RI WKH TXHVWLRQQDLUH DVNHG WKH RFFXSDQWV WR LGHQWLI\ DQ\ KHDOWK V\PSWRPV WKH\ KDG
H[SHULHQFHGLQWKHSDVWZHHNDQGWKRVHWKDWWKH\ZHUHSUHVHQWO\H[SHULHQFLQJ7KHHQWLUHTXHVWLRQQDLUHWRRNDSSUR[LPDWHO\
PLQXWHVWRILOORXW$WRWDORIVXUYH\VZHUHFRPSOHWHGDQGEDVHGRQWKHDYHUDJHQXPEHURIRFFXSDQWVSUHVHQWLQWKH
GHVLJQVWXGLRVWKHUHVSRQVHUDWHZDV
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
118
ASHRAE Transactions
Thermal Perceptions
2FFXSDQWVZHUHDVNHGWRUDWHWKHLULQGLYLGXDOWKHUPDOSHUFHSWLRQVRIWHPSHUDWXUHGD\OLJKWQRLVHOHYHODLUPRYHPHQW
HOHFWULF OLJKW DQG KXPLGLW\ DW WKHLU UHVSHFWLYH ZRUNVSDFHV RQ D VFDOH RI WR 2QO\ WHPSHUDWXUH DQG DLU PRYHPHQW DUH
UHOHYDQWWRWKHGLVFXVVLRQKHUH)LJXUHVKRZVWKHGLVWULEXWLRQRIWHPSHUDWXUHDQGDLUPRYHPHQWVDWLVIDFWLRQYRWHVIRUD
WRWDO RI UHVSRQGHQWV RI DOO UHVSRQGHQWV ZHUH VDWLVILHG ZLWK WKH WHPSHUDWXUH LQ WKH VSDFH $ PDMRULW\ RI WKH
RFFXSDQWVZHUHQRWHQWLUHO\VDWLVILHGZLWKWKHDLUPRYHPHQWLQWKHVSDFH0RVWRIWKHPIHOWWKDWWKHDLUPRYHPHQW
ZDVWRROLWWOHZKHUHDVRFFXSDQWVIHOWWKDWWKHDLUPRYHPHQWZDVRQWKHKLJKHUVLGH
)LJXUH 6DWLVIDFWLRQZLWKWHPSHUDWXUHDQGDLUPRYHPHQWLQWKHZRUNVSDFH
Occupant Control
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
6RPH TXHVWLRQV LQ WKH VXUYH\ LQTXLUHG DERXW WKH RFFXSDQWV· FRQWURO RYHU WKHLU HQYLURQPHQW ,Q RQH TXHVWLRQ WKH
RFFXSDQWVZHUHDVNHGDERXWWKHLUDELOLW\WRRSHQZLQGRZV5HVXOWVVKRZHGWKDWRIWKHRFFXSDQWVZHUHVDWLVILHGZLWK
WKHFRQWUROWKH\KDGRYHUZLQGRZRSHUDWLRQ0DMRULW\RIWKHRFFXSDQWVZKRRSHQHGDZLQGRZLQWKHSDVWZHHNGLGVRWROHW
IUHVKDLULQRUWRFRROGRZQWKHURRP)RUWKRVHWKDWGLGQ·WRSHQDZLQGRZFLWHGWKHFRROLQFRPLQJDLUDVRQH
RIWKHSULPDU\UHDVRQVIRUQRWGRLQJVRRULQDELOLW\WRRSHUDWHWKHZLQGRZV$ODUJHQXPEHURIRFFXSDQWVIHOW
QRQHHGWRRSHQDQ\ZLQGRZV
Illness Symptoms
,Q WKH VHFRQG SDUW RI WKH TXHVWLRQQDLUH WKH UHVSRQGHQWV ZHUH DVNHG LI WKH\ H[SHULHQFHG D VSHFLILF V\PSWRP HLWKHU
ZKLOHZRUNLQJLQWKHVWXGLRDWWKHWLPHRIWKHVXUYH\RUDWDQ\WLPHGXULQJWKHSUHFHGLQJZHHN7KLVDSSURDFKZDVRQHRI
WKH PRGLILFDWLRQV PDGH YHUVXV WKH ZD\ WKH TXHVWLRQ ZDV DVNHG LQ WKH 86 (3$ %$6( VWXG\ 7KH EXLOGLQJUHODWHG
V\PSWRPV LQFOXGH ORZHU UHVSLUDWRU\ HJ ZKHH]LQJ VKRUWQHVV RI EUHDWK RU FRXJKLQJ XSSHU UHVSLUDWRU\ DQG PXFRVDO
PHPEUDQHHJGU\H\HVVWXII\RUUXQQ\QRVHVRUHWKURDWRUVQHH]LQJXQXVXDOIDWLJXHGLIILFXOW\FRQFHQWUDWLQJGL]]LQHVV
RUOLJKWKHDGHGQHVVDQGVNLQV\PSWRPVHJGU\QHVVLUULWDWLRQRULWFKLQJ5HVSRQGHQWVZHUHDVNHGWRFKHFNDOOWKDWDSSO\
RFFXSDQWVRXWRIWKHUHVSRQGHQWVUHSRUWHGDWOHDVWRQHRUPRUHLOOQHVVV\PSWRPVGXULQJWKHSDVWZHHNRU
ZKLOHWDNLQJWKHVXUYH\2QDYHUDJHDUHVSRQGHQWFODLPHGLOOQHVVV\PSWRPV)LJXUHVKRZVDGLVWULEXWLRQRIWKHQXPEHU
RI V\PSWRPV UHSRUWHG E\ D SHUVRQ DW WKH WLPH RI WKH VXUYH\ DQG LQ WKH SUHFHGLQJ ZHHN :H QRWH WKDW PRVW RI WKH
UHVSRQGHQWVKDGPRUHWKDQRQHV\PSWRPZKHUHDVSHRSOHDOVRUHSRUWHGDWOHDVWLOOQHVVV\PSWRPV5HVXOWVLQGLFDWH
WKDWWKHRFFXSDQWVZHUHH[SHULHQFLQJSURQRXQFHGUDWHVRILOOQHVVV\PSWRPV
$6+5$(


)LJXUH 2FFXSDQWV\PSWRPRFFXUUHQFHVLQ$UFKLWHFWXUH+DOO
7DEOHVKRZVWKHSUHYDOHQFHRILOOQHVVV\PSWRPVLQ$UFKLWHFWXUH+DOOUHSRUWHGIRUWKHSUHYLRXVZHHNDQGDWWKHWLPH
RIWKHVXUYH\DQG86(3$%$6(VWXG\EXLOGLQJV:HQRWHWKDWWKHSUHYDOHQFHRIKHDOWKV\PSWRPVLQ$UFKLWHFWXUH+DOO
IRUSDVWZHHNLVFRQVLGHUDEO\KLJKHUWKDQWKH%$6(EXLOGLQJVIRUDOPRVWDOOV\PSWRPVH[FHSWVKRUWQHVVRIEUHDWK7KH
SUHYDOHQFH IRU WKH ´5LJKW QRZµ FDWHJRU\ HLWKHU H[FHHGV RU LV VLPLODU WR WKH UHVXOWV IURP WKH %$6( EXLOGLQJV IRU PRVW
V\PSWRPV
Table 3 Comparison of Prevalence (%) of Illness Symptoms in Architecture Hall and BASE Buildings
$UFKLWHFWXUH+DOO $UFKLWHFWXUH+DOO
,OOQHVV6\PSWRP %$6(%XLOGLQJV
3DVW:HHN 5LJKW1RZ
7LUHGRUVWUDLQHGH\HV
'U\LWFK\RULUULWDWHGH\HV
+HDGDFKH
8QXVXDOIDWLJXHRUWLUHGQHVV
6WXII\RUUXQQ\QRVH
6QHH]LQJ
6RUHRUGU\WKURDW
&RXJK
'LIILFXOW\UHPHPEHULQJWKLQJV
'U\RULWFK\VNLQ
'LIILFXOW\FRQFHQWUDWLQJ
'L]]LQHVVRUOLJKWKHDGHGQHVV
1DXVHDRUXSVHWVWRPDFK
:KHH]LQJ
6KRUWQHVVRIEUHDWK

$QRWKHUVWXG\$SWHDQG(UGPDQQXVHGWKH86(3$%$6(VWXG\GDWDWRTXDQWLI\WKHUHODWLRQVKLSEHWZHHQLQGRRU

&2 FRQFHQWUDWLRQV DQG XSSHU UHVSLUDWRU\ DQG PXFRVDO PHPEUDQH 00 DQG ORZHU UHVSLUDWRU\ /5 EXLOGLQJ UHODWHG
V\PSWRPV7KHVWXG\IRXQGVWDWLVWLFDOO\VLJQLILFDQWFRUUHODWLRQVRI00DQG/5EXLOGLQJUHODWHGV\PSWRPVZLWKLQFUHDVLQJ
LQGRRU &2 FRQFHQWUDWLRQV 7KH VPDOO VDPSOH VL]H RI OLPLWV WKH DSSOLFDWLRQ RI VLPLODU VWDWLVWLFDO PHWKRGV WR ILQG
FRUUHODWLRQVLQ$UFKLWHFWXUH+DOO+RZHYHUDKLJKIUHTXHQF\RIRFFXUHQFHRI00SDVWZHHNDQGULJKWQRZDQG
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
120
ASHRAE Transactions
/5V\PSWRPVSDVWZHHNDQGULJKWQRZDQGWKHSUHVHQFHRIHOHYDWHG&2OHYHOVLQWKHEXLOGLQJLVDFDXVHIRU
FRQFHUQ
CONCLUSION
7KH UHVXOWV IURP WKLV VWXG\ VXJJHVW WKDW WKH DLU TXDOLW\ LQ WKH QDWXUDOO\ YHQWLODWHG GHVLJQ VWXGLRV LV TXHVWLRQDEOH
$OWKRXJKWKHGDWDGRQRWVXSSRUWDFDVXDOUHODWLRQVKLSEHWZHHQKLJK&2OHYHOVORZDLUPRYHPHQWDQGLOOQHVVV\PSWRPV
WKH UHVXOWV DUH FRQVLVWHQW ZLWK UHVXOWV IURP RWKHU VWXGLHV ZKLFK KDYH XVHG VLPLODU VXUYH\ PHWKRGV DQG HQYLURQPHQWDO
PHDVXUHV HJ 86 (3$ %$6( VWXG\ ,W ZRXOG EH XVHIXO WR KDYH UHVXOWV IURP H[SHULPHQWDO VWXGLHV WKDW FOHDUO\ VKRZ
V\PSWRPVUHODWHGWRVSHFLILFHQYLURQPHQWDOSDUDPHWHUV1RQHWKHOHVVWKLVVWXG\VKRZVWKDWWKHUHLVUHDVRQWREHOLHYHWKDW
DQLQFUHDVLQJXVHRIQDWXUDOYHQWLODWLRQWRUHGXFHEXLOGLQJHQHUJ\FRQVXPSWLRQPD\EHTXHVWLRQDEOHLQVRPHFLUFXPVWDQFHV
:KHQWKHQDWXUDOYHQWLODWLRQVWUDWHJ\UHOLHVRQRFFXSDQWVWRRSHQZLQGRZVUDWKHUWKDQDXWRPDWHGFRQWUROVZHFDQH[SHFW
WR ILQG D FRQIOLFW EHWZHHQ FRPIRUW PDLQWHQDQFH DQG YHQWLODWLRQ ZKHQ RXWGRRU WHPSHUDWXUHV DUH ORZ 8QGHU WKHVH
FRQGLWLRQVRSHQLQJZLQGRZVFUHDWHVGLVFRPIRUWWKDWLVUHVROYHGE\FORVLQJWKHZLQGRZVDJDLQ7KLVOHDGVWREXLOGXSRI
&2OHYHOVDQGUHGXFHGDLUPRYHPHQWWKDWDUHDVVRFLDWHGZLWKLOOQHVVV\PSWRPV

REFERENCES
$16,$6+5$(6WDQGDUG9HQWLODWLRQIRU$FFHSWDEOH,QGRRU$LU4XDOLW\
$SWH0* (UGPDQQ&$0XFRXVPHPEUDQHDQGORZHUUHVSLUDWRU\EXLOGLQJUHODWHGV\PSWRPVLQUHODWLRQWR
LQGRRUFDUERQGLR[LGHFRQFHQWUDWLRQVLQWKHEXLOGLQJ%$6(GDWDVHW,QGRRU$LU
$SWH 0 *)LVN : - 'DLVH\ - 0 $VVRFLDWLRQVEHWZHHQ ,QGRRU&DUERQ 'LR[LGH &RQFHQWUDWLRQV DQG 6LFN
%XLOGLQJ6\QGURPH6\PSWRPVLQ862IILFH%XLOGLQJV$Q$QDO\VLVRIWKH%$6(6WXG\'DWD,QGRRU$LU

%UDJHU * 6 3DOLDJD * 5LFKDUG G ' 2SHUDEOH :LQGRZV 3HUVRQDO &RQWURO DQG 2FFXSDQW &RPIRUW
$6+5$(7UDQVDFWLRQV3DUW
%ULJKWPDQ+0LOWRQ':\SLM'%XUJH+ 6SHQJOHU-(YDOXDWLQJEXLOGLQJUHODWHGV\PSWRPVXVLQJWKH86
(3$%$6(6WXG\UHWXOV,QGRRU$LU
(YROD* 3RSRY9&RPSXWDWLRQDODQDO\VLVRIZLQGGULYHQQDWXUDOYHQWLODWLRQLQEXLOGLQJV(QHUJ\DQG%XLOGLQJV

*RGLVK76LFN%XLOGLQJVGHILQLWLRQGLDJQRVLVDQGPLWLJDWLRQ%RFD5DWRQ/HZLV3XEOLVKHUV
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
+HHUZDJHQ'9LWDO6LJQV2EVHUYLQJDLUIORZLQEXLOGLQJV
+RUDQ-0 )LQQ'36HQVLWLYLW\RIDLUFKDQJHUDWHVLQQDWXUDOO\YHQWLODWHGDWULXPVSDFHVXEMHFWWRYDULDWLRQVLQ
H[WHUQDOZLQGVSHHGDQGGLUHFWLRQ(QHUJ\DQG%XLOGLQJV
,O\DV 6 (PHU\ $ +HHUZDJHQ ' $VVHVVPHQWV RI WKH 1DWXUDO 9HQWLODWLRQ )XQFWLRQ LQ D 8QLYHUVLW\ %XLOGLQJ
XVLQJ&20HDVXUHPHQW$6+5$(7UDQVDFWLRQV3DUW
/DUVHQ76 +HLVHOEHUJ36LQJOHVLGHGQDWXUDOYHQWLODWLRQGULYHQE\ZLQGSUHVVXUHDQGWHPSHUDWXUHGLIIHUHQFH
(QHUJ\DQG%XLOGLQJV
/DZUHQFH706HOHFWLQJ&2&ULWHULDIRU2XWGRRU$LU0RQLWRULQJ$6+5$(-RXUQDO1R
6WRZLMN-$-6LFN%XLOGLQJ6\QGURPH&RPPHQWDU\(QYLURQPHQWDO+HDOWK3HUVSHFWLYHV
86(QYLURQPHQWDO3URWHFWLRQ$JHQF\(3$$6WDQGDUGL]HG(3$3URWRFROIRU&KDUDFWHUL]LQJ,QGRRU$LU4XDOLW\
LQ/DUJH2IILFH%XLOGLQJV
9LVDJDYHO. 6ULQLYDVDQ3$QDO\VLVRIVLQJOHVLGHGYHQWLODWHGDQGFURVVYHQWLODWHGURRPVE\YDU\LQJWKHZLGWKRI
WKHZLQGRZRSHQLQJXVLQJ&)'6RODU(QHUJ\
:DUJRFNL3:\RQ'36XQGHOO-&ODXVHQ* )DQJHU327KH(IIHFWVRI2XWGRRU$LU6XSSO\5DWHLQDQ
2IILFH %XLOGLQJ RQ3HUFHLYHG $LU 4XDOLW\ 6LFN EXLOGLQJ 6\QGURPH6%6 6\PSWRPV DQG 3URGXFWLYLW\,QGRRU$LU

$6+5$(

SA-12-C016
What We Have Learned From 20 Years of

HVAC System Commissioning
E. Thomas Lillie, P.E.

ASHRAE Life Member
ABSTRACT
As most are aware by now ASHRAE Guideline 0 – 2005 states that “the Commissioning Process is a quality-oriented process for
achieving, verifying, and documenting that the performance of the facilities, systems, and assemblies meets defined objectives and criteria”. After 20
years of providing this service, it is appropriate to look back and identify some of the most beneficial lessons we have learned that can help building
designers, owners and operators.
Over the years we have found numerous issues such as missing return air ducts, missing supply and return air duct insulation, misapplied
diffusers, misapplied terminal units, and system effect issues. While these are not all that unusual, there have been issues found that have a
significant impact on performance and energy consumption.
This paper will identify issues found during the commissioning process that were energy significant, comfort compromising, and most
importantly re-occurring. It will also provide resolutions of the issues and the resulting benefits. Many of these issues were found on projects where
the commissioning provider was not engaged until construction was almost complete or on retro-commissioning projects. Issues found in these cases
are typically profound. Examples include: improper control of return air fans and building pressure; improper use of VAV plenum return air
systems; improper designed or installed ducted return air systems; improper cooling coil piping connections; questionable strategies for controlling
discharge air pressure in VAV systems; inadequate control of chilled water flow in primary / secondary pumping systems of large campus
systems; and improper training of O&M staff.
INTRODUCTION
Today facility owner groups such as Universities, Healthcare, Laboratories, Mission Critical, Developers and Large
Retail are committed to the commissioning of their buildings. Design engineers, as well as contractors, have developed a
positive regard for the importance of commissioning, especially in large and complicated projects. Recently, a design
engineer related some of his experiences that solidified his belief that commissioning was needed.
1. “Reviewing a balance report with 150+ circuit setters – only to get on-site and find that none of them were
ever installed.”
2. “Arriving on-site after an equipment start-up to find that the cooling coil had been “sucked” 3 inches closer
to the fan inlet because the contractor didn’t open any return air dampers when they started the unit (a lab
unit with almost 10” of total static).”
E. Thomas Lillie, PE is a former Secretary/ Treasurer of the Building Commissioning Association and Partner at Horizon Engineering Associates, LLP, St.
Louis, MO.
122
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---

ASHRAE Transactions
3. “I was once asked to price a design on a make-up air handler for an existing lab… when I arrived on-site I
found a modular home filled with lab hoods – easily breaking about 100 OSHA regulations. Whenever
anyone in upstate NY received mail that they thought may contain Anthrax, authorities would send the
sample to this “lab” to process the threat – I didn’t stay long.”
4. “While performing an energy study on a school in the Bronx the maintenance staff asked that I look at a new
building that was nearing construction completion – air handlers as large as 25,000 CFM were installed with
filter and coil pull areas that had no more than 42” of headroom.”
5. “While performing the same energy study in the Bronx I found a multitude leaks in the hot water systems
throughout campus. The maintenance staff had also found the leaks – and fashioned funnels under them,
which they attached to hoses that went to the nearest drain. The estimated energy loss on these leaks was
hundreds of thousands of dollars over the life of the systems.”
The late 1960s began an era of the design/build construction delivery of commercial buildings. In many cases, the
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
consulting engineer was left out. The result of which were many buildings that did not perform per the Owner’s
expectations. Many of these buildings had comfort issues as well as high energy consumption. This, coupled with the energy
crisis of the early 1970s, shed light on the fact that buildings could be made more energy efficient without sacrificing
comfort and, in many cases, improving it. Building commissioning (Cx) has emerged as the ideal method of verifying and
documenting that building systems are installed and operating to provide the performance envisioned by the designer and
Owner. Building Cx is relatively new, having its beginning in 1977 when Public Works Canada began to use Cx in its
project delivery system. In 1984, the ASHRAE HVAC Commissioning Guideline Committee was formed. In 1988,
ASHRAE published the HVAC Commissioning Guideline. Subsequently, the University of Michigan established a facilities
evaluation and Cx group in 1988. In the first half of the 1990s, government agencies such as the General Services
Administration (GSA), U.S. Corp of Engineers, Environmental Protection Agency (EPA) and Department of Energy
(DOE) adopted Cx guidelines; and the first annual National Conference on Building Commissioning (NCBC) was held. In
1996, ASHRAE published Guideline 1-1996: The HVAC Guideline Commissioning Process.
The Building Commissioning Association (BCA) was incorporated in 1998. The same year, the United States Green
Building Council (USGBC) LEED criteria included Cx. In 2004, the USGBC introduced LEED-EB that includes a
prerequisite for retro-commissioning. The following year, ASHRAE published Guideline 0-2005: The Commissioning
Process. In 2006, the ASHRAE Sustainability Roadmap sets an organization-wide agenda to lead the advancement of
sustainable building design and operations and announces plans for a certification program for Cx and integrated building
design. In 2008, USGBC restructures LEED®-EB as LEED® for Existing Buildings: Operations & Maintenance, allowing
up to 4-points for EBCx and 2-points for ongoing commissioning.
The impetus leading to the evolution of Cx has been Federal Agencies, USGBC and ASHRAE. What started out as
commissioning new building HVAC construction has expanded into existing building and ongoing or continuous
commissioning. Building envelope commissioning is now common. Many Architectural/Engineering firms who specialize
in critical mission facilities and laboratories require commissioning to be a part of the building process.
Commissioning today serves the construction industry to:

1. Ensure that the building will perform per the Owner’s expectations.
2. Reduce performance liability for Owners, Designers and Contractors
3. Reduce the number of Change Orders.
4. Be a positive influence to the project schedule.
$6+5$(

5. Shorten designers and contractors time at the end of the project.
THE COMMISSIONING PROCESS

My experience over the course of the past 20 years includes over 1,100 projects ranging from $250,000,000 and
$100,000. I have first-hand experience to attest that the Cx process has been instrumental in identifying significant issues of
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
performance and high energy consumption. These projects were new construction Cx, existing building Cx, and continuous
Cx.
It is important to recall the Owner’s issues that were the genesis of Cx. The Cx industry likes to use the phrase
“buildings that don’t work” when discussing needs. While this is true, it is an oversimplification. As it was, there were
buildings that failed to meet the operational expectations of the Owner. Operational failures are more often than not
multifaceted issues resulting from disconnect between Owner’s expectations, unrealistic Owner’s expectations, design
issues, installation issues, operating issues and the finished product. Thus, the documented Design Intent and Criteria, or as
it is now called, Owner’s Project Requirements (OPR) is obviously essential. Owners who have an established building
program with a planning, design and construction group are generally not susceptible to unrealistic expectations. The OPR
ideally developed by the Owner or Owner/Architect/Engineer should be detailed enough to identify the design solutions;
all design parameters; all local codes; all applicable ASHRAE guidelines and standards; and all process codes, guidelines and
standards. The OPR should have information that provides expectations for all phases of the project. This document thus
becomes the governing document for use throughout the project, including operations. As such, it is the measuring stick
for project success. Thus, a well developed and documented OPR defines the project with all subsequent commissioning
activities supporting it.
An ‘Issues List’ is developed by the Commissioning Authority (CxA) with input from the Cx team. All issues are
documented; meaning they are dated when entered and dated when resolved with details regarding the action taken and
responsible party identified. It is important to understand that an item is only an issue if it is not in accordance with the
OPR. With that said, it is rare when all issues are clearly referenced in the OPR. Thus, it becomes the CxA’s responsibility
to resolve any ambiguity with the Cx team.
In new construction, Cx issues generally fall into these categories:
1. Resulting from lack of or poorly defined OPR.
2. Design issues resulting from documents lacking clear definition of scope; poor system flow and
instrumentation diagrams, poorly selected equipment, clear and concise details, clear and thorough
specifications; and a concise description of system operating procedures and goals.
3. Construction issues resulting from conflicting details to standard practice, guidelines, and standards;
misinterpretation of sequence of operations; misunderstanding of system functionality; mislabeled vendor
supplied equipment; shoddy workmanship; and a lack of qualified technical management.
4. Operational issues resulting from system performance issues, lack of ‘record documentation’, ineffective
training and system understanding.
Below is a list of common issues found on projects:

1. Return air fans over pressurizing mixed air chambers. This condition can occur intermittently and is typical of
volumetric control return fans.
2. Cross connection of chilled water piping. This occurs at wall penetrations and coil connections. In some cases
labeling on the coil is incorrect.
3. Incorrect fill pressure. Incorrect diaphragm pressure in expansion tanks.
124
ASHRAE Transactions
4. Instrument calibration not addressed.
5. Air monitoring stations installed in small mechanical rooms resulting in inaccurate readings.
6. AM and PM are switched on HVAC control schedules (or on the computer that the controls reside on) –
meaning that the equipment is scheduled to be occupied just as people are leaving and scheduled to
unoccupied settings just before they arrive.
7. Controls contractors not tuning control loops. Recently, a controls contractor didn’t tune a humidification
loop (therefore ramping up humidity too fast) and repeatedly flooded the discharge air duct detector – causing
repeated nuisance trips and need for the detector to be replaced.
8. Safeties, such as freeze protection, wired to the building automation system (BAS) instead of hard wired.
9. TAB contractors not testing multiple pumps in parallel but rather individually. Parallel pump and system
curves are not developed.
10. Wiring issues such as reverse phases. Incorrect size of overload protection.
11. Missing or inadequate record documents.
12. Poorly developed and inadequate O&M manuals from contractors. Lacking specific project related equipment
information, and little system information. In most cases system information is left to the controls and TAB
contractor.
Examples of issues that are more complex or rare are:

1. A chilled water plant was renovated to a capacity of 6,400 tons. The plant supplied chilled water to 11
buildings of a performing arts center. The new design called for Flow Limiting Devices. These were spool
pieces that had a number of orifices that were spring loaded. The orifices would open or close based on a
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
pressure drop across the spool piece. If the pressure drop got too big, spring loaded orifices would start to
close. The VFD controlled pumps would, in turn, ramp down until the pressure drop was in range. These
were designed and correctly installed on the condenser water return side of five chillers and the supply side 10
cooling towers. Each condenser water pipe from the chillers were between 12” and 16” in diameter. The main
condenser water header was 30” in diameter and 100+ feet long. The pumps, chillers and towers were
respectively in parallel to each other and connected by headers. The headers were connected in series. During
pre-functional testing, the flow was checked with chillers off. With one chiller, pump, tower combination the
system worked well. When a second combination of chiller, pump and tower were started the piping began to
shake and move violently. After a considerable effort between the CxA, the design engineer and orifice
manufacturer, it was determined that the flow limiting devices could be installed either in series or parallel but
not both. The five devices were removed from the chillers and the system worked well.
2. A 36,000sf, BSL-3 bioresearch center with vivarium was nearing completion. Ownership decided that it would
be prudent to commission the facility given problems with past projects of this nature. The facility was within
several months of scheduled opening thus OPR review, design review, submittal review and construction
review was moot. A heat recovery system was added by addendum shortly after drawings were released. Initial
testing determined that the laboratory air handler would not provide the necessary air changes in the vivarium
animal holding room under dirty filter conditions. Other testing found that the laboratory exhaust flow was
substantially less than designed. Further testing determined that the air handler pressure was compromised by
a system effect issue directly downstream of the unit. It was the result of a structural member that limited the
placement of the air handler thus shortening the run out length of the discharge duct. In addition, the fan
motor size was maximized. After reviewing all possible solutions including reducing the air exchange amount
$6+5$(

it was decided to remove the heat recovery coils in the air handler and exhaust duct. The result of which was
higher energy usage.
3. A $220,000,000 maximum security correctional center with 2,000 beds and 18 buildings, included a central
mechanical plant and an above ground one mile long piping loop. The balancing of the secondary chilled
water loop proved to be virtually impossible. In a correction center, the internal load moves throughout the
day almost in mass. The simile “like a dog chasing his tail” was appropriate. The result was the secondary flow
would at times exceed the primary and flow backwards in the bypass for abnormally longer periods than
expected resulting in the inability of the chillers to “see” the load. After a month of adjustments, the loop
balance was improved so that the length of time in reverse bypass flow was reduced. As it was, there was
overflowing of building flow control valves as building demand increased due to the pressure dependent
pumping system. The solution was to add pressure independent flow valves the building take offs of the loop.
ISSUES FOUND IN EXISTING BUILDING COMMISSIONING

Recently Existing Building Cx (EBCx) has become popular and is gaining significant momentum. The popularity is
the result of Owners realizing government incentives to engage a commissioning engineer to analyze high energy
consumption. The benefit of which is to identify operational issues, indoor air quality issues and comfort concerns. Existing
building commissioning is also known as Retro-Commissioning and Recommissioning. The more popular RetroCx is the
commissioning of a building that was not originally commissioned. Typically these buildings have issues that have
knowingly existed for years or even decades. ReCx pertains to a building that was previously commissioned.
Initially, in areas of the country where high energy cost was not an immediate concern and incentives were non-
existent, EBCx primarily addressed nonperformance concerns resulting in resolving comfort issues, improving performance
and energy efficiency. Other reasons for employing EBCx include indoor air quality (IAQ) issues.
Examples of issues found in EBCx:

1. During the initial walk-through of a 15 year old, 1,000,000 sf county operated hospital, it was noted that the
chillers were operating at a two degree delta T across the chiller bundle. After discussing with the operating
personnel, it was noted that the operators were told by their supervisor to maintain the two degree delta. This
apparently was passed down from supervisor to supervisor over the years. If the delta T increased they were
told to increase pump flow. Notification alarms were set to notify operators during off hours when the delta
got greater. The hospital was one of three hospitals operated by the county and operators were routinely
rotated. The matter was confirmed by the new manager, as he noted that he has discussed this with operators
to increase the delta to 10 degrees. He added that there was some reluctance.
2. A three-story, three year old 500,000 sf office building was to be commissioned as part of the Owner’s desire
to attain a LEED-EB certification. The building was a design/build project. It utilized a under floor air
distribution (UFAD) system with seven large packaged air handling units located on the roof. Chilled water
was provided by a central chiller plant, also located on the roof. During the initial walk-through, it was noted
that there was a large amount of return air flowing up the open stairwell and entering the return plenum on
the third floor. Further investigation revealed that the UFAD system was not performing properly as the
temperature profile of the space indicated even temperatures from floor to ceiling. The system was acting as a
typical mixing system.
Another issue was the under floor pressure. The building pressure was maintained at 0.08 inches w.g. The
under floor pressure was maintained at 0.05 inches w.g. Each was using the outside air as a reference. Thus
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
126
ASHRAE Transactions
airflow from the plenum to the space was not uniformed. This was improved by changing the reference for
the unfloor pressure control to reference the space pressure. Other recommendations to the under floor duct
runs were offered but not implemented due the cost.
Further investigation revealed that the roof mounted unit air return was not connected to the return shaft. As
a result the return shaft was in essence a transfer duct however it was not sized to perform as a transfer duct.
Thus the return air was returning via the open stairwells. Other significant issues were leaks at the under floor
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
duct penetrations to the plenum. This resulted in supply air bypassed to the return shaft. This was not initially
a problem as the return shaft was not a negative pressure shaft.
The return shaft issue was eventually resolved by installing two low rpm propeller fans in the return shaft
above the second floor return openings. Once this was in place the UFAD operation worked as required as
return air move to the shaft and not up the stairwell. The UFAD temperature profile improved but leakage at
the duct penetrations t the under floor plenum got worst. These areas were caulked.
3. A 5 year old elementary school located near the confluence of the Ohio River and Mississippi River found
mold growing throughout the building. It is not uncommon for the wet bulb temperature to reach as high as
78 to 80 degrees in the late spring, summer and fall months. Walk throughs and interviews confirmed that
mold was found behind hanging pictures and at floor / wall junctures. The staff and students were
complaining of sinus issues and several students with immune deficiencies were constantly sick and missing
classes. The school district was looking at several lawsuits.
There were approximately 22 classrooms. The classrooms were cooled, ventilated and heated with self-
contained wall units. Thirteen ventilators were 3.5 tons and the remaining were 3 tons. Hot water coils in the
unit ventilators provided heat. Outside air was introduced through the ventilators. Several constant volume air
handling units served the kitchen /dining area, gymnasium, library, and administration offices. The building
was a single story, slab on grade, block and steel construction with a combination of carpet and tile flooring.
Cellular insulation sheets were installed along the building perimeter under the concrete slab. The insulation
extended 6 feet into the building.
A small BAS controlled scheduling and setbacks. The building was setback at night and during unoccupied
periods of the summer. Integral thermostats controlled the unit ventilators. During occupied hours the
ventilator fans were operated and outside air dampers were opened. The thermostats responded to the cooling
and heating needs.
Investigation identified that the 3 ton units were sixed to handle 2 tons of outside air load and 1 ton of the
remaining load. It was determined that the 3.5 ton units were oversized. It was also determined that units were
not satisfying the code required amount of outside air. There was no dehumidification mode as the hot water
coil was in the preheat position. During the spring, late spring, late summer and fall months wet bulb
temperatures were typically high. In a typical classroom on a call for cooling the refrigeration cycle would
engage, with the outside air damper in the open position. When there was no call for cooling the fan would
remain running and outside air damper would remain open for ventilation purposes. This condition didn’t
appear to be problematic as they (unconditioned air introduction) were thought to be short in duration and as
soon as the introduction of warm air raised the space temperature the refrigeration cycle would engage.
Unfortunately, there were two other situations that occurred. The classrooms occupancy would change during
the day with periods of non-occupancy as students would be out for recess, library, gym, lunch and labs. The
other was that during the spring and fall there were periods of rain when the outside temperatures were in the
high 60s and low to mid 70s. The unconditioned air would enter and condense on the tile floor at the wall
$6+5$(

junctures. Forty eight hours later mold would appear. Mold will grow if there is moisture, organic material and
time. When dry the mold would spread.
The solution was to install an outside air make up unit with a total energy recovery wheel and deliver air to the
space at room temperature. The unit ventilators were modified to eliminate the outside air intake and change
the compressors out to 1 ton. Energy usage was slightly higher.
4. A $125,000,000 Casino/Hotel expansion. Originally the facility was a small casino and a connected 9 story
hotel. Two centrifugal chillers located in the facility served the casino’s constant volume systems and the hotel
fan coil units. The expansion tripled the casino floor and doubled the hotel space. A new standalone chilled
water plant was installed on-site with underground hot and chilled water piping serving the building and
connecting to the existing piping system. New large air handlers served the new floor with the smaller original
air handlers relocated and serving smaller areas.
During the first two years of operations the casino was very negative in the winter resulting in a cold floor and
requiring the casino to shut down tables near the main doors. The top three floors of the 9 story new hotel
could not be used in the winter. The construction delivery method utilized a Contract Manager, a
Construction Manager and multiple trade contractors. The project was phased for the purpose of keeping the
casino operating. There were multiple piping, sheet metal and control contractors.
During the initial investigation the major issues became evident. The chilled water underground piping was
leaking. It was discovered that the contractor had changed the specified fiberglass pipe to a C900 PVC push
joint pipe. The underground hot water pipe was leaking. The contractor had used the specified fiberglass pipe
but had not installed the specified anchor system. The chilled water and hot water systems were not operating
at the required fill pressure and pump pressure. The chilled water plant had 5chillers and 4 cooling towers.
The cooling towers had no flow control at the towers thus the flow was unbalanced.
The new hotel major issues were insufficient chilled and hot water pressure; air in the piping system; and toilet
exhaust without balancing dampers. The toilet exhaust fans were located on the roof. The makeup air unit was
located at the first floor and also did not have balancing dampers at the floors. This combination of issues
resulted in the top floors very negative with outside air entering the floor from the elevator shafts and
stairwells and consequently cold. Thus the floors could not be used.
The casino negative pressure issue was complex but in essence it was a combination of issues that included a
lack of sufficient outside air delivered by the air handling units. The new air handling units had a energy
recovery wheel and as such had limited ability to introduce outside air. Some of the existing airs handling units
were not capable to handle outside air. Other issues were that there were multiple TAB reports completed as
phases were completed but there was no TAB performed with all phase completed. Controls and the BAS had
the same issue.
The major issue of pipe leakage remains. Completion of EBCx was limited as a result.
Other examples of issues found during EBCx are:

1. While conducting EBCx on a large retail store, a gas pipe that was running along the roof was kicked and it
fell apart – the pipe had never been threaded, it was just laying inside the elbow – dry-fit. (That particular
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
store had one of the highest gas usages per square foot in NY, and was one of the oldest in upstate NY). The
day that this was found the roof was being redone…with torches.
2. While conducting EBCx on a high-rise building in NYC, 30-year-old induction units with filters that had
never been changed were discovered – some of the nozzles were completely solid with dust and debris,
128
ASHRAE Transactions
building was typically seeing about 30% of their designed airflow (some spaces were seeing -0- airflow).
3. A university residence hall was in its second year of operation. It had been commissioned. The administration
offices were very uncomfortable. The chilled water system was composed of 2 loops one to air handler
serving the administration offices and the other serving the 5 story residence hall fan coil units. The
distribution pump was controlled by a pressure sensor located on the fifth floor riser. The problem was the
air handler which was located about 25 feet from the pump had a higher pressure drop.
4. A historic six-story seminary was renovated into offices. About 350 water source heat pumps were installed.
After three years of operation, compressors on the top floor began to fail. It was found that the fill pressure
in the water loop was low and incorrect.
5. Attached are photos of other issues found during EBCx.
LESSONS LEARNED
The list below does not include lessons learned that would be best served in passing onto design engineers. That is,
answers or considerations for such topics as: What is the most ideal location for a VAV discharge air pressure sensor?
Where is two thirds or three fourths of the duct length? Why there? Which method of controlling return air fans works the
best?
1. Owner commitment to the commissioning process is essential to a successful project. A commitment of

involvement by the Owner’s team includes an understanding of the commissioning process and expectations.
Occasionally Owners or Owners’ representatives take an inactive role in the commissioning process. When
the Owner becomes inactive while maintaining a semblance of Owner involvement with an uninformed or
misguided Owner’s representative the Cx process will be ineffective and the benefits not realized It is a rare
project where this management style is successful. Without Owner commitment it is highly possible that
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
there will be no value added from the Cx process and as such is not a wise investment..
If the responsibility is assigned to a member of the project team the success rate improves but is not ideal as
there exist a conflict of interest. The Building Commissioning Association (BCA) recognizes this conflict
and considers it an essential attribute which is so fundamental to effective building commissioning that all
members agree in writing to adhere to them whenever they serve as a project’s Commissioning Authority.
The attribute as stated is “The CxA is an objective, independent advocate of the Owner. If the CxA’s firm
has other project responsibilities, or is not under direct contract to the Owner, a conflict of interest exists.
Wherever this occurs, the CxA discloses, in writing, the nature of the conflict and the means by which the
conflict shall be managed.”(1) Conflict of interest exists when the CxA is trying to serve the owner’s interest
while under contractual obligation responsible for design, construction management, equipment supply, or
construction, whose interests may include making a profit, meeting a schedule, or protecting their reputation.
A conflict of interest is inherent to such an arrangement. Ethical professionals may be able to manage it to
the mutual satisfaction of the concerned parties. The BCA believes that the conflict of interest and the
means of managing it should be clearly identified, understood, and documented. The BCA also believes that
uncensored communication between the CxA, the owner, and the other members of the commissioning
team is an essential component of effective conflict of interest management.
With LEED splitting the Cx process into Fundamental (Pre requisite) and Enhanced a more common
method is to split the Cx responsibility into two contracts with Fundamental Cx contracted to the CM and
Enhanced Cx to the Owner. Fundamental Cx typically represents 65% to 80% of the total cost of Cx. Thus
$6+5$(

Fundamental Cx cost is assigned to the project construction cost managed by the CM. The CxA becomes
responsible to the CM and organizationally appears the same as the trade contractors. Thus the CxA is
selected on lowest bid without a strenuous vetting of qualifications. Today some Federal agencies along with
some Universities are employing this methodology. Inherently this jeopardizes the Cx process as a “quality-
oriented process” with the unintentional result to water it down and thus become ineffective.
The common denominator of all successful projects over the last 20 years is a high degree of Owner
involvement in the Cx process.
2. Equally important to the success of the Cx process is the Owners Project Requirements. It is rare to find
failure when a detailed and thorough OPR has been developed and approved. As with all such plans it must
be revised and monitored throughout the project. Most importantly it is the measure of project success.
The quality of the document is typically lacking especially in performance detail. Any and all assumptions
that the Architect and Engineer make must be identified including Owner and User requirements; design
parameters for load calculations; equipment selection criteria; energy efficiency goals; environmental and
sustainability goals; indoor air quality requirements; equipment and system operational expectations; building
occupant and O&M requirements; and interpretation of applicable standards, guidelines and codes.
The OPR is initially used by the CxA during the Construction Documents review. Any omissions, variances,
ambiguities or confusing interpretations of the document is identified and resolved with the Cx team. In
effect this is a potential problem exercise with the purpose of strengthening the documents so that they meet
the intent OPR.
No better example of the importance of an OPR is exemplified then during the Retro or EBCx 0f a building.
Without exception the common denominator of poorly performing buildings is a vague OPR or lack thereof.
3. As mentioned, commissioning is a quality-oriented process, in essence a quality management process. Many
commissioning efforts are minimized because there is a lack of technical expertise on of the CxA. Knowledge
of the technical aspects of the project is paramount to the commissioning success of the projects.
Some Commissioning firms mistakenly address only the functionality of equipment or systems meaning they
checkout the Controls and Sequence of Operations but they do not consider the performance requirements
thus they miss the operational requirements of the OPR. In other words how do the systems and subsystems
perform as the load moves from one design extreme to the other? Confirming that a damper opens but
missing considerations of continuous operations of the damper. Confirming the TAB results of a pump
balance but not investigating pump performance during parallel pumping. Single event testing alone is not
performance testing. With technical knowledge of system performance single event testing can provide
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
information that assists the CxA.
The industry must do a better job in establishing qualifications of firms not only in the application of the
process but in the technical understanding of systems and their applications.
4. To quote one of our Project Engineers “What is remarkable is how minor the items are that I can name on
projects I’ve commissioned – probably because being on site mitigates so many ridiculous deficiencies from
ever happening. The worst stuff that I’ve ever found was as a design engineer that could have been avoided
by performing commissioning”:
It has become evident that projects where Owners have employed commissioning services there are the
unexpected benefits of reduced issues or deficiencies. In 2001 a 70,000 sf plant science research lab had 858
construction issues. A more recent 40,000 sf bioresearch lab completed in 2010 had 132 construction issues.
This is not an anomaly it is consistent with other projects. It is a significant trend among Owners such as
130
ASHRAE Transactions
Universities who have committed to the Cx process over multiple projects.
The majority of issues eliminated are installation details of accessibility, maintainability, piping details such as
reducing pipe sizing from a control valve to a coil connection. Duct details such as duct fabrication details,
sealant, insulation and duct support and connections. Temperature control issues related to inadequate point
to point documentation and vague temperature control diagrams.
5. The impact of commissioning is intended to positively influence building operations and maintenance. There
are several areas where this has been lacking, record ‘as built’ documentation, O&M training, and Lessons
Learned meetings. This is true in spite of close out specifications and various guidelines. Typically, the close-
out specifications of a project are not enforced and ignored or lack specific details. ASHRAE Guideline 4-
2008 provides O&M close out details that are very seldom specified.
In the case of O&M staff training the commissioning provider is not involved in actual training but is to
review lessons, and confirm that it has been provided by the contractors. The O&M manual in nothing more
than a collection of vendor supplied equipment O&M manuals. Occasionally the manual will include
approved equipment submittals. While equipment training by vendors is typical there is generally no system
training with exception to BAS / Temperature controls training by the Temperature Control Contractor.
Improvement in this is needed. First the commissioning provider must have a direct involvement including
providing system training with assistance from the design engineer. The design engineer must tighten the
specifications and provide more detail of contractor expectations.
Lessons Learned meetings are an important way to pass on the valuable information learned as the result of
the commissioning process. Unfortunately they are seldom used. Obviously there have been lessons learned
by all participants in the course of the commissioning exercise.
THE FUTURE OF HVAC COMMISSIONING

I think that, generally, while we’ve learned a lot over the last 20 years (as an industry) there is still a long way to go. It
seems that systems have become more complex with the changing technology. A visit to the tri-annual AHR Expo trade
show in Chicago provides the visitor a glimpse of the enormity of the industry as well as the changing technology. The era
of sustainable design including a refocus on energy efficiency, indoor air quality, and smart grid technologies has raised the
bar. Furthermore, the construction industry is constantly searching for new or improved project delivery methods to
improve productivity and reduce builders’ risk.
The more complex systems get –complicated DDC, etc. – the more expertise it takes to install the systems and make
them work correctly…which means they are more often mis-applied, specified incorrectly, installed poorly, or not
maintained properly. Some of the corrections that we make every day seem extremely subtle because we are catching
controls issues or tuning buildings rather than just finding gross errors that make buildings unusable – but their economic
impacts are often significant.
When the commissioning process was introduced it was popular to think of it as a “management process”, not
necessarily performed by engineers. In other words, if the process is followed then personnel with unsophisticated HVAC
technical expertise could successfully implement it. Commissioning training was centered on the process. It then followed
that when existing building commissioning became popular the focus was also on the process. We have learned that this is
not the case if value to the Owner is indeed the desired result.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

Thus, it has become more and more evident that commitment of continuing training of commissioning personnel is
imperative if commissioning is to be of value to the Owner. Commissioning personnel should undergo continuous internal
and external training in the commissioning process of both new and existing buildings, mechanical and electrical building
system analysis, BAS and temperature control systems, new technologies, sustainable design, ASHRAE Standards, and
equipment and systems operations and maintenance.
At the turn of the century commissioning was primarily employed on new construction. Building commissioning was
focused on mechanical systems, indoor air quality and ancillary electrical systems such as emergency or stand by power.
Ten years later new building construction includes building envelope, and Existing Building Commissioning has move to
the forefront with new challenges. On the horizon is Continuous Commissioning and data monitoring which will be driven
by the sustainable design movement and smart grid technology. Systems will become more complicated and the technical
demands of the commissioning provider will increase.
Initially, the building commissioning process was developed to address ‘buildings that didn’t work’ as the Owner
expected. Today we can say the when the Commissioning Process is implemented properly it works – building
commissioning is here to stay!
FOOTNOTES
(1)BCA Essential Attribute number 2
REFERENCES
ASHRAE Guideline 0 – 2005: The Commissioning Process
ASHRAE Guideline 4 – 2008: Preparation of Operating and Maintenance Documentation for Building Systems
BCA Essential Attributes of Building Commissioning
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
132
ASHRAE Transactions
SA-12-C017
0RGHOLQJDQG2SWLPL]DWLRQRI+9$&
6\VWHPV8VLQJ$UWLILFLDO,QWHOOLJHQFH
$SSURDFKHV

1DELO1DVVLI3K'3(
$6+5$(0HPEHU
$%675$&7
,QWHOOLJHQW HQHUJ\ PDQDJHPHQW FRQWURO V\VWHP (0&6 LQ EXLOGLQJV RIIHUV DQ H[FHOOHQW PHDQV RI UHGXFLQJ HQHUJ\
FRQVXPSWLRQV LQ +9$& V\VWHPV ZKLOH PDLQWDLQLQJ RU LPSURYLQJ LQGRRU HQYLURQPHQWDO FRQGLWLRQV 7KLV FDQ EH DFKLHYHG
WKURXJK WKH XVH RI FRPSXWDWLRQDO LQWHOOLJHQFH DQG RSWLPL]DWLRQ 7KH SDSHU WKXV SURSRVHV DQG HYDOXDWHV PRGHOEDVHG
RSWLPL]DWLRQ SURFHVV IRU +9$& V\VWHPV XVLQJ HYROXWLRQDU\ DOJRULWKP DQG DUWLILFLDO QHXUDO QHWZRUNV 7KH SURFHVV FDQ EH
LQWHJUDWHG LQWR WKH (0&6 WR SHUIRUP VHYHUDO LQWHOOLJHQW IXQFWLRQV DQG DFKLHYH RSWLPDO ZKROHV\VWHP SHUIRUPDQFH 7KH
SURSRVHG PRGHOV DQG WKH RSWLPL]DWLRQ SURFHVV DUH WHVWHG XVLQJ GDWD FROOHFWHG IURP DQ H[LVWLQJ +9$& V\VWHP 7KH WHVWLQJ
UHVXOWVVKRZWKDWWKHPRGHOVFDQFDSWXUHYHU\ZHOOWKHV\VWHPSHUIRUPDQFHDQGWKHRSWLPL]DWLRQSURFHVVFDQUHGXFHHQHUJ\
FRQVXPSWLRQE\DERXWZKHQFRPSDUHGWRWKHWUDGLWLRQDORSHUDWLQJVWUDWHJLHVDSSOLHG
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
,1752'8&7,21
*UHDWHIIRUWVKDYHEHHQLQYHVWHGLQPLQLPL]LQJWKHHQHUJ\FRVWVDVVRFLDWHGZLWKWKHRSHUDWLRQRI+9$&V\VWHPV,QWHOOLJHQW
(QHUJ\ 0DQDJHPHQW DQG &RQWURO 6\VWHPV (0&6 FDQ SURYLGH DQ HIIHFWLYH ZD\ RI GHFUHDVLQJ HQHUJ\ FRVWV LQ +9$&
V\VWHPV ZKLOH PDLQWDLQLQJ LQGRRU HQYLURQPHQWDO FRQGLWLRQV $6+5$( 7KLV (0&6 FDQ LQFOXGH VHYHUDO LQWHOOLJHQW
IXQFWLRQVVXFKDVRSWLPXPVHWSRLQWVDQGRSHUDWLQJPRGHV$6+5$(1DVVLIHWDO:DQJDQG-LP=KHQJ
DQG=DKHHU8GGLQDQGIDXOWGHWHFWLRQDQGGLDJQRVLV 6HHPDQG /HHHWDO7KHLQWHOOLJHQW(0&6FDQEH
DFKLHYHGWKURXJKWKH XVHRIWKHFRPSXWDWLRQDOLQWHOOLJHQFHDQG RSWLPL]DWLRQ+DQL7KHSDSHUWKXVSURSRVHV PRGHO
EDVHG RSWLPL]DWLRQ SURFHVV IRU +9$& V\VWHPV XVLQJ JHQHWLF DOJRULWKP DQG DUWLILFLDO QHXUDO QHWZRUNV 7KH SURFHVV FDQ EH
LQWHJUDWHGLQWRWKH(0&6WRSHUIRUPVXFKLQWHOOLJHQWIXQFWLRQVDQGDFKLHYHRSWLPDOZKROHV\VWHPSHUIRUPDQFH7KH+9$&
RSWLPL]DWLRQSUREOHPVDUHVROYHGXVLQJWKHJHQHWLFDOJRULWKPLQVSLUHGE\QDWXUDOHYROXWLRQZKLFKLVVXFFHVVIXOO\DSSOLHGWRD
ZLGHUDQJHRIDSSOLFDWLRQVLQFOXGLQJ+9$&V\VWHPV'HE*ROGEHUJ;XHWDO0RVVROO\HWDO7KH
+9$&RSWLPL]DWLRQSUREOHPVDUHG\QDPLFDQGWKHSUREOHPFKDQJHVRYHUWKHFRXUVHRIWKHRSWLPL]DWLRQ7KXVWKH*$ZLWK
SURSHUHQKDQFHPHQWDQGWKHDELOLW\RIFRQWLQXRXVO\WUDFNWKHPRYHPHQWRIWKHRSWLPXPRYHUWLPHDUHGHYHORSHGDQGXVHG
&RPSRQHQWV PRGHOV DUH UHTXLUHG IRU WKH RSWLPL]DWLRQ SURFHVV DQG IRU DQ\ RWKHU IXQFWLRQV 'HSHQGLQJ RQ WKH W\SH RI
IXQFWLRQVDQGWKHDFFXUDF\UHTXLUHGWKHPRGHOVFDQYDU\IURPVLPSOHWRPRUHVRSKLVWLFDWHGFDOFXODWLRQV+RZHYHULWLVRI
SUDFWLFDOLPSRUWDQFHWRGHYHORSDVLPSOH\HWDFFXUDWHDQGUHOLDEOHPRGHOWREHWWHUPDWFKWKHUHDOEHKDYLRURIWKHVXEV\VWHPV
DQG RYHUDOO V\VWHP RYHU WKH HQWLUH RSHUDWLQJ UDQJH 0RGHOV FDQ EH GHYHORSHG E\ WZR GLVWLQFW PHWKRGV $6+5$(
*UHHQVERUR1&
$6+5$(

IRUZDUGPRGHOVDQGGDWDGULYHQPRGHOV)RUZDUGPRGHOVPD\QHHGGHWDLOHGSK\VLFDOLQIRUPDWLRQWKDWLVQRWDOZD\VDYDLODEOH
0RUHRYHULWLVDOPRVWLPSRVVLEOHWRGHYHORSDPRGHOEDVHGRQSK\VLFDONQRZOHGJHWKDWSHUIHFWO\VLPXODWHVWKHUHDOV\VWHP
EHKDYLRU,QDGGLWLRQWKHFKDUDFWHULVWLFVRIWKHUHDOV\VWHPFDQFKDQJHRYHUWLPHWKHSK\VLFDOPRGHOZLWKIL[HGSDUDPHWHUV
PD\QRORQJHUEHDEOHWRWUDFNFRUUHFWO\WKHUHDOSHUIRUPDQFH$OWHUQDWLYHO\GDWDGULYHQPRGHOVDUHDQRWKHUDSSURDFKWREH
XVHG IRUH[LVWLQJV\VWHPV,Q WKLVSDSHU ZH ZLOOH[SORUHWKHXVHRIDUWLILFLDO QHXUDO QHWZRUNDVGDWDGULYHQDQGVHOIWXQLQJ
PRGHOV LQ +9$& DSSOLFDWLRQV IRFXVLQJ RQ RSWLPL]DWLRQ DQG FRQWURO 7KRVH PRGHOV KDYH DGYDQWDJHV RYHU FRQYHQWLRQDO
RQOLQHLGHQWLILFDWLRQPHWKRGLQWHUPVRIUREXVWQHVVDQGDFFXUDF\RIPRGHOSUHGLFWLRQV$KPHGHWDO/HHHWDO
7KHSURSRVHGPRGHOVDQGWKHRSWLPL]DWLRQSURFHVVDUHERWKWHVWHGHYDOXDWHGXVLQJGDWDFROOHFWHGRYHUWKUHHVXPPHUPRQWKV
IURPDW\SLFDOH[LVWLQJ+9$&V\VWHP
6<67(0&21),*85$7,21
)LJXUHVKRZVDVFKHPDWLF GLDJUDPRID W\SLFDO+9$& V\VWHPDQGWKHRSWLPL]DWLRQSURFHVVLQWHJUDWHGLQWR(0&6
,QFOXGHGLQWKLVVFKHPDWLFDUHWKHSURSRVHGDUWLILFLDOQHXUDOQHWZRUN$11PRGHOVDQGKRZWKH\LQWHUDFWZLWKLQWKHRYHUDOO
V\VWHP$W\SLFDO+9$&V\VWHPWKDWXVHVYDULDEOHDLUYROXPH9$9FRQWUROLVLOOXVWUDWHGLQ)LJXUH,WFRQVLVWVPDLQO\RI
LWKHUHWXUQDQGVXSSO\IDQVLLWKHRXWGRRUGLVFKDUJHDQGUHFLUFXODWLRQGDPSHUVLLLWKHDLUKDQGOLQJXQLW$+8ZLWK
FRPSRQHQWVVXFKDVILOWHUDQGFRROLQJDQGKHDWLQJFRLOVLYWKHSUHVVXUHLQGHSHQGHQW9$9WHUPLQDOER[HVDQGYWKHORFDO
ORRSFRQWUROOHUVLH&&DQG&7KHVXSSO\DLUWHPSHUDWXUHLVFRQWUROOHGE\WKHFRQWUROOHU&7KHGXFWVWDWLFSUHVVXUH
LVFRQWUROOHGE\WKHFRQWUROOHU&7KH]RQHDLUWHPSHUDWXUHDWDQ\SDUWLFXODU]RQHQLVFRQWUROOHGE\WKHFRQWUROOHU&Q
7KH(0&6FROOHFWVWKHPHDVXUHGGDWD³UHDOGDWDÍURPFRPSRQHQWVRUVXEV\VWHPV7KH$11PRGHOVDUHFRQWLQXRXVO\
WUDLQHGXVLQJWKHUHDOGDWDWREHWWHUPDWFKWKHUHDOEHKDYLRURIWKHVXEV\VWHPVDQGRYHUDOOV\VWHP7KHWUDLQLQJDOJRULWKPXVHG
LV/HYHQEHUJ0DUTXDUGW6LQFHWKHPRGHODUHDOZD\VXSGDWHGE\XVLQJRQOLQHUHDOGDWDWKHRSWLPDOYDOXHVWKDWSURGXFHWKH
EHVWHVWLPDWHGSHUIRUPDQFHZLOOOHDGWRWKHEHVWSHUIRUPDQFH$VUHFRPPHQGHGLQWKLVSDSHUIRURSWLPDOFRQWUROVWUDWHJ\DW
HDFKWLPHLQWHUYDOHJPLQWKH$11PRGHOVSURYLGHRSWLPDOZKROHV\VWHPSHUIRUPDQFHWKURXJKGHWHUPLQLQJRSWLPDO
VHWSRLQWVDQGRSHUDWLRQVHTXHQFHV
*HQHWLF DOJRULWKP LV XVHG WR VROYH WKH RSWLPL]DWLRQ SUREOHP 7KH SUREOHP YDULDEOHV DUH WKH FRQWUROOHU VHWSRLQWV DQG
RSHUDWLRQPRGHV,QWKLVSDSHUWKHVXSSO\DLUWHPSHUDWXUHGXFWVWDWLFSUHVVXUHDQGRXWGRRUDLUIORZUDWHDUHRQO\WKHSUREOHP
YDULDEOHVFRQVLGHUHG7KHREMHFWLYHIXQFWLRQLVWKHWRWDOHQHUJ\FRQVXPSWLRQRYHUWKHRSWLPL]DWLRQSHULRGGHWHUPLQHGE\WKH
$11 FRPSRQHQW PRGHOV 7KH *$ VWDUWV ZLWK D UDQGRP JHQHUDWLRQ RI WKH LQLWLDO SRSXODWLRQ LQLWLDO VROXWLRQ 7KH SUREOHP
YDULDEOHVFRQWUROOHUVHWSRLQWVDUHHQFRGHGWRIRUPDFKURPRVRPHDVWULQJRIYDULDEOHVWKDWUHSUHVHQWVDQLQGLYLGXDORQH
VROXWLRQLQWKHSRSXODWLRQ7KHSHUIRUPDQFHRUREMHFWLYHIXQFWLRQ IRIHDFKLQGLYLGXDORIWKHILUVWJHQHUDWLRQLVHVWLPDWHG
7KH VHFRQG JHQHUDWLRQ LV JHQHUDWHG XVLQJ RSHUDWLRQV RQ LQGLYLGXDOV VXFK DV VHOHFWLRQ FURVVRYHU DQG PXWDWLRQ LQ ZKLFK
LQGLYLGXDOV ZLWK KLJKHU SHUIRUPDQFH ILWQHVV KDYH D JUHDWHU FKDQFH WR VXUYLYH 'HE 7KH SHUIRUPDQFH RI HDFK QHZ
LQGLYLGXDO LV DJDLQ HYDOXDWHG 7KH SURFHVV LV UHSHDWHG XQWLO WKH PD[LPXP QXPEHU RI JHQHUDWLRQV *PD[ LV UHDFKHG 7KH
VLPXODWHGELQDU\FURVVRYHU6%;LVXVHGWRFUHDWHWZRRIIVSULQJVIURPWZRSDUHQWVROXWLRQV7KHUDQGRPVLPSOHVWPXWDWLRQ
RSHUDWRULVDSSOLHGWRFUHDWHDQHZVROXWLRQIURPWKHHQWLUHVHDUFKVSDFH7KLVDOJRULWKPXVHVWKHHOLWHSUHVHUYLQJRSHUDWRU
ZKLFKIDYRUVHOLWHVRIDSRSXODWLRQE\JLYLQJWKHPDQRSSRUWXQLW\WREHGLUHFWO\FDUULHGRYHUWRWKHQH[WJHQHUDWLRQ$IWHUWZR
RIIVSULQJVDUHFUHDWHGXVLQJWKHFURVVRYHUDQGPXWDWLRQRSHUDWRUVWKH\DUHFRPSDUHGZLWKERWKRIWKHLUSDUHQWVWRVHOHFWWZR
EHVW VROXWLRQV DPRQJ WKH IRXU SDUHQWRIIVSULQJ VROXWLRQV 6WRFKDVWLF XQLYHUVDO VDPSOLQJ686 YHUVLRQ RI SURSRUWLRQDWH
VHOHFWLRQLVXVHG'HE$SHQDOW\LVLPSRVHGRQWKHREMHFWLYHIXQFWLRQVEDVHGRQWKHFRQVWUDLQWV7KHFRQVWUDLQWVDUH
UHODWHGWRWKHUHVWULFWLRQVRQWKHRSHUDWLRQRIWKH+9$&V\VWHPDQGWKHUDWHRIYDULDEOHFKDQJHV7KH\FRYHUWKHORZHUDQG
XSSHU OLPLWV RI YDULDEOHV WKH GHVLJQ FDSDFLW\ RI FRPSRQHQWV 7KH IDQ DQG ]RQH DLUIORZ UDWHV IRU LQVWDQFH DUH UHVWULFWHG
ZLWKLQWKHPD[LPXPDQGPLQLPXPOLPLWV7KHPLQLPXP]RQHDLUIORZUDWHVDUHUHVWULFWHGWREHDOZD\VKLJKHUWKDQRI
GHVLJQDLUIORZUDWHV)RUWKHUHVXOWVRIWKLVVWXG\WKHFRPSXWDWLRQVZLWKWKHRSWLPL]DWLRQSURFHVVDUHSHUIRUPHGZHOOZLWKLQD
WLPH OHQJWK RI WKUHH PLQXWHV RQ D SHUVRQDO FRPSXWHU 7KLV FRPSXWDWLRQ WLPH DOORZV WKH RSWLPL]DWLRQ SURFHVV WR EH
LPSOHPHQWHGRQOLQH7KHWLPHFRXOGEHDOVRGHFUHDVHGXVLQJDPRUHPRGHUQFRPSXWHU

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
134
ASHRAE Transactions

)URPRWKHU]RQHV
'DPSHU 9$9ER[ =RQH

0RWHU
)LOWHUDQGFRROLQJ
DQGKHDWLQJFRLOV 6XSSO\IDQ 7RRWKHU]RQHV
6XSSO\DLUWHPSHUDWXUH
=RQHDLUWHPSHUDWXUH
'XFWVWDWLFSUHVVXUH
7RRWKHU]RQH
&RQWUROOHU FRQWUROOHUV
& &Q
&RQWUROOHU &RQWUROOHU
7R&KLOOHU
& &
&RQWUROOHU
(0&6 5HDOGDWD
2SWLPL]DWLRQSURFHVV
2SWLPDOVHWSRLQWVDQG (UURU
0RGHOSDUDPHWHUV (YROXWLRQDU\DOJRULWKP

)DXOWGHWHFWLRQ
UHSRUWV 7UDLQLQJDOJRULWKP (VWLPDWHGGDWD
/HYHQEHJ
'LIIHUHQWDSSOLFDWLRQV 0DUJXDUGW
$110RGHOV
:HLJKWVDQGELDV
)LJXUH 7\SLFDO+9$&V\VWHPDQG(0&6DORQJZLWKWKHSURSRVHG$11PRGHOVDQGRSWLPL]DWLRQSURFHVV
&20321(1702'(/6
7KH +9$& FRPSRQHQW PRGHOV EDVHG RQ DUWLILFLDO QHXUDO QHWZRUNV ZLWK µVHOIOHDUQLQJ¶ FDSDELOLW\ DUH GHYHORSHG DQG
XWLOL]HGLQWKH(0&6WRSHUIRUPWKHDGYDQFHGDQGLQWHOOLJHQWIXQFWLRQV$QLQILQLWHYDULHW\RIQHWZRUNDUFKLWHFWXUHVFDQEH
XVHG IRU WKLV SXUSRVH EXW LQ WKH LQWHUHVW RI FRQVHUYLQJ FRPSXWHU WLPH WKH VLPSOHVW VWUXFWXUH QHHG WR EH FRQVLGHUHG (DFK
PRGHO KDV RQO\ RQH KLGGHQ OD\HU ZLWK WZHQW\ QHXURQV DQG ZLWK K\SHUEROLF DFWLYDWLRQ IXQFWLRQ WDQFK 7KH PRGHO KDV WKH
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

RXWSXWOD\HUZLWKRQHQHXURQDQGDQDFWLYDWLRQIXQFWLRQEDVHGRQWKHVXPRIWKHZHLJKWHGKLGGHQOD\HUQHXURQV(DFKQHXURQ
KDVDOVRELDV
7KH PRGHOV GHVFULEHG KHUH DUH WKH IDQ FRROLQJ FRLO DQG FKLOOHU PRGHOV 7KRVH PRGHOV FDQ EH XVHG IRU YDULRXV
DSSOLFDWLRQVEXWWKHLQSXWVDQGWKHRXWSXWVKDYHWREHFOHDUO\GHILQHG7KHLQSXWVRIWKHPRGHOVIRUDVSHFLILFDSSOLFDWLRQPD\
EHFRPH RXWSXWV LQ RWKHU DSSOLFDWLRQ )RU LQVWDQFH LQ WKH FRQWURO DSSOLFDWLRQ WKH FRQWUROOHG YDULDEOHV VXFK DV GXFW VWDWLF
SUHVVXUH DQG VXSSO\ DLU WHPSHUDWXUH DUH WKH RXWSXWV RI WKH IDQ PRGHO DQG FRROLQJ FRLO PRGHO UHVSHFWLYHO\ )RU WKH
RSWLPL]DWLRQ WKH PRGHO RXWSXWV DUH WR HVWLPDWH WKH HQHUJ\ FRQVXPSWLRQV WKH REMHFWLYH IXQFWLRQ VXFK DV IDQ SRZHU DQG
FRPSUHVVRUSRZHU$OWKRXJKWKHPRGHOVZHUHYDOLGDWHGDQGWHVWHGIRUERWKFDVHVWKHGLVFXVVLRQVLQWKLVSDSHUZHUHOLPLWHG
IRU WKH RSWLPL]DWLRQ DSSOLFDWLRQ 7KXV WKH LQSXWV RI WKH IDQ PRGHO DUH WKH V\VWHP DLUIORZ UDWH DQG VWDWLF SUHVVXUH DQG WKH
RXWSXWLVIDQSRZHU7KHLQSXWVRIWKHFRROLQJFRLOPRGHODUHIDQDLUIORZUDWHHQWHULQJOLTXLGWHPSHUDWXUHHQWHULQJDLUGU\
EXOEWHPSHUDWXUHDQGKXPLGLW\UDWLROHDYLQJDLUGU\EXOEWHPSHUDWXUHDQGWKHRXWSXWLVFRROLQJORDG7KHLQSXWVRIWKHFKLOOHU
PRGHODUHWKHFRROLQJFRLOORDGFKLOOHGZDWHUVXSSO\WHPSHUDWXUHDQGWKHFRQGHQVHUZDWHUWHPSHUDWXUHDQGWKHRXWSXWLVWKH
FRPSUHVVRU SRZHU $GGLWLRQDO EDVLF FDOFXODWLRQV DUH DOVR UHTXLUHG IRU WKH RSWLPL]DWLRQ LQFOXGLQJ D ]RQH PRGHO 7KH ]RQH
PRGHOLVWRGHWHUPLQH]RQHDLUIORZUDWHVDQGORFDOKHDWLQJHQHUJ\XVHDQGUHWXUQDLUFRQGLWLRQVEDVHGRQWKHUPDOORDGV7KH
FDOFXODWLRQVDUHEDVHGRQVWHDG\VWDWHKHDWEDODQFHHTXDWLRQIRUHDFK]RQHLQZKLFKWKHVHQVLEOHORDGLVDIXQFWLRQRIDLUIORZ
UDWHDQGWKHGLIIHUHQFHVEHWZHHQWKHVSDFHDQGVXSSO\DLUWHPSHUDWXUHV6LPLODUO\WKHKXPLGLW\LVGHWHUPLQHGXVLQJWKHODWHQW
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
ORDG7KHORDGVDUHGHWHUPLQHGIURPWKHVDPHPRGHOEXWZLWKDQLQYHUVHIRUPXVLQJPHDVXUHGGDWDRISUHYLRXVSHULRGDQG
WKHQWKHORDGVDUHDVVXPHGWREHFRQVWDQWGXULQJWKHFXUUHQWRSWLPL]DWLRQSHULRG7KHHOHFWULFUHKHDWLVFRQVLGHUHGKHUHDQGLW
WXUQV RQ RQO\ ZKHQ WKH DLUIORZ UDWH UHDFKHV LWV PLQLPXP OHYHO HJ RI GHVLJQ DLUIORZ UDWH DQG VSDFH WHPSHUDWXUH
EHFRPHVORZHUWKDQKHDWLQJVHWSRLQW7KHV\VWHPDLUIORZUDWHXVHGDVDQLQSXWIRUFRROLQJRUIDQPRGHOLVHTXDOWRWKHVXPRI
]RQHDLUIORZUDWHVIRXQGIURPWKH]RQHPRGHO,WHUDWLRQSURFHVVVKRXOGEHDSSOLHGWRHVWLPDWHWKHUHWXUQDLUFRQGLWLRQVLQLWLDO
FRROLQJ FRLO OHDYLQJ DLU KXPLGLW\ UDWLR LV DVVXPHG DQG WKH QHZ YDOXH LV FDOFXODWHG DQG UHXVHG 7KLV LWHUDWLYH SURFHVV
FRQWLQXHVFDOFXODWLQJWKURXJKWKHORRSVHYHUDOWLPHVXQWLOWKHYDOXHVRIFRROLQJFRLOOHDYLQJDLUKXPLGLW\UDWLRVWDELOL]HZLWKLQ
DVSHFLILHGWROHUDQFH7KHFDOFXODWLRQSUHVHQWHGDERYHLVVLPLODUWRWKDWIRUD9$9PRGHOLQ+9$&WRRONLW%UDQGHPXHKOHW
DO 7KH PLQLPXP DLU IORZ UDWH EDVHG RQ WKH $6+5$( VWDQGDUG $6+5$( LV LQFOXGHG LQ WKH
RSWLPL]DWLRQFDOFXODWLRQV7KHRXWGRRUDLULVGHWHUPLQHGE\WKHPXOWL]RQHSURFHGXUHRI$6+5$(VWDQGDUGEDVHGRQWKH
DFWXDO ]RQH DLUIORZ UDWHV 7KH DGYDQWDJH RI LQFOXGLQJ WKH PLQLPXP RXWGRRU VWDQGDUG SURFHGXUH LQ WKH ZKROH RSWLPL]DWLRQ
SURFHVVLVWRPLQLPL]HWKHHQHUJ\XVHZKLOHUHVSHFWLQJWKHYHQWLODWLRQUHTXLUHPHQWVE\WKHFXUUHQWVWDQGDUG
02'(/7(67,1*
'DWD IURP DQ H[LVWLQJ 9$9 V\VWHP DUH FROOHFWHG RYHU WKUHH VXPPHU PRQWKV -XQH -XO\ DQG $XJXVW DW RQHPLQXWH
LQWHUYDOV7KHV\VWHPSURYLGHVFRQGLWLRQHGDLUWRFODVVURRPVDQGRIILFHVDWRWDORI]RQHV7KHILUVWGD\VGDWDRIHDFK
PRQWKGDWDSRLQWVZHUHUDQGRPO\GLYLGHGLQWRWZRW\SHVRIVDPSOHVIRUWUDLQLQJGDWDSRLQWVDQG
IRUYDOLGDWLRQGDWDSRLQWV7KHODVWGD\VRIHDFKPRQWKGDWDSRLQWVLVNHSWIRUWKHPRGHOWHVWLQJDQGWRGHILQH
WKH PRGHO DFFXUDF\ 'XULQJ WKH WUDLQLQJ WKH QHWZRUN LV DGMXVWHG DFFRUGLQJ WR LWV HUURU XVLQJ /HYHQEHUJ0DUTXDUGW
RSWLPL]DWLRQ7KHYDOLGDWLRQ VDPSOHVDUH XVHG WR PHDVXUH QHWZRUNJHQHUDOL]DWLRQDQGWRKDOWWUDLQLQJ ZKHQ JHQHUDOL]DWLRQ
VWRSVLPSURYLQJ7KHWHVWLQJ GDWDKDYH QRHIIHFWRQWUDLQLQJDQGSURYLGHDQLQGHSHQGHQW PHDVXUHRI QHWZRUNSHUIRUPDQFH
DIWHU WUDLQLQJ 7KH PRGHO SHUIRUPDQFH LV PHDVXUHG E\ WKH FRHIILFLHQW RI YDULDQFH &9 ZKLFK LV GHILQHG DV WKH UDWLR RI WKH
VWDQGDUGGHYLDWLRQWRWKHPHDQ
)LJXUHVKRZVWKHUHVXOWVRIWKH$11IDQPRGHOWUDLQLQJYDOLGDWLRQDQGWHVWLQJ7KHVWUDLJKWOLQHLVDRQHWRRQHOLQH
LQGLFDWLQJDJUHHPHQWEHWZHHQWKHPHDVXUHGWDUJHWDQGVLPXODWHGRXWSXWIDQSRZHU7KHPRGHOVOLJKWO\RYHUHVWLPDWHVWKH
SRZHU DW HOHYDWHG YDOXHV GXULQJ WKH WUDLQLQJ SHULRG DQG XQGHUHVWLPDWHV WKH SRZHU GXULQJ WKH WHVWLQJ SHULRG7KH UHVXOWV LQ
WHUPRI&9VDUHVKRZQLQ)LJXUH7KH&9LVFRPSDULQJWRGXULQJWKHWUDLQLQJSHULRG7KHUHVXOWVDUHDYHUDJHG
IRUDSHULRGRIWKUHHPRQWKVRQHZHHNIRUHDFKPRQWK7KHLQSXWVRIWKH$11IDQPRGHODUHWKHPHDVXUHGWRWDOSUHVVXUH
GLIIHUHQFHDFURVVWKHIDQDQGDLUIORZUDWH,QWKHRSWLPL]DWLRQSUREOHPWKHHQHUJ\XVHLQIDQQHHGVWRFRUUHODWHZLWKWKHGXFW
VWDWLFSUHVVXUHVHWSRLQWQRWZLWKWKHWRWDOSUHVVXUHGLIIHUHQFH7KXVWKHSUHVVXUHGURSVEHWZHHQWKHGXFWVWDWLFVHQVRUDQGIDQ
136
ASHRAE Transactions
RXWOHWDQGEHWZHHQWKHRXWGRRUDLUGDPSHUDQGIDQLQOHWQHHGWREHFRQVLGHUHGDQGFDQEHVLPSOLILHGDVSURSRUWLRQDOWRWKH
VTXDUHRIWKHIORZUDWH
,QWKHFRROLQJFRLO PRGHOWKHUHTXLUHGRXWSXWIRUWKHRSWLPL]DWLRQLVFRROLQJFRLOORDGWKDWLVLQWXUQWKHLQSXWWRWKH
FKLOOHUPRGHO7KHPHDVXUHGFRROLQJFRLOORDGVUHTXLUHGIRUWUDLQLQJDQGWHVWLQJDUHFDOFXODWHGE\WKHPHDVXUHGDLUIORZUDWH
DQGGLIIHUHQFHEHWZHHQ WKH PHDVXUHGLQOHWDQGRXWOHWHQWKDOS\)LJXUH VKRZVWKHUHVXOWVRI WKH $11FRROLQJFRLO PRGHO
WUDLQLQJ YDOLGDWLRQ DQG WHVWLQJ %\ FRPSDULQJ EHWZHHQ WKH PHDVXUHG WDUJHW DQG VLPXODWHG RXWSXW FRROLQJ ORDG IRU D
WHVWLQJ SHULRG RI WKUHH ZHHNV î WKH FRHIILFLHQW RI YDULDQFH &9 LV 7KH PRGHO VOLJKWO\ RYHUHVWLPDWHV WKH
RXWSXWVDWHOHYDWHGYDOXHVDQGXQGHUHVWLPDWHVDWUHODWLYHO\ORZHUYDOXHVGXULQJWKHWUDLQLQJSHULRG+RZHYHUWKHRXWSXWVDUH
VRPHZKDWVFDWWHUHGGXULQJWKHWUDLQLQJSHULRG

7UDLQLQJ2XWSXWVYV7DUJHW 9DOLGDWLRQ2XWSXWVYV7DUJHW

&9 &9

2XWSXWN:
2XWSXWN:

7UDLQLQJGDWDSRLQWV 9DOLGDWLRQGDWDSRLQWV
7DUJHW 2XWSXW 7DUJHW 2XWSXW

7DUJHWN: 7DUJHWN:
7HVWLQJ2XWSXWVYV7DUJHW

&9

2XWSXWN:

7HVWLQJGDWDSRLQWV
7DUJHW 2XWSXW

7DUJHWN:
)LJXUH 7HVWLQJRIWKH$11IDQPRGHO

7KH FKLOOHG ZDWHU DQG FRQGHQVHU ZDWHU VXSSO\ WHPSHUDWXUHV FROOHFWHG IURP WKH H[LVWLQJ V\VWHP DUH DOZD\V IL[HG DQG
WHVWLQJRIWKHFKLOOHUPRGHOLVQRWYDOLGRYHUDZLGHUUDQJHRIRSHUDWLRQ7KXVWKH$11FKLOOHUPRGHOLVHYDOXDWHGDJDLQVWWKH
(QHUJ\3OXV¶VHOHFWULFFKLOOHUPRGHOEDVHGRQFRQGHQVHUOHDYLQJWHPSHUDWXUHGHYHORSHGE\+\GHPDQHWDO+\GHPDQHWDO
7KH FKLOOHU SHUIRUPDQFH FXUYHV DUH JHQHUDWHG E\ ILWWLQJ PDQXIDFWXUHU¶V FDWDORJ GDWD 7KH FRROLQJ FRLO ORDG FKLOOHG
ZDWHUWHPSHUDWXUHFRQGHQVHUZDWHUWHPSHUDWXUHDUHWKHQWKHLQSXWVIRUFKLOOHUPRGHO7KHRXWSXWLVWKHFRPSUHVVRUSRZHU
)LJXUHVKRZVWKHWHVWLQJUHVXOWVRIWKHFKLOOHUPRGHO7KHFKLOOHUHQHUJ\XVHDVDIXQFWLRQRISDUWORDGUDWLR3/5FRROLQJ
ORDGWRUDWHGRQHLVLOOXVWUDWHGDWWZRGLIIHUHQWYDOXHVRIFKLOOHGZDWHUVXSSO\WHPSHUDWXUH R&DQG R&R)R)
DQGHQWHULQJFRQGHQVHUZDWHUWHPSHUDWXUHRIR& R)8QGHUWKHVHWZRFRQGLWLRQVDQGZLWKDQLQWHUYDORI3/5
RSHUDWLQJFRQGLWLRQVWKHDFFXUDF\RIWKHPRGHOLQWHUPRIWKHFRHIILFLHQWRIYDULDQFH&9LV7KHVHWHVWLQJUHVXOWV
VKRZ WKDW WKH PRGHOV FDSWXUH YHU\ ZHOO WKH V\VWHP SHUIRUPDQFH DQG FDQ EH XVHG IRU WKH FDOFXODWLRQV UHTXLUHG IRU WKH
RSWLPL]DWLRQSURFHVVRUDQ\RWKHUDSSOLFDWLRQV

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

7UDLQLQJ7DJHWYV2XWSXW 9DOLGDWLRQ7DUJHWYV2XWSXW

&9 &9

2XWSXWN:
2XWSXWN:

7UDLQLQJGDWDSRLQWV 9DOLGDWLRQGDWDSRLQWV
7DUJHW 2XWSXW 7DUJHW 2XWSXW

7DUJHWN: 7DUJHWN:

7HVWLQJ7DUJHWYV2XWSXW

&9

2XWSXWN:

7HVWLQJGDWDSRLQWV
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
7DUJHW 2XWSXW

7DUJHWN:
)LJXUH 7HVWLQJRIWKH$11FRROLQJFRLOPRGHO

$11&KLOOHGZDWHUVXSSO\WHPSHUDWXUH R&
(QHUJ\3OXV
V0RGHO&KLOOHGZDWHUVXSSO\WHPSHUDWXUH R&
$11&KLOOHGZDWHUVXSSO\WHPSHUDWXUH R&

(QHUJ\3OXV
V0RGHO&KLOOHGZDWHUVXSSO\WHPSHUDWXUH R&
&KLOOHUHQHUJ\GHPDQGN:

3/5U
)LJXUH 7HVWLQJRIWKH$11FKLOOHUPRGHO
6<67(0237,0,=$7,21
7KH RSWLPL]DWLRQ SURFHVV LQFOXGLQJ WKH $11 PRGHOV DV VKRZQ LQ )LJXUH LV HYDOXDWHG XVLQJ GDWD IURP DQ H[LVWLQJ
9$9V\VWHPVHUYLQJFODVVURRPVDQGRIILFHVZLWKDWRWDORI]RQHV'HWDLOHGLQIRUPDWLRQRQWKHV\VWHPFDQEHIRXQGLQWKH
UHIHUHQFH 1DVVLI 7KH RSWLPL]DWLRQ SURFHVV SUHGLFWV WKH V\VWHP SHUIRUPDQFH RYHU D SHULRG RI PLQ RSWLPL]DWLRQ
SHULRG'XULQJWKLVVKRUWRSWLPL]DWLRQSHULRGWKHORDGVDQGRXWGRRUDLUFRQGLWLRQVDUHDVVXPHGWREHFRQVWDQWDQGHVWLPDWHG
IURP WKH PHDVXUHG GDWD FROOHFWHG GXULQJ WKH SUHYLRXV SHULRG 7KH $11 PRGHOV DUH XVHG WR ILQG WKH HQHUJ\ XVH LQ HDFK
138
ASHRAE Transactions
FRPSRQHQW DQG WKHQ WKH WRWDO HQHUJ\ XVH LQ UHVSRQVH WR WKH FRQWUROOHU VHW SRLQWV DQG RSHUDWLQJ PRGHV 7KH LQSXWV DUH WKH
FRQWUROOHU VHW SRLQWV SUREOHP YDULDEOHV DQG WKH RXWSXW LV WKH HQHUJ\ XVH REMHFWLYH IXQFWLRQ 7KH JHQHWLF DOJRULWKP *$
VHQGVDVHWRILQGLYLGXDOVROXWLRQVFRQWDLQLQJWULDOFRQWUROOHUVHWSRLQWVDQGWKHPRGHOVWKHQHVWLPDWHWKHREMHFWLYHIXQFWLRQ
WRWDOHQHUJ\XVHDQGVHQGEDFNWRWKH*$WRHOLPLQDWHHYROYHDQGSDVVWKLVVROXWLRQWRWKHQH[WJHQHUDWLRQ7KLVSURFHVV
FRQWLQXHVXQWLOUHDFKRSWLPDORUQHDURSWLPDOVROXWLRQV7KHVXSSO\DLUWHPSHUDWXUHGXFWVWDWLFSUHVVXUHDQGRXWGRRUDLUIORZ
UDWH DUH RQO\ FRQVLGHUHG DQG WKH RSWLPL]DWLRQ LV GRQH RQO\ IRU WKH VXPPHU SHULRG ,Q WKH H[LVWLQJ V\VWHP WKH VXSSO\ DLU
WHPSHUDWXUHVHWSRLQWLVDXWRPDWLFDOO\UHVHWEDVHGRQWKHVXSSO\DLUIORZUDWHDQGRXWGRRUDLUGU\EXOEWHPSHUDWXUHQRFRQWURO
VWUDWHJ\LVDSSOLHGWRUHVHWWKHGXFWVWDWLFSUHVVXUHVHWSRLQWDQGLWLVDOZD\VFRQVWDQWDW3DLQ:*DQGRXWGRRUDLULV
FRQWUROOHGE\&2VHQVRUORFDWHGLQDUHWXUQDLUGXFW7KHZKROHV\VWHPRSWLPL]DWLRQILQGVWKHVROXWLRQWKDWSURGXFHVWKHOHDVW
WRWDOHQHUJ\XVH)LJXUHVKRZVWKHDFWXDODQGRSWLPDOWRWDOHQHUJ\XVHIRUIRXUGD\V7KHWRWDOHQHUJ\XVHLVHTXDOWRWKH
VXPRIIDQHOHFWULFUHKHDWDQGFKLOOHUSRZHUV,QWKHH[LVWLQJV\VWHPVHYHUDODLUKDQGOLQJXQLWVUHFHLYHFKLOOHGZDWHUIURP
RQHFRPPRQFKLOOHUHTXLSSHGZLWKDFRQVWDQWVSHHGFKLOOHGZDWHUSXPS$VRQO\RQHDLUKDQGOLQJXQLW$+8LVFRQVLGHUHGLQ
WKLV VWXG\ RQH GHGLFDWHG FKLOOHU LV DVVXPHG DQG WKH SRZHU LV VLPXODWHG EDVHG RQ (QHUJ\ 3OXV¶V HOHFWULF FKLOOHU PRGHO
+\GHPDQHWDOZLWKWKHUDWLQJFRROLQJFDSDFLW\DFFRUGLQJWRWKHLQYHVWLJDWHG$+8FDSDFLW\$VDUHVXOWE\DSSO\LQJ
WKHRSWLPL]DWLRQSURFHVVWKHWRWDOHQHUJ\VDYLQJLVIRUWKRVHIRXUGD\VDQGIRUWKUHHVXPPHUPRQWKV-XQH-XO\
DQG$XJXVW7KHVDYLQJFRXOGYDU\GHSHQGLQJRQWKHV\VWHPW\SHVEXLOGLQJW\SHDQGORFDWLRQVH[LVWLQJHQHUJ\HIILFLHQF\
RSSRUWXQLW\DQGFXUUHQWFRQWUROVWUDWHJLHV

2SWLPDO
$FWXDO

(QHUJ\XVHN:

'D\V
)LJXUH 2SWLPDODQGDFWXDOHQHUJ\XVH

&21&/86,21
$UWLILFLDOLQWHOOLJHQFHDSSURDFKHVDUHSURSRVHGWRXVHLQ+9$&SURFHVVDQGWRDGYDQFHWKH(0&66HOIWXQLQJ+9$&
FRPSRQHQW PRGHOV EDVHG RQ WKH DUWLILFLDO QHXUDO QHWZRUN ZHUH GHYHORSHG DQG YDOLGDWHG DJDLQVW GDWD FROOHFWHG IURP DQ
H[LVWLQJ+9$&V\VWHP7KHWHVWLQJUHVXOWVVKRZHGWKDWWKHPRGHOVH[KLELWJRRGDFFXUDF\DQGILWZHOOWKHLQSXWRXWSXWGDWD
$QLQILQLWHYDULHW\RIQHWZRUNDUFKLWHFWXUHVFDQEHXVHGIRUWKLVSXUSRVHEXWLQWKHLQWHUHVWRIFRQVHUYLQJFRPSXWHUWLPHWKH
VLPSOHVWVWUXFWXUHZLWKRQHKLGGHQOD\HUDQGWZHQW\QHXURQVLVXVHG7KHHUURUVRIWKHIDQFRROLQJFRLOFKLOOHUPRGHOVLQWHUP
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
RI FRHIILFLHQW RI YDULDQFH &9V DUH ZLWKLQ 7KH PRGHOV FRXOG EH LQFRUSRUDWHG LQWR WKH (0&6 WR SHUIRUP VHYHUDO
$6+5$(

LQWHOOLJHQW IXQFWLRQV LQFOXGLQJ HQHUJ\ PDQDJHPHQW DQG RSWLPDO FRQWURO 7R WKDW HQG D ZKROH V\VWHP RSWLPL]DWLRQ SURFHVV
EDVHGRQJHQHWLFDOJRULWKPZDVGHYHORSHGDQGWHVWHG7KHWHVWLQJUHVXOWVLQGLFDWHGWKDWWKHRSWLPL]DWLRQSURFHVVFDQSURYLGH
DWRWDOHQHUJ\ VDYLQJRI7KHVDYLQJFRXOG YDU\GHSHQGLQJRQWKH V\VWHPW\SHVEXLOGLQJW\SHDQGORFDWLRQVH[LVWLQJ
HQHUJ\HIILFLHQF\RSSRUWXQLW\DQGFXUUHQWFRQWUROVWUDWHJLHV
5()(5(1&(6
$KPHG 2 -: 0LWFKHOO DQG .OHLQ 6 $SSOLFDWLRQ RI *HQHUDO 5HJUHVVLRQ 1HXUDO 1HWZRUN *511 LQ +9$&
3URFHVV,GHQWLILFDWLRQDQG&RQWURO$6+5$(7UDQVDFWLRQV
$6+5$( $6+5$( +DQGERRN+9$& $SSOLFDWLRQV $PHULFDQ 6RFLHW\ RI +HDWLQJ 5HIULJHUDWLQJ DQG $LU
&RQGLWLRQLQJ(QJLQHHUV$WODQWD&KDSWHU
$6+5$( $6+5$( +DQGERRN)XQGDPHQWDOV $PHULFDQ 6RFLHW\ RI +HDWLQJ 5HIULJHUDWLQJ DQG $LU&RQGLWLRQLQJ
(QJLQHHUV$WODQWD&KDSWHU
$6+5$($6+5$(6WDQGDUG9HQWLODWLRQIRU$FFHSWDEOH,QGRRU$LU4XDOLW\$WODQWD$PHULFDQ6RFLHW\RI
+HDWLQJ5HIULJHUDWLQJDQG$LU&RQGLWLRQLQJ(QJLQHHUV,QF
%UDQGHPXHKO0-6*DEHODQG,$QGHUVHQ$7RRONLWIRU6HFRQGDU\+9$&6\VWHP(QHUJ\&DOFXODWLRQ$PHULFDQ
'HE.0XOWL2EMHFWLYH2SWLPL]DWLRQ8VLQJ(YROXWLRQDU\$OJRULWKPV1HZ<RUN-RKQ:LOH\ 6RQV
*ROGEHUJ'(*HQHWLF$OJRULWKPVLQ6HDUFK2SWLPL]DWLRQDQG0DFKLQH/HDUQLQJ5HDGLQJ$GGLVRQ:HVOH\
+DQL + (PSOR\LQJ &RPSXWDWLRQDO ,QWHOOLJHQFH WR *HQHUDWH 0RUH ,QWHOOLJHQW DQG (QHUJ\ (IILFLHQW /LYLQJ 6SDFHV
,QWHUQDWLRQDO-RXUQDORI$XWRPDWLRQDQG&RPSXWLQJ
+\GHPDQ01:HEE36UHHGKDUDQDQG6%ODQF'HYHORSPHQWDQG7HVWLQJRID5HIRUPXODWHG5HJUHVVLRQ%DVHG
(OHFWULF&KLOOHU0RGHO$6+5$(7UDQVDFWLRQV
/HH :< -0 +RXVH & 3DUN DQG *( .HOO\ )DXOW 'LDJQRVLV RI DQ $LU +DQGOLQJ 8QLW 8VLQJ $UWLILFLDO 1HXUDO
1HWZRUNV$6+$5(7UDQVDFWLRQV
0RVVROO\ 0 . *KDOL DQG 1 *KDGGDU 2SWLPDO FRQWURO VWUDWHJ\ IRU D PXOWL]RQH DLU FRQGLWLRQLQJ V\VWHP XVLQJ D
JHQHWLFDOJRULWKP(QHUJ\9RO1RSS
1DVVLI16.DMODQG56DERXULQ2SWLPL]DWLRQRI+9$&&RQWURO6\VWHP6WUDWHJ\8VLQJ7ZR2EMHFWLYH*HQHWLF
$OJRULWKP+9$& 55HVHDUFK
1DVVLI 1 2SWLPL]DWLRQ RI +9$& FRQWURO V\VWHP XVLQJ WZRREMHFWLYH JHQHWLF DOJRULWKP 3K' 7KHVLV (FROH GH
WHFKQRORJLHVXSHULHXU0RQWUHDO&DQDGD
6HHP-(8VLQJLQWHOOLJHQWGDWDDQDO\VLVWRGHWHFWDEQRUPDOHQHUJ\FRQVXPSWLRQLQEXLOGLQJV(QHUJ\DQG%XLOGLQJV
±
:DQJ 6 DQG ; -LQ 0RGHO%DVHG 2SWLPDO &RQWURO RI 9$9 $LU&RQGLWLRQLQJ 6\VWHP 8VLQJ *HQHWLF $OJRULWKP
%XLOGLQJDQG(QYLURQPHQW
;X;+6::DQJ =:6XQDQG);LDR$0RGHOEDVHG2SWLPDO9HQWLODWLRQ&RQWURO6WUDWHJ\RI0XOWL]RQH
9$9$LUFRQGLWLRQLQJ6\VWHPV8VLQJ*HQHWLF$OJRULWKP$SSOLHG7KHUPDO(QJLQHHULQJ9±
=KHQJ*5DQG0=DKHHU8GGLQ2SWLPL]DWLRQRIWKHUPDOSURFHVVHVLQDYDULDEOHDLUYROXPH+9$&V\VWHP(QHUJ\

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
140
ASHRAE Transactions
SA-12-C018
Beyond the Shoebox: Thermal Zoning

Approaches for Complex Building Shapes
Lillian Smith, AIA

Member ASHRAE
ABSTRACT
Current architectural trends as well as advanced building information modeling (BIM) software facilitate the design and construction of building
forms that go far beyond the typical shoebox models commonly used for energy simulation. ASHRAE 90.1 Appendix G provides two-
dimensional guidelines for how to model thermal zones in buildings where HVAC systems and zones have not yet been designed. What is the best
way to zone unusual three-dimensional shapes that present issues beyond what is covered in the ASHRAE guidelines? This paper will examine
this question and determine the impact on energy model outputs by studying 3 different unusually shaped buildings of 2 different building use types
in 4 different climate zones. Several zoning approaches beyond the ASHRAE guidelines will be simulated and compared using DOE-2
simulation engine and TMY2 climate data.
INTRODUCTION
During the conceptual phase of building design, decisions such as basic form and orientation can have a significant
effect on energy use, yet whole building energy analysis is rarely used to inform this design process. Typically, these studies
have not been done because the tools and methods are too difficult and time consuming to use at early stages and models
aren’t built correctly for energy simulations and rapid design iteration. The 2008 Autodesk/AIA Green Index reports that
architects believe that using the design process to reduce building energy consumption includes an increased reliance on
design software. Leveraging existing software and making it easy to learn and use is important because nearly three-fourths
of architects (72%) are concerned that clients are not willing to pay the added first costs of green designs (Autodesk/AIA
2008). McGraw/Hill reports that while fewer than 20% of firms are simulating energy performance, in two years 80% of
firms see it as very important and want to simulate whole building energy use. (McGraw/Hill 2010). Underscoring the
importance of early energy use studies, in September 2011, the USGBC introduced a new LEED credit encouraging teams
to conduct analyses, including energy load reduction, so that they can understand key issues before design decisions are
made (USGBC 2011).
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
In most cases, simple extruded representations of building geometry are used for energy modeling. Energy modeling
programs, such as eQUEST, provide templates for zoning many different floor plan configurations; however, changing this
configuration from floor to floor is difficult and time consuming. Advances in building modeling software, as well as
design trends, are making more complex shaped buildings easier to design and build. Advances in energy modeling
software are also making complex geometries easier to analyze for energy use. Recent advances in software provide
Lillian Smith is an Architect and Principal User Experience Designer at Autodesk, Inc. in Waltham MA.
$6+5$(

functionality to automatically create an energy model from a conceptual model. (Smith, Bernhardt, and Jezyk 2011).
Additionally, new algorithms can automatically divide more complex geometry into zones specific to the geometry at each
floor based on ASHRAE 90.1 Appendix G modeling requirements of thermal blocks (Autodesk® Project Vasari 2011).
Automatic zoning is important for conceptual stage energy modeling because design decisions about layout have not been
made yet and the architects and engineers do not necessarily know yet how a building will be zoned. Providing a quick and
automatic way to turn an architect’s geometry into an appropriately zoned energy model without additional modeling is
critical for doing early stage energy simulations. The goal of this paper is to demonstrate the importance of dividing a
conceptual stage building into zones based on ASHRAE guidelines and to explore the sensitivity of choices required for
automatic zoning beyond ASHRAE recommendations that were discovered in the development of these automated
approaches.
METHODOLOGY
To investigate the various impacts of zoning approaches on energy model results, 3 different building geometries of 2
building use types were simulated using 4 different climate zones and building envelopes. The goal was to provide a
representative range of building parameters for the study. ASHRAE guidelines for envelope settings and zoning
configurations were used to simulate and compare the three geometrical options using a DOE-2 simulation engine and
TMY2 climate data. Autodesk® Project Vasari software was used to generate the energy model geometry because of its
capability to automatically divide complex geometries into energy models.
Building Geometries
Three different building geometries were chosen for this study, as shown in Figure 1. Each option has a different
floor plan at each level due to the shape of the building. Option A is a deep plan configuration with an angled front face.
Option B is a tall narrow building with a bowed front and Option C is a long thin building with a large sloped roof.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 1 Building geometries and their associated square footages.
Building Types
Two different building types were selected for the simulations. Each of the three geometries was simulated as an
office and as a healthcare facility. Assumptions for these building types are based on ASHRAE guidelines. The biggest
difference in these 2 building types is the people density with the Healthcare facility having a much greater density – 10
people/100sq M (1076.391 ft²) vs 3.5 people/100 sq. M. (1076.391 ft²) for the office. The Healthcare facility is also
142
ASHRAE Transactions
operated for more hours than the office building.
Climates
Four different climates were chosen to represent a variety of different conditions as shown in Table 1.
Table 1 Climates
City Region Zone HVAC
Minneapolis Northern 6A Cold/Humid Mostly Heating
Baltimore North/Central 4A Mixed/Humid Heating & Cooling
Atlanta South/Central 3A Mixed/Humid Heating & Cooling
Phoenix South 2B Hot/Dry Mostly Cooling
Weather data
TMY2 (Typical Meteorological Year) data was used for weather data to represent a range of weather conditions that
was also consistent with long term averages for the location. Studies have shown that TMY2 data is appropriate for this
kind of study (Crawley and Huang 1997).
Building Envelope
Different envelopes (Table 2 and Table 3) were specified for each location, based on ASHRAE 90.1 2007 Table 5.5-1-
6 Building Envelope Requirements for Climate Zones, and the choices offered by Project Vasari. Minneapolis and
Baltimore had the same settings as did Atlanta and Phoenix. All climate zones used the same glazing values.
Table 2 Building Envelop Settings Per Location

Location Surface R-Value Unit Density Heat Capacity
(ft²-hr ºF/Btu) (lbm/ ft²) (Btu/( ft²•°F)
Exterior Wall 17 21 4
Roof 22 10 3
Minneapolis
Floor 29 4 1
Slab on Grade 16 123 25
Roof 22 10 3
Baltimore
Floor 29 4 1
Roof 22 10 3
Atlanta
Floor 21 4 1
Roof 22 10 3
Phoenix
Floor 21 4 1
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

Table 3 Glazing Settings
Location Surface U-Value Solar Heat Gain Visible
Btu/(ft²·°F·h) Coefficient Transmittance
All Glazing 0.56 0.69 0.78
Zoning Configurations
Eight zoning configurations were chosen for testing (Table 4). The first three configurations purposely do not follow
ASHRAE recommendations in order to illustrate the importance of the zoning guidelines. Zoning configuration 1 is
simply one zone per floor. Zoning configuration 2 is one zone per floor with a 15’ deep perimeter zone, but no further
zone divisions. Zoning configuration 3 takes the centroid of the geometry and creates a cut based on 4 cardinal directions
after dividing the building into floors and a 15' offset perimeter zone. This is a very simple way to algorithmically cut many
different shapes, but can create zones with multiple exposures. Zoning configuration 4 and 5 explore different perimeter
zone depths of 10’ and 20’ instead of the ASHRAE recommended 15’. Zoning configuration 6 explores questions that
arise with geometries that are not extruded; for example, walls that splay out or a large sloped roof as in Geometry B. This
zoning configuration offsets the perimeter depth at the floor, rather than half way between levels which is the Project
Vasari default behavior. Zoning configuration 7 simulates examples with no specified maximum zone length. ASHRAE
does not specify a maximum zone length; however, common practice is to separate zones that are greater than 70’ and this
is the Vasari default. The final configuration is the comparison case with the ASHRAE standard 15’ perimeter depth and
Vasari default setting to cut the plan at 5’ vertically from the floor and impose a 70’ maximum zone length.
Table 4 Zoning Configurations

Zoning Configuration Geometry A Geometry B Geometry C
1. One zone per floor
2. 15' perimeter depth that

are not divided into further
zones
3. Cardinal direction zoning
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
144
ASHRAE Transactions
4. 10’ perimeter depth
5. 20’ perimeter depth
6. Level cut at floor instead

of mid-way between levels
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
7. 15' perimeter depth, no
maximum zone length
8. 15' perimeter depth,

perimeter divided per
ASHRAE 90.1. Plan cut at
5’ above floor with a
maximum 70’ zone length.
This is the comparison
case.
Simulation Engine
Simulations were run using Autodesk® Green Building Studio® software and DOE-2.2-44e4.
RESULTS
Simulation results are shown in Figure 2. Percentage difference of total fuel and electric use are compared to zoning
configuration 8. Fuel and electric use values were used because they are commonly understood values, reflect where the
differences mattered most, and remove any further complications involving utility tariffs. Percentage differences of more
than 3% are considered significant because changes of +/-2% can mean an entire LEED energy point (USGBC 2009).
$6+5$(

Figure 2 Results shown in percentage difference from the base case (zoning configuration 8) for fuel and electric use over all climates,
geometries, and building types per zoning configuration. Solid bars show the average percentage difference and background
grey lines plot each run.
Zoning configurations 1, 2, and 3, which do not follow ASHRAE guidelines, display up to 40% difference from the
comparison case. Configuration 4 (10’ divided perimeter zone) displays a consistent overestimation of fuel use averaging
over 8%. Configuration 5 (20’ divided perimeter zone) shows a consistent underestimation of fuel use averaging -5%.
Results for electric use average differences in both cases are negligible. The differences for configuration 6, level cut at
floor instead of mid-way between levels, show negligible average differences for electric use, however, in these cases, the
results could still differ by as much as 3% to -4.5% for fuel use. The last case, configuration 7, which tested no maximum
zone length, showed a negligible average difference in both fuel and electric use.
Looking in depth at percentage difference in fuel and electric use per building geometry (Figure 3), we can see that the
results from the different configurations are consistently over or under estimated use, except for Geometry A configuration
3 (Table 4 Zoning Configurations), where the deep plan of this geometry option reacts strongly to cardinal direction
zoning due to zones wrapping around 3 orientations, and overestimates fuel use by 25% and electric use by 7%.
Configuration 6 (level cut at floor instead of mid-way between levels) shows negligible differences, except in the angled
geometry case C where the differences range from -5% to 3%.
146
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
ASHRAE Transactions
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 3 Percentage difference in fuel and electric use per zoning configuration and geometry option. Solid bars show average percentage
differences as compared to configuration 8 and gray background lines plot the percentage differences of each run.
CONCLUSION
The results demonstrate that zoning configurations 1-3, which do not follow ASHRAE guidelines, can significantly
affect energy simulations and should not be used because they primarily underestimate energy use which is problematic for
early studies. Overestimation is arguably preferred so that designers won’t be encouraged to follow designs that use more
energy. Configuration 3 is especially problematic in that it can both under and overestimate energy use and due to the fact
that in deep plan buildings zones can actually wrap around multiple orientations. Sensitivities beyond ASHRAE
recommendations, as seen in results from configurations 4-7, are also significant because changes in 2% can mean an entire
LEED energy point. Configuration 4 (10’ divided perimeter zones) consistently overestimates use by an average of over
8% and configuration 5 (20’ divided perimeter zones) underestimates use by an average of -5%. Comparisons of these
configurations suggest that perhaps a better default for perimeter offset would be closer to 16’. Changing the location of
the plan cut where the perimeter offset is measured (configuration 6) are negligible in this study, but can make a significant
difference especially in cases like Geometry C where results range from -5% to 3% due to the significant sloping geometry.
$6+5$(

This suggests that perhaps the level of the plan cut should be a user controllable option. For configuration 7, no maximum
zone length, all runs are within a 2% difference suggesting that this setting may be unnecessarily creating more zones and
complexity than required. This study presents a limited set of options and climates due to feasibility of conducting the
simulation runs. Future studies could extend the selected options to further validate these conclusions.
Overall, these results demonstrate the importance of reasonable zoning in conceptual models. Using software to
automatically cut geometric forms with a different floor plan at each level into valid energy models can be a huge benefit for
studying form and orientation at the earliest stage of design when changes of this nature can be made and will have a big
effect on energy use. Automatic zoning algorithms that can be trusted to return reliable early stage energy simulation results
can enable rapid design iteration, which allows for more studies and options to be considered and intelligently analyzed.
ACKNOWLEDGMENTS
Special thanks to John Kennedy and Ian Molloy, Energy Analysis Engineers for help in specifying simulation runs and
analyzing the results.
REFERENCES
U.S. Green Building Council 2009. LEED-BD+C Green Building Rating System for New Construction & Major
Renovations v.3.0. http://www.usgbc.org/DisplayPage.aspx?CMSPageID=1970.
U.S. Green Building Council, 2011 LEED Pilot Credit 48: IP – Discovery: Analysis to Support Integrative Process.
http://www.usgbc.org/ShowFile.aspx?DocumentID=10101.
Autodesk, Inc. and AIA, 2008. The 2008 Autodesk/AIA Green Index, 4
McGraw Hill Construction 2010 Green BIM How Building Information Modeling is Contributing to Green Design and
Construction. 6,26.
Smith, Lillian, Bernhardt, Kyle and Jezyk, Matthew, 2011, Automated Energy Model Creation for Conceptual Design.
SimAUD, 2011. http://www.simaud.org/proceedings/.
Crawley, Drury B. and Huang, Y. Joe, Does it Matter Which Weather Data You Use in Energy Simulations? Building
Energy Simulation User News, Spring 1997 – Vol. 18, No.1.
Autodesk, Inc. Autodesk Project Vasari Software http://www.projectvasari.com
Autodesk, Inc. Autodesk Green Building Studio Software. https://gbs.autodesk.com/
ASHRAE. 2007. ASHRAE 90.1-2007. Energy Standard for Buildings Except Low-Rise Residential Buildings. American
Society of Heating, Refrigerating, and Air-Conditioning Engineers, Inc.
ASHRAE 2007. ASHRAE 62.1-2007. Ventilation for Acceptable Indoor Air Quality. American Society of Heating,
Refrigerating, and Air-Conditioning Engineers, Inc.
DOE-2. 2006. Lawrence Berkeley National Laboratory. University of California, Berkeley.
DOE-2. 2007. eQUEST the Quick Energy Simulation Tool. http://doe2.com/equest.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
148
ASHRAE Transactions
SA-12-C019
A Simulation and Experimental Study of

the Impact of Passive and Active Façade
Systems on the Energy performance of
Building Perimeter Zones
Ying-Chieh Chan Athanasios Tzempelikos, PhD

ABSTRACT
2SWLPDOFRQILJXUDWLRQRIIDFDGHVDFURVVGLIIHUHQWEXLOGLQJVDQGORFDWLRQVLQYROYHVDFRPSOH[GHFLVLRQPDNLQJSURFHVV7KLVSDSHU
SUHVHQWV D PHWKRG IRU DVVHVVLQJ WKH LQWHJUDWHG HQHUJ\ SHUIRUPDQFH RI SDVVLYH DQG DFWLYH PXOWLVHFWLRQ IDoDGH V\VWHPV FRPELQHG ZLWK
OLJKWLQJDQGWKHUPDOFRQWUROVRISHULPHWHUEXLOGLQJ]RQHVXVLQJDQRSHQVRXUFHODQJXDJH$WKHUPDOQHWZRUNDSSURDFKLVXVHGWRSUHGLFW
LQGRRUWKHUPDOHQYLURQPHQWDOFRQGLWLRQVDQGDQQXDOHQHUJ\FRQVXPSWLRQRISHULPHWHU]RQHVHTXLSSHGZLWKFRPELQDWLRQVRISDVVLYHDQG
DFWLYH IDoDGH V\VWHPV VXFK DVVHOHFWLYH JOD]LQJV WUDQVOXFHQWSDQHOV PRWRUL]HG VKDGHV DQG EOLQGV LQ FRQMXQFWLRQ ZLWK GD\OLJKWOLQNHG
OLJKWLQJFRQWUROV7KHPRGHOXVHVDQLVRWURSLFVN\PRGHOVIRUDQDFFXUDWHSUHGLFWLRQRIVRODUJDLQVYDULDEOHDQJXODUJOD]LQJSURSHUWLHVDQG
QRQOLQHDULQWHULRUDQGH[WHULRUFRQYHFWLRQDQGUDGLDWLRQKHDWWUDQVIHUFRHIILFLHQWVWRJHWKHUZLWKWUDQVLHQWLQWHUQDOJDLQVREWDLQHGIURP
WUDQVLHQW OLJKWLQJ VLPXODWLRQ $ VLPSOH SULYDWH RIILFH ZLWK RQH H[WHULRU ZLQGRZ LV XVHG DV DQ LQLWLDO FDVH VWXG\ ([SHULPHQWDO
PHDVXUHPHQWVLQIXOOVFDOHODERUDWRULHVZLWKDQGZLWKRXWVKDGLQJZHUHXVHGWRFRPSDUHWRVLPXODWLRQUHVXOWVLQRUGHUWRYDOLGDWHWKH
GHYHORSHGPRGHOV7KHQWKHPRGHOLVH[SDQGHGDQGJHQHUDOL]HGWRLQFOXGHWKHLPSDFWRIPXOWLVHFWLRQDOIDFDGHVLQRUGHUWRHYDOXDWHWKH
FRPELQHGLPSDFWRIJOD]LQJDQGVKDGLQJFRQILJXUDWLRQVRQHQHUJ\GHPDQGDQGWKHUPDOFRPIRUW7KLVVWXG\SUHVHQWVDEDVLVIRUDIXWXUH
PRGHOLQJWRROWKDWFDQEHXVHGWRRSWLPL]HWKHGHVLJQRIPXOWLIXQFWLRQDOG\QDPLFIDoDGHV\VWHPV
INTRODUCTION
)DoDGHGHVLJQLVDNH\SRLQWIRUDFKLHYLQJKLJKSHUIRUPDQFHEXLOGLQJV,WKDVDVLJQLILFDQWLPSDFWRQKHDWLQJ
FRROLQJDQGOLJKWLQJORDGVDVZHOODVRQWKHUPDODQGYLVXDOFRPIRUWLQSHULPHWHU]RQHVRIFRPPHUFLDOEXLOGLQJV
6RODU HQHUJ\ XWLOL]DWLRQ FDQ UHGXFH KHDWLQJ ORDG LQ ZLQWHU DQG DW QLJKW UHGXFH OLJKWLQJ HQHUJ\ FRQVXPSWLRQ
7]HPSHOLNRV DQG $WKLHQLWLV DQG LPSURYH OLJKWLQJ TXDOLW\ 6HONRZLW] DQG /HH EXW DOVR KDV VRPH
GUDZEDFNVVXFKDVLQFUHDVLQJSHDNFRROLQJORDGFRROLQJHQHUJ\FRQVXPSWLRQDQGWKHUPDOGLVFRPIRUWGXHWRKLJK
UDGLDQWWHPSHUDWXUHV$WPDFDHWDO
'HVLJQ RI SDVVLYH IDoDGH V\VWHPV LQFOXGHV WKH VHOHFWLRQ RI W\SHV VL]HV PDWHULDOV DQG FRQILJXUDWLRQV IRU
JOD]LQJVKDGLQJLQVXODWLRQDQGWKHUPDOPDVVV\VWHPVLQWKHSHULPHWHU]RQH$FWLYHV\VWHPVUHIHUWRWKHGHVLJQRI
FRQWURO DOJRULWKPV DQG WKH FRQWURO RSHUDWLRQV WKDW GLUHFWO\ LPSDFW WKH IXQFWLRQ SRVLWLRQ SHUIRUPDQFH DQG WKH
SK\VLFDOSURSHUWLHVRIIDFDGHVWKDWDIIHFWWKHZKROHEXLOGLQJHQHUJ\FRQVXPSWLRQ6XFKDFWLYHG\QDPLFV\VWHPV
LQFOXGHPRWRUL]HGVKDGLQJJOD]LQJSURGXFWVZLWKFRQWUROODEOHSURSHUWLHVYDULDEOHVHWSRLQWVFRQWUROOHGOLJKWLQJHWF
Ying-Chieh Chan is a Ph.D student and Athanasios Tzempelikos is an Assistant Professor of Architectural Engineering in the School
of Civil Engineering, Purdue University, West Lafayette, Indiana, USA.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

WKDW KDYH WKH SRWHQWLDO RI VDYLQJ HQHUJ\ RU LPSURYLQJ FRPIRUW IRU WKH RFFXSDQWV YDQ 0RHVHNH HW DO
-RQVVRQ DQG 5RRV '\QDPLF IDFDGHV ZLWK DSSURSULDWH DGMXVWPHQW WR FKDQJLQJ RXWVLGH FRQGLWLRQV DOORZ
IOH[LELOLW\ LQ WKH LQWHJUDWLRQ RI GHVLJQ FRQVLGHUDWLRQV 7KLV FRQFHSW PD\ OHDG WR D EDODQFH EHWZHHQ SRVLWLYH DQG
QHJDWLYH LPSDFWV RI VRODU HQHUJ\ XWLOL]DWLRQ LQ SHULPHWHU ]RQHV DQG LPSURYH HQHUJ\ SHUIRUPDQFH DQG FRPIRUW
FRQGLWLRQVIRUWKHRFFXSDQWV7KHRSWLPDOVROXWLRQIRUHDFKEXLOGLQJZRXOGEHVSHFLILFWRLWVXQLTXHFKDUDFWHULVWLFV
VXFK DV W\SH VL]H SURSHUWLHV FOLPDWH DQG RULHQWDWLRQ $ IOH[LEOH DQG DFFXUDWH PRGHO WKDW FDQ XWLOL]H LQWHJUDWHG
V\VWHP RSWLRQV DQG GLIIHUHQW FRQWURO DOJRULWKPV LV QHHGHG WR SUHGLFW WKH HQHUJ\ SHUIRUPDQFH DQG RFFXSDQW
FRPIRUW )UDQ]HWWL HW DO 7KLV LQGLFDWHV WKH LPSRUWDQFH RI LQWHJUDWLRQ RI OLJKWLQJ DQG WKHUPDO PRGHOV LQ
RUGHU WR JHW D FRPSUHKHQVLYH VROXWLRQ /HH HW DO +RZHYHU VWDWLF DQG G\QDPLF IDoDGH V\VWHPV DUH RIWHQ
HQJDJHG LQ WKH FRQIOLFW EHWZHHQ OLJKWLQJ DQG DLUFRQGLWLRQLQJ VDYLQJV $ W\SLFDO H[DPSOH LV DXWRPDWHG VKDGLQJ
DOORZLQJPRUHOLJKWZRXOGUHGXFHOLJKWLQJFRQVXPSWLRQZKLOHDWWKHVDPHWLPHLQFUHDVLQJSHDNFRROLQJGHPDQG
DQGHQHUJ\FRQVXPSWLRQ$QDOWHUQDWLYHVROXWLRQWRWKHDERYHSUREOHPLVWKHGHYHORSPHQWRIPXOWLVHFWLRQDODQG
PXOWLIXQFWLRQDO IDFDGHV ZLWK HPEHGGHG LQWHOOLJHQW FRPSRQHQWV 7]HPSHOLNRV HW DO 7KHVH IDoDGHV
FRQVLVWRI GLIIHUHQW V\VWHPVDORQJ WKHLUKHLJKWHJD ORZHU ÝLVLRQµ SDUWSURWHFWHG IURP WKH VXQ DQGDWRS SDUW
SULPDULO\XVHGIRUGD\OLJKWLQJWKDWFDQEHFRQWUROOHGLQGHSHQGHQWO\LQRUGHUWRRSWLPL]HLQGRRUFRQGLWLRQV
([SHULPHQWDO PHDVXUHPHQWV DUH DOZD\V LPSRUWDQW LQ RUGHU WR YDOLGDWH DQG ILQHWXQH VSHFLILF PRGHOV DQG
DOJRULWKPV &ODUNH $OWKRXJK H[SHULPHQWV DUH WLPHFRQVXPLQJ DQG H[SHQVLYH FRPSOH[ FDVHV FDQ EH
SHUIRUPHGDQGDQDO\]HGHDVLHUWKDQZLWKDQDO\WLFDOYDOLGDWLRQ0DQ]HWDO7KHREMHFWLYHVRIWKLVSDSHUDUH
WR GHYHORS D G\QDPLF WUDQVLHQW WKHUPDO PRGHO LQ FRQMXQFWLRQ ZLWK D OLJKWLQJ PRGHO WR HYDOXDWH WKH HQHUJ\
SHUIRUPDQFH RI GLIIHUHQW SDVVLYH DFWLYH DQG PXOWLIXQFWLRQDO IDoDGH V\VWHPV DQG FRPSDUH WKH UHVXOWV ZLWK
H[SHULPHQWDO PHDVXUHPHQWV DQG WR VHW WKH EDVLV IRU EXLOGLQJ D WRRO WKDW FRXOG EH XVHG IRU RSWLPL]HG IDoDGH
GHVLJQXVLQJLQQRYDWLYHFRPSRQHQWVDQGFRQWUROVDWWKHHDUO\VWDJH
INTEGRATED THERMAL AND DAYLIGHTING SIMULATION

7KHELJFKDOOHQJHRILQWHJUDWLQJDGD\OLJKWLQJDQGWKHUPDOPRGHOVLVWRGHWHUPLQHG\QDPLFOLQNLQJSDUDPHWHUV
ZKLFK LQIOXHQFH ERWK WKH GD\OLJKWLQJ PRGHO DQG WKH WKHUPDO PRGHO 7]HPSHOLNRV HW DO DQG WKH
IHHGEDFN ORRS EHWZHHQ WZR PRGHOV 7KLV PHDQV WKDW WKH RXWSXW RI WKH GD\OLJKWLQJ PRGHO ZLOO EH LQSXW WR WKH
WKHUPDO PRGHO DQG YLFH YHUVD LQ D VHTXHQWLDO RU FRXSOHG VLPXODWLRQ DSSURDFK ,Q DQ\ FDVH WKH WZR PRGHOV DUH
LQWHUGHSHQGHQW)LJXUH
Baseline Building Description

7KHEDVHOLQHEXLOGLQJSULYDWHRIILFHLVORFDWHGLQ:HVW/DID\HWWH,QGLDQDDQGFRQVLVWVRIDVLQJOHURRPZLWK
ODUJHZLQGRZVIDFLQJVRXWK7KHWRWDOIORRUDUHDLVP PE\PDQGWKHFHLOLQJKHLJKWLVP:DOOVIORRU
DQGFHLOLQJDUHFRQVWUXFWHGZLWKKLJKUHVLVWDQFHPDWHULDOV5DQG5DQGZLWKWKHUPDOPDVVFRQFUHWHVODEV
LQERWKIORRUDQGFHLOLQJ7KHWRSSDUWRIWKHIDFDGHPKLJKLVGRXEOHJOD]HGORZHJODVV8YDOXH
:P& DQG WKH ERWWRP SDUW P KLJK LV GRXEOH JOD]HG ZLQGRZV ZLWK D VHOHFWLYH ORZH FRDWLQJ 8YDOXH
:P& 7KH LQILOWUDWLRQ UDWH LV DFKKU ZKLFK YDULHV EDVHGRQWKH URRP FRQVWUXFWLRQGHWDLOVHJ
RSHQLQJVDQGRXWVLGHFRQGLWLRQVHJZLQGVSHHGDQGZLQGGLUHFWLRQ7KHURRPLVHTXLSSHGZLWKDFRQWUROODEOH
OLJKWLQJV\VWHPDQGDXWRPDWHGLQWHULRUUROOHUVKDGHV
Solar Radiation Model

7\SLFDO0HWHRURORJLFDO<HDU70<GDWDLVLQWHUSRODWHGWRUHSUHVHQWDW\SLFDOZHDWKHUSDWWHUQRIWKLVFLW\DV
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
WKHLQSXWYDOXHIRUERWKWKHUPDODQGGD\OLJKWLQJPRGHOV)RUWKHGLUHFWLOOXPLQDQFHDQGVRODUUDGLDWLRQWKH\FDQEH
FDOFXODWHGGLUHFWO\IURPEHDPQRUPDOVRODUUDGLDWLRQVXUIDFHWLOWDQJOHRULHQWDWLRQDQGVRODUDQJOHZKLFKYDU\ZLWK
WLPH )RU WKH GLIIXVH SDUW WKH 3HUH] HW DO PRGHO LV XVHG ZKLFK SURYLGHV D PHWKRG WR FDOFXODWH VRODU

150
ASHRAE Transactions
UDGLDWLRQRQWLOWHGVXUIDFHV

>ŝŐŚƚŝŶŐ
ƵŝůĚŝŶŐĂƚĂ
WĂƌĂŵĞƚĞƌƐ
>ŽĐĂƚŝŽŶ͕ ŽƌŝĞŶƚĂƚŝŽŶ͕ ůŝŵĂƚĞĂƚĂ
^ƉĂĐĞůĂǇŽƵƚ ͕ ƐƵƌĨĂĐĞ
ŐĞŽŵĞƚƌǇ͕ ĐŽŶƐƚƌƵĐƚŝŽŶ͕ dDzϯĂŶĚƉƌŽĐĞƐƐĞĚĚĂƚĂ
ƉƌŽƉĞƌƚŝĞƐ͕ǁŝŶĚŽǁůŽĐĂƚŝŽŶ͕
ƐĐŚĞĚƵůĞƐ͕
͕
ůŝŐŚƚŝŶŐƐƉĞĐƐ͕ĂŶĚĐŽŶƚƌŽůƐ
ĂŶĚŝŶƚĞƌŶĂůŐĂŝŶƐ
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
WĂƐƐŝǀĞ&ĂĕĂĚĞ^ǇƐƚĞŵƐ
'ůĂǌŝŶŐƚǇƉĞĂŶĚƉƌŽƉĞƌƚŝĞƐ
tŝŶĚŽǁƐŝǌĞ
&ŝǆĞĚƐŚĂĚŝŶŐƉƌŽƉĞƌƚŝĞƐ ĂǇůŝŐŚƚŝŶŐ^ŝŵƵůĂƚŝŽŶ
dŚĞƌŵĂů^ŝŵƵůĂƚŝŽŶ
ŶŝƐŽƚƌŽƉŝĐƐŬǇŵŽĚĞůƐ
&ŝŶŝƚĞĚŝĨĨĞƌĞŶĐĞƚŚĞƌŵĂů
ĂŶĚƌĂĚŝŽƐŝƚǇͬƌĂǇͲƚƌĂĐŝŶŐ
ŶĞƚǁŽƌŬĂŶĚŝŵƉůŝĐŝƚ
ĐƚŝǀĞ&ĂĕĂĚĞ^ǇƐƚĞŵƐ ŵŽĚĞůƐĨŽƌŝŶƚĞƌŝŽƌǁŽƌŬ
ƐŽůƵƚŝŽŶ DŽƚŽƌŝǌĞĚƐŚĂĚŝŶŐƉƌŽƉĞƌƚŝĞƐĂŶĚĐŽŶƚƌŽů
>ŝŐŚƚŝŶŐĂŶĚĚĂǇůŝŐŚƚŝŶŐĐŽŶƚƌŽůƐ
ƉůĂŶĞŝůůƵŵŝŶĂŶĐĞ
KĐĐƵƉĂŶĐǇƐĞƚƚŝŶŐƐ
ŽŶƚƌŽůůĂďůĞŐůĂǌŝŶŐƐ
sĂƌŝĂďůĞƐĞƚƉŽŝŶƚƐ
DƵůƚŝͲƐĞĐƚŝŽŶĂůĂŶĚŵƵůƚŝͲĨƵŶĐƚŝŽŶĂů
ĚǇŶĂŵŝĐĨĂĕĂĚĞƐǇƐƚĞŵƐ
/ŶƚĞƌŝŽƌƚĞŵƉĞƌĂƚƵƌĞƐ
tŽƌŬƉůĂŶĞŝůůƵŵŝŶĂŶĐĞ
dŚĞƌŵĂůůŽĂĚƐ
ĂǇůŝŐŚƚĂƵƚŽŶŽŵǇ
ͬŶĞƌŐǇĐŽŶƐƵŵƉƚŝŽŶ
'ůĂƌĞŝŶĚŝĐĞƐ
WĞĂŬĚĞŵĂŶĚ
>ŝŐŚƚŝŶŐĞŶĞƌŐǇƵƐĞ

)LJXUH )ORZFKDUWRILQWHJUDWHGWKHUPDOGD\OLJKWLQJVLPXODWLRQZLWKG\QDPLFIDoDGHV\VWHPV
Absorbed Solar Gains

6RODUUDGLDWLRQWKDWVWULNHVWKHH[WHULRUZDOOLVDEVRUEHGDQGUHIOHFWHGWKHSDUWWKDWLVLQFLGHQWRQWKHZLQGRZV
LV DEVRUEHG UHIOHFWHG DQG WUDQVPLWWHG LQWR WKH URRP 7KH DQJOHGHSHQGHQW JOD]LQJ SURSHUWLHV WUDQVPLWWDQFH
UHIOHFWDQFHDQGDEVRUSWDQFHDUHREWDLQHGIURP:,1'2:DQGPDWFKHGZLWKH[SHULPHQWDOPHDVXUHPHQWV7KH
WUDQVPLWWDQFHRIUROOHUVKDGHVGHSHQGVRQWKHIDEULFPDWHULDOSURSHUWLHVIRUEOLQGVDQGOLJKWVKHOYHVLWGHSHQGVRQ
WKHSURILOHDQJOHVODWRUVKHOIWLOWDQJOHDQGVODWFRQILJXUDWLRQDQGSURSHUWLHV:KHQWKHUROOHUVKDGHVFORVHWKH
LQWHUUHIOHFWLRQVEHWZHHQWKHJOD]LQJDQGWKHIDEULFDUHWDNHQLQWRFRXQWZKHQFDOFXODWLQJWKHWUDQVPLWWDQFHRIWKH
HQWLUHIHQHVWUDWLRQV\VWHP
τ ƐŚĂĚĞƐ ⋅ ρ ƐŚĂĚĞƐ ⋅ ρĚŝĨ −ŐůĂƐƐ
τ Ěŝƌ − ƐǇƐƚĞŵ = τ Ěŝƌ −ŐůĂƐƐ ⋅τ ƐŚĂĚĞƐ + τ Ěŝƌ −ŐůĂƐƐ ⋅
ϭ − ρ ƐŚĂĚĞƐ ⋅ ρĚŝĨ −ŐůĂƐƐ
ZKHUH τ Ěŝƌ − ƐǇƐƚĞŵ LVWKHGLUHFWWUDQVPLWWDQFHRIWKHIDoDGHV\VWHP τ Ěŝƌ − ŐůĂƐƐ LVWKHGLUHFWWUDQVPLWWDQFHRIWKHJODVV
τ ƐŚĂĚĞƐ LVWKHWUDQVPLWWDQFHRIWKHUROOHUVKDGHV ρ ƐŚĂĚĞƐ LVWKHVKDGHUHIOHFWDQFHDQG ρĚŝĨ −ŐůĂƐƐ LVWKHGLIIXVHJODVV
UHIOHFWDQFH ,Q WKH DEVHQFH RI VKDGLQJ WUDQVPLWWHG VRODU UDGLDWLRQFDQ EH WUDFNHG RQWKH GLIIHUHQW URRP LQWHULRU
VXUIDFHV XVLQJ EDVLF UD\WUDFLQJ WHFKQLTXHV +RZHYHU WKLV LV D FRPSXWDWLRQDOO\ H[SHQVLYH SURFHGXUH HVSHFLDOO\
ZKHQD\HDUO\DQDO\VLVLVLQYROYHG$VLPSOHDVVXPSWLRQZDVXVHGWRDFFRXQWIRUWKHGLUHFWVRODUUDGLDWLRQDEVRUEHG
LQURRPVXUIDFHVRIGLUHFWVRODUUDGLDWLRQFRPHVLQWRWKHURRPLVDVVXPHGWRIDOORQWKHWKUHHZDOOVDQG
LVDVVXPHGWRIDOORQWKHIORRU7KLVDVVXPSWLRQDOWKRXJKFUHDWLQJHUURUVVKRXOGFKDQJHGHSHQGLQJRQWKHURRP
JHRPHWU\$IWHUWKHILUVWDEVRUSWLRQVRODUUDGLDWLRQZLOOEHUHIOHFWHGWRRWKHUVXUIDFHV7KHDPRXQWRIGLIIXVHVRODU
UDGLDWLRQ DQG UHIOHFWHG VRODU UDGLDWLRQ DEVRUEHG E\ HDFK VXUIDFH LV HVWLPDWHG E\ DUHD DQG UHIOHFWDQFH RI HDFK

$6+5$(

VXUIDFH7KHHTXDWLRQIRUWKHDEVRUEHGUDGLDWLRQLQWKHIORRUFDQEHZULWWHQDV
ĨůŽŽƌα ĨůŽŽƌ
^ ĨůŽŽƌ = Ϭ͘ϳ/Ěŝƌα ĨůŽŽƌ + ª¬ Ϭ͘ϯ/Ěŝƌ ;ϭ − α ǁĂůů Ϳ + Ϭ͘ϳ/Ěŝƌ ;ϭ − α ĨůŽŽƌ Ϳ + /ĚŝĨĨƵƐĞ º¼ Ŷ

¦ α
ŝ =ϭ
ŝ ŝ + ǁŝŶ ;ϭ − ρǁŝŶ Ϳ
ZKHUHơLVWKHDEVRUSWDQFHƱLVUHIOHFWDQFH$LVDUHD / LVVRODUUDGLDWLRQDQGQLVWKHWRWDORSDTXHVXUIDFHQXPEHU
:KHQ WKH VKDGHV DUH FORVHG WKH DEVRUEHG GLIIXVH UDGLDWLRQ LV FDOFXODWHG DV FRPELQHG DUHD DQG DEVRUWSWDQFH
DYHUDJHGFRQVLGHULQJDOOLQWHULRUVXUIDFHV
Thermal Model Description

$WKHUPDOQHWZRUNDSSURDFKLVXVHGWRSUHGLFWLQGRRUWKHUPDOHQYLURQPHQWDOFRQGLWLRQVDQGDQQXDOHQHUJ\
FRQVXPSWLRQ RQ SHULPHWHU ]RQHV HTXLSSHG ZLWK FRPELQDWLRQV RI SDVVLYH DQG DFWLYH IDoDGH V\VWHPV VXFK DV
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
VHOHFWLYHJOD]LQJVWUDQVOXFHQWSDQHOVPRWRUL]HGVKDGHVDQGEOLQGVLQFRQMXQFWLRQZLWKGD\OLJKWOLQNHGOLJKWLQJDQG
VKDGLQJ FRQWUROV (DFK FRPSRQHQW RI WKH EXLOGLQJ V\VWHP LV SUHVHQWHG DV D QRGH H[ LQWHULRU VXUIDFH WKHUPDO
PDVVH[WHULRUVXUIDFHVKDGLQJOD\HUVDLU)RUHDFKQRGHDKHDWEDODQFHJRYHUQLQJHTXDWLRQFDQEHZULWWHQEDVHG
RQLWVFRQQHFWLRQVZLWKVXUURXQGLQJQRGHVE\FRQGXFWLRQFRQYHFWLRQDQGUDGLDWLRQDQGWKHKHDWJHQHUDWLRQDQG
KHDW VWRUDJH LQFRQWURO YROXPH 7KH FRQYHFWLRQV IRU ERWK RXWVLGH DQG LQVLGH VXUIDFHV DUH REWDLQHG E\ XVLQJ WKH
IROORZLQJ$6+5$(UHFRPPHQGHGUHODWLRQ$6+5$(7KHUDGLRVLW\PHWKRGZLWKQRQOLQHDUFRHIILFLHQWVLV
XVHGIRUUDGLDWLRQH[FKDQJH7KHFRQYHFWLRQKHDWWUDQVIHUEHWZHHQWKHVKDGHVDQGJDSDLULVFRPSXWHGXVLQJDQ
DGYDQFHGPHWKRG(QHUJ\3OXV7KHHTXDWLRQVHWVDUHILQDOO\VROYHGE\WKHLPSOLFLWILQLWHGLIIHUHQFHPHWKRG
7KHJRYHUQLQJHTXDWLRQIRUWKHWHPSHUDWXUHRIWKHLWKQRGHDWDWLPHVWHSSFDQEHZULWWHQDV

Δƚ Ɖ+ϭ
dŝ Ɖ+ϭ = ;Ƌŝ + ¦hŝũ ;dũ Ɖ+ϭ − dŝ Ɖ+ϭ Ϳ + dŝ Ɖ Ϳ
ŝ ũ
:KHUH Δƚ LVWKHWLPHVWHS ŝ LVWKHWKHUPDOFDSDFLW\RIWKHFRQWUROYROXPH hŝũ LVWKH8YDOXHVEHWZHHQQRGHL

Ɖ +ϭ
DQGVXUURXQGLQJQRGHVM dũ LVWKHWHPSHUDWXUHRIWKHVXUURXQGLQJQRGHVM ƋŝƉ +ϭ LVWKHKHDWLQSXWHJDEVRUEHG
VRODU UDGLDWLRQ RQ QRGH L DWWKH VDPH WLPH VWHS DQG dŝ Ɖ LV WKH WHPSHUDWXUH DW SUHYLRXV WLPH VWHS 7R PDLQWDLQ
DFFXUDF\ QXPHULFDO VWDELOLW\ DQG FRPSXWLQJ HIILFLHQF\ D VXLWDEOH WLPH VWHS LV GHWHUPLQHG IURP EXLOGLQJ GHWDLOV
HIILFLHQF\DQGVHQVLWLYLW\DQDO\VLVDQGDPRGLILHG*DXVV6LHGHOLWHUDWLRQDOJRULWKPLVHPSOR\HGWRVROYHWKHLWHUDWLRQ
RI HTXDWLRQ VHWV ,W UHTXLUHV ODUJHUFRPSXWDWLRQDOPHPRU\EXWLQFUHDVHVWKHFRQYHUJHQFHUDWHDQGVSHHGVXSWKH
FRQYHUJHQFHWLPH)RUJLYHQLQILOWUDWLRQUDWHDQGLQWHUQDOKHDWJDLQVWKHORDGDWHDFKWLPHVWHSLV
Ŷ
Y = ¦ Śŝ −Ăŝƌ ⋅ ;dŝ − dĂŝƌ Ϳ +hŝŶĨŝůƚƌĂƚŝŽŶ ⋅ ;dŽ − dĂŝƌ Ϳ + YŝŚŐ
ŝ =ϭ
ZKHUHQLVWKHWRWDOQXPEHURILQWHULRUVXUIDFHQRGHV7LLVWKHLQWHULRUVXUIDFHWHPSHUDWXUHIRUDOOVXUIDFHVL7DLU
LV WKH LQGRRU DLU WHPSHUDWXUH 7R LV WKH RXWVLGH DLU WHPSHUDWXUH Śŝ −Ăŝƌ LV WKH FRQYHFWLRQ FRHIILFLHQW RI VXUIDFH L
hŝŶĨŝůƚƌĂƚŝŽŶ LV WKH DLU FRQGXFWDQFH GXH WR LQILOWUDWLRQ DQG YŝŚŐ LV LQWHUQDO FRQYHFWLYH KHDW JDLQ ZKLFK FRPHV IURP
OLJKWLQJSHRSOHDQGRWKHUHTXLSPHQW7KHORDGFDQEHSRVLWLYHFRROLQJRUQHJDWLYHKHDWLQJ
,QWKHPXOWLVHFWLRQDOIDFDGHFDVHVWKHWRSDQGERWWRPVHFWLRQVDUHWUHDWHGDVGLIIHUHQWVXUIDFHV)RUHDFKRI
WKHPVHSDUDWHQRGHVDUHXVHGIRURXWVLGHDQGLQVLGHVXUIDFHWHPSHUDWXUHVDQGDVZHOODVVHSDUDWHVKDGLQJGHYLFH
SURSHUWLHVVXUIDFHWHPSHUDWXUHVDQGFRQWUROV
Lighting Model Description

152
ASHRAE Transactions
$ SUHYLRXVO\ GHYHORSHG GD\OLJKWLQJ PRGHO 6KHQ DQG 7]HPSHOLNRV LV XVHG WR SUHGLFW ZRUN SODQH
LOOXPLQDQFHOHYHOVDQGOLJKWLQJGHPDQG)RUGLIIXVHLOOXPLQDQFHWKHZLQGRZLVDVVXPHGWREHDSHUIHFWGLIIXVHU
)RU GLUHFW LOOXPLQDQFH ZH WUDFN WKH H[DFW ORFDWLRQV DQG DUHDV WKDW GD\OLJKW IDOOV RQ DQG XVH WKRVH DV VHSDUDWHG
VXUIDFHV7KHYLHZIDFWRUVEHWZHHQLQWHULRUVXUIDFHVDUHFDOFXODWHGILUVWDQGWKHUDGLRVLW\PDWUL[LVVROYHGWRJHWWKH
LOOXPLQDQFHRQHDFKVXUIDFH&RQILJXUDWLRQIDFWRUVEHWZHHQHDFKVXUIDFHDQGVHOHFWHGSRLQWVRQWKHZRUNSODQHDUH
WKHQ FDOFXODWHG DQG WKH ILQDO ZRUN SODQH LOOXPLQDQFH GLVWULEXWLRQ LV GHWHUPLQHG E\ WKH FRQILJXUDWLRQ IDFWRU DQG
LOOXPLQDQFHRQHDFKVXUIDFH7KHOLJKWLQJHOHFWULFLW\GHPDQGWKHQFDQEHGHWHUPLQHGIURPWKHLOOXPLQDQFHYDOXHV
IRU HDFK WLPH VWHS ,I WKH GD\OLJKW VXSSOLHV HQRXJK DPRXQW RI LOOXPLQDQFH IRU WKH ZRUNHUV WKH PRGHO ZLOO
DXWRPDWLFDOO\GLPRUWXUQRIIWKHHOHFWULFOLJKWV8QGHUGLUHFWVXQOLJKWWKHPRGHOZLOODFWLYDWHWKHVKDGHV7KHVH
FRQWURO DFWLRQV DIIHFW WKH VWDWXV RI WKH WKHUPDO QRGHV DQG WKH LQWHUQDO KHDW JDLQV WKHUHIRUH WKLV LQIRUPDWLRQ LV
SDVVHGWRWKHWKHUPDOPRGHODWWKHVDPHWLPHVWHS
SIMULATION RESULTS
)LJXUHVKRZVDUHSUHVHQWDWLYHDQQXDOORDGSURILOHIRUWKHEDVHOLQHPRGHOLQ:HVW/DID\HWWHIRUWKHFDVHZLWK
DFRQVWDQWVHWSRLQW&DQGRSHQVKDGHV7KHURRPLVH[SRVHGWRKLJKVRODUJDLQVWKURXJKRXWWKH\HDUVRWKH
DQQXDO FRROLQJ ORDG LV KLJK HYHQ LQ WKH ZLQWHU 3HDN FRROLQJ GHPDQG UHDFKHV .: LQ 6HSWHPEHU ZKLOH SHDN
KHDWLQJGHPDQGUHDFKHV.:LQ)HEUXDU\
ϱϬϬϬ
ϰϬϬϬ
ϯϬϬϬ
dŚĞƌŵĂů>ŽĂĚ;tĂƚƚƐͿ
ϮϬϬϬ
ϭϬϬϬ
ͲϭϬϬϬ
ͲϮϬϬϬ
ͲϯϬϬϬ
dŝŵĞ
)LJXUH 6LPXODWHGDQQXDOORDGSURILOHIRUWKHEDVHOLQHPRGHOZLWKFRQVWDQWVHWSRLQWDQGRSHQVKDGHV
Effects of Set Point and Shading Operation

7KH LPSDFW RI VHW SRLQW FRQWURO DQG VKDGLQJ RSHUDWLRQ RQ HQHUJ\ SHUIRUPDQFH ZDV LQYHVWLJDWHG XVLQJ WKH
G\QDPLFWKHUPDOOLJKWLQJPRGHO)LJXUHSUHVHQWVUHVXOWVIRUUHSUHVHQWDWLYHFDVHVZLWKYDULDEOHDQGFRQVWDQWVHW
SRLQWVDVZHOODVZLWKFORVHGRSHQDQGFRQWUROOHGUROOHUVKDGHV7KHEDVHOLQHEXLOGLQJKDVILYHH[WHULRUVXUIDFHVDQG
DODUJHJODVVIDoDGHWKHUHIRUHWKHKHDWLQJDQGFRROLQJHQHUJ\LVPXFKKLJKHUWKDQWKHOLJKWLQJHQHUJ\GHPDQG2SHQ
VKDGHVDOORZXWLOL]DWLRQRIVRODUHQHUJ\DQGUHGXFHWKHKHDWLQJORDGUHVXOWLQJLQRYHUKHDWLQJLQDOPRVWDOOVHDVRQV
1RWH WKDW WKLV LV QRW D UHDOLVWLF FDVH VKRZHG RQO\ IRU FRPSDULVRQ SXUSRVH VLQFH WKH VKDGHV QHHG WR FORVH WR
SURWHFW IURP JODUH ZKHQHYHU WKHUH LV VXQOLJKW LQFLGHQW RQ WKH IDFDGH &ORVHG VKDGHV VLJQLILFDQWO\ UHGXFH FRROLQJ
HQHUJ\ XVH UHIOHFWLYH IDEULFV ZHUH FRQVLGHUHG 7KH FRQWUROOHG VKDGLQJ UHVXOWV SUHVHQWHG LQ )LJXUH UHIHU WR
DXWRPDWLFFORVLQJRIVKDGHVZKHQWKHUHLVLQFLGHQWGLUHFWUDGLDWLRQRQWKHIDoDGHKLJKHUWKDQDVPDOOWKUHVKROG
:P7KLVLVDVWULFWFULWHULRQXVHGDVDEDVHFDVH²WKHUHDUHQRVWDQGDUGVLQVKDGLQJFRQWURORSHUDWLRQDWWKH
PRPHQW ,I WKH WKUHVKROG LV LQFUHDVHG WKH HQHUJ\ SHUIRUPDQFH LV FOHDUO\ LPSURYHG DOVR GXH WR OLJKWLQJ HQHUJ\
VDYLQJV $ GHWDLOHG DQDO\VLV RI WKH LPSDFW RI VKDGH SURSHUWLHV DQG FRQWURO FDQ EH IRXQG LQ 7]HPSHOLNRV DQG

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

$WKLHQLWLVDQGLQ6KHQDQG7]HPSHOLNRV
&RQVLGHULQJWKHVHWSRLQWFRQWUROLWLVREYLRXVWKDWDFRQVWDQWVHWSRLQWGXULQJGD\DQGQLJKWLVQRWHIILFLHQW
)LJXUHSUHVHQWVUHVXOWVIRUDIHZGLIIHUHQWFDVHVLQFOXGLQJYDULDEOHVHWSRLQWVGXULQJQRQRIILFHKRXUV&
DQG&DVZHOODVV\VWHPVKXWGRZQGXULQJQLJKWWLPHZKHQWKHWHPSHUDWXUHVDUHQRWYHU\KLJKRUYHU\ORZ
7KHODWWHUPLJKWEHSUREOHPDWLFVLQFHWKHUHZLOOEHKLJKSHDNVDQGWKHUPDOGLVFRPIRUWZKHQWKHV\VWHPLVDFWLYDWHG
HDUO\LQWKHPRUQLQJ5DPSVKDSHGVHWSRLQWLQFUHDVHDQGPRGHOSUHGLFWLYHFRQWUROVSURYLGHEHWWHUVROXWLRQV

ŶŶƵĂůĞŶĞƌŐǇĐŽŶƐƵŵƉƚŝŽŶ;ŬtŚͬŵϮǇĞĂƌͿ ϯϱϬ
ϯϬϬ
ϮϱϬ
ϮϬϬ
ϭϱϬ >ŝŐŚƚŝŶŐ
ϭϬϬ ŽŽůŝŶŐ
,ĞĂƚŝŶŐ
ϱϬ
Ϭ
ĐŽŶƚƌŽů
ĐŽŶƚƌŽů
ĐŽŶƚƌŽů
ĐŽŶƚƌŽů
ŽƉĞŶ
ŽƉĞŶ
ŽƉĞŶ
ŽƉĞŶ
ĐůŽƐĞ
ĐůŽƐĞ
ĐůŽƐĞ
ĐůŽƐĞ
^ŚĂĚŝŶŐŽƉĞƌĂƚŝŽŶ
Ϯϯ ϮϮͲϮϰ ϭϴͲϭϲ ƐŚƵƚͲĚŽǁŶ EŽŶͲŽĨĨŝĐĞŚƌƐƐĞƚƉŽŝŶƚ

Ϯϯ ϮϮͲϮϰ ϮϮͲϮϰ ϮϮͲϮϰ KĨĨŝĐĞŚƌƐĞƚƉŽŝŶƚ
)LJXUH $QQXDOHQHUJ\SHUIRUPDQFHUHVXOWVXQGHUGLIIHUHQWVHWSRLQWDQGVKDGLQJFRQWUROVWUDWHJLHV
Model Validation
([SHULPHQWDOPHDVXUHPHQWVLQIXOOVFDOHRXWGRRUWHVWRIILFHVSDFHVDW3XUGXH8QLYHUVLW\ZHUHFRQGXFWHGLQ
RUGHUWRFRPSDUHZLWKVLPXODWLRQUHVXOWV)LJXUHVKRZVDVFKHPDWLFVHFWLRQYLHZRIWKHH[SHULPHQWDOVHWXSDQGD
SLFWXUH RI WKH IXOOVFDOH RIILFH VSDFHV ZLWK PXOWLVHFWLRQ IDoDGH 0HDVXUHG TXDQWLWLHV LQFOXGH LQFLGHQW DQG
WUDQVPLWWHGVRODUUDGLDWLRQDQGLOOXPLQDQFHVXUIDFHWHPSHUDWXUHVDQGLQGRRUDQGRXWGRRUDLUWHPSHUDWXUHV

)LJXUH ([SHULPHQWVHWXSVFKHPDWLFLQVHFWLRQYLHZDQGIXOOVFDOHWHVWRIILFHV
5HSUHVHQWDWLYH FRPSDUDWLYH LQGRRU DLU WHPSHUDWXUH UHVXOWV IUHHIORDWLQJ DUH VKRZQ LQ )LJXUH ZLWK ERWK
RSHQ DQG FORVHG VKDGHV LQ 'HFHPEHU WRJHWKHU ZLWK WKH UHVSHFWLYH RXWGRRU DLU WHPSHUDWXUH DQG LQFLGHQW VRODU
--`,`,``,,`,`,```,`,,``,`,``,,,-

154
ASHRAE Transactions
UDGLDWLRQ YDOXHV )RU FORVHG VKDGHV D FRQVLVWHQW WUHQG LV REVHUYHG EHWZHHQ VLPXODWHG DQG H[SHULPHQWDO UHVXOWV
ZLWK D PD[LPXP GLIIHUHQFHRI &)RURSHQ VKDGHV RQ VXQQ\ GD\V VLPXODWHGDQG H[SHULPHQWDO UHVXOWV DOVR
KDYHJRRGDJUHHPHQW+RZHYHULQSDUWO\FORXG\GD\VWKHPD[LPXPGLIIHUHQFHH[FHHG &7KLVLVGXHWRVPDOO
XQFHUWDLQWLHVLQWKHFRQVWUXFWLRQDQGDOVRGXHWRDVVXPSWLRQVPDGHLQWKHPRGHOHGDEVRUEHGVRODUJDLQVDQGIDEULF
SURSHUWLHV 7KH GLVWLQFWLRQ LQ WZR FDVHV LQGLFDWHVWKDW WKH PRGHO LV YHU\ VHQVLWLYH WR VRODU UDGLDWLRQ 2QH RI WKH
FKDOOHQJHVGHDOLQJZLWKVRODUUDGLDWLRQLVKRZWRVSOLWGLIIXVHDQGGLUHFWLRQSRUWLRQ
,QERWKFDVHVWKHORZHVWWHPSHUDWXUHRFFXUVDWDPEHIRUHVXQULVHDQGWKHKLJKHVWWHPSHUDWXUHRFFXUV
EHWZHHQSPDQGSPZKHQVRODUUDGLDWLRQDQGRXWGRRUWHPSHUDWXUHVUHDFKWKHLUKLJKHVWOHYHOV7KHFRQFUHWH
VODEVKDYHDQHIIHFWRQWKHSHDNVKLIWKUVDVH[SHFWHGZLWKPHGLXPWKHUPDOPDVV

)LJXUH &RPSDULVRQRIPHDVXUHGDQGVLPXODWHGUHVXOWVIRUWKHLQGRRUDLUWHPSHUDWXUH
)LJXUH SUHVHQWV DQRWKHU FRPSDULVRQ RI PHDVXUHG YHUVXV H[SHULPHQWDO UHVXOWV IRU GLIIHUHQW WHPSHUDWXUH
UDQJHVXVLQJPLQGDWDFROOHFWHGGXULQJWKUHHGD\VRIPHDVXUHPHQWV$JDLQIRUFORVHVKDGHVWKHUHVXOWVDUHLQ
JRRGDJUHHPHQW%XWWKHUHVXOWVDUHGLYHUVHZLWKRSHQVKDGHV0RUHH[SHULPHQWVIRUGLIIHUHQWFDVHVDUHSODQQHG
LQFOXGLQJYDULDWLRQVRIPXOWLVHFWLRQDOIDFDGHVZLWKLQGHSHQGHQWO\FRQWUROOHGVHFWLRQV
Close Shades Open Shades
21 30
Measured Air Temperature (deg C)
Measured Air Temperature (deg C)
19
25
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
17 12:00pm
15 20 12:00pm
13

11 1:00am 15
9 1:00am
10
7
7:00am
5 5 7:00am
5 7 9 11 13 15 17 19 21 5 10 15 20 25 30
Simulated Air Tmeperature (deg C) Simulated Air Tmeperature (deg C)

)LJXUH $LUWHPSHUDWXUHUHVXOWVFRPSDULVRQIRUGLIIHUHQWWHPSHUDWXUHUDQJHVXVLQJPLQFROOHFWHGDQG
VLPXODWHGGDWD
CONCLUSION
7KLVSDSHUSUHVHQWVDPHWKRGIRUDVVHVVLQJWKHLQWHJUDWHGHQHUJ\SHUIRUPDQFHRISDVVLYHDQGDFWLYHPXOWL
VHFWLRQ IDoDGH V\VWHPV FRPELQHG ZLWK OLJKWLQJ DQG WKHUPDO FRQWUROV RI SHULPHWHU EXLOGLQJ ]RQHV XVLQJ DQ RSHQ
VRXUFH ODQJXDJH $ WKHUPDO QHWZRUN DSSURDFK LV XVHG WR SUHGLFW LQGRRU WKHUPDO HQYLURQPHQWDO FRQGLWLRQV DQG

$6+5$(

DQQXDOHQHUJ\FRQVXPSWLRQRISHULPHWHU]RQHVHTXLSSHGZLWKFRPELQDWLRQVRISDVVLYHDQGDFWLYHIDoDGHV\VWHPV
7KHPRGHOXVHVDQLVRWURSLFVN\PRGHOVIRUDFFXUDWHSUHGLFWLRQRIVRODUJDLQVYDULDEOHDQJXODUJOD]LQJSURSHUWLHV
DQG QRQOLQHDU LQWHULRU DQG H[WHULRU FRQYHFWLRQ DQG UDGLDWLRQ KHDW WUDQVIHU FRHIILFLHQWV WRJHWKHU ZLWK WUDQVLHQW
LQWHUQDO JDLQV REWDLQHG IURP WUDQVLHQW OLJKWLQJ VLPXODWLRQ $ SULYDWH SHULPHWHU RIILFH ZLWK D JODVV IDoDGH DQG
HTXLSSHGZLWKDXWRPDWHGVKDGLQJDQGOLJKWLQJV\VWHPVZDVXVHGDVDEDVHFDVH7KHUHVXOWVVKRZWKDWFKDQJHVRI
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
FRQWURO VWUDWHJLHV DQG FRPSRQHQWV SURSHUWLHV PD\ UHVXOW LQ ODUJH DPRXQWV RI HQHUJ\ VDYLQJV 7KHUH LV DOZD\V D
WUDGHRIIEHWZHHQOLJKWLQJKHDWLQJDQGFRROLQJGHPDQG([DPLQDWLRQRIWKHRYHUDOOHQHUJ\SHUIRUPDQFHEDVHGRQ
EXLOGLQJ JHRPHWU\ DQG ZHDWKHU FRQGLWLRQV LV XVHIXO IRU ERWK GHVLJQHUV DQG HQJLQHHUV WR DFKLHYH D FRPIRUWDEOH
HQYLURQPHQW DQG DW WKH VDPH WLPH VDYLQJ PRQH\ )XUWKHU FRQVLGHUDWLRQV VXFK DV FRVW DQG IHDVLELOLW\ VKRXOG EH
WDNHQ LQWR DFFRXQW LQ WKH UHDO GHFLVLRQ PDNLQJ SURFHVV )RU H[DPSOH VRPH DGYDQFHG FRQWURO VWUDWHJLHV PD\ EH
LPSOHPHQWHGE\LQVWDOOLQJH[SHQVLYHGHWHFWLQJDQGPRQLWRULQJV\VWHPV
0XOWLIXQFWLRQDODQGPXOWLVHFWLRQDOG\QDPLFIDoDGHV\VWHPVFDQSURYLGHVROXWLRQVWRSUREOHPVUHODWHGWRWKH
EDODQFHEHWZHHQDLUFRQGLWLRQLQJDQGOLJKWLQJHQHUJ\GHPDQG(DFKVHFWLRQFRXOGEHLQGHSHQGHQWO\FRQWUROOHGVR
DVWRRSWLPL]HHQHUJ\XVHDQGPDLQWDLQFRPIRUWIRUWKHRFFXSDQWV'HWDLOHGDQGIXOOVFDOHH[SHULPHQWDOVWXGLHVFDQ
KHOSLQYDOLGDWLQJDQGILQHWXQLQJPRGHOVDQGLQLQYHVWLJDWLQJZD\VRIXWLOL]LQJVRODUHQHUJ\LQSHULPHWHUVSDFHVRI
FRPPHUFLDOEXLOGLQJV0RUHZRUNLVQHHGHGIRUGHYHORSPHQWRIV\VWHPDWLFGHVLJQDQGRSWLPL]DWLRQPHWKRGVIRU
LQWHJUDWHGIDoDGHFRQFHSWV
REFERERENCES
$6+5$( $6+5$( +DQGERRN³)XQGDPHQWDOV $WODQWD $PHULFDQ 6RFLHW\ RI +HDWLQJ 5HIULJHUDWLQJ DQG
$LU&RQGLWLRQLQJ(QJLQHHUV,QF
$WPDFD , .D\QDNOL 2 <LJLW $ (IIHFWV RI UDGLDQW WHPSHUDWXUH RQ WKHUPDO FRPIRUW %XLOGLQJ DQG
(QYLURQPHQW
&ODUNH-$(QHUJ\VLPXODWLRQLQEXLOGLQJGHVLJQ2[IRUG%XWWHUZRUWK+HLQHPDQQS
)UDQ]HWWL & )UDLVVH * $FKDUG * ,QIOXHQFH RI WKH FRXSOLQJ EHWZHHQ GD\OLJKW DQG DUWLILFLDO OLJKWLQJ RQ
WKHUPDOORDGVLQRIILFHEXLOGLQJV(QHUJ\DQG%XLOGLQJV
-RQVVRQ $ 5RRV $ (YDOXDWLRQ RI FRQWURO VWUDWHJLHV IRU GLIIHUHQW VPDUW ZLQGRZ FRPELQDWLRQV XVLQJ
FRPSXWHUVLPXODWLRQV6RODU(QHUJ\
/HH ( 6 'L%DUWRORPHR ' / 6HONRZLW] 6 ( 7KHUPDO DQG GD\OLJKWLQJ SHUIRUPDQFH RI DQ DXWRPDWHG
YHQHWLDQEOLQGDQGOLJKWLQJV\VWHPLQDIXOOVFDOHSULYDWHRIILFH(QHUJ\DQG%XLOGLQJV
0DQ] + /RXW]HQKLVHU 3 )UDQN 7 6WUDFKDQ 3 $ %XQGL 5 0D[ZHOO * 6HULHV RI H[SHULPHQWV IRU
HPSLULFDOYDOLGDWLRQRIVRODUJDLQPRGHOLQJLQEXLOGLQJHQHUJ\VLPXODWLRQFRGHV³([SHULPHQWDOVHWXSWHVWFHOO
FKDUDFWHUL]DWLRQVSHFLILFDWLRQVDQGXQFHUWDLQW\DQDO\VLV%XLOGLQJDQG(QYLURQPHQW
6HONRZLW]6/HH(6$GYDQFHGIHQHVWUDWLRQV\VWHPVIRULPSURYHGGD\OLJKWSHUIRUPDQFH Daylighting ’98
Conference Proceedings, Ottawa, Ontario, Canada.
Tzempelikos, A., Athienitis, A.K., 2005. Integrated daylighting and thermal analysis of office buildings. ASHRAE
Transactions 111(1), 227-238.
Perez, R., Ineichen, P., Seals, R., 1990. Modeling daylight availability and irradiance components from direct and
global irradiance. Solar Energy 44(5), 271-289.
6KHQ+7]HPSHOLNRV$'D\OLJKWLQJDQGHQHUJ\DQDO\VLVRISULYDWHRIILFHVZLWKDXWRPDWHGLQWHULRUUROOHU
VKDGHV6RODU(QHUJ\LQSUHVV
7]HPSHOLNRV$$WKLHQLWLV$.7KHLPSDFWRIVKDGLQJGHVLJQDQGFRQWURORQEXLOGLQJFRROLQJDQGOLJKWLQJ
GHPDQG6RODU(QHUJ\
7]HPSHOLNRV $ $WKLHQLWLV $ . 1D]RV $ ,QWHJUDWHG 'HVLJQ RI 3HULPHWHU =RQHV ZLWK *ODVV )DFDGHV
$6+5$(7UDQVDFWLRQV
YDQ0RHVHNH*%UX\qUH,'H+HUGH$,PSDFWRIFRQWUROUXOHVRQWKHHIILFLHQF\RIVKDGLQJGHYLFHVDQG
IUHHFRROLQJIRURIILFHEXLOGLQJV%XLOGLQJDQG(QYLURQPHQW
:,1'2:7+(501)5&6LPXODWLRQ0DQXDO/DZUHQFH%HUNHOH\1DWLRQDO/DERUDWRU\
(QHUJ\3OXV(QJLQHHULQJ'RFXPHQWWKHUHIHUHQFHWR(QHUJ\3OXVFDOFXODWLRQV86'HSDUWPHQWRI(QHUJ\

156
ASHRAE Transactions
SA-12-C020
Uncertainties in Achieving Energy Savings from

HVAC Maintenance Measures in the Field
Kristin Heinemeier 1, Marshall Hunt2,
Marc Hoeschele3, Elizabeth Weitzel3, Brett Close4
1
Western Cooling Efficiency Center, UC Davis, Davis, CA, USA
2
Pacific Gas & Electric, San Francisco, CA, USA
3
Davis Energy Group, Davis, CA, USA
4
Southern California Edison, Irwindale, CA, USA
ABSTRACT
HVAC maintenance measures in residential and small commercial buildings have been demonstrated
in the laboratory to have the potential to save a significant amount of energy. This significant potential for
savings has prompted utilities across the nation to include HVAC maintenance measures in energy
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
efficiency programs. This is currently seen as the cutting edge of utility HVAC programs. However,
evaluation, measurement and verification (EM&V) studies of these programs have shown mixed
results. This paper presents analysis of the sources of uncertainty in delivering and measuring these
programs. The gaps that can account for the discrepancy between the potential and the measured savings
are described. By identifying the range of issues, program planners can address as many as possible of the
potential sources of uncertainty.
Measurement issues are of particular focus. An analysis is done of the uncertainties in the
measurements of common variables as measured in the laboratory, by EM&V teams, by participants in
maintenance programs, and by typical contractors. These uncertainties were combined to identify the
resulting uncertainty in the calculated subcooling, superheat, EER values and annual kWh. The remainder
of this paper presents recommendations for improving maintenance measures, based upon the uncertainties
identified.
INTRODUCTION
Maintenance measures have a significant potential for energy savings in residential and commercial
unitary HVAC systems, prompting utilities across the nation to include HVAC maintenance measures in
energy efficiency programs. With encouragement from the California Long Term Energy Efficiency
Strategic Plan (CPUC 2007), California investor owned utilities (IOUs) have been embarking on an
extensive program of incenting maintenance measures through rebates. These programs include measures
such as refrigerant charge adjustment (RCA), airflow adjustment, duct leakage testing and sealing. Other
proposed measures include condenser and evaporator coil cleaning, economizer retro-commissioning, and
HVAC controls. The programs involved contractors and their technicians who do the maintenance work,
Verification Service Providers (VSPs) who develop protocols and verify that the work is done correctly
(note that the current round of programs is not using VSPs), and Evaluation, Measurement and Verification
(EM&V) contractors who verify that expected savings are realized. This is currently seen in California as
the cutting edge of utility HVAC programs.
HVAC maintenance measures have been demonstrated in the laboratory to have the potential to save a
significant amount of energy. For example, Mowris et al. 2012 shows that by combining charge
adjustment, airflow improvement, duct sealing, and elimination of non-condensables, over 30% of HVAC
energy use can be saved.
However, EM&V studies of these programs have shown mixed and sometimes disappointing results.
KEMA 2010 shows that savings range from greater than expected to much less than expected. While
1
Kristin Heinemeier is a Principal Development Engineer at the Western Cooling Efficiency Center.
$6+5$(

EM&V teams acknowledge significant unmet challenges in measuring the savings from this type of
program, there is clearly a gap between potential and reality.
What is the source of this gap? There are multiple sources of uncertainty in maintenance measures that
result in savings found in the field differing from savings found in the lab. This paper describes a study
conducted to analyze these sources of uncertainty (Hunt et al, 2010). These sources include programmatic,
process, measurement, system, and human factors. The gaps that can account for the discrepancy between
the expected and the measured savings are described. These issues are presented in the following
categories: factors that affect potential for savings from the measure, factors that affect the delivery of the
savings in the field, factors that affect the persistence of savings in the field, and factors that affect the
potential to measure the realized savings. By identifying the range of issues, program planners can address
the potential sources of uncertainty.
Measurement issues were of particular focus. An analysis was done of the stated uncertainties in the
measurements of common variables (such as dry bulb temperature, wet bulb temperature, refrigerant line
temperatures, refrigerant pressures, airflow, and power) as measured in the laboratory, by EM&V teams, by
participants in maintenance programs, and by typical contractors. These stated uncertainties were
combined to identify the resulting uncertainty in the calculated subcooling, superheat, EER, and annual
kWh values.
UNCERTAINTIES IN MAINTENANCE MEASURES
Uncertainties in maintenance measures go far beyond the inaccuracies in measurements. They include
a host of factors including programmatic, process, instrumentation, system, and human factors. Table 1
shows a number of the factors that come into play in maintenance measures and the programs that provide
them. What are the factors that can jeopardize savings in maintenance measures? They are described in
the next few sections.
Table 1: Sources of Uncertainty in Maintenance Measures and Programs
Human Factors
While HVAC maintenance is primarily a technical service, the service is provided by people who work
for other people, and is performed for people who own and occupy the building--all of whom may behave
in ways that help or hurt the success of the program.
Technicians: A maintenance measure is only effective if technicians have the tools, skills and training
to implement it properly. While the maintenance tasks themselves may be quite simple to accomplish, the
technician will need to exercise significant judgment in evaluating the data and the condition of the system,
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
158
ASHRAE Transactions
and deciding whether it makes sense to proceed with maintenance, which tasks to perform, and how to
perform them and verify the effectiveness. Technicians need to have practical training that prepares them
for the diagnostic and remediation tasks they must perform. Even when the maintenance measures are
achievable in the lab by someone with a typical technician’s skill set, actually achieving savings in the field
is primarily in the hands of the technicians performing the work, using procedures provided and enforced
by their employers. That is, assuming the technician can do the work, the question becomes whether they
will do the work properly. The likelihood of high-quality work is affected by several factors including
whether or not the technician understands and follows the specified processes and whether the technician
has calibrated his/her instruments.
Contractors: Even when technicians are well trained and capable, however, they may not be given
enough time to complete the necessary tasks at the job site because their employer (the contractor) has
scheduled them to complete a certain number of service calls per day. The contractor’s business model can
be a limiting factor in the potential for savings. Activities that support changes in the contractor business
model will help in this regard—e.g., by encouraging customer demand and willingness to pay for quality
maintenance, reducing customer call-backs, and increasing the perceived value related to this type of
service.
Building Owners: Building owners have a significant impact on the savings, by choosing to perform
(or not) more comprehensive routine maintenance, and could make more informed decision with a basic
knowledge of system maintenance requirements.
Occupants: Other behavioral challenges can be harder to overcome. Occupants will often “take back”
a fraction of the savings from a measure by increasing their comfort level (and HVAC energy use) once the
system is operating more efficiently.
State of an Air Conditioning System
The specific air conditioning system being serviced affects the estimated and measured energy savings,
and their persistence. The range of savings from one unit to another can vary by an order of magnitude due
to the mechanical status of each unit, the sizing and loads imposed on the unit, and, especially in residential
applications, the behavior of the occupant. Maintenance programs must walk a fine line between a mass-
market approach that is easily managed, and a specialized approach that addresses each individual system
on a custom basis. The development of a measure that is a probabilistic package (menu) of technician
quality maintenance (QM) activities or treatments shows promise.
If multiple faults are present, any single adjustment alone, such as a standard refrigerant charge
adjustment process, has a limited potential for savings. Given the likelihood of multiple faults and their
impact on the potential for savings, a standard procedure or protocol is needed to support the technician
diagnostic efforts. If a unit and its duct system were badly designed or installed, then HVAC maintenance
measures are likely to be more effective than assumed. A common residential system problem is ducts that
are too small for the required airflow. Ducts with disconnects and other major, visible leaks can be fixed
first because they are the dominant cause of system inefficiency. If ducts are inaccessible, then they cannot
be adequately sealed using methods that most contractors would employ. If an air conditioner’s design or
installation leaves little room for accessing a (dirty) evaporator coil, the technician can optimize the charge
of the system and savings can occur, although they will not be as great as if the underlying problem were
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
addressed. If there is a leak in the refrigerant line, then simply adding charge will not address the problem
and the low charge fault will reoccur. More research needs to be done to establish the likelihood of
multiple faults and remediation techniques.
Diagnostic and Remediation Process Uncertainties
Even with high-quality instruments, the process used to take measurements and adjust air conditioning
systems will determine whether or not savings can be achieved. This process must be effective, efficient,
well-defined, clearly specified, and well-carried out.
The definition of a measure and the process used to implement that measure have a huge impact on the
potential to save energy. For example, coil cleaning is a common practice but was not required by most
past programs and, therefore, contractors did not receive an incentive for implementing it—so it was often
neglected. The cost of cleaning a severely fouled evaporator coil can be prohibitive in cases where it is in a
$6+5$(

confined space and is not equipped with an access panel, yet some would say that it is a necessary part of
the Air Conditioner Contractors of America Quality Maintenance standards, Standard 4 and Standard 180
(ACCA, 2008; ASHRAE, 2008).
MEASUREMENT UNCERTAINTY
The measurements that are taken during air conditioning maintenance are critical to saving energy
because they dictate the remediation steps that should be taken. The accuracy of instrumentation required
by the 2006-2008 California IOU program specifications (AEC, 2004) has been reexamined and there is a
general consensus that some of the specifications should be revised. If the required specifications are
inadequate, then the contractor may be unable to service the system effectively. Highly accurate
instruments can support the technician in achieving savings, but the cost and fragility of laboratory
instrumentation make them impractical for use in the field. Instrumentation sensitivity and cost must be
balanced: how accurate do instruments need to be to attain the desired level of confidence, and what is the
most cost-effective way to reach this precision? Advances in digital instrumentation such as digital
refrigerant pressure gauges with 1% accuracy instead of the 3% achieved with analog dial gauges are
making it possible to require improved accuracy—whether or not this is warranted has not been
established.
Instrumentation is the first root of uncertainty of measurement. When considering specifying an
instrument to measure a physical property, the instrument’s accuracy must be taken into account in the
reported reading. Commonly used instruments can vary in their levels of accuracy. Higher accuracy units
are not in wide use because they cost more, are often more fragile, and must be sent off to be calibrated.
Calibration errors add a second layer of uncertainty in measurement, which can only be addressed by
implementing a consistent and regular calibration protocol. The current California refrigerant charge and
airflow protocol (AEC 2004) calls for annual or monthly recalibration of instruments. A review of
accuracy and calibration requirements on an instrument-by-instrument would identify the most cost
effective means of calibrating technician instruments. The third layer of uncertainty relates to how the
measurement is conducted. There is a need for a detailed description of where and how sensors are to be
installed, including system diagrams, installation detail drawings, and photographs—especially for
temperature measurements. The fourth layer of uncertainty occurs when measurements on a system are
taken at different times and with different instruments. Measurements taken by the HVAC technician and
then later by the EM&V technician are conducted under different ambient, space, and attic conditions, and
without looking at the effect of these differences we cannot ascribe a major significance to the different
findings.
Uncertainties in Temperature Measurement
Measurement of refrigerant line temperature (later used in superheat and subcooling determination) is
performed by connecting a temperature measurement device to the bare and clean copper line, with
insulation. Haorong Li’s doctoral dissertation (Li 2004) summarized the evaluation of the difference in
measuring line temperature with a resistance thermal device (RTD) versus a thermocouple. The factors
affecting measurement include the presence of insulation, mechanics of heat transfer, and the temperature
differential of line and ambient. Li summarized the heat transfer to an RTD in an equivalence resistance
model, and concluded that even under the best of installations, the error in measurement is 20% of the total
differential temperature, or 20 °F if there is a 100 °F differential temperature. A thermocouple in the same
condition is capable of an error of 1.6%, or 1.6 °F if there is a 100 °F differential temperature. These
uncertainties illustrate the need to consider not only the published accuracy but also the accuracy that is
dependent on the application.
Suction line surface temperature is higher on the tube bottom due to the returning compressor oil,
indicating that the probe must be located on the top of the tube. Technicians are taught to measure at the 3
or 9 o’clock positions to avoid this problem but protocols are needed to make it a standard practice.
Exploratory testing done at the Pacific Gas & Electric laboratory shows the impact of using different
sensors mounted in different manners on the accuracy of pipe surface temperature measurements. The best
sensor and mounting method was the insulated and calibrated bead thermocouple which resulted in a 3%
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
160
ASHRAE Transactions
error. This contrasts with a measured 7% error when using a clamp-on thermocouple as commonly used
and as allowed by the 2004 RCA specifications (Davis, 2007).
Proctor Engineering Group has conducted additional testing on 5 commonly used temperature probes in
support of California Energy Code (Title 24) revisions (Hairrell et al, 2010). Of particular concern is the
time it takes for the relatively high mass thermistor probe to reach the terminal temperature. This research
also found that one clamp-on thermocouple of the type preferred by technicians worked very well, while in
general thermistors performed poorly. The work of Li, Davis and Proctor support the assertion that just
considering the published accuracy of sensors is not sufficient. Additional testing is needed to develop the
measurement and instrument specifications to support technician service work.
Uncertainties in Humidity Measurement
Moisture can be measured in three different ways, all of which have their own challenges and
uncertainties. Dewpoint measurement is the preferred method in laboratory settings (although wetted
temperature sensors are used), but it is not feasible for field measurement. Digital readout relative humidity
(RH) instruments are common and reasonably priced, although these sensors have accuracy issues and tend
to drift over time. Wet bulb temperature measurements are made with a wet cotton sock over a temperature
probe, which can be a thermocouple, a mercury bulb thermometer, or a RTD. The sock must be kept moist
and airflow is needed to keep evaporation maximized. Measurement of supply air moisture is difficult
because most methods lose accuracy when RH is over 90%, which is a common condition.
UNCERTAINTY IN CALCULATED MEASURES
The research team analyzed the uncertainties in calculation of various system variables, as a function
of the accuracy of primary measurements (Hunt et al, 2010). Uncertainties in the measured variables are
published in instrument documentation, or can be obtained by testing the instrument. Performance metrics
are obtained from a system equation that involves direct measurements of measured variables. The total
uncertainty in the system variable is related to sensitivity coefficients—the calculated partial derivatives of
the system variable with respect to the measured variables. For ease of calculation, the research team input
system equations into an equation solver with built-in psychrometric and refrigerant charts.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Subcooling and Superheat
Subcooling temperature is dependent on two measurements, liquid line surface temperature and
discharge pressure. Hence the uncertainty in subcooling is determined from the partial derivatives with
respect to these measurements and the stated accuracies of both devices.
Energy Efficiency Ratio (EER)
In contrast to the onetime measurements that go into superheat and subcooling, the uncertainty of in-
field determination of total EER and sensible EER makes a onetime determination of system performance
suspect, especially given that improvements from fixing system faults, as measured under ideal laboratory
conditions, do not rise to the 20% level. With monitoring over time, these gaps in uncertainty can be
reduced with an analysis of time-series data that has been collected on an actual HVAC system. The two
analytic techniques used to address this type of data are generalized additive models and the calculation of
autocorrelation functions for each of the lower level (directly measured) variables. This sophisticated
statistical analysis is documented in the project’s final report (Hunt et al, 2010).
The conclusion of this analysis was that time series data must be collected to reduce measurement
uncertainties in EER values, and due to computational complexity, the benefit of additional data decreases
sharply after approximately 100 data points. A residential scale air conditioner is deemed to reach
equivalent steady state in 15 minutes of operation after which a valid set of superheat and subcooling
measurements can be taken. If 100 data points are needed and they are taken on a 1 minute interval then
steady state operation needs to occur for almost two hours. In real-world applications with cooling
equipment that is oversized even at peak conditions, two hours of operation will rarely occur. This makes
it necessary to perform a field study to determine if a set of separated periods of steady state operation can
$6+5$(

be aggregated into a data set of 100 points. If this is not possible other methods will need to be developed
such as ways to force the two hours of operation.
Annual Energy Use and Savings
Calculation of annual energy savings has considerably more uncertainty than even EER. These
estimates are typically arrived at using building simulation based upon measured or assumed EER values.
As one example of this type of analysis, five increasingly detailed levels of tuned simulations were
performed using data from six buildings in Southern California (Alereza and Faramarzi 1994). “Level 1”
simulations used building specific data for the inputs to the model. “Absolute estimation errors for HVAC
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
End Use Intensity (EUI) ranged from 1 to 27 percent.” The average error was 17.8%. “Level 5” reduced
the average error in HVAC EUI to 11.6% and “three buildings had reduced absolute errors and three had
increased absolute errors.” One can expect that estimates of annual energy usage based upon even the best
EER measurements and calculations is on the order of 20%, potentially masking the savings which are
themselves on the order of 20%.
Resulting Uncertainties
Table 2, taken from the Hunt study (2010), presents the results of this statistical analysis in the context
of measurements from a number of sources:
x A review of published instrument accuracies from several laboratory-testing facilities: PG&E (Davis
and D’Albora 2001), Intertek (Mowris et al., 2010) and Purdue University Herrick Laboratory (Shen et
al. 2006), (“Lab”)
x A review of the accuracies claimed in EM&V reports by KEMA (2010) and Robert Mowris (2004).
(“EM&V”),
x 2006-2008 Program Specifications, AEC 2004 (“AEC 2004”),
x A review of the instrumentation used by contractors working for Verified Service Providers (“VSP”),
x A review of instrumentation typically used by contractors from a 2006 AEC report for PG&E, AEC
2006 (“Contractor Current”), and
x A recommended revision to the Program Specifications, described later (“Revised AEC 2010”).
Table 2: Uncertainties in Measured Values (from Hunt et al, 2010)

y
Lab EM&V AEC 2004 VSP Contractor Revised AEC
Measured Variables Units Min Max Min Max Min Max Current 2010 Tech
Supply Air (Dry Bulb) F ± 0.05 ± 2.3 ± 0.18 ± 0.50 ± 1.5 ± 0.7 ± 1.5 ± 2.1 ± 1.5
Return Air (Dry Bulb) F ± 0.05 ± 1.8 ± 0.18 ± 0.80 ± 1.5 ± 0.7 ± 1.5 ± 2.1 ± 1.5
Outside Air (Dry Bulb) F ± 0.05 ± 1.8 ± 0.18 ± 1.00 ± 1.5 ± 0.7 ± 1.5 ± 2.1 ± 1.5
Supply Air (Wet Bulb) F ± 0.05 ± 1.8 ± 1.5 ± 2.1
Return Air (Wet Bulb) F ± 0.05 ± 1.8 ± 1.5 ± 2.1
Outside Air (Wet Bulb) F ± 0.05 ± 1.8 ± 1.5 ± 2.1
Supply Air (RH) (DEWPOINT) %|F ± 0.005 ± 0.4 ±1% ±2% ±3% ± 2.0 % ±3% ±3% ±2%
Return Air (RH) (DEWPOINT) %|F ± 0.005 ± 0.4 ±1% ±2% ±3% ± 2.0 % ±3% ±3% ±2%
Outside Air (RH) (DEWPOINT) %|F ± 0.005 ± 0.4 ±1% ±2% ±3% ±3%
Condenser Discharge (Dry Bulb) F ± 0.05 ± 0.9 ± 0.2 ± 1.0
Suction Line (Dry Bulb) F ± 0.30 ± 0.9 ± 0.5 ± 1.5 ± 1.5 ± 1.5 ± 3.2 ± 3.5 ± 1.5
Liquid Line (Dry Bulb) F ± 0.30 ± 0.9 ± 1.0 ± 1.5 ± 1.5 ± 1.5 ± 3.2 ± 3.5 ± 1.5
Suction Pressure psig ± 0.3 ± 1.1 ± 0.13 ± 1.35 ± 4.04 ± 1.35 ± 4.04 ± 4.04 ± 1.00
Discharge Pressure psig ± 1.0 ± 3.2 ± 0.40 ± 4.05 ± 12.14 ± 4.05 ± 12.14 ± 12.14 ± 1.00
Condenser Power (True RMS) W ± 0.2 ± 10.0 ± 1.5 ± 3.0 ± 3.0 ± 3.0
Compressor Power (True RMS) W ± 2.9 ± 14.3 ± 28.6 ± 57.2 ± 57.2 ± 57.2
AHU Power (True RMS) W ± 0.5 ± 10.0 ± 4.5 ± 9.0 ± 9.0 ± 9.0
AHU Flow Rate CFM ± 6.1 ± 12.1 ± 36 ± 85 ± 85 ± 36 ± 85 ± 85
Atmospheric Pressure Pa ± 0.03 ± 0.3 ± 1.0 ± 1.0 ± 1.0 ± 0.05
Calculated Variables Units

Superheat F ± 0.33 ± 1.02 ± 0.50 ± 1.61 ± 2.32 ± 0.50 ± 3.20 ± 3.92 ± 1.56
Subcooling F ± 0.36 ± 1.09 ± 1.00 ± 1.68 ± 2.74 ± 1.00 ± 3.20 ± 4.19 ± 1.51
Condensing Over Ambient F ± 0.19 ± 1.90 ± 0.79 ± 1.00 ± 2.74 ± 0.70 ± 2.74 ± 3.11 ± 1.51
EER (Total) ± 0.20 ± 2.17 ± 0.39 ± 0.95 ± 1.69 ± 1.46
EER (Sensible) ± 0.03 ± 1.04 ± 0.15 ± 0.38 ± 0.78 ± 0.77
162
ASHRAE Transactions
From Table 2, we can see that lab measurements are more accurate than field measurements, (as
expected) and that the EM&V measurements are more accurate than the measurements made by VSPs.
Some VSPs are not meeting the program specifications, and the instrumentation typically used by
contractors currently has a high degree of uncertainty. The Current Contractors column does not apply to
contractors working with VSP that require instrumentation that meets or exceeds the AEC specification.
The problem with EER uncertainties is that the impact of measures is often less that the uncertainty of the
one-time measurement. This can be addressed by time-series monitoring, as discussed earlier.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
RECOMMENDATIONS
In order to address uncertainties outlined above, carefully designed and implemented pilot programs
with EM&V monitoring are needed to establish the potential energy and demand savings from
implementing the diagnostic protocol. These programs will initially establish a baseline performance of
representative systems as they are found. Residential split systems and commercial packaged units will be
included in the monitoring. The performance of air conditioners with various changes in parameters has
been established in lab testing and more of that needs to be done. We also need to determine what faults
exist in the field—their prevalence and the distribution of the magnitude of the errors. Long-term, detailed
monitoring of hundreds of sites is required. Additional laboratory testing needs to be part of the integrated
research plan.
The minimum instrumentation requirements for maintenance programs need to continue to be
examined and upgraded based on experience and data. A “Technical Forum” was held in August, 2010 in
Stockton, CA, which included 31 of the most prominent researchers, contractors, manufacturers, VSPs, and
others in the industry. A set of recommended changes to the 2006-2008 California IOU instrumentation
and diagnostic protocol was drafted by the participants, shown in the last column in Table 2. While the
proposals were discussed extensively, it was not the proper venue to adopt changes. Further work should be
done, bringing stakeholders together to discuss the issues and develop a consensus.
A fault detection and diagnostic protocol needs to be developed that can deal with multiple system
variables and multiple faults. This protocol will need to include an inventory and assessment component.
Without IOU incentive programs, technicians use simple checklists to show the customer that the suite of
tasks has been done. Technicians have little or no historical information on the unit being serviced. This
can be the case even when a technician from the same company was the last one servicing the unit. ACCA
Standards 4 and 180 require documentation to address this problem. The IOU RCA programs of necessity
require documentation to qualify for incentive payments. But there is not a generic performance standard.
The foundation has also been laid for developing a generic, performance based measurement and
diagnostic protocol that can be implemented by VSPs, instrument manufacturers, and other entities whose
business is to support the work of HVAC technicians. The ACCA standards and the experience gained
from implementing RCA programs provide the raw material from which stakeholders can hammer out the
protocols. This will be a process that will first be tested in meetings and then be tested in the field.
SUMMARY
The California IOUs have shown a great deal of leadership in initiating maintenance-based HVAC
programs. These and other energy efficiency programs have been in existence since the 1980s, and have
reached millions of homes and small businesses. Despite their success in reaching the market, however, the
energy savings attributable to HVAC maintenance programs have been called into question. For example,
one evaluation of savings for RCA programs in the 2006-2008 California IOU program cycle found quite
low savings rates, but also found wide variations in the different program performance metrics. These
studies raised the possibility that some of the EM&V questions being asked and answered have such large
uncertainties that conclusions and recommendations based on it should be considered carefully.
In reviewing the preceding sections, it is notable that the uncertainties in measurements made in
maintenance services are large, and the savings from current programs are asserted to be small in relation to
the uncertainties. We conclude that a holistic approach to both the design and implementation of the
HVAC programs with integrated measurement and evaluation methods can reduce the uncertainties and
increase the savings, such that investments in expanded HVAC QM programs can show prudent use of
$6+5$(

ratepayer funds. A holistic approach does not necessitate that the program implementing it be complicated,
time consuming and therefore cost ineffective. Future programs will build upon the knowledge gained from
previous programs to deliver a holistic set of site specific services to achieve significant savings that are
needed to meet CPUC goals and which are verifiable.
In the short run, maintenance-based programs continue and continue to be refined, improved, and
redesigned. The focus of this study was on how the industry can be moved from current programs to future
even better programs. In the long run, achieving the ambitious California Public Utilities Commission
(CPUC) “Big-Bold” HVAC goal of 50% improvement in residential and small commercial HVAC system
efficiency will require new, more comprehensive programs that have the potential for greater impact. For
these programs to constitute a prudent use of ratepayer money, however, they must be designed based on a
good understanding of the impacts and interrelationships of individual and combined system faults (i.e.,
abnormal conditions that may lead to system performance degradation or failure) and maintenance
measures. A simple “widgets” approach that focuses on individual measures that save 10% here and 5%
there will not achieve the level of savings that is needed to meet this ambitious goal. HVAC technologies
benefit from a broad based systems approach with a thorough understanding of the associated uncertainties.
REFERENCES
ACCA. 2008. Standard 4: Maintenance of Residential HVAC Systems. Air Conditioning Contractors of
America, Arlington, VA.
Alereza, T. and R. Faramarzi. 1994. “More Data is Better, But How Much is Enough for Impact
Evaluations?” ACEEE Summer Study Proceedings.
Architectural Energy Corporation (AEC). 2004. “Residential Air Conditioner Charge and Air Flow
Verification Study, Task 4 Report, Technical Specifications and Best Practices for Charge and Air
Flow Verification Services”, Contract # 4600010737, Revision 2, PG&E, August 19.
Architectural Energy Corporation (AEC). 2006. Commercial HVAC Refrigerant Charge & Airflow and
Economizer Verification Study, Contract #4600016485, Job I.D. #06 CEE-T-3589, Task 3:
Conduct Additional Research, Draft, December 15.
ASHRAE. 2008. Standard 180: Standard Practice for Inspection and Maintenance of Commercial
Building HVAC Systems.
California Public Utilities Commission (CPUC). 2008. California Long Term Energy Efficiency Strategic
Plan, September.
Davis, R. 2007. Experimental Analysis of Tube Surface Temperature Measurements, Laboratory Test
Report # 491-07.6, Applied Technology Services, PG&E, San Ramon, December 6.
Davis, R., and E. D'Albora. 2001. Influence of Expansion Device and Refrigerant Charge on the
Performance of a Residential Split-System Air Conditioner using R-410a Refrigerant. Report No.:
491-01.7. San Francisco, Calif.: Pacific Gas and Electric.
Hairrell, A. 2010. “Laboratory Report on Refrigerant Line Temperature Probes.” Proctor Engineering
Group.
Hunt, M., K. Heinemeier, M. Hoeschele, and E. Weitzel. 2010. “HVAC Energy Efficiency Maintenance
Study.” Report to SCE. CALMAC SCE0293.01.
KEMA. 2010. “Evaluation Measurement and Verification of the California Public Utilities Commission
HVAC High Impact Measures and Specialized Commercial Contract Group Programs, 2006-2008
Program Year”, Volume 1, KEMA, CADMUS and Summit Blue, CPUC, February 10, and
Volume 2, Appendices.
Li, H. 2004. A Decoupling-based Unified Fault Detection and Diagnosis Approach for Packaged Air
Conditioners, Ph.D. Thesis, West Lafayette, IN: Purdue University.
Mowris R, E. Jones, and R. Eshom. 2012. “ Laboratory Measurements of HVAC Installation and
Maintenance Faults.” ASHRAE Transactions.
Mowris, R., and A. Blankenship. 2004, "Field Measurements of Air Conditioners With and Without
TXV's," ACEEE Summer Study Proceedings.
Mowris, R. 2010. Draft Report – Evaluation Measurement and Verification of Air Conditioner Quality
Maintenance Measures, prepared by VERIFIED, Incorporated, James J. Hirsch & Associates, and
Intertek Testing Services (North America) under the auspices of the California Public Utilities
Commission, April.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
164
ASHRAE Transactions
SA-12-C021
Laboratory Measurements of HVAC

Installation and Maintenance Faults
Robert Mowris, PE, MS Ean Jones, BS Robert Eshom, AS

Member ASHRAE Associate Member ASHRAE Associate Member ASHRAE
ABSTRACT HEADING
Residential and commercial HVAC consumption in the United States accounts for 30% of average summer peak-day electricity loads,
13% of total electricity use, and 44% of total natural gas use. A 2002 study published by the Hewlett Foundation indicates that improved
HVAC installation and maintenance represents one of the largest achievable opportunities for energy efficiency savings. Laboratory test results are
provided of a new 3-ton split-system 13-SEER air conditioner using R-22 refrigerant. Test conditions differ from those used to rate cooling
systems (ANSI/AHRI Standard 210/240) to match typical installations in California. The equipment was set up in three chambers to
simulate standard test conditions and more realistic hot attic conditions. Tests were conducted using both thermostatic expansion valve (TXV)
and piston metering device (non-TXV). Test results are provided for the following HVAC faults that occur due to installation and maintenance
deficiencies and degradation: low airflow, coil blockage, refrigerant under and over charge, duct leakage, ducts and equipment located in hot attic
conditions, improper TXV sensing bulb installation, non-condensables, and restrictions. Baseline tests using the “code tester” instead of forced air
unit (FAU) are only 2.9% less than the 13 SEER rating under the ARI 210/240 test conditions (TXV, ducts and equipment located
in conditioned space). When the FAU is used, however, the SEER is reduced by 8.1%. The hot attic tests reduce EER* by 11% and SEER*
by 23%. Moderate to severe non-condensables (0.3 to 1% by weight) reduce EER* and SEER* by 12 to 38% and increase electric power by 8
to 22%. Refrigerant restrictions reduce EER* and SEER* by 30 to 59%. The combination of multiple faults such as low airflow, undercharge,
duct leakage, and condenser coil blockage reduce EER* by 54% and SEER* by 67%. Laboratory tests are used to develop methods to
differentiate non-condensables from over charge or condenser coil blockage and restrictions from under charge and avoid false charge diagnostics.
This is not possible using standard charge procedures specified in the California building energy efficiency standards.
INTRODUCTION
Residential and commercial heating, ventilating, and air conditioning, (HVAC) consumption in the United States
accounts for 30% of average summer peak-day electricity loads, 13% of total electricity use, and 44% of total natural gas use
[USEIA 2009]. A 2002 study published by the Hewlett Foundation indicates that improved HVAC installation and
maintenance represents one of the largest economically achievable opportunities for energy efficiency savings [Rufo 2002].
This paper provides laboratory test results of a new 3-ton split-system 13-SEER air conditioner using R-22 refrigerant. Test
conditions differ from those used to rate cooling systems to match typical installations in California. The equipment was set
up in three chambers to simulate both ARI 210/240 indoor and outdoor conditions and more realistic hot attic conditions.
Laboratory test results are provided for HVAC faults that occur due to installation and maintenance deficiencies and
degradation. Tests were conducted using TXV and non-TXV piston metering devices. Test results are provided for the
following faults: uninsulated TXV sensing bulb, low airflow, ducts and equipment in hot attic conditions, evaporator coil
blockage, duct leakage, condenser coil blockage, improper refrigerant charge, non-condensables, and restrictions.
Robert Mowris is a professional engineer at Verified, Inc., Olympic Valley, CA. Ean Jones is an engineering associate at Verified, Inc., Reno,
NV. Robert Eshom is a NATE certified HVAC technician/trainer at Verified, Inc., Corona, CA.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

TEST EQUIPMENT AND LABORATORY SETUP
The tested split-system air conditioning equipment is a nominal 3-ton (36,000 Btu/hr) unit with a Seasonal Energy
Efficiency Rating (SEER) of 13 and an Energy Efficiency Rating (EER) of 11.2 when equipped with a hard shut-off (HS)
TXV, time delay relay (TDR), and R-22 refrigerant.1 A manifold with isolation valves was used to test the HS-TXV and
non-TXV (piston) expansion devices commonly found on older equipment. Outlet tubes from the three expansion devices
merged to supply liquid refrigerant to the evaporator coil. The outdoor unit consists of a condenser, compressor, and
condenser fan. The indoor unit consists of an evaporator coil, FAU, and appropriate supply and return ducts to connect the
unit to measurement equipment and the indoor chamber conditions. The equipment was manufactured in 2010, but is
currently unavailable in the United States due to the phase out of R-22 refrigerant.
Laboratory tests were conducted according to ARI Standard 210/240-2008 with modifications to obtain “application”
Energy Efficiency Ratio* (EER*) and Seasonal Energy Efficiency Ratio (SEER*) [ANSI/AHRI 2008].2 Test modifications
include locating ducts, evaporator, and FAU in hot attic conditions of 118°F dry-bulb and 78°F wet-bulb temperature to
simulate typical peak cooling applications in California and other hot climates where ducts are located in unconditioned
attics. Test modifications also included duct leakage on supply and return ducts, longer ducts with bends typical of field
installations, and line-set lengths of 25 and 50 feet between the condenser and evaporator. The unit was also tested with
equipment located in conditions of 80°F dry-bulb and 67°F wet-bulb (per the AHRI 210/240 test). The “application”
efficiency ratings are the combined equipment plus distribution system efficiency typical of California residential
applications, but not equivalent to published AHRI ratings. The air conditioning equipment was tested with its AHRI-rated
configuration including HS-TXV with copper sensing bulb/strap, 25-feet line set and “code tester” fan to verify the rating.
The unit was also tested with an HS-TXV with stainless-steel sensing bulb/strap and non-TXV. The tests were performed
to support energy efficiency programs promoting quality installation and maintenance [ANSI/ACCA 2007].
The tests were performed at an AHRI-certified laboratory located in the United States. The laboratory is used by
manufacturers to certify air conditioners and heat pumps for AHRI testing. The test facility consists of climate-controlled
indoor, outdoor, and hot attic chambers where ducts, evaporator, and FAU are located. The air conditioning equipment,
liquid-line filter drier, metering devices (TXV and non-TXV), sight glasses, and standard test equipment were assembled
and installed in the test chambers by laboratory technicians. Prior to charging with refrigerant, the system was pressurized
to 300 psig with nitrogen and held for 60 minutes to absorb moisture and check for leaks. After the nitrogen leak test, a
vacuum pump was used to evacuate the system to below 500 micron mercury (ƬHg) vacuum held for 30 minutes.
Baseline Tests
Table 1 provides a comparison of test findings to the AHRI rating for a typical residential installation. Test 318-2 was
performed with the code-tester fan (no FAU), HS-TXV with copper bulb/strap, and ducts and evaporator located in
conditioned space at 80°F dry-bulb and 67°F wet-bulb temperatures. Test 318-2 measured 11.3 EER and 12.63 SEER
which are within 0.5 to 2.9% of the 11.2 EER and 13 SEER rated values. Test 310 was performed with an HS-TXV with
stainless-steel bulb/strap and ducts, evaporator, and FAU in conditioned space causing a 4.7% reduction in EER*A and
8.5% reduction in SEER* compared to the AHRI rating. Test 303 was performed with the same equipment as test 310 but
with ducts, evaporator, and FAU in hot attic conditions causing a 15.8% reduction in EER*A and 29.2% reduction in
SEER* compared to the AHRI rating. Comparing Test 303 to 310 shows that locating equipment in hot attic conditions
reduces EER* by 11% and SEER* by 23%.

1
AHRI Rating is for the condenser and evaporator pair tested with no FAU, hard shut-off TXV, and time delay relay (TDR) that continues fan operation after
compressor turns off to recover latent cooling from evaporator and increase efficiency rating.
2
The ARI 210/240 EERA and EERB indoor air dry-bulb temperature is 80°F and the wet-bulb is 67°F. The EERA outdoor air dry-bulb is 95qF. The EERB
outdoor air dry-bulb is 82°F. The SEER “C” and “D” outdoor air dry-bulb is 82°F, indoor air dry-bulb is 80°F, and indoor air wet-bulb is 57°F.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
166
ASHRAE Transactions
Table 1. Baseline TXV Tests with Equipment in Conditioned Space and Hot Attic
EER*A
Duct EER*A SEER*
Description Capacity EER*A SEER* Test ID
Leakage Impact % Impact %
(Btuh)
AHRI Rating (no FAU, HS-TXV, TDR) 2% 33,800 11.2 13.0 n/a
Code tester, HS-TXV copper bulb/strap,
2% 35,607 11.3 0.5% 12.63 -2.9% 318-2
ducts/evap in conditioned space, no TDR
FAU, HS-TXV stainless-steel bulb/strap,
2% 35,030 10.7 -4.7% 11.9 -8.5% 310
ducts/evap in conditioned space, no TDR
Above + ducts/evap/FAU in 118F hot attic 2% 31,054 9.5 -15.8% 9.2 -29.2% 303
Baseline and multiple fault tests for a non-TXV expansion valve are presented in Table 2. Test 189-4 was performed
with ducts, evaporator, and FAU in the hot attic chamber with a 10.5% reduction in EER*A and 17.6% reduction in SEER*
compared to conditioned space efficiency (exclusive of duct leakage). Test 409 was performed with multiple faults including
25% low airflow 10% undercharge, 30% duct leakage, and 50% condenser coil blockage with a 53.5% reduction in EER*A
and 66.8% reduction in SEER*. The baseline for developing diagnostic tests for non-condensables is referenced to ducts,
evaporator, and FAU in hot attic conditions (TXV Test 303 and non-TXV Test 189-4).
Table 2. Baseline and Multiple Fault non-TXV Tests with Equipment in Conditioned Space and Hot Attic
EER*A
Duct EER*A SEER*
Description Capacity EER*A SEER* Test ID
Leakage Impact % Impact %
(Btuh)
FAU, non-TXV, no TDR, Ducts, Evap, FAU
2% 34,542 10.5 NA 10.8 NA 300
in Conditioned Space
Above + Ducts, Evap, FAU in Hot Attic
2% 31,050 9.4 -10.5% 8.9 -17.6% 189-4
118°F, 50' line set
Above + 25% low airflow, 10% undercharge,
30% 13,731 4.4 -53.5% 2.94 -66.8% 409
30% duct leakage, 50% condenser coil block
Thermostatic Expansion Valve Tests

TXV tests were conducted to evaluate sensing bulb insulation. Table 3 provides insulated and uninsulated TXV
sensing bulb test results with correct charge of 102 ounces and 10% to 40% overcharge with 50 feet line set and equipment
located in hot attic conditions. The manufacturer target subcooling (SC) is 7°F.3 The uninsulated TXV sensing bulb causes
improper metering of refrigerant and false diagnostics for Tests 22, 35, 38, and 39. The refrigerant charge diagnostic for
Test 42 (+40% charge) is a “slight overcharge.” For proper operation the TXV sensing bulb should be installed at the
correct orientation with one copper strap and one wrap of R-1 closed-cell foam insulation.
Table 3. Insulated and Uninsulated TXV Sensing Bulb Tests with Equipment in Hot Attic Conditions
EER*A
EER*A Actual SC Refrigerant Charge
Description Capacity EER*A Test ID
Impact % (°F) Diagnostic
(Btuh)
TXV R-1 Insulation, Correct Charge, 102 oz. 31,420 9.60 NA 6.3 Correct Charge 23
TXV Uninsulated, Correct Charge, 102 oz. 30,873 9.41 NA 1.5 False Undercharge 22
TXV Uninsulated, +10% Charge, 112.2 oz 30,237 9.19 -2.4% 3.6 False Undercharge 35
TXV Uninsulated, +20% Charge, 122.4 oz 30,075 8.97 -4.7% 3.7 False Undercharge 38
TXV Uninsulated, +30% Charge, 132.6 oz 28,676 8.44 -10.3% 6.2 False Correct Charge 39
TXV Uninsulated, +40% Charge, 142.8 oz 27,754 7.99 -15.1% 13.7 Slight Overcharge 42

3
Target subcooling is based on manufacturer data. Subcooling measures the heat removed from refrigerant after it changes to liquid and is defined as the
difference between condenser saturation temperature and liquid line temperature. Delta subcooling is the difference between actual and target subcooling.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

Airflow Tests
Airflow test results are shown in Table 4 for the non-TXV unit with equipment in hot attic conditions. All tests are
performed with proper refrigerant charge of 108 ounces. Typical in situ airflow is 250 to 350 cfm/ton which reduces
EER*A by 3.2 to 11.6%. The non-TXV tests demonstrated that low airflow down to 250 cfm/ton does not cause false
undercharge diagnostics (i.e., delta SH within +/-5°F).4 However, airflow below 250 cfm/ton can cause false over charge
diagnostics due to icing of the indoor coil which will block airflow.
Table 4. Low Airflow Impact on EER* (non-TXV) and Hot Attic Conditions
EER*A
Airflow EER*A Delta SH Actual SC
cfm/ton Impact % (°F) (°F)
(Btuh)
Base Case Correct Airflow 31,302 391 9.49 NA -1.4 9.7 53
10% Low Airflow 29,501 351 9.19 -3.2% 3.2 9.2 64
23% Low Airflow 28,538 302 9.04 -4.7% 0.7 9.4 65
36% Low Airflow 26,174 250 8.39 -11.6% -4.4 8.7 66
Evaporator Coil Blockage Tests
Table 5 provides test results for evaporator coil blockage for the non-TXV with equipment in conditioned space.5
With refrigerant charge of 78.4 ounces and 50% evaporator coil blockage, airflow is reduced by 16%, EER* is reduced by
4.6%, and delta superheat (SH) is outside the acceptable range of +/-5°F. These non-TXV test results show that 50%
evaporator coil blockage causes false overcharge diagnostics (i.e., delta SH less than -5°F). With an 8 ounce charge removal
(total charge of 70.4 oz.), delta SH is increased to -1.2°F which is within +/-5°F, but the EER* is reduced by 7.2%.
Table 5. Evaporator Coil Blockage Impact on EER* (non-TXV) with Equipment in Conditioned Space
EER*A
Airflow EER*A Delta SH Actual SC
(Btuh)
Base non-TXV 78.4 oz. 33,652 400 10.40 NA 1.1 7.3 8
50% Evap. Coil Block 78.4 oz. 31,281 335 9.92 -4.6% -11.1 6.2 10
50% Evap. Coil Block 70.4 oz. 30,531 336 9.66 -7.2% -1.2 5.4 11
Table 6 provides evaporator coil blockage test results for the HS-TXV unit. With 50% coil blockage airflow is
reduced by 15% and EER* is reduced by 4.2%.
Table 6. Evaporator Coil Blockage Impact on EER* (HS-TXV) with Equipment in Conditioned Space
EER*A
Airflow EER*A Delta SC Actual SC
(Btuh)
Base TXV 86.4 oz. charge 34,205 395 10.24 NA 2.8 9.8 1
50% Evap. Coil Block 86.4 oz. 31,239 335 9.81 -4.2% 0 7.0 12
Duct Leakage Tests

Table 7 provides duct leakage test results for the non-TXV system with ducts, evaporator, and FAU located in hot
attic conditions maintained at 118°F. The base 2% duct leakage is measured at 25 Pascal as the percentage of total system

4
Target superheat is based on return air wetbulb and condenser entering air drybulb. Delta superheat is the difference between actual and target superheat.
Superheat measures the heat added to refrigerant after it changes to vapor and is defined as the difference between suction line temperature and evaporator
saturation temperature.
5
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Tests in conditioned space were performed with 25 feet line-set requiring less refrigerant charge than tests in hot attic conditions with 50 feet line-set.
Evaporator coil was blocked with plastic corrugated cardboard covering 50% of the upstream side of the evaporator cross-sectional area.
168
ASHRAE Transactions
airflow of 300 cfm/ton per the modified test procedure. The 6% duct leakage is per California Energy Commission
building energy efficiency standards (CEC 2008). The 15% duct leakage is per Energy Star and the 30% duct leakage is
common of many US homes [Parker 1998, Roberts 2010, USEPA 2011]. Tests were performed with 50% of duct leakage
on the return plenum and 50% on the supply plenum. Modified ARI 210/240 tests were conducted to obtain the following
unit plus distribution system application efficiencies: EER*A, EER*B, and SEER*.
Table 7. Duct Leakage Impacts on EER* and SEER* for non-TXV and Hot Attic Conditions
EER*A EER*A EER*B SEER*
Description Capacity EER*A Impact EER*B Impact SEER* Impact Test ID
(Btu/hr) % % %
Base 2% Duct Leakage 28,670 9.27 NA 10.62 NA 9.17 NA 400
6% Duct Leakage 26,804 8.65 -6.7% 9.81 -7.6% 7.93 -13.5% 401
15% Duct Leakage 23,708 7.59 -18.1% 8.47 -20.2% 6.95 -24.3% 402
30% Duct Leakage 18,736 5.96 -35.7% 6.5 -38.8% 5.33 -41.8% 403
Condenser Coil Blockage Tests

Condenser coil blockage test results are shown in Table 8. Condenser coil blockage impacts EER*A by -5.8% to -
29.7% and increases total air conditioner power by 6.9% to 26.6%.
Table 8. Condenser Coil Blockage Impact on EER* (non-TXV) and Hot Attic Conditions
EER*A
EER*A
Description Capacity EER*A kWA kWA Impact % Test ID
Impact %
(Btu/hr)
Base Clean Condenser 32,335 9.82 NA 3.292 NA 189-2
50% Condenser Coil Blockage 31439 8.94 -5.8% 3.52 6.9% 191-2
80% Condenser Coil Blockage 27,806 6.67 -29.7% 4.168 26.6% 190-2
Refrigerant Charge Tests

Table 9 provides refrigerant charge test results for the HS-TXV unit and hot attic conditions of 118°F dry bulb and
78°F wet bulb temperatures. The HS-TXV EER* performance is severely impacted by under charge. The condenser over
ambient (COA) is defined as the condenser saturation temperature minus ambient air temperature.
Table 9. Refrigerant Charge Impacts on EER* and kWA for HS-TXV and Hot Attic Conditions
EER*A EER*A
kWA
Description Capacity EER*A Impact kW*A COA (°F) Test ID
Impact %
(Btu/hr) %
Baseline Charge 102 oz. 31,420 9.60 NA 3.28 NA 14.6 23
+10% Charge 112.2 oz. 31,796 9.24 -3.7% 3.44 5.0% 22.4 36
+20% Charge 122.4 oz. 31,730 8.84 -7.9% 3.59 9.6% 27.8 37
+30% Charge 132.6 oz. 31,321 8.12 -15.4% 3.86 17.8% 31.6 40
+40% Charge 142.9 oz. 30,796 8.01 -16.5% 3.84 17.3% 34.3 41
-10% Charge 91.8 oz. 29,200 9.22 -3.9% 3.17 -3.3% 10.7 45
-20% Charge 81.6 oz. 25,826 8.28 -13.8% 3.12 -4.7% 9.1 48
-30% Charge 71.4 oz. 21,170 6.98 -27.3% 3.03 -7.4% 6.7 49
-40% Charge 61.2 oz. 12,242 4.24 -55.9% 2.89 -11.8% 2.6 52
Table 10 provides refrigerant charge test results for the non-TXV piston and hot attic conditions of 118°F dry bulb
and 78°F wet bulb temperatures. Non-TXV EER* performance is severely impacted by undercharge similar to the TXV.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

Table 10. Refrigerant Charge Impacts on EER* and kWA for non-TXV and Hot Attic Conditions
EER*A EER*A
kWA
Description Capacity EER*A Impact kW*A COA (°F) Test ID
Impact %
(Btu/hr) %
Baseline Charge 108.2 oz. 31,302 9.49 NA 3.30 NA 17.7 53
+10% Charge 119 oz. 31,600 9.29 -1.3% 3.40 3.9% 20.1 60a
+20% Charge 129.8 oz. 31,013 9.08 -3.5% 3.42 4.3% 20.4 61a
+30% Charge 140.6 oz. 30,765 8.91 -5.3% 3.45 5.5% 21 62
+40% Charge 151.5 oz. 30,526 8.57 -8.9% 3.56 8.7% 21.9 63
-10% Charge 97.4 oz. 23,352 7.47 -21.3% 3.13 -5.2% 10.1 55
-20% Charge 86.6 oz. 20,371 6.70 -29.4% 3.04 -7.8% 8.4 56a
-30% Charge 75.8 oz. 12,629 4.30 -54.7% 2.94 -11.0% 4.7 57a
-40% Charge 64.9 oz. 9,186 3.20 -66.3% 2.87 -13.0% 2.9 58
Non Condensable Tests

If proper vacuum is not achieved at installation the refrigerant system will be contaminated with non condensable air
and water vapor which can cause compressor failure. Non condensables decrease condenser heat transfer and cooling
capacity and increase condenser pressure and power input. Table 11 provides laboratory test results for 0.3 oz (~0.3% of
system charge) of non-condensable nitrogen on the unit operating with the HS-TXV and hot attic conditions. The impact is
-12.2% for the EER*A, +6.13% for kW*A, and -13.4% for SEER*.
Table 11. Non Condensable Impacts on EER* and SEER* for TXV and Hot Attic Conditions
EER*A EER*A kWA SEER*
Description Capacity EER A * Impact kW*A Impact SEER * Impact Test ID
(kBtuh) % % %
Baseline 31,054 9.48 NA 3.28 NA 9.21 NA 303
0.3% non-condensable 27,373 8.27 -12.2% 3.48 6.13% 7.98 -13.4% 505
Table 12 provides results for 0.3 oz (~0.3% system charge) of non-condensable nitrogen on the non-TXV unit. The
impact is -18.2% for EER*A, +7.6% for kW*A, and -18.5% for SEER*. The impact of ~1% non-condensables (Test 501X)
is -37.7% for the EER*A test and 21.6% kWA increase. Most field installations are evacuated using a time-based procedure
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
without a vacuum pressure gauge. Many vacuum pumps have contaminated oil. Changing oil after every evacuation is
required to achieve proper evacuations (JB 2007). Performing a deep vacuum to 240 ƬHg held at or below 500 ƬHg for 30
minutes will remove non condensables from the air conditioning system (ASHRAE 2010).
Table 12. Non Condensable Impacts on EER* and SEER* for non-TXV and Hot Attic Conditions
EER*A EER*A SEER*
* kWA Impact *
Description Capacity EER A Impact kW*A SEER Impact Test ID
%
(kBtuh) % %
Baseline 31,050 9.42 NA 3.30 NA 8.86 NA 189-4
0.3% non-condensable 27,373 7.71 -18.2% 3.55 +7.6% 7.22 -18.5% 501
~1% non-condensable 20,486 5.87 -37.7% 4.01 +21.6% 501X
Refrigerant Restriction Tests

Moisture, copper particles, flux and/or brazing residue, and particulates left inside the system damage the compressor,
clog metering devices, or make the metering device function improperly. Liquid line filter driers are recommended to
remove moisture, acid, and particulates to prevent refrigerant restrictions on field-charged split systems (Carrier 2010,
Lennox 2008). Table 13 provides laboratory test results for refrigerant restrictions on the non-TXV unit (introduced with
an adjustable valve on the liquid line). The impact is -29.7% for EER*A, -45.4% for EER*B, and -35.4% for SEER*. Power
decreased by 100 W, or 3% similar to an under-charge.
170
ASHRAE Transactions
Table 13. Refrigerant Restriction Impacts on EER* and SEER* for non-TXV and Hot Attic Conditions
EER*A
EER*A EER*B SEER*
Description Capacity EER*A EER*B SEER* Test ID
Impact % Impact % Impact %
(kBtuh)
Base no restriction non-
TXV 32,759 9.42 NA 10.64 NA 8.86 NA 189
Restriction non-TXV 22,385 6.62 -29.7% 5.81 -45.4% 5.72 -35.4% 701
Table 14 provides test results for restrictions on the TXV unit. The impact is -36.1% for EER*A, -54.9% for EER*B,
and -59% for SEER*.
Table 14. Refrigerant Restriction Impacts on EER* and SEER* for TXV and Hot Attic Conditions
EER*A
EER*A EER*B SEER*
Description Capacity EER*A EER*B SEER* Test ID
Impact % Impact % Impact %
(kBtuh)
Base no restriction TXV 32,764 9.48 NA 11.14 NA 9.21 NA 303
Refrig. Restriction TXV 19,812 6.06 -36.1% 5.02 -54.9% 3.78 -59% 801
If refrigerant restrictions are present in the system and an acid test indicates sludge caused the restrictions, then
standard cleanup procedures must be followed using oversize suction and liquid line filter driers specified to remove sludge.
Vacuum pumps are not designed to remove sludge. If a split system older than 10 years contains sludge, then it might be
more cost effective to install a new system with new filter drier and clean or replace the line set.
Refrigerant Charge Diagnostic Test Matrix

The California Energy Commission (CEC) Standard Charge Measurement Procedure requires verification of
subcooling for TXV units or verification of superheat for non-TXV units (CEC 2008). These simple tests can only be used
to diagnose refrigerant charge faults if no other faults are present. Laboratory test data are used to develop comprehensive
refrigerant charge diagnostic test procedures that include superheat, subcooling, EST (evaporator saturation temperature),
and COA (condenser saturation minus ambient temperature). Table 15 provides an example refrigerant charge diagnostic
test matrix.
Table 15. Refrigerant Charge Diagnostic Test Matrix
Low Superheat High Superheat
Refrigerant under charge or liquid line
restriction SH, SC and low EST thresholds
Low Subcooling TXV or TXV sensing bulb failure
g(OAT, SH, SC) or compressor valve failure
(check compressor Watts)
Non-condensables check SH, SC and high COA
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Refrigerant over charge, blocked condenser coil,
low airflow, or non condensables check SH, SC thresholds g(OAT, SH, SC) or liquid line
High Subcooling
restriction check SH, SC and low EST
and high COA thresholds g(OAT, SH, SC)
thresholds g(OAT, SH, SC)
The diagnostic matrix can be used to differentiate non-condensables from overcharge, blocked condenser, or low
airflow, and to differentiate refrigerant restrictions from refrigerant undercharge conditions or TXV failure. If TXV failure
is detected then technicians should verify whether or not the sensing bulb is properly attached to the suction line in the
correct orientation and tightly strapped with at least one copper strap and one wrap of closed-cell foam insulation with 50%
overlap. Copper is 4 to 10 times more thermally conductive than brass or stainless steel.
$6+5$(

CONCLUSION
The tested unit efficiency is within +0.5% of 11.2 EER AHRI rating and within -2.9% of the 13 SEER AHRI rating.
Without considering duct leakage, hot attic conditions reduce EER*A by 11% and SEER* by 23%. These differences are
typical since current California building efficiency standards do not include component and installation differences which
cause lower efficiencies. Other common installation deficiencies such as duct leakage, undercharge, low airflow, non
condensables, or refrigerant restrictions cause lower operating efficiencies. The combination of multiple deficiencies such as
low airflow, undercharge, duct leakage, and condenser coil blockage can reduce EER* by 54% and SEER* by 67%. Low
airflow reduces EER* by 3 to 12%. Tests of uninsulated TXV sensing bulb installation indicate failure to properly meter
refrigerant with correct charge or over charge causing false diagnostics. Evaporator coil blockage of 50% reduces EER* by
5 to 7%. Duct leakage reduces EER* by 7% to 39% and SEER* by 14% to 42%. Condenser coil blockage of 50 to 80%
reduces EER* by 6 to -30% and increases power use by 7 to 27%. Improper refrigerant charge reduces EER* by 1 to 66%
and SEER* by 2 to 56%. Moderate non-condensables (0.3%) reduces EER* by 12 to 18% and SEER* by 13 to 19% and
increase power use by 6 to 8%. Severe non-condensables (1%) reduces EER* by 38% and increases power use by 22%.
Liquid line refrigerant restrictions reduce EER* by 30 to 55% and SEER* by 35% to 59%. Liquid line filter driers are
required to remove moisture, acid, and particulates to prevent refrigerant restrictions on field-charged split systems.
Laboratory test data are used to develop methods to differentiate non-condensables from over charge or condenser coil
blockage and restrictions from under charge and avoid false charge diagnostics. This is not possible using standard charge
measurement procedures specified in the California building energy efficiency standards.
ACKNOWLEDGMENTS
Laboratory tests were funded by James J. Hirsch & Associates and Verified, Inc., under the auspices of the California
Public Utilities Commission. Thanks to John Hill, PhD, Rocky Bacchus, Anne Premo, CPUC, and Peter Jacobs.
REFERENCES
American National Standards Institute (ANSI)/Air Conditioning Contractors of America (ACCA). 2007. ANSI/ACCA
Standard 5: HVAC Quality Installation Specification and Standard 4: Maintenance of Residential HVAC Systems.
ANSI/AHRI. 2008 Standard for Performance Rating of Unitary Air-Conditioning and Air-Source Heat Pump Equipment
Standard 210/240.
ASHRAE. 2010. ASHRAE Handbook-Refrigeration. Page 8.2. Table 1. American Society of Heating Refrigeration and Air
California Energy Commission (CEC). 2008. Reference Appendices for the 2008 Building Energy Efficiency Standards for
Residential and Nonresidential Buildings. Appendix RA3 - Residential Field Verification and Diagnostic Test Protocols.
Carrier Corporation. 2010. Commercial Packaged Engineering Standard Work Procedure: System Evacuation and
Dehydration. Carrier A United Technologies Company.
JB Industries. Deep Vacuum: Its Principle and Application. 2007. www.jbind.com.
Lennox Industries, Inc., 2008. Application and Design Guidelines: Lennox Refrigerant Piping Design and Fabrication
Guidelines. See page 2. One. Corp. 9351îL9. February 25, 2008. www.lennox.com.
Parker, D., Sherwin, J., "Comparative Summer Attic Thermal Performance of Six Roof Constructions," The 1998
ASHRAE Annual Meeting, Toronto, Canada, June 20-24, 1998.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Roberts, D., Winkler, J. 2010. Ducts in the Attic? What Were They Thinking? National Renewable Energy Laboratory, 2010
ACEEE Summer Study on Energy Efficiency in Buildings. American Council for an Energy-Efficient Economy.
Rufo, M., Coito F. 2002. California’s Secret Energy Surplus: The Potential for Energy Efficiency. Prepared by Xenergy, Inc.
Prepared for the Hewlett Foundation. http://www.p-2.com/PEERS/Hewlett-Foundation-Report-9-23-02.pdf.
United States Energy Information Agency (USEIA). 2009. Residential Energy Consumption Survey. Washington, D.C.:
http://205.254.135.7/consumption/residential/data/2009/.
United States Environmental Protection Agency (USEPA). 2011. Duct Sealing Brochure. Available online:
http://www.energystar.gov/ia/products/heat_cool/ducts/DuctSealingBrochure04.pdf.
172
ASHRAE Transactions
SA-12-C022
(QHUJ\0RGHOLQJ%DVLFV$%UDYH1HZ
:RUOGIRU<RXQJ%XLOGLQJ0RGHOLQJ
3URIHVVLRQDOV

-DUHG$+LJJLQV3(
$6+5$(0HPEHU
$%675$&7
:LWKVHYHUDOEXLOGLQJSURJUDPVUHTXLULQJYDOLGDWLRQLQGHVLJQHQHUJ\XVHSULRUWRFRQVWUXFWLRQHQHUJ\
PRGHOLQJSURIHVVLRQDOVNLOOVHWVDUHFXUUHQWO\LQGHPDQGQRZPRUHWKDQHYHU+RZHYHUVHYHUDO\RXQJFROOHJH
JUDGXDWHVWKDWZLOOHQWHUWKHHQHUJ\PRGHOLQJILHOGZLOOKDYHUHFHLYHGOLWWOHWRQRWUDLQLQJSULRUWREHJLQQLQJWKHLU
FDUHHU%HIRUHWKHVH\RXQJSURIHVVLRQDOVMXPSLQWRDGYDQFHGDQGLQWULFDWHSURMHFWVWKH\ZLOOQHHGWROHDUQWKH
IXQGDPHQWDOVRIFUHDWLQJPDQLSXODWLQJDQGILQDOL]LQJHQHUJ\PRGHOLQJVNLOOVLQWRDSHUIHFWHGFUDIW7KHVHVNLOOV
LQFOXGHXQGHUVWDQGLQJHQYHORSHFRPSRQHQWFDOFXODWLRQVV\VWHPVL]LQJLQWHUQDODQGSURFHVVORDGVEXLOGLQJXVHU
RSHUDWLQJSURILOHVDQGHQHUJ\UDWHVWUXFWXUHVDVZHOODVJHQHUDOHQHUJ\PRGHOVWDUWXSUHTXLUHPHQWV%\
XQGHUVWDQGLQJWKHEDVLFVRIWKHHQHUJ\PRGHOLQJFRQFHSWQHZEXLOGLQJHQHUJ\PRGHOLQJSURIHVVLRQDOVZLOOJDLQ
H[SHULHQFHDQGFRQILGHQFHLQWRDEUDYHQHZPDUNHWVHFWRU

,1752'8&7,21

2YHUWKHSDVWWZRGHFDGHVHQHUJ\PRGHOLQJKDVEHFRPHDSRSXODUDQGHIIHFWLYHPHWKRGWRSUHGLFWWKHHQHUJ\
XVDJH LQ EXLOGLQJV ,Q DGGLWLRQ WR FRQVWUXFWLQJ WKH HQHUJ\ PRGHO EDVHG RQ GHVLJQ FKDUDFWHULVWLFV VHYHUDO NH\
DVVXPSWLRQV KDYH WR EH PDGH WKURXJKRXW DQ HQHUJ\ PRGHO WR DWWHPSW WR DFFXUDWHO\ SUHGLFW WKLV XVDJH VXFK DV
EXLOGLQJ RSHUDWLRQ PDLQWHQDQFH FOLPDWH DQG DQWLFLSDWHG RFFXSDQW EHKDYLRU %\ ZRUNLQJ FORVHO\ ZLWK WKH GHVLJQ
WHDPDQGEXLOGLQJRZQHUWKURXJKRXWWKHSURMHFWWKHHQHUJ\PRGHOLQJSURIHVVLRQDOZLOOEHDEOHWRGHYHORSDSUHFLVH
HQHUJ\ PRGHO FDSDEOH RI HVWLPDWLQJ WKH EXLOGLQJ¶V HQHUJ\ XVDJH )XUWKHUPRUH E\ VWDUWLQJ WKH PRGHOLQJ SURFHVV
HDUO\LQWKHGHVLJQSKDVHVHYHUDONH\GHFLVLRQVDIIHFWLQJWKHEXLOGLQJ¶VHQHUJ\XVDJHFDQEHPDGHWKDWZLOOKDYHOHVV
RIDGHVLJQLPSDFWWKDQLIPDGHODWHULQWKHGHVLJQSURFHVV
7KHWKUHHPDLQXVHVIRUHQHUJ\PRGHOLQJDUHIRUFRGHFRPSOLDQFHGHVLJQSHUIRUPDQFHDQGPHDVXUHPHQWDQG
YHULILFDWLRQ(DFKRIWKHVHXVHVUHTXLUHGLIIHUHQWOHYHOVRIHQHUJ\PRGHOLQJDQGDOVRDUHVWDUWHGDWGLIIHUHQWWLPHVRI
WKH GHVLJQ SKDVH &RGH FRPSOLDQFH HQHUJ\ PRGHOLQJ XVHV D ZKROHEXLOGLQJ HQHUJ\ PRGHO WR YHULI\ WKH DFWXDO
GHVLJQHGEXLOGLQJ¶VHQHUJ\XVHFRPSDUHGWRDSUHGHWHUPLQHGEDVHOLQHEXLOGLQJ¶VHQHUJ\ XVHLQRUGHUWRGHWHUPLQH
WKDW PLQLPXP FRGH UHTXLUHPHQWV ZHUH PHW 7KLV PHWKRG LV W\SLFDOO\ SHUIRUPHG DW WKH HQG RI WKH GHVLJQ SKDVH
KRZHYHU LW PD\ EH SHUIRUPHG HDUOLHU LQ WKH SURMHFW WR GHWHUPLQH LI DQ\ HQHUJ\ FRQVHUYDWLRQ PHDVXUHV QHHG WR EH
LQFOXGHG WR KHOS ORZHU HQHUJ\ XVDJH 'HVLJQ SHUIRUPDQFH HQHUJ\ PRGHOLQJ LV SHUIRUPHG WKURXJKRXW WKH GHVLJQ
SKDVH RI D SURMHFW LQ RUGHU WR GHWHUPLQH HQHUJ\ XVH RI D YDULHW\ RI SDUDPHWHUV DSSOLHG WR WKH EXLOGLQJ 7\SLFDOO\
-DUHG$+LJJLQV3(LVDFRQVXOWLQJHQJLQHHUZLWK3DUNKLOO6PLWK &RRSHU,QF/XEERFN7H[DV
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

PXOWLSOH DOWHUQDWLYHV DUH FUHDWHG WR FRPSDUH VHOHFW EXLOGLQJ V\VWHPV DQG FRPSRQHQWV LQ RUGHU WR GHWHUPLQH WKH
DVVRFLDWHGDPRXQWRIHQHUJ\UHGXFWLRQZLWKHDFKDOWHUQDWLYHWRDVVLVWLQWKHGHFLVLRQRILQFRUSRUDWLQJWKHDOWHUQDWLYH
LQWR WKH SURMHFW (YHQ WKRXJK WKLV PHWKRG FDQ XVH OHVV FRPSOH[ PRGHOV DQG KDYH D TXLFNHU DQDO\VLV WLPH LW LV
UHFRPPHQGHG WR VWDUW WKLV W\SH RI HQHUJ\ PRGHO DV HDUO\ DV SRVVLEOH (QHUJ\ PRGHOV IRU PHDVXUHPHQW DQG
YHULILFDWLRQ DUH FRQVWUXFWHG DIWHU D EXLOGLQJ KDV EHHQ RFFXSLHG 7KH PRGHO LV FRQVWUXFWHG EDVHG RQ WKH DFWXDO
HTXLSPHQWDQGEXLOGLQJV\VWHPVLQVWDOOHGDQGWKHDQDO\WLFDOUHVXOWVDUHFRPSDUHGWRWKHDFWXDOPHDVXUHGXWLOLW\GDWD
7KLVW\SHRIPRGHOLQJPXVWEHYHU\ILQHWXQHGLQRUGHUWRDFFXUDWHO\SUHGLFWWKHDFWXDOHQHUJ\XVHIRUWKHEXLOGLQJ
RZQHUWRYHULI\WKDWWKHEXLOGLQJV\VWHPVDUHZRUNLQJSURSHUO\
7KHUHDUHPDQ\UHVRXUFHVWKDWZLOODVVLVWDQHZHQHUJ\PRGHOLQJSURIHVVLRQDOLQGHWHUPLQLQJWKHEHVWFRXUVHRI
DFWLRQLQFUHDWLQJDQHQHUJ\PRGHO7KHUHDUHMXVWDVPDQ\HQHUJ\PRGHOLQJSURJUDPVWKDWKDYHEHHQWHVWHGDQG
KDYH EHHQ DFFHSWHG E\ UDWLQJ DXWKRULWLHV 7\SLFDOO\ WKH QHZ PRGHOLQJ SURIHVVLRQDO ZLOO OHDUQ IURP WKH ILUP¶V
FXUUHQW HQHUJ\ PRGHOHUV DQG ZLOO XVH WKH SURJUDP WKDW WKH ILUP SURYLGHV 7KHUH LV D ODUJH DPRXQW RI RQ WKH MRE
WUDLQLQJUHTXLUHGKRZHYHUWKLVLVDQH[FLWLQJILHOGWRVWDUWDFDUHHUE\GHWHUPLQLQJLQIRUPDWLRQWKDWZLOOFKDQJHWKH
IXWXUH

&5($7,1*7+((1(5*<02'(/

*HWWLQJVWDUWHGRQWKHHQHUJ\PRGHOUHTXLUHVDWWHQWLRQWRIRXUPDLQDUHDVFRQVWUXFWLQJWKHEXLOGLQJHQYHORSH
GHWHUPLQLQJ LQWHUQDO DQG SURFHVV ORDGV +9$& V\VWHP VL]LQJ DQG XVHU RSHUDWLQJ SURILOHV (DFK DUHD UHTXLUHV
H[WHQVLYH GHWDLO DQG GLVFXVVLRQ ZLWK WKH GHVLJQ WHDP RU EXLOGLQJ RZQHU LQ RUGHU WR FRUUHFWO\ FRQVWUXFW WKH HQHUJ\
PRGHO+RZHYHU PDQ\ PRGHOLQJVRIWZDUHSURJUDPV KDYHJXLGHVDQG VDPSOHV WR KHOS JHW VWDUWHG RQ WKH SURFHVV
0RGHOLQJ VRIWZDUHSURJUDPV DOVRW\SLFDOO\ KDYHVHYHUDOSUHVHOHFWHGUHJLRQDOZHDWKHUILOHVLQFOXGHGKRZHYHULID
PRUHGLVWLQFWZHDWKHUILOHLVUHTXLUHGLWPD\KDYHWREHUHWULHYHGIURPDSSURYHGZHEVLWHV

&RQVWUXFWLQJWKH%XLOGLQJ(QYHORSH

7KHILUVWVWHSLQFUHDWLQJDQHQHUJ\PRGHOLVWRDFWXDOO\FRQVWUXFWWKHEXLOGLQJLQWKHPRGHOLQJVRIWZDUH7KLV
ZLOO UHTXLUH LQSXWWLQJ SDUDPHWHUV RI WKH RYHUDOO VTXDUH IRRWDJH ZDOO KHLJKWV JODVV H[SRVXUH IORRU WR IORRU KHLJKW
LQIRUPDWLRQDQGRWKHUDGGLWLRQDOIDFWRUV7KHFULWLFDOHOHPHQWLQWKLVDVSHFWLVGHWHUPLQLQJWKHRYHUDOOKHDWWUDQVIHU
UDWHRIWKHEXLOGLQJVXUIDFHVWKDWKDYHDQH[WHULRUH[SRVXUH2QHRIWKHEHVWSODFHVWRILQGLQIRUPDWLRQRQKRZWR
GHWHUPLQHKHDWORDGFDOFXODWLRQVLQEXLOGLQJDVVHPEOLHVLV&KDSWHU+HDW$LUDQG0RLVWXUH&RQWUROLQ%XLOGLQJ
$VVHPEOLHV±)XQGDPHQWDOVLQWKH$6+5$()XQGDPHQWDOV+DQGERRN7KLVFKDSWHUSURYLGHVGHWDLOHGJXLGDQFHLQ
GHWHUPLQLQJKRZWRFDOFXODWHD8IDFWRURIDEXLOGLQJFRPSRQHQW,QRUGHUWRDFFRPSOLVKWKLVWKHPRGHOHUZLOOKDYH
WRZRUNZLWKWKHEXLOGLQJGHVLJQDUFKLWHFWWRREWDLQZDOODQGURRIEXLOGLQJVHFWLRQVDVZHOODVVSHFLILFLQIRUPDWLRQ
RQWKHW\SHRIIHQHVWUDWLRQVXVHG
2QFHDOOWKHPDWHULDOVRIDQDVVHPEO\KDYHEHHQGHWHUPLQHG&KDSWHU+HDW$LUDQG0RLVWXUH&RQWUROLQ
%XLOGLQJ $VVHPEOLHV ± 0DWHULDO 3URSHUWLHV RI WKH $6+5$( )XQGDPHQWDOV +DQGERRN SURYLGHV VHYHUDO WDEOHV
FRQWDLQLQJ UHVLVWDQFH DQG FRQGXFWDQFH LQIRUPDWLRQ IRU VHOHFW PDWHULDOV 7DEOHV DQG RI &KDSWHU LQFOXGH
LQIRUPDWLRQ RQ GHWHUPLQLQJ UHVLVWLYLW\ DQG HPLVVLYLW\ RI DLU VSDFHV 7DEOH LQFOXGHV LQIRUPDWLRQ RQ UHVLVWDQFH
FRQGXFWLYLW\ DQG VSHFLILF KHDW IRU VHYHUDO GLIIHUHQW FRPPRQ EXLOGLQJ DQG LQVXODWLQJ PDWHULDOV ,I DQ H[DPSOH LV
QHHGHGWRGHWHUPLQHWKHEHVW FRXUVHRIDFWLRQLQFDOFXODWLQJWKLV LQIRUPDWLRQ &KDSWHU +HDW$LUDQG0RLVWXUH
&RQWUROLQ%XLOGLQJ$VVHPEOLHV±([DPSOHVRIWKH$6+5$()XQGDPHQWDOV+DQGERRNSURYLGHVVHYHUDOGLIIHUHQW8
IDFWRUFDOFXODWLRQV,QDGGLWLRQWRWKH$6+5$()XQGDPHQWDOV+DQGERRN6WDQGDUGSURYLGHVVHYHUDOWDEOHVRI
DVVHPEO\8IDFWRUVIRUFRPPRQW\SHVRIURRIVZDOOVIORRUVDQGJOD]LQJLQ$SSHQGL[$$OWKRXJKWKLVLQIRUPDWLRQ
PD\ DOUHDG\ EH GHWHUPLQHG WKH HQHUJ\ PRGHOHU PD\ EH FRQVXOWHG WR SHUIRUP VHYHUDO PRGHOLQJ VFHQDULRV WR
GHWHUPLQHWKHFRVWHIIHFWLYHQHVVRIDGGLQJLQVXODWLRQWRWKHEXLOGLQJHQYHORSHRULQVWDOOLQJPRUHHIILFLHQWJOD]LQJ

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
174
ASHRAE Transactions
'HWHUPLQLQJ,QWHUQDODQG3URFHVV/RDGV

:LWKWKHEXLOGLQJ IORRUSODQDQGHQYHORSHHQWHUHGLQWRWKHPRGHOLQJ VRIWZDUH WR GHWHUPLQH WKH H[WHUQDO KHDW
ORDGV WKH QH[W SKDVH RI FRQVWUXFWLQJ WKH PRGHO ZLOO EH WR GHWHUPLQH ZKDW LQWHUQDO KHDW ORDGV DUH LQ WKH EXLOGLQJ
,QWHUQDOORDGVDUHSULPDULO\FRPSULVHGRISHRSOHOLJKWLQJDQGHTXLSPHQW,QRUGHUWRREWDLQWKLVLQIRUPDWLRQWKH
EXLOGLQJRZQHUDQGHOHFWULFDOGHVLJQHQJLQHHUZLOOEHFUXFLDO
7KHEXLOGLQJRZQHUZLOOEHDEOHWRDVVLVWZLWKGHWHUPLQLQJWKHQXPEHURISHRSOHWKDWZLOORFFXS\HDFK]RQH
6HYHUDOPRGHOLQJVRIWZDUHKDYHWKHDELOLW\WRDSSO\DYDULDWLRQIDFWRUWRWKHKHDWORDGJHQHUDWHGE\DSHUVRQEDVHG
RQ WKH RFFXSDQF\ W\SH +RZHYHU LI WKH VRIWZDUH EHLQJ XVHG GRHV QRW KDYH WKLV DELOLW\ 7DEOH LQ &KDSWHU
1RQUHVLGHQWLDO &RROLQJ DQG +HDWLQJ /RDG &DOFXODWLRQV RI WKH $6+5$( )XQGDPHQWDOV +DQGERRN SURYLGHV WKH
VHQVLEOHDQGODWHQWKHDWJDLQIRUGLIIHUHQWW\SHVRIDFWLYLWLHVRIKXPDQEHLQJV7KHEXLOGLQJRZQHUZLOODOVREHDEOH
WR KHOS GHWHUPLQH ZKDW W\SH RI HTXLSPHQW ZLOO EH ORFDWHG LQ HDFK ]RQH 7KH RZQHU PD\ DOVR EH DEOH WR SURYLGH
LQIRUPDWLRQRQWKHKHDWORDGJHQHUDWHGE\WKHHTXLSPHQWKRZHYHULIWKLVLVQRWWKHFDVHWKHUHLVDQRWKHUUHVRXUFH
2QFHDJDLQ&KDSWHURI$6+5$()XQGDPHQWDOV+DQGERRNVXSSOLHVVHYHUDOWDEOHVRIKHDWJHQHUDWLRQHTXLSPHQW
7DEOHV$WKURXJK(SURYLGHLQIRUPDWLRQIRUFRRNLQJDSSOLDQFHV7DEOHVDQGSURYLGHLQIRUPDWLRQRQPHGLFDO
DQGODERUDWRU\HTXLSPHQW7DEOHWKURXJK7DEOHSURYLGHLQIRUPDWLRQRQW\SLFDOFRPSXWHUDQGRIILFHHTXLSPHQW
,Q RUGHU WR GHWHUPLQH WKH LQWHUQDO ORDGV GXH WR OLJKWLQJ WKH HOHFWULFDO GHVLJQ HQJLQHHU ZLOO EH DQ H[WUHPHO\
KHOSIXOUHVRXUFH%DVHGRQWKHOLJKWLQJGHVLJQWKHHQJLQHHURUGHVLJQHUVKRXOGEHDEOHWRSURYLGHDOLJKWLQJSRZHU
GHQVLW\ IRU HDFK ]RQH DV ZHOO DV WKH EXLOGLQJ H[WHULRU +RZHYHU LI W\SLFDO YDOXHV DUH QHHGHG IRU OLJKWLQJ SRZHU
GHQVLWLHV7DEOHLQ&KDSWHURIWKH$6+5$()XQGDPHQWDOV+DQGERRNDQG7DEOHLQ$6+5$(6WDQGDUG
SURYLGH OLJKWLQJ SRZHU GHQVLWLHV EDVHG RQ WKH VSDFHE\VSDFH PHWKRG 7DEOH LQ 6WDQGDUG DOVR
SURYLGHV OLJKWLQJ SRZHU GHQVLWLHV XVLQJ WKH EXLOGLQJ DUHD PHWKRG 7DEOH LQ 6WDQGDUG SURYLGHV OLJKWLQJ
SRZHUGHQVLWLHVIRUEXLOGLQJH[WHULRUV7KHHQJLQHHURUGHVLJQHUZLOODOVREHDEOHWRDVVLVWZLWKDQ\W\SHRIOLJKWLQJ
FRQWUROV EHLQJ SURYLGHG VXFK DV RFFXSDQF\ VHQVRUV DXWRPDWLF OLJKWLQJ FRQWURO V\VWHP RU D GD\OLJKWLQJ FRQWURO
V\VWHP 6SHFLILFDWLRQV IRU GD\OLJKWLQJ FRQWURO V\VWHPV VKRXOG DOVR LQFOXGH LQIRUPDWLRQ RQ FRQWLQXRXV GLPPLQJ
V\VWHPVRUVWHSFRQWUROV\VWHPV

+HDWLQJ9HQWLODWLQJDQG$LU&RQGLWLRQLQJ+9$&6\VWHP6L]LQJ

7KHUHDUHSULPDULO\WZRVFHQDULRVWKDWWKHHQHUJ\PRGHOHUZLOOHQFRXQWHUFRQFHUQLQJ+9$&V\VWHPVL]LQJ,Q
HLWKHU VFHQDULR WKH HQHUJ\ PRGHOHU ZLOO ZRUN FORVHO\ ZLWK WKH PHFKDQLFDO HQJLQHHU RU +9$& V\VWHP GHVLJQHU LQ
RUGHU WR GHWHUPLQH KRZ WKH EXLOGLQJ LV ]RQHG DQG ZKDW W\SH RI HTXLSPHQW DQG HIILFLHQFLHV ZLOO EH XVHG ,Q RQH
VFHQDULRWKHGHVLJQHUZLOOSURYLGHLQIRUPDWLRQWRWKHHQHUJ\PRGHOHULQFOXGLQJWKHVL]HRIWKHHTXLSPHQWDVKHRU
VKHZLOOKDYHDOUHDG\SHUIRUPHGDORDGDQDO\VLVHTXLSPHQWHIILFLHQFLHVEDVHGRQWKHHTXLSPHQWWKDWWKHGHVLJQHUKDV
VHOHFWHGIRUWKHSURMHFWZLOODOVREHSURYLGHG+RZHYHUWKHRWKHUVFHQDULRLQYROYHVWKH+9$&GHVLJQHUWXUQLQJWR
WKHHQHUJ\ PRGHOHUIRUDVVLVWDQFHLQGHWHUPLQH ZKDWW\SH RI+9$&V\VWHPWRXVHIRUWKHEXLOGLQJ7KLVVFHQDULR
W\SLFDOO\ RFFXUV HDUO\ LQ WKH GHVLJQ SURFHVV WR GHWHUPLQH OLIH F\FOH FRVWV DQDO\VHV IRU GLIIHUHQW W\SHV RI +9$&
V\VWHPV ,I WKH GHVLJQ FRQWDLQV D ODUJH QXPEHU RI GLYHUVLILHG ]RQHV WKH GHVLJQHU PD\ DOVR VHHN LQSXW IURP WKH
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
HQHUJ\PRGHOHUZKHQGHWHUPLQLQJWKHPRVWHIIHFWLYHDSSURDFKWR]RQHWKHHTXLSPHQW
7KHGHVLJQHUPD\KDYHDOUHDG\SHUIRUPHGDORDGDQDO\VLVWRDVVLVWWKHHQHUJ\PRGHOHU,IWKLVLVQRWWKHFDVH
WKHHQHUJ\ PRGHOHU PD\ KDYHWRSHUIRUPDORDGDQDO\VLVEDVHGRQWKHLQIRUPDWLRQWKDWKDVDOUHDG\EHHQREWDLQHG
6RPH PRGHOLQJ VRIWZDUH KDYH WKH FDSDELOLW\ RI SHUIRUPLQJ D ORDG DQDO\VLV LQ DGGLWLRQ WR UXQQLQJ HQHUJ\ PRGHOV
KRZHYHULIWKHVRIWZDUHGRHVQRWKDYHWKDWFDSDELOLW\DQDGGLWLRQDOSURJUDPZLOOKDYHWREHXVHGIRUWKLVSXUSRVH
2QFH WKH KHDWLQJ DQG FRROLQJ ORDGV IRU WKH EXLOGLQJ KDYH EHHQ FDOFXODWHG GLIIHUHQW +9$& V\VWHPV PXVW EH
HYDOXDWHG 7KH $6+5$( +9$& 6\VWHPV DQG (TXLSPHQW +DQGERRN SURYLGHV GHWDLOHG LQIRUPDWLRQ DERXW VHYHUDO
GLIIHUHQW W\SHV RI +9$& V\VWHPV WKDW FDQ EH XVHG 7KLV KDQGERRN JLYHV VHYHUDO VFKHPDWLFV DQG H[DPSOHV RQ
FRPSRQHQW VL]LQJ DQG RSHUDWLRQ WR JLYH D EHWWHU XQGHUVWDQGLQJ RI WKH V\VWHP ,Q DGGLWLRQ WR WKH +DQGERRN
$6+5$(

$6+5$( 6WDQGDUG SURYLGHV LQIRUPDWLRQ RQ PLQLPXP HTXLSPHQW HIILFLHQFLHV LQ 6HFWLRQ +HDWLQJ
9HQWLODWLQJDQG$LU&RQGLWLRQLQJ7DEOHV$WKURXJK-SURYLGHPLQLPXPHTXLSPHQWHIILFLHQFLHV IRUDLU
FRQGLWLRQLQJXQLWVKHDWSXPSVFKLOOHUVDQGERLOHUVEDVHGRQHTXLSPHQWFDSDFLW\
6HYHUDO PRGHOLQJ VRIWZDUH SURJUDPV SURYLGH WXWRULDOV RQ KRZ WR VHW XS +9$& V\VWHPV LQ WKHLU VRIWZDUH
6RIWZDUH XVHU PDQXDOVDUHDOVRJRRGUHIHUHQFHVLQSURYLGLQJLQIRUPDWLRQ7KHUHDUHDOVRJRRGRQOLQHIRUXPV IRU
HQHUJ\PRGHOLQJSURIHVVLRQDOVWRH[SUHVVTXHVWLRQVDQGJLYHIHHGEDFNRQGLIIHUHQWVRIWZDUHSURJUDPVDQGPRGHOLQJ
PHWKRGV
$QDGGLWLRQDOFRQVLGHUDWLRQWKDWPXVWEHWDNHQLQWRDFFRXQWRQWKH+9$&V\VWHPLVKRZIUHVKDLUYHQWLODWLRQ
UHTXLUHPHQWVDUHEHLQJKDQGOHG7KLVLVEDVHGRQWKHFODVVLILFDWLRQRIWKHVSDFHEHLQJVHUYHGWKHQXPEHURISHRSOH
LQWKHVSDFHDQGWKHRYHUDOOVTXDUHIRRWDJHRIWKHVSDFH6HFWLRQ3URFHGXUHVRI6WDQGDUGSURYLGHVGHWDLOHG
HTXDWLRQV IRU GHWHUPLQLQJ WKH YHQWLODWLRQ UDWHV DQG 7DEOH LQ $6+5$( 6WDQGDUG SURYLGHV WKH PLQLPXP
YHQWLODWLRQUDWHEDVHGRQRFFXSDQF\FDWHJRU\6HYHUDOVRIWZDUHSURJUDPVKDYHDOVRDOUHDG\LQFRUSRUDWHGWKHVHUDWHV
DQGHTXDWLRQVDVDVWDQGDUGIHDWXUH+RZHYHUGHSHQGLQJRQKRZWKHIUHVKDLULVLQWURGXFHGLQWRWKHEXLOGLQJ ZLOO
SOD\DKXJHUROHLQWKHHQHUJ\PRGHO7KH+9$&V\VWHPGHVLJQHUZLOOQHHGWREHLQYROYHGWRGHWHUPLQHLIWKHIUHVK
DLUZLOOEHLQWURGXFHGWKURXJKWKH+9$&V\VWHPDQHQHUJ\UHFRYHU\V\VWHPRUWKURXJKQDWXUDOYHQWLODWLRQ

2SHUDWLQJ3URILOHV

2QFHDJDLQWKHEXLOGLQJRZQHUZLOOQHHGWREHFRQVXOWHGWRGHWHUPLQHKRZWKHEXLOGLQJLVJRLQJWRRSHUDWH
7KLVZLOOLQFOXGHSURILOHVVXFKDVZKHQWKHEXLOGLQJLVJRLQJWREHRFFXSLHGDQGDWZKDWGHQVLW\OLJKWLQJRSHUDWLRQ
HQHUJ\ PDQDJHPHQW FRQWURO V\VWHPV FRRNLQJ RSHUDWLRQV YHQWLODWLRQ VFKHGXOHV DQG PLVFHOODQHRXV HTXLSPHQW
RSHUDWLRQVXFKDVFRPSXWHUVSULQWHUVDQGFRSLHUV7KHEXLOGLQJRZQHUPD\RUPD\QRWEHDEOHWRSURYLGH\RXDOORI
WKLV LQIRUPDWLRQ +RZHYHU DOPRVW DOO HQHUJ\ PRGHOLQJ VRIWZDUH SURJUDPV SURYLGH W\SLFDO RSHUDWLQJ SURILOHV WR
DVVLVWWKHHQHUJ\PRGHOHUZLWKTXLFNUHVXOWV
2SHUDWLQJSURILOHVDOVRSOD\DYHU\FULWLFDOUROHZKHQXVLQJHQHUJ\PRGHOVIRUGHWHUPLQLQJPHDVXUHPHQWDQG
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
YHULILFDWLRQ RQ H[LVWLQJ IDFLOLWLHV %\ HVWDEOLVKLQJ D EDVHOLQH SHUIRUPDQFH HQHUJ\ PRGHO HQHUJ\ FRQVHUYDWLRQ
PHDVXUHV WKDW KDYH EHHQ LPSOHPHQWHG LQ D IDFLOLW\ FDQ KDYH WKHLU SHUIRUPDQFH PHDVXUHG %\ XVLQJ HQHUJ\
PRGHOLQJ DQDO\VHV WR GHWHUPLQH WKH DPRXQW RI HQHUJ\ WKDW D IDFLOLW\ VKRXOG EH FRQVXPLQJ GXULQJ D UHWUR
FRPPLVVLRQLQJSURMHFWRSHUDWLQJDGMXVWPHQWVFDQEH PDGHLQWKHIDFLOLW\WRILQHWXQHWKHEXLOGLQJVRSHUDWLRQDQG
UHGXFHRSHUDWLQJFRVW

(QHUJ\5DWH6WUXFWXUHV

7KHODVWSLHFHRIWKHSX]]OHLVWRHQWHUWKHXWLOLW\UDWHVWUXFWXUHVLQWRWKHPRGHOLQJVRIWZDUH7KLVLVW\SLFDOO\
REWDLQHG IURP WKH EXLOGLQJ RZQHU DV D GRFXPHQW WKDW SURYLGHV LQIRUPDWLRQ RQ SUHYLRXV VWUXFWXUHV ,W JLYHV
LQIRUPDWLRQRQDXWLOLW\FRQVXPSWLRQUDWHVXFKDVDNLORZDWWKRXUIRUHOHFWULFDOFRQVXPSWLRQDQGKXQGUHGFXELFIHHW
RU FXELF PHWHUV IRU QDWXUDO JDV ,W DOVR JLYHV D GHPDQG FKDUJH LI RQH LV SUHVHQW +RZHYHU LQ VRPH FDVHV WKH
EXLOGLQJPD\KDYHSXUFKDVHGFKLOOHGZDWHUKHDWLQJZDWHURUVWHDPLQDGGLWLRQWRDVWDQGDUGHOHFWULFRUQDWXUDOJDV
UDWH7KLVLQIRUPDWLRQLVYHU\FUXFLDOLQGHWHUPLQLQHQHUJ\FRVWVDYLQJVDVZHOODVSD\EDFNLQIRUPDWLRQRQYDULRXV
HOHPHQWV RI WKH GHVLJQ :KHQ FRQVLGHULQJ SD\EDFN SHULRGV WKH HQHUJ\ PRGHOHU ZLOO DOVR QHHG WR SURYLGH DQ
LQIODWLRQUDWHWRHVFDODWHWKHFRVWRIXWLOLWLHVRYHUWKHVWXG\SHULRG

176
ASHRAE Transactions

&21&/86,216

$V VKRZQ WKURXJKRXW WKLV SDSHU HQHUJ\ PRGHOLQJ LV YHU\ FRPSOH[ DQG LQFRUSRUDWHV VHYHUDO GLIIHUHQW
GLVFLSOLQHV WR REWDLQ DQ DFFXUDWH ILQLVKHG SURGXFW 7KH HQHUJ\ PRGHOLQJ SURIHVVLRQDO ZLOO SOD\ D FUXFLDO UROH
UHJDUGOHVV RI WKH OHYHO RI HQHUJ\ PRGHOLQJ :KHWKHU XVHG IRU FRGH FRPSOLDQFH GHVLJQ SHUIRUPDQFH RU
PHDVXUHPHQW DQG YHULILFDWLRQ WKH UHVXOWV SURGXFHG E\ WKH HQHUJ\ PRGHO ZLOO EH XVHG WR PDNH GHFLVLRQV RQ WKH
FXUUHQW SURMHFW DV ZHOO DV IXWXUH SURMHFWV ,Q FRPSOH[ IDFLOLWLHV HQHUJ\ PRGHOLQJ PD\ EH WKH RQO\ GHVLJQ WRRO
DYDLODEOHIRUGHFLVLRQPDNLQJ
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
7KH EDVLV RI HDFK HQHUJ\ PRGHO LV WR VHW XS WKH EXLOGLQJ DV GHWDLOHG DV SRVVLEOH WKURXJK FRQVWUXFWLQJ WKH
EXLOGLQJHQYHORSHGHWHUPLQLQJLQWHUQDODQGSURFHVVORDGVVL]LQJWKH+9$&V\VWHPHVWDEOLVKLQJRSHUDWLQJSURILOHV
DQG HQWHULQJ WKH XWLOLW\ UDWHV 7KHVH UHTXLUHPHQWV ZLOO UHTXLUH LQSXW IURP WKH EXLOGLQJ GHVLJQHUV DV ZHOO DV WKH
EXLOGLQJRZQHU:KHUHLQSXWLVQRWUHDGLO\DYDLODEOHVHYHUDO$6+5$(6WDQGDUGVDQG+DQGERRNVDUHDYDLODEOHIRU
DGGLWLRQDOJXLGDQFH6HYHUDORIWKHPRGHOLQJVRIWZDUHSURJUDPVDOVRSURYLGHJXLGDQFHDQGWXWRULDOLQIRUPDWLRQLI
DGGLWLRQDO TXHVWLRQV DULVH GXULQJ WKH ILUVW IHZ HQHUJ\ PRGHOV 7KURXJK RQ WKH MRE WUDLQLQJ DQG SUDFWLFH HQHUJ\
PRGHOLQJFDQEHPDVWHUHGLQWRDSHUIHFWHGFUDIW

120(1&/$785(
+9$& KHDWLQJYHQWLODWLQJDQGDLUFRQGLWLRQLQJ
$6+5$( $PHULFDQ6RFLHW\RI+HDWLQJ5HIULJHUDWLRQDQG$LU&RQGLWLRQLQJ(QJLQHHUV

5()(5(1&(6

$16,$6+5$( $6+5$( 6WDQGDUG 7KHUPDO (QYLURQPHQWDO &RQGLWLRQV IRU +XPDQ 2FFXSDQF\
$WODQWD$PHULFDQ6RFLHW\RI+HDWLQJ5HIULJHUDWLRQDQG$LU&RQGLWLRQLQJ(QJLQHHUV,QF
$16,$6+5$( $6+5$( 6WDQGDUG 9HQWLODWLRQ IRU $FFHSWDEOH ,QGRRU $LU 4XDOLW\ $WODQWD
$PHULFDQ6RFLHW\RI+HDWLQJ5HIULJHUDWLRQDQG$LU&RQGLWLRQLQJ(QJLQHHUV,QF
$16,$6+5$($6+5$(6WDQGDUG(QHUJ\6WDQGDUGIRU%XLOGLQJVH[FHSW/RZ5LVH5HVLGHQWLDO
%XLOGLQJV$WODQWD$PHULFDQ6RFLHW\RI+HDWLQJ5HIULJHUDWLRQDQG$LU&RQGLWLRQLQJ(QJLQHHUV,QF
$6+5$($6+5$(+DQGERRN)XQGDPHQWDOV$WODQWD$PHULFDQ6RFLHW\RI+HDWLQJ 5HIULJHUDWLRQDQG $LU
&RQGLWLRQLQJ(QJLQHHUV,QF
$6+5$( $6+5$( +DQGERRN ± +9$& 6\VWHPV DQG (TXLSPHQW $WODQWD $PHULFDQ 6RFLHW\ RI +HDWLQJ
5HIULJHUDWLRQDQG$LU&RQGLWLRQLQJ(QJLQHHUV,QF

$6+5$(

SA-12-C023
DOE Commercial Building Energy Asset

Rating: An Application of Centralized
Modeling Tool
Na Wang, PhD Will Gorrissen Viraj Srivastava, PhD
Cody Taylor
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
ABSTRACT
This paper presents a novel approach used to develop the U.S. Department of Energy (DOE) commercial building energy asset rating tool. Asset
rating, a national standard for a voluntary energy rating system, is intended to help building owners better understand the installed system
performance and the total energy use. The asset rating tool allows users to benchmark their buildings against peers and other market players to
understand the relative efficiency of different buildings in a way that is distinct from their operations and occupancy. A simplified data collection
integrated with full-scale energy-modeling method is employed to disaggregate building energy information and will include a mechanism for
identifying energy improvement opportunities. A more detailed modeling approach to formulate an asset rating would most likely provide the
greatest and accuracy; while a simplified model approach requires less user investment for collecting data. However, our market research suggested
that an asset rating program needs to consider not only the applicability and accuracy across the breadth of commercial buildings but also ease of
use. To take the above design drivers into account, we are developing an asset rating tool, a web-based application with a simplified user interface
built on an inference engine and a centralized modeling engine. The method presented in this paper separates model inputs into categories based on
their overall energy impact, difficulty level of data collection, and variability among buildings. We outline an approach that will allow great
flexibility in terms of how many and which of the different categories of variables must be found to produce an accurate energy model. The approach
will allow all key variables to be inferred from some reduced set of variables while at the same time allowing a user to enter many more variables if
he or she has reliable details on them. The asset rating tool is not just a rating tool, but is aimed at providing a cost-effective means for building
owners and operators to gain insight into the energy efficiency potential of their buildings. The development of such a tool enables reduced modeling
time and expertise requirements while maintaining accuracy and the ability to support the variability and complexity that exist in buildings.
INTRODUCTION
This paper discusses the energy modeling methodology for the U.S. Department of Energy (DOE) commercial building
energy asset rating program. The main objectives of the program are to enable commercial building stakeholders to directly
compare as-built energy performance among similar buildings, irrespective of occupant behavior and building operation, to
drive further investment in energy efficiency, and thus to reduce energy use in the commercial building sector. The rating
will include physical characteristics of the building as built and its overall energy efficiency, independent of occupancy and
operational choices. The physical characteristics will include the building envelope, the mechanical and electrical systems,
Na Wang, Will Gorrissen, and Viraj Srivastava are engineers in the Electricity Infrastructure and Buildings Division, Pacific Northwest National
Laboratory, Richland, Washington. Cody Taylor is a project manager in the Office of Energy Efficiency and Renewable Energy, U.S.
Department of Energy, Washington, D.C.
178
ASHRAE Transactions
and other major energy-using equipment. DOE is considering using source energy use intensity (EUI) as the primary metric
with which to generate the asset ratings. Although site energy use can be measured directly and apparently reflects the
energy use over which building owners have control, using source energy is less likely to unintentionally favor or penalize
the use of one energy fuel type over another. In addition, source energy use is a proxy for energy cost. Source energy metric
is also aligned with ENERGY STAR Portfolio Manager (U.S. Environmental Protection Agency, 2012). DOE is also
considering the inclusion of other metrics as optional indicators, such as site energy use separated out by fuel types,
greenhouse gas emissions, peak demand, energy cost, etc. Unlike an operational rating, which can use utility information to
measure the total outcome of building energy efficiency, operation, and maintenance, asset rating needs to disaggregate
building energy use information by simulating building performance under standard operating and occupancy conditions.
By focusing on building characteristics and removing occupancy and operational variations, “apples-to-apples” comparisons
can be made.
To establish a national building energy rating standard, the basic criteria include the consistency, repeatability, and
accuracy of the modeled results. These criteria become extremely important when asset rating is used during a real estate
transaction and building energy efficiency is considered as part of the property asset value. Another consideration of setting
up a national rating system is the time and resources required to obtain a rating. With energy expenditures in U.S.
commercial buildings averaging $1.43/ft2 ($15.39/m2) (U.S. Energy Information Administration, 2003), a 20%
improvement in efficiency could yield savings of $0.29/ft2 ($3.12/m2). A comprehensive energy audit and modeling analysis
can cost up to $0.50/ft2 ($5.38/m2), (California Energy Commission, 2000). Therefore, detailed audits and modeling can
often be cost-prohibitive for all but the largest buildings and commercial building owners. Any cost burden related to data
collection and modeling becomes a barrier to the implementation of the asset rating program.
The usability of the asset rating result is another critical criterion. Unlike large institutional investors who are actively
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
looking at benchmarking their portfolios to improve the market value of their properties, smaller-building owners/investors
and owner-occupied building owners may lack motivation to obtain an asset rating, especially when their real estate
exposure is in less desirable markets. For this group of building owners, an easy, ready access to suggestions for energy
efficiency improvement is more valuable.
Based on the above considerations, the asset rating project team decided to develop a centralized modeling tool to
facilitate ease of application, reduce cost, and increase standardization allowing for consistent and reliable comparisons. The
tool not only generates a standard building energy asset rating but also aims to provide added value in the first step of
assessment of a building by describing some possible upgrade opportunities. It is not intended to replace a full energy audit
of a building but rather produce a preliminary energy report. The objective of the tool is to give property owners the means
of gauging the efficiency of their properties as compared to the potential efficiency, providing information on key action
steps and motivating them to make reasoned and value-conscious investments enabling them to target limited capital
resources toward those areas that will produce the greatest return.
EXPERIENCE FROM ESTABLISHED RATING TOOLS AND AUDITING PROTOCOLS
The building characteristics or inputs considered by asset rating tools affect the accuracy of their rating. In December
2010, the Buildings Performance Institute Europe (BPIE) published a review of the European Union experience under the
Energy Performance Building Directive (EPBD). BPIE (2010) discusses certain measures can be taken to increase the
effectiveness of implementation and public acceptance of the rating systems. Their research suggests that model
reproducibility is improved by simplifying the data acquisition and relying on a larger number of default values required for
the calculation, and found that overall inaccuracy is reduced from ±20% to ±15%; the deviation from calculated
performance to actual building performance is correspondingly reduced from ±30% to ±10%. More simplified data
acquisition also requires a lower level of expertise, less time, and less effort from the assessors, thus resulting in lower costs
for the assessment.
$6+5$(

Existing auditing procedures (American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc., 2004)
specify a walkthrough analysis for a Level 1 assessment. A Level 2 analysis requires measurement of key operating
parameters and comparing them to design, as well as a breakdown of the total annual energy use into end-use components.
The deliverable from such an analysis includes both a discussion of the existing situation and reasons for excess energy use
and an outline of the recommended measures. Such auditing guidelines set out generalized procedures to guide the analyst
and require that each analyst exercise a substantial amount of judgment. The asset rating data requirement is aimed to
require less time and effort to collect than an ASHRAE Level 2 audit.
Different approaches have been taken to standardize energy modeling. ISO 13790:2008 (ISO, 2008), which is widely
applied in Europe for the calculation of building energy performance, defines the calculation recipe according to a set of
normative statements about functional building category, assumed usage scenario, system efficiency, and so on. Three types
of methods are covered in the ISO standard; they are monthly quasi-steady-state calculation method, simple hourly dynamic
calculation method, and calculation procedures for detailed dynamic simulation methods. The fully prescribed monthly
quasi-steady-state method was adopted by United Kingdom Department of Communities and Local Government and
developed into a simplified asset-based calculation procedure, Simplified Building Energy Model (SBEM). SBEM is also
used by Ireland to calculate energy use of simple, non-residential buildings for energy performance certificate (asset rating).
Dynamic simulation software is used for complex-non-residential buildings, for example in Ireland and Portugal. The
monthly calculation method gives more accurate results on an annual basis, but large relative errors occur in the months
close to the beginning and the end of the heating and cooling season (ISO 2008, section 5.3). However, the accuracy of the
calculated energy use becomes less relevant because a standardized expression of performance does not need to predict the
actual energy consumption but to guarantee the results of asset rating, which is, in the case of European asset rating system,
the energy performance coefficient of the rated building and the comparator building (Lee, Fei, & Augenbroe, 2011).
In the U.S., existing guidance (Deru, 2007) for modeling and inspecting energy-efficient property in commercial
buildings for certification of the energy and power cost savings specifies that the energy modeling must be completed in
accordance with the Performance Rating Method (PRM) presented in ANSI/ASHRAE/IESNA Standard 90.1-2004,
Appendix G, but the models must comply with the requirements of Standard 90.1-2001. COMNET Commercial Buildings
Energy Modeling Guidelines and Procedures also provide modeling rules for building energy modeling professionals.
The above systems are some examples of auditing protocols and approaches varying in terms of the time and
expertise required of the user. These factors influence the real-world applicability of the tools based on the cost of
implementation and the accuracy of the result. Seeking the balance of the competing drivers of ease of use and accuracy of
results is the essential for developing a reliable and useful rating system.
MODELING APPROACH
All buildings are different, and each modeler needs to exercise a substantial amount of judgment. To avoid modeler’s
bias and reduce the implementation cost, an asset rating tool is being developed for the DOE asset rating system. This tool
sets out generalized procedures by employing a uniform method of estimating building performance while following the
modeling requirements specified in ASHRAE Standard 90.1-2007, Appendix G. In selecting the building energy modeling
approach a number of options were considered. These included
• pre-simulation with regression and/or interpolation,

• time series energy use data based analysis,
• normative calculation method, and
• dynamic energy simulation.
Pre-simulation would have provided a robust modeling approach and allowed for relatively simple tool development
but would not have provided the flexibility that the tool would need to accurately represent the diversity of buildings that
180 --`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---

ASHRAE Transactions
could potentially be addressed using the tool. Time series data analysis, on the other hand, was discounted as an option
because it is linked so closely to the operation and maintenance characteristics of a specific building and thus would not
allow the interbuilding comparisons for which the tool was being designed. Normative calculation method utilizes a set of
algebraic equations and observable or empirical parameters. One disadvantage is that special energy saving technologies
may not be properly considered due to the fact that normative calculations assume the building performance as a series of
steady states and ignore the dynamics between the steady states. Therefore some countries, such as Ireland, allow using
approved dynamic simulation software. This dual approach complicates the rating system and decreases the standardization.
Two different real-time dynamic building energy modeling options were considered as a means to calculate building
energy use. The first was a simplified energy modeling approach that would use a number of approximations and
simplifications. This approach was discounted because of potential users’ discomfort with results based on such
simplifications and approximations. It was ruled out also because of the difficulty such an approach would have simulating
more complex building energy components and systems, especially those present in today’s high-performance buildings.
The other real-time energy modeling approach examined was a highly detailed, sub-hourly whole-building energy
model. This approach would provide the level of detail required to model the most complex buildings being built today and
would produce results in which the end users would presumably have greater confidence (assuming that an established tool
were used). The drawback of the detailed modeling approach is that the required number of inputs would make the tool
unusable to all but the most experienced users.
To overcome the inherent issues in each of the approaches examined, we build the asset rating tool on a combination
of a simplified user interface, a reference engine, and a detailed energy modeling engine. The user interface enables the
building of a simplified building geometry and the collection of a reduced set of model inputs. EnergyPlus, a widely
accepted building energy modeling tool, is used to simulate a whole-building energy model. One consideration of using a
detailed energy model is to enable users store their building information in a standardized model, which can be downloaded
by the users and used for other purposes. For example, a third-party may develop a building retrofit tool that modifies the
standardized model generated by the asset rating tool. Although such a sub-hourly simulation may provide information that
may not be used for asset rating at this stage, the approach provides opportunities for future expansion.
To link a simplified user interface with a detailed energy model input, it was necessary to develop an inference engine,
which maps out one-to-many relationship between the different building characteristics. This was accomplished by building
on existing work (Pacific Northwest National Laboratory, 2008) that derived these relationships from a number of sources,
including
• dummy variable ordinary least squares regression of building survey data,
• equipment efficiency standards,
• building energy codes and adoption rates,
• ASHRAE handbooks,
• energy model internal system sizing algorithms, and
• previous research.
These relationships (previously developed to provide inference capabilities for the Facility Energy Decision System) allowed --`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
the tool builders to build the required detailed inputs from a small subset of user inputs. The reduced set of user inputs was
described as the simple-level input set. This input level will always be required by tool users, and so has been developed to
be relatively simple to collect accurately. As the users include more detailed inputs on the way to the advanced-level set, the
energy model results will reflect the added level input by becoming more tailored to the each user’s specific building. This
system of simplified inputs is then used to predict the remaining building characteristics, providing the ease of use,
flexibility, and accuracy required to make the tool useful to many disparate user groups.
The combination of the simplified user interface, inference engine, and modeling engine would enable us to make the
final tool user friendly in order to encourage broad adoption and provide the accuracy, detail, and extensibility needed for
applicability across the wide range of variation that exists within the built environment. The keys to this approach are data
collection and determining how to categorize the parameters into a simple-level input set and an advanced-level input set.
$6+5$(

The asset rating data inputs are outlined in the following sections.
ASSET RATING DATA INPUT REQUIREMENTS
Given that building performance is determined by multiple factors, including building usage and design, occupant
behaviors, equipment degradation and failure, and structural deficiencies, development of an energy modeling methodology
for asset rating included the development of a consistent set of metrics and standards in which key energy-related data
factors that impact asset value proposition are identified. The focus is on factors that drive the most significant changes.
Interviews and feedback received during the development of the asset rating tool reflected a mix of responses from
stakeholders; although there is a concern over additional burden of time and expense, some stakeholders also desire more
detailed energy modeling. In the following sections, we describe the inputs required for the asset rating tool, with
consideration given to such stakeholder feedback.
General Input Classification
To determine the required inputs that asset rating tool users would be expected to provide, classification of the input
variables was necessary. The first step was to collect a full list of building characteristics that play a role in building energy
consumption. This list of candidate inputs were derived from a number of sources, including
• building energy modeling inputs,

• simplified building load analysis inputs,
• previous research analyzing actual building energy use data,
• previous research analyzing modeled building energy use data, and
• building energy auditing and modeling professionals.
These sources provided the initial full list of potential user inputs that were considered. The potential variables were then
assessed based on three characteristics: ease of collection by target user, impact on energy consumption, and expected
variability between buildings. Each variable was rated either as low, medium, or high for each of the three characteristics.
Ease of collection was ascertained through preliminary interviews with experienced building energy professionals. The
following questions were asked:
• In your experience, what are the most difficult building characteristics to determine when auditing a building?
• Which difficult-to-obtain variables are due to the time/tedium required (example: counting lights)?
• Which difficult-to-obtain variables are due to the expertise required (example: ascertaining HVAC control
strategies)?
• Which variables are difficult to obtain through nondestructive examination (example: wall cavity insulation
thickness)?
Responses to the preliminary interview allowed for the initial categorization of potential inputs by ease of collection,
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
which is considered a proxy for cost to obtain for this assessment. Interviewees were not asked to provide great detail but
rather a general assessment based on experience. Examples of simple to collect building characteristics include geographic
location, floor area, and the year built; more difficult variables include HVAC system details and envelope thermal
characteristics. The effort required for each aspect of data collection and input will be further examined by filed assessment
during the DOE asset rating pilot project.
In addition to ease of collection, variables were categorized based on their likely impact on the energy use of a
building. To assess the potential impact of a specific variable, two main sources were used. The first source, as in ease of
collection, was domain expert interviews. In the case of potential energy use impact, the domain experts consulted were
experienced building energy analysts at the Pacific Northwest National Laboratory (PNNL). The interviewees were asked to
182
ASHRAE Transactions
assess the different variables selected in terms of the likelihood that a small change would lead to a large overall change in
total energy consumption. These variables included building characteristics such as air infiltration, HVAC equipment
efficiency, wall and ceiling R-values, etc. In addition to interviews, an examination of existing literature on the topic of
building energy modeling inputs was undertaken to supplement the information found through expert evaluation. Much of
the literature that was examined pointed more to the inconsistency of variables impact across building use type,
configuration, and location (ASHRAE, 2006). For example, some studies found building energy use to be highly dependent
on internal loads and only mildly affected by envelope loads, while other studies found the converse to be true. Therefore,
to maintain broad applicability of the tool, when a variable’s impact could not be conclusively agreed upon, expected
variable impact was judged of limited use when considering the exclusion of a specific variable.
The third metric used to categorize the selected potential asset rating inputs was the expected variability between
buildings. The expected variability of an input from one building to another was determined by consulting both building
energy auditors and building energy modeling experts at PNNL. They were asked to rank, based on their experience, the
potential inputs as either low, medium, or high expected variability. Often, inputs that were deemed highly variable, such as
occupancy patterns and outdoor air levels, were those that were intended be held constant as a result of the specific
requirements of asset rating, and so are to be considered low variability inputs.
Table 1. Classification of input variables

Step 1 Step 2 Step 3 Variable Examples S* A* BA*
Ease of Collection Impact on Energy Use Variability Type
Easy Low Low 1 floor plate type,
Easy Medium Low 1 thermal zone layout
Easy High Low 1
Easy Low Medium 1
Easy Medium Medium 2 floor area,
Easy High Medium 2 building vintage,
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Easy Low High 2 wall type
Easy Medium High 2
Easy High High 2
Moderate Medium Low 3 insulation thickness,
Moderate High Low 3 window solar heat gain
Moderate Medium Medium 3 coefficient
Moderate High Medium 4 shading dimension
Moderate Medium High 4 HVAC efficiency
Moderate High High 4
Moderate Low Low 5 wall insulation
Moderate Low Medium 5 thickness
Moderate Low High 5
Difficult Low Low 6 fan blade efficiencies
Difficult Medium Low 6
Difficult High Low 6
Difficult Low Medium 6
Difficult Low High 6
Difficult Medium Medium 7 air infiltration rates
Difficult High Medium 7 wall insulation R-value
Difficult Medium High 7
Difficult High High 7
*Note: S=Simple level, A=Advanced level, BA=Beyond advanced level
$6+5$(

From the three-step classifications described above, a user input list was developed by removing all input variables
found to exhibit any of the following characteristics:
• high level of time commitment and/or level of expertise required to collect,

• low impact on final energy consumption of buildings, and
• low variability between buildings.
The remaining input variables were then classified by the different combinations of ease of collection, impact, and
variability. Seven input types were selected based on the combination of characteristics and the corresponding suitability for
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
making the input required, based on the competing drivers of ease of use and accuracy of results. These seven variable types
are described in matrix form in Table 1.
Levels of Input Sets
Once the inputs had been separated into the seven categories as outlined above, the final list was then broken into
three levels (shown in Table 1) based on the designation of each variable; the levels were defined as follows:
• Simple level – Variables of types 2 and 4. (These variables are easy or moderately difficult to collect. They have
significant impact on energy use and the values vary from one building to another.)
• Advanced level – Variables of types 1 through 5. (Compared with the simple level, the additional variables
required, types 1, 3, and 5, are easy or moderately difficult to collect, their impact on energy use may be high even
though their variability may not high. These are required to obtain simulation results of acceptable accuracy.)
• Beyond advanced level – Variables of types 1 through 7. (The additional variables required, type 6 and 7, are
difficult to collect.)
The grouped variables correspond to the input thresholds for three application cases, which have different purposes
of using the asset rating tool and the different levels of requirements for data accuracy. The first application corresponds to
the simple level variables listed above. The application for this set of inputs represents a preliminary analysis of building
performance and guidance in finding potential areas for building performance upgrades. These variables are generally quick
to collect and do not required a high level of building energy domain expertise to accurately ascertain. If a variable deemed
slightly more time consuming to collect is placed into this category, it is because it is considered to be highly important in
accurately assessing a building’s total energy consumption. Examples of simple-level data are shown in Table 1. The full list
will not be finalized until after a testing period where users are able to respond to the usefulness of the results and the
difficulty of the inputs. Any of the other variable types (1, 3, 5–7) can be entered to refine the result of the simple-level
application, up to the complete set of simple and advanced levels of variables, at which point there is sufficient detail to
meet the needs of an advanced-level application.
The advanced-level application requires that all inputs that fall into categories 1 through 5 be entered by the user.
These inputs have been selected to produce more robust predictions of building energy use and likely areas for cost-
effective asset upgrades. Table 1 gives examples of the current advanced-level. The advanced-level users of this tool will
have the ability to refine their results by filling in any of the remaining tool inputs; that is, those that fall into categories 6
and 7 and so make up the inputs beyond the minimum requirements for the advanced level. Plus inputs. Example inputs
include air infiltration rates and fan blade efficiencies. Further, these inputs will be difficult to capture and will not be
required but could potentially provide the expert user with added detail and thus more insight into the performance of the
building being examined. The inferability of different variable types, as described in the section Modeling Approach, is
outlined by variable type in Table 2.
184
ASHRAE Transactions
Table 2. Inferability of Variable Types
Inferable for Simple
Variable Type Inferable for Advanced Level
Level
1 Yes No
2 No No
3 Yes No
4 No No
5 Yes No
6 Yes Yes
7 Yes Yes
Asset Rating Application Process
The primary users of this tool are expected to include commercial property owner/investors and owner/operators.
Secondary users are anticipated to be financing sources, valuation experts, designers/engineers (to do a preliminary
modeling in support of more in-depth analysis and recommendation), and municipalities for disclosure requirements.
Owners of larger properties or portfolio owners may use the tool as a first pass, essentially a phase one energy report to
assess their buildings and prioritize which buildings should be assessed further using a more detailed energy audit. Smaller
property owners may be expected to use the tool as a cost effective means of evaluating energy efficiency and identifying
specific actions that may be taken to improve building performance.
Although it is beneficial for the user of the asset rating tool to have level of expertise or training in an area related to
building energy auditing, at a minimum the individual collecting such information may need to be experienced with building
systems and the process of extracting information from drawings and equipment cut sheets, or have ready access to people
who do.
In addition to the input variable classification described in the previous sections, we also classify the process of
collecting data based on the person likely to have that information, as well as the time estimated to collect the information.
We estimate that some information will be known to the facility manager immediately (e.g., number of floors, floor area,
HVAC system type), whereas collecting other inputs may require referring to the architectural or mechanical construction
drawings or equipment cut sheets. These inputs may include window-to-wall ratios or fan airflows. We further classify these
inputs as immediate, short, and long, based on the time required to collect the information as described in Table 3. The
estimated average time for collecting data of the immediate, short, and long variable types is less than 2 minutes, 5–10
minutes, and 10–30 minutes, respectively, given appropriate level of expertise and access to building systems or data. We
estimated the total required time to be less than 6 hours for the simple level and less than 20 hours for the advanced level.
These estimations are based on the interviews with the experienced energy auditors. They are only used to guide the tool
development. The field assessment during the pilot project will further examine the time required for each level of data
collection.
Table 3. Estimation of Data Collection Time

Data Collection Time Data Description
Information known to a person experienced with the building immediately, e.g., number of
Immediate (Easy) floors, HVAC system type.
Information that may be obtained immediately after referring to the building drawings, e.g.,
Short (Moderate) wall construction, thermal zoning.
Information that may be obtained after studying the building drawings or equipment
Long (Difficult) specifications for a period of time, or through onsite measurement, e.g. air infiltration,
cooling tower fan power.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

Use-Dependent Operational Settings
To evaluate the building energy use under typical operation, maintenance, and occupancy conditions, inputs related to
building operation and maintenance are standardized. Operating assumptions include thermostat settings, number of
occupants, receptacle loads, process loads, and hot water loads, as well as schedules of operation for HVAC systems,
lighting systems and other systems. All buildings of a similar type are assumed to have identical hours of operation and
occupancy patterns. In determining use-dependent inputs, we examined variables specified in existing protocols, such as
COMNET (Architectural Energy Corporation, 2010). In addition to COMNET, the fixed input parameters in the asset
rating tool derived from other sources (such as building operation schedules and setpoints) follow the inputs as specified in
the DOE commercial reference buildings models (National Renewable Energy Laboratory, 2011). These reference models
provide consistency in modeling approaches and implementation across commercial buildings. Inputs for these reference
models are determined from ASHRAE Standards 90.1-2004, 62.1-2004, and 62-1999 for new construction and Standard
90.1-1989 for post-1980 construction. Others were determined from studies of data and standard practices—for example,
the Facility Energy Decision System.
ASSET RATING TOOL OUTPUTS
The primary modeling output of the asset rating tool is the energy use intensity (EUI), which will be used to generate
the asset rating score. No comparator buildings are needed because the calculated EUI will be placed on a fixed scale, which
is under development. The asset rating scale should reflect the current variability with the building stock and recognize the
energy efficiency improvement of both low-efficient and high-performance buildings. Other useful outputs include energy
use breakout and system evaluations of lighting, heating plant, cooling plant, air distribution system, roof, wall, window, and
service hot water systems. This information can help users identify the part of the building in need of attention. For two
buildings with the same asset rating, the subsystem evaluation can help users gain insights into the problems and potential
of the two buildings.
The asset rating tool will also generate identified upgrade opportunities based on life-cycle cost (LCC) analyses of
energy efficiency measures (EEMs). Unlike a checklist or best-practice guide, the EEMs provided to the users are based on
a building’s specific situation. Using a checklist, building owners can receive general recommendations like increasing roof
insulation or installing a more efficient cooling plant. However, a more detailed recommendation—for example, increasing
roof insulation by 2 inches or installing an air-cooled electric chiller—cannot be obtained unless a professional engineering
firm is hired to perform an in-depth analysis. The asset rating is not intended to replace engineering evaluation, and the
results are not intended to be used to purchase equipment or materials. The asset rating tool will fill in the gap between an
untailored checklist and a professional engineering analysis. A preliminary decision on whether it is worthwhile to retrofit a
building and how to prioritize the activities can rely on the preliminary analysis provided by the asset rating at a low cost.
The LCC function will be developed based upon the existing algorithm and database of Facility Energy Decision
System (FEDS). FEDS uses the federally mandated LCC methodology, as specified in 10 CFR Part 436, to combine the
retrofit capital cost, O&M cost changes, and changes in the energy costs to determine the cost effectiveness of potential
retrofits. When the minimum LCC configuration of generation and end-use technologies is determined, all interactive
effects between energy systems are explicitly modeled. For example, when considering a lighting retrofit, the model
evaluates the change in energy consumption in all building energy systems rather than just the change in the lighting energy.
The FEDS cost data consists of state averages. Data sources vary and include RS Means and information from vendors,
suppliers, and contractors.
SUMMARY AND PATH FORWARD
The asset rating tool is not just a rating tool but is aimed at providing a cost-effective means for owners and operators
to gain insight into potential for capital improvements to increase energy efficiency and reduce costs. A full-scale detailed
186 --`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---

ASHRAE Transactions
energy model is more accurate but would require a substantial investment from the user in collecting energy audit data and
hiring of a professional to perform energy modeling and analysis. In comparison, a simplified energy model requires fewer
input combinations, which could reduce opportunities for data entry error and allow an inexperienced user to quickly
develop energy models. To balance these tradeoffs, PNNL is developing a full-scale energy modeling approach with
simplified data entry by classifying the inputs and employing an inference engine to provide different levels of analysis
based on the input levels. The inference engine can effectively reduce the data-gathering complexity. It provides a more
accurate estimate of a building parameter than a single default value such as one specified in an energy or equipment
efficiency code or an expected sector average value because it considers ages, use type, equipment type, size, and climate.
To ensure the repeatability of the modeling results, PNNL is developing an Internet-based online tool to guide data
collection, store building information, and generate the asset rating and report. Once a user submits building information,
the asset rating tool will run a real-time energy simulation, using EnergyPlus as the selected modeling engine. The user is
expected to receive the result within a reasonable amount of time, depending on the complexity of the building systems.
The rollout of the asset rating program has been organized in a phased approach, focusing first on building types with
typically fewest complexities and about which we have more information to establish a rating system. Phase one, which will
be included in the initial rollout, includes buildings in the office, school, retail, unrefrigerated warehouse, and assembly
categories. Phase two includes mixed-use types of buildings that incorporate phase-one use types, lodging buildings, food
service, and food sales. Phase-three building types are those with either more complex systems or those for which we
currently have a limited body of information, such as data centers, laboratories, refrigerated warehouse, and health-care
facilities.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
The asset rating tool can be used for both new constructions and existing buildings. For new constructions, it is a
benchmark tool for preconstruction evaluation providing comparison to a peer group. For existing buildings, it serves as a
tool to obtain system evaluation and measures to improve performance. The asset rating tool will be tested further during
the pilot project in 2012. The pilot project will collect feedback on the estimated data collection time and resources, test the
accuracy of asset rating energy model, and examine the relevance of the energy efficiency recommendations.
REFERENCES
American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc. (2004). Procedures for Commercial Building
Energy Audit. Atlanta, GA: ASHRAE.
ANSI/ASHRAE Standard 62.1. (2004). ANSI/ASHRAE Standard 62.1-2004 Ventilation for Acceptable Indoor Air Quality.
Atlanta, GA: American Society of Heating, Refrigerating and Air-Conditioning Engineers, ASHRAE.
ANSI/ASHRAE/IESNA Standard 90.1. (2007). ANSI/ASHRAE/IESNA Standard 90.1-2007 Energy Standard for Buildings
Except Low-Rise Residential Buildings. Atlanta, GA: American Society of Heating, Refrigerating and Air-Conditioning
Engineers,.
Architectural Energy Corporation. (2010). Commercial Buildings Energy Modeling Guidelines and Procedures. Oceanside, CA:
RESNET Publication.
ASHRAE Standard 62. (1999). ASHRAE Standard 62-1999 -- Ventilation for Acceptable Indoor Air Quality. Atlanta, GA:
American Society of Heating, Refrigerating and Air-Conditioning Engineers, ASHRAE.
Building Performance Institute Europe. (December 2010). Energy Performance Certificates across Europe: From design to
implementation. Brussels, Belgium.
California Energy Commission. (2000). How to hire an energy auditor to identify energy efficiency projects.
Deru, M. (2007). Energy Savings Modeling and Inspection Guidelines for Commercial Building Federal Tax Deductions, Second Edition.
Golden, Colorado: National Renewable Energy Laboratory.
ISO. (2008). Energy Performance of Buildings -- Calculation of Energy Use for Space Heating and Cooling. ISO 13790:2008. Geneva:
International Organization for Standardization.
Lee, S., Fei, Z., & Augenbroe, G. (2011). The Use of Normative Energy Calculation Beyond Building Performance Rating
Systems. Proceedings of Building Simulation 2011 (pp. 2753-2760). Sydney, Australia: International Building Performance
Simulation Association.
$6+5$(

National Renewable Energy Laboratory. (2011). U.S. Department of Energy Commercial Reference Building Models of the National
Building Stock. Golden, Colorado: U.S. Department of Energy, Office of Energy Efficiency & Renewable Energy.
Pacific Northwest National Laboratory. (2008). Facility Energy Decision System User’s Guide Release 6.0. Richland, Washington:
Battelle Memorial Institute.
Technical Committee CEN/TC BT/WG173. (2006). Energy performance of buildings — Overall energy use, CO2 emissions and
definition of energy ratings. Technical Committee CEN/TC BT/WG173 “Energy performance in buildings directive”.
U.S. Energy Information Administration. (2003). Energy Use in Commercial Buildings. Retrieved from
http://www.eia.gov/emeu/cbecs/cbecs2003/detailed_tables_2003/2003set9/2003pdf/c4.pdf
U.S. Environmental Protection Agency. (2012). ENERGY STAR Portfolio Manager. Retrieved from
http://www.energystar.gov/index.cfm?c=evaluate_performance.bus_portfoliomanager
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
188
ASHRAE Transactions
SA-12-C024
Calibration of a Building Energy Model

Considering Parametric Uncertainty
Zheng O’Neill, PhD, PE Bryan Eisenhower, PhD Vladimir Fonoberov, PhD

Member ASHRAE Member ASHRAE
Trevor Bailey, PhD
ABSTRACT
This article presents the calibration of a building energy model of a historic office building, developed using the EnergyPlus simulation
program. The building under study is the Fleet and Family Support Center located at Naval Station Great Lakes, IL. It was built in 1901
and renovated multiple times. This building has a total floor area of 36,843 ft 2 (3424 m2) and mainly consists of offices and conference rooms.
An extensive sensitivity study that efficiently perturbs more than two thousand model parameters is employed for model calibration. Those
parameters that most affect the building’s energy end-use are selected and automatically refined to calibrate the model by applying an analytic meta-
model based optimization. Real time data including weather and energy meter data in 2010 is used for model calibration and 2011 data is used
for model verification. The modeling process, calibration and verification results, as well as implementation issues encountered throughout the model
calibration process from a user’s perspective are discussed. The total facility and plug electricity consumption predictions from the calibrated
EnergyPlus model match the actual measured monthly data within ±5%.
INTRODUCTION
The total energy consumption for US commercial buildings was 17.43 quads (CBECS, 2003), approximately 18% of
the total U.S. energy consumption. The Department of Energy (DOE), the International Energy Agency (IEA),
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Intergovernmental Panel on Climate Change (IPCC) and other agencies have declared a need for commercial buildings to
become 70-80% more energy efficient. Building energy modeling recently has received increased attention as a tool to
reduce building energy consumption. Most of time, energy models provide hourly calculations of building energy
consumption, HVAC (Heating, Ventilation and Air-conditioning), and lighting systems performance, taking into account
the dynamic interactions among the building envelope, airflow, weather, internal loads, building usage, equipment, and
controls. Energy models can be used for evaluations of different concepts during the building design stage and reference
points for building real time performance monitoring and energy diagnostics during the operation stage. The performance
generated by the energy model, which represents “design intent” or ideal performance, can be compared with real time
measured data from the building. The performance deviation will indicate sub-optimal operation or faults.
Model calibration and validation is essential to ensure that the building and HVAC systems are properly modeled and
integrated together to predict the building energy performance. Model based assessment of different Energy Conservation
Measures (ECM) has been widely applied in building energy community. Kaplan and Canner (1992) have made
recommendations for the maximum allowable difference between predicted and monitored data. For instance, the
Zheng O’Neill is research engineer. Trevor Bailey is project leader at United Technologies Research Center, East Hartford, CT. Bryan
Eisenhower is associate director at the Center for Energy Efficient Design, University of California, Santa Barbara, CA. Vladimir Fonoberov
is chief scientist at Aimdyn, Inc., Santa Barbara, CA.
$6+5$(

prediction of energy use for interior loads such as lighting, receptacles, or domestic hot water is satisfactory when the
difference is within 5% on a monthly basis and 15% on a daily basis. However, the acceptable difference may increase up to
15%–25% (monthly) and 25%–35% (daily) for the simulation of HVAC systems. The annual simulated energy use should
be within 10% of collected information, while a difference of less than 25% is acceptable on a seasonal basis. These
recommendations are well aligned with conclusions from a recent workshop at the ASHRAE existing building conference
(Tuluca, 2010). In this paper, an EnergyPlus model for a historic office building was calibrated and validated with real time
measured data by using sensitivity analysis and automated tuning of input parameters. The adequacy of this calibration was
evaluated against the ASHRAE Guideline 14-2002. The purpose of this study is to provide a calibrated model as a reference
point for model based performance monitoring and energy diagnostics.
The building used in this case study is Building 26, Fleet and Family Support Center (FFSC)/Navy Marine Corps
Relief Society (NMCRS), at Naval Training Center, Great Lakes, IL. It is a two-storey office building with basement. The
gross area of this building is approximately 37,000 ft2 (3,437 m2).This building has special historical masonry construction
and is one of the original barracks of Naval Station Great Lakes dating back to 1901. Since then, there were several major
renovations including windows upgrades, exterior repairs and a HVAC system retrofit in 2006. Figure 1 shows the front
view of this building.
Figure 1 Building 26 at Naval Station Great Lakes
The Building 26 HVAC system consists of two airside systems and two separate waterside systems. The office and
administrative area on the first and second floors is served by two variable-air volume (VAV) Air Handling Units (AHU)
with VAV terminal units (with hot water reheat). These AHUs have both heating and cooling capability. Operation of these
units depends on the occupancy of the building. The chilled water system consists of one 54.5-ton air-cooled rotary-screw
type chillers with fixed-speed primary pumping. Heating is supplied from the existing base-wide steam system through a
steam-to-water heat exchanger. The hot water serves unit heaters, VAV box reheating coils, and air handling unit heating
coils while electric unit heaters and baseboards are used to provide heating to stairwells and restrooms. The communication
service room is served by one dedicated split system. Table 1 lists major HVAC equipment used in Building 26. A
distributed Direct Digital Control (DDC) system is installed in this building which monitors all major environmental
systems. In addition to this, building electric and water meters are being read by the DDC system. Operator workstations
provide graphics with real-time status for all DDC input and output connections.
Table 1. Major Equipment Used in Building 26
Equipment Number
Duct free split system 1
Air cooled screw chiller 1
Variable volume air handler unit 2
Hydronic unit heater 4
Electric unit heater 2
Electric baseboard 4
VAV box with hot water reheat coil 38
Pumps 6
Additional meters and sensors are required to calibrate models and accurately measure energy consumption to validate
results. An on-site weather station, including a pyranometer, aspirated wet and dry bulb temperature sensors, and wind
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
190
ASHRAE Transactions
speed and direction sensors, is installed on the roof. BTU meters (a matched pair of supply and return water temperature
sensors, water flow meters) are installed for the chiller and hot water loop. Lighting load power, plug load power and
individual chiller power are also monitored through sub meters. These sensors and meters are integrated into the existing
building Energy Management and Control System (EMCS).
ENERGYPLUS MODEL
A whole-building EnergyPlus simulation model representing the performance of the envelope, HVAC, lighting, water,
and control systems was developed in EnergyPlus (EnergyPlus, 2011), which is a whole-building simulation program
developed by the United States Department of Energy. An EnergyPlus model takes as input a description of the building
(e.g., geometry, materials, roof type, window type, shading geometry, location, orientation etc.), its usage and internal heat
loads (as a scheduled function of time), and the HVAC equipment and system description (e.g., chiller performance), and
then computes the energy flows, zonal temperatures, airflows, and comfort levels on sub-hourly intervals for periods of
days to years. There are 2063 input parameters (parameters with numerical values) for this EnergyPlus model.
The EnergyPlus interface used for this study is DesignBuilder (DesignBuilder, 2011), which allows for a graphical
display of all the three-dimensional geometry. After the geometry is entered into DesignBuilder, an IDF file with all
geometry information is exported, and then EnergyPlus IDF Editor is used to create the HVAC system model. The image
in Figure 2a contains rendered geometry outline generated by DesignBuilder.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 2 (a) Building 26 rendered geometry (b) Floor 2 zoning in Building 26
The version of EnergyPlus used in this study was 6.0 (build 6.0.0.023). Annual real time weather data in for 2010,
including outside dry bulb temperature, wet bulb temperature, wind direction and speed, direct normal solar radiation and
diffuse horizontal radiation etc., were used as ambient conditions for the simulation. In order to keep the size of the model
and computation time manageable, zoning simplifications were made when entering the building geometry. All the rooms
served by the same VAV box were integrated into one thermal zone. The building model consists of 44 conditioned zones
(5, 17, and 22 zones for the basement, first, and second floors respectively). Some zones represent a physical room in the
building while other zones represent adjacent multiple rooms operating under similar energy usage/requirements. Each
zone includes an "internal mass" that represents the thermal storage capacity of the room(s) (e.g., interior walls, furnishings,
books, etc.). Figure 2b indicates the zoning used for the second floor in Building 26. This includes 4 unconditioned zones.
HVAC System Model

HVAC Zone Setup. For the zones served by VAV boxes with reheating coils, the EnergyPlus Object
(AirTerminal:SingleDuct:VAV:Reheat) is used. EnergyPlus Objects (ZoneHVAC:Baseboard:Convective:Electric and
ZoneHVAC:Baseboard:Convective:Water) are used to model the electric baseboard and hydronic unit heater. Thermostat
schedules for all zones are as follows:
Cooling set point: 76 °F (24.4°C) occupied, 86 °F (30°C) unoccupied
Heating set point: 70 °F (21.1°C) occupied, 59 °F (15°C) unoccupied
Hot Water and Chilled Water Distribution Loops. The heating water distribution loop in the building is modeled
as a variable flow system including variable speed drives on the pumps, which are modeled with premium efficiency motors.
$6+5$(

The chilled water distribution loop in the building is modeled as constant flow system including constant speed drive on the
pump. The chilled water loop is modeled with a set-point temperature of 45 °F (7.2 °C). The heating-water loop is modeled
with a set-point temperature of 180 °F (82.2°C). Pump power consumption is described by the following part load
performance curve.
Fraction Full Load Power = C1 +C2PLR+C3PLR2+C4PLR3 (1)
The coefficients C1, C2, C3, and C4 are constant parameters and PLR is the Part Load Ratio (i.e., PLR=water flow
rate/design water flow rate).
Chiller Model. One 54.5-ton air cooled chiller (Carrier 30RAN055-61PK) is used in the chiller plant. This chiller
model is an empirical model, which is commonly used in the EnergyPlus building energy simulation program (DOE, 2010).
The model uses performance information at reference conditions along with three curve fits for cooling capacity and
efficiency to determine chiller operation at off-reference conditions. Chiller performance curves are generated by fitting
information from a manufacturer’s catalog. Cooling is available from April 15 th to October 15th, and whenever the outside
air temperature is greater than 58 °F (14.4 °C), the chiller is turned on. Whenever the outside air temperature is less than 56
°F (13.3 °C), the chiller is turned off.
MODEL CALIBRATION
Originally, the EnergyPlus model for Building 26 was created and selection of its 2063 parameters was performed
using the best information that was available at the time. For example, the plug and lighting load profiles were generated by
using real time submetered data for two months (March and April) in 2010. HVAC system operation schedules and
setpoints were taken from actual building EMCS data. During model calibration, a smaller subset of the 2063 parameters
that are most critical was identified using sensitivity analysis and subsequently automatically tuned so that the model better
matches measured data. Model calibration was performed using annual measured data from 2010, and the outputs of the
model were then compared to measured data for a few months in 2011 in which the model was not specifically tuned to
match (model verification).
Predominate historical data included monthly metering data for total building electricity, plug load electricity and
steam usage. This data had been recorded for 2009, 2010, and part of 2011 at the time of this study. Since there were issues
with condensate meter, the steam usage data was not selected for the calibration analysis.
Calibration Approach
The model calibration process relies heavily on characterizing parametric influences on the outputs of the model.
This analysis is performed by sampling all parameters of the model around their nominal value to create a database of
output data which is used to calculate the sensitivity of these outputs to parameter variation as well as to derive an analytic
meta-model based on this model data (Eisenhower et al. 2011a). Once the most influential parameters (on the order of 10
to 20) of the model are identified, an optimization can be performed (using the meta-model) in order to identify parameter
combinations that produce the best fit to measured data. Only 10-20 of the most influential input parameters instead of
thousands were optimized during the optimization/calibration process to avoid the issue of over fitting the model.
This type of work has been performed in a manual way in a previous project (O’Neill et al., 2011). A schematic of the
automated process used in this paper is presented in Figure 3. Although the case study presented in this paper is focusing
on the calibration of an EnergyPlus model, the proposed calibration approach can be implemented for any other building
simulation programs.
Uncertainty Analysis (UA) and Sensitivity Analysis (SA)
The purpose of the uncertainty and sensitivity analysis is to identify which of the entire list of parameters (i.e., 2063
parameters in this study) are best to use for calibration purposes. In order to do this, a list of the parameters and their

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---

Create energy
model Perform UA/SA
Identify uncertain
Calculate reduced-
parameters,
order meta-model
perform sampling
Perform numerous Perform

simulations, pre- calibration/optimiz
process output ation
Calculate full order

Verify calibration
meta-model
Figure 3 Schematic of the proposed calibration process
nominal values is collected and a range is then created which spans +/- 25% of the nominal value using a uniform
distribution if the nominal value is nonzero and an exponential distribution otherwise. 6500 samples are created in this
range, which are concurrently perturbed using a deterministic sampling approach (Aimdyn, 2010). The EnergyPlus models
associated with these samples are then simulated in parallel on a 184 CPU Linux cluster to generate output data for each of
these instances.
Once the output data (e.g., monthly plug and total electricity consumption in this study) is generated, sensitivity
analysis is performed to rank-order the parameters in terms of their influence on the output. An example of the sensitivity
calculation for total electricity in March, 2010 is presented in Figure 4. In this figure, all 2063 sensitivity indices are plotted
while the top 10 most influential parameters that influence total electricity in March are highlighted in the legend. The
sensitivity index is the relative influence that each parameter has on a particular output. For months in the cooling season,
different input parameters including chiller capacity and chiller reference COP are identified to be most influential
parameters using the same approach.
March Total Electricity [kW]
0.18
0.16 All Data

FL2_Zone5 plug load power density
0.14
Sensitivity Index
0.12 Hot water pump rated power

0.1
0.08 FL1_Zone1 plug load power density
0.06
Hot water loop maximum water temp
0.04 FL1_Zone11 plug load power density
Office lighting schedule fraction weekend
0.02
0
0 200 400 600 800 1000 1200 1400 1600 1800 2000
Parameter Index
Figure 4 Sensitivity indices for March total electricity illustrating the top ten parameters that influence this model
particular model output.
Optimization (Calibration)
To perform the actual calibration, a mathematical optimization problem is defined in which parameters are varied in
order to minimize the difference between the model output and measured data. Optimization using whole-building energy
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

models is often challenging due to the computationally expensive nature of the simulation as well as discontinuities that are
often found in the cost surface (Wetter and Wright, 2003). In order to circumvent these issues, an analytic model of the full
EnergyPlus model (meta-model) is created using a machine learning regression technique (Eisenhower et al. 2011b).
With an analytic representation of the building dynamics, rapid optimization is performed on the meta-model using an
interior point method (a gradient based optimizer) and confirmed using the full EnergyPlus model. Since the function
evaluation is so rapid, optimization experiments can be performed with thousands or just a few key parameters of the
model. In order to perform the optimization, a cost function is defined as the following:
¦ (model data ) 2
where the two variables under the radical are either monthly or annual energy consumption in this study.
An issue that arose during the calibration process was the significant disparity between the un-calibrated model and
measured data. In the first subplot of Figure 5, the uncertainty distribution for January total electricity consumption is
presented along with the prediction of the baseline un-calibrated model (red dot) as well as measured data (blue dot). It is
evident in this plot that the measured data is significantly far from the baseline model such that changing the parameters by
+/- 25% does not move the output into the range of the measured data. This is an issue for the calibration process because
the meta-model that is used for calibration is most accurate where the uncertainty data was generated and becomes less
accurate on the tails of the distribution (under the black curve in Figure 5a). To alleviate this concern, optimization was
performed with constraints on the output variables. Optimal parameters were defined such that the output did not leave an
ellipse (in black color) that encompasses the data (as illustrated in Figure 5b).
4
x 10 March 2010
0.12
2.4
Sampled Data Sampled Output
0.1 Measured Data 2.3 Sensor Data
Nominal Model Baseline Model
Total Electricity [kWh]
2.2
0.08 Calibrated Meta
2.1 Calibrated E+
Probability
0.06 2
0.04 1.9
1.8
0.02
1.7
0 1.6
1.6 1.7 1.8 1.9 2 2.1 2.2 1.1 1.2 1.3 1.4 1.5 1.6
Total Elec in kWh for January 2010 4 Plug Electricity [kWh] 4
x 10 x 10
Figure 5 (a) Raw uncertainty distributions with measured data (b) Results from constrained optimization
The optimization results for monthly energy consumption are presented in Figure 6 where it can be seen that there is
significant improvement with respect to the ability of the model to represent measured data after the automated calibration.
The total facility and plug electricity consumption predictions from the calibrated EnergyPlus model match the actual
measured monthly data within ±5%. The calibrated model outputs and measured data are also compared in terms of Mean
Bias Error (MBE) and coefficient of variation of Root Mean Squared Error (CV(RMSE)). Table 2 shows that both MBE
and CV(RMSE) are far less than the limits defined by ASHRAE guideline 14-2002, respectively, ±5% and ±15%
(ASHRAE 2002). In each case, only the top 10-20 significant parameters were used for the calibration. Table 3 shows final
values for the top 10 input parameters (in March, 2010) identified by the optimization/calibration process. It should be
noted that these results are from the optimization with output constraints and the error can be reduced if these constraints
are lifted. These constraints can be lifted by moving the cloud of sampled data (with which the meta-model is derived)
closer to the measured data. This can be done by either larger perturbations on the sampled inputs, or by moving the
nominal value closer to the measured data. One recommended way to achieve this is to perform SA on the entire list of
parameters with few samples to determine strong parameters, and then hand tune these strong parameters to get the model
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
194
ASHRAE Transactions
output mean closer to measured data. Another SA will be followed only for these strong parameters with larger ranges. The
proposed automated calibration then will be performed.
16000 40000
14000
Plug Electricity 35000
Total Electricity
12000 30000
10000 25000
kWh
kWh
8000 20000
6000 15000
4000 10000
2000 5000
0 0
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec 1 2 3 4 5 6 7 8 9 10 11 12
2010 meter data E+ after calibration E+ before calibration 2010 meter data E+ after calibration E+ before calibration
Figure 6 Calibration results for each month in 2010
Table 2. Calibration Results for Building 26 EnergyPlus Model

Error Plug Electricity Total Electricity
MBE 0.02% -2.31%
CV(RMSE) 0.47% 2.80%
*Equations to calculate MBE and CV(RMSE) are from ASHRAE Guideline 14-2002 (ASHRAE 2002).
Table 3. Top 10 parameters values before and after calibration

Parameters Before Calibration After Calibration
FL2 Zone5 plug load power density W/m2 20 17.513
Hot water pump rated power consumption W 2238 2797.498
FL1 Zone4 plug load power density W/m2 14.5 18.118
FL1 Zone1 plug load power density W/m2 14.5 18.106
FL0 Zone5 plug load power density W 600 652.766
Hot water loop max water temperature °C 100 85.758
Office lighting schedule fraction weekend 0.26 0.250
*Top 10 parameters were selected based on the sensitivity indices (see the Figure 4).
Validation
The above calibration process illustrates how well a model can be tuned to fit a pre-described measured data set.
Once the model is calibrated, it is desirable for the model to have prediction capability for future data. This validation
process was performed using three months of measured data from 2011 (recall that the model was calibrated for 2010 data)
and presented in Figure 7. This Figure illustrates that the error in prediction using the calibrated model (Verification with
2011 weather in blue bar) is acceptable and better than the un-calibrated model (Nominal with 2011 weather in green bar).
There is a significant error in the month of May which is due to a chiller failure (confirmed with facility team), which did
not occur when the model was calibrated with 2010 data. In a sense, this verification test for May, 2011 illustrated an
unexpected excursion in the data due to an equipment fault which is well predicted by the validated model. This
demonstrates that a well calibrated energy model can be a means for fault detection and diagnostics.
CONCLUSIONS
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`
$6+5$(

25
Verification with 2011 weather
20 Nominal with 2011 weather
Calibration with 2010 weather
15
10
Error in %
5
-5
-10
March April May
-15
1 2 3
Figure 7 Verification results for 2011
We presented a systematic and automated way to calibrate building energy models with a large number of uncertain
input parameters. A subset of these parameters was identified using sensitivity analysis and subsequently automatically
tuned so that the model better matches measured data. Using the proposed approach, the total facility and plug electricity
consumption predictions from the calibrated EnergyPlus model match the actual measured monthly data within ±5% for a
historic office building. The calibrated model gives 2.80% of CV(RMSE) and -2.31% of MBE for the whole building
electricity use, which is acceptable based on the ASHRAE Guideline 14.
ACKNOWLEDGMENTS
This work was performed under the project EW09-29 administered by ESTCP (Environmental Security Technology
Certification Program) technology program of the Department of Defense. We would like to thank ESTCP program
manager, Dr. James Galvin and facility manager at Great Lakes, Mr. Peter Behrens, for their support. We would also like to
thank Satish Narayanan (UTRC) and Professor Igor Mezic (UCSB) for insightful discussions on this topic. Views,
opinions, and/or findings contained in this paper are those of the authors and should not be construed as an official
Department of Defense position or decision unless so designated by other official documentation.
REFERENCES
Aimdyn GoSUM Software, 2010. Global Optimization, Sensitivity and Uncertainty in Models
(GoSUM) [online]. Santa Barbara, CA: Aimdyn Inc. Available from: http://aimdyn.com [Accessed 11 August 2011].
ASHRAE. 2002. ASHRAE Guideline: Measurement of energy demand and savings. ASHRAE Guideline 14-2002.
CBECS. 2003. http://www.eia.gov/emeu/cbecs/
DesignBuilder, 2011. http://www.designbuilder.co.uk/
DoE. 2010. EnergyPlus documentation: engineering reference.
Eisenhower, B., O’Neill, Z., Fonoberov, V. A., and Mezic, I. 2011a. Uncertainty and sensitivity decomposition of building
energy models. Journal of Building Performance Simulation, In Press. DOI: 10.1080/19401493.2010.549964
Eisenhower, B., O’Neill, Z., Narayanan, S., Fonoberov, V. A., and Mezic, I., 2011b. A methodology for meta-model based
optimization in building energy models. Energy and Buildings. In Press doi:10.1016/j.enbuild.2011.12.001.
EnergyPlus. 2011. http://apps1.eere.energy.gov/buildings/energyplus/
Kaplan, M., and P. Canner.1992. Guideline for Energy Simulation of Commercial Buildings. Portland: Bonneville Power
Administration.
O'Neill, Z., Eisenhower, B., Yuan, S., Bailey, T., Narayanan, S. and Fonoberov, V. 2011. Modeling and calibration of
energy models for a DoD Building. ASHRAE Transactions, 117(2): 358–365.
Tuluca, A. 2010. Modeling for Existing Buildings. ASHRAE Existing Building Conference. New York, NY.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Wetter, M. and J, Wright. 2003. Comparison of a generalized pattern search and a genetic algorithm optimization method.
Proceedings of the Eighth International IBPSA Conference, Eindhoven, Netherlands, pp. 1401–1408.
196
ASHRAE Transactions
SA-12-C025
Use of Flexible Research Platforms (FRP)

for BIM and Energy Modeling Research
Piljae Im, PhD Mahabir Bhandari, PhD

ABSTRACT
Two light commercial building flexible research platforms (FRP) are being constructed on the Oak Ridge National Laboratory (ORNL)
main campus as part of an American Recovery and Reinvestment Act (ARRA) funded project for a Maximum Energy Efficiency Laboratory.
The FRPs will offer a unique capability of exposing ‘test buildings’ to natural weather conditions for purposes of R&D leading to building system
and component -level advanced energy efficiency solutions for new and retrofit applications. The platforms provide the opportunity to prove solutions
in a low risk environment so they can be deployed with speed and scale in actual commercial buildings. Initially each platform will accommodate
one test building with simulated occupancy and miscellaneous electric loads (MELs) to serve as a baseline. The test building will then be modified
to test a series of retrofits by replacing building envelope systems and HVAC systems over time. Performance will be monitored with an extensive
suite of sensors, with data collected at 15 minute intervals. Comparisons can then be made between the measured performance of the whole building
(and its individual components) and the corresponding energy modeling performance. This data will be also used to calibrate the building energy
models, validate several EnergyPlus algorithms, and estimate the energy savings pre- and post-retrofit.
This paper introduces two major research areas of the FRP project; 1) Evaluation to facilitate seamless integration of Building Information
Modeling (BIM) into building energy models (BEM), and 2) Calibration and validation of the building energy model with the detailed
monitoring data. The BIM to BEM evaluation will explore the benefit of forming a direct connection between BIM and energy analysis software.
The overall work flow of the integration process, as well as interoperability between BIM and energy analysis software, will be evaluated using
existing building simulation packages having import/export capability. In energy modeling research, the modeling of the FRPs, measurement and
verification (M&V), and further calibration of models will be investigated. This work could then serve as reference material for the building
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
modeling community and building owners. The calibration - envisaged in this project will serve to generate confidence in the use of energy modeling
tools and their predicted energy savings for small commercial buildings. In the long term, accurate modeling and M&V will result in increased
energy savings, better data to assess financing for retrofits, as well as more accurate emissions predictions.
INTRODUCTION
In general, light commercial buildings are defined as those 50,000 ft2 (4,645 m2)and less in size, which represent 95%
of the nation’s commercial buildings and 50% of the nation’s commercial building floor space (EIA 2003, Table A1). Light
commercial buildings are responsible for 47% of commercial building energy bills (EIA 2003, Table C2A), or about $80
billion annually (DOE 2009, Table 3.3) at today’s energy rates. A reasonable range of potential energy retrofit savings in
existing commercial buildings is between 10 and 20 percent of present energy consumption using today’s technology
(Belzer 2009). Assuming a five year simple payback on average and 10 percent of the existing building stock addressed
annually, today’s technology could support a $4 to $8 billion per year industry focused on energy retrofits in light
commercial buildings. DOE’s goal is to significantly reduce energy use in existing and new light commercial buildings with
Piljae Im and Mahabir Bhandari are R&D Staffs in the Building Technology Research & Integration Center (BTRIC), Oak Ridge National
Laboratory, Oak Ridge, TN.
$6+5$(

technologies that are affordable, reliable, and deployable with speed and scale. Building Information Modeling (BIM)
(Eastman et al, 2008) and Building Energy Modeling (BEM) could play a major role in reducing the energy use in light
commercial buildings.
Due to the severe cost and schedule pressures in construction and retrofit of light commercial buildings, the service
providers (i.e., architects, engineers, contractors) in this market segment are likely to embrace new ways to reduce expenses.
Software interoperability has the opportunity to add great benefit because it enhances labor productivity and reduces costs.
Whether BIM/BEM tools will be feasible for use on each and every light commercial project versus being used to pre-
establish standardized prescriptive solutions remains to be seen. In either case, it is necessary to benchmark where these
tools are today. It is therefore necessary to utilize these tools in an actual design, construction, and retrofit project to
determine whether current pre-establish, standardized prescriptive methods are sufficient and cost effective, and to develop
better tests and methodologies than the current prescriptive method. Those could be simplified modeling methods from
detailed drawings and BIM files (e.g., automatic generation of simulation input), building geometry generated from building
photographs, or smart templates for a building energy model of a retrofit case.
The validation of the building energy modeling tools is extremely important in order to give the user confidence in
these tools and the predicted results. Judkoff et al (2008) discuss the relative merits of three different validation techniques
used in building energy simulation software. These techniques include: 1) analytical validation (comparing program results
to an analytical solution), 2) comparative validation (program-to-program comparisons), and 3) empirical validation
(comparing results with an actual experiment). Analytical and comparative validations seem to be the most common
approaches for validation due to the high cost involved in empirical validation. In addition, empirical techniques using
laboratory test cells have been carried out by various researchers. Generally, these experiments are geared towards validating
particular building components or algorithms. There is a need for a representative real building exposed to the natural
environment to measure and predict the performance of energy saving measures and to validate building energy models,
including energy model algorithms for various components. One main barrier for this empirical validation with real
buildings is the uncertainty due to building occupant behavior. In real buildings, controlling occupancy and miscellaneous
electric loads (MELs) are almost impossible. With these needs, two Flexible Research Platforms (FRPs) are being
constructed at ORNL to expose these test facilities to natural weather conditions with controlled (simulated) occupancy and
MELs, so that these platforms represent real light commercial buildings with greater certainty. The FRPs will be equipped
with more than 150 sensors for extensive measurements and a dedicated weather station to create the actual weather file.
The platforms provide a unique opportunity to prove solutions in a low risk environment so they can be deployed with
speed and scale in actual commercial buildings, and to validate the whole building energy models and selected building
modeling algorithms. More than 18 industry partners from building envelope, HVAC, control, BIM and energy modeling
industries will be participating in this project as of December 2011. This paper focuses only on the use of these unique
FRPs to demonstrate Building Information Modeling (BIM) in design, energy modeling, construction, model validation and
commissioning and calibration of building energy models and validation of several algorithms within DOE’s flagship
building energy analysis tool, EnergyPlus (EnergyPlus, 2011).
FLEXIBLE RESEARCH PLATFORMS

Flexible Research Platforms, consisting of slabs and steel, are part of a multi-year project. This project, with the goal
being to install and monitor temporary, instrumented baseline ‘test buildings’, was established as part of the ARRA-funded
Maximum Energy Efficiency Laboratory (MAXLAB) project. The test buildings will be exposed to natural weather, with
data used to characterize baseline energy performance. Then energy-saving tune-ups or retrofits to the test buildings will be
implemented and exposed to natural weather, with measured data used to characterize the tune-up or retrofit energy-
savings. The single story FRP, having a footprint of 40’x60’ (12 m x 18 m), and the two-story FRP, having a footprint of
40’x40’ (12 m x 12 m) (Figure 1a and 1b), will be used to physically simulate light commercial buildings common in the
nation’s existing building stock, energy-saving tune-ups and retrofits that can be implemented on such buildings. The FRPs
are unoccupied research apparatuses where occupancy will be simulated by process control of lighting and other internal
loads to minimize human interference with the building, since human interference is one of the main factors of uncertainty

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---

Permanent Apparatus
(a)
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Permanent Apparatus With Test Buildings
(b)
Figure 1 Sketches of Flexible Research Platforms (FRPs) with (a) Permanent apparatus and (b) Test building.
in building energy use. The ORNL research team will be able to force faults in order to test fault detection and diagnostics
(FDD) capabilities embedded in HVAC systems or other active load components including the use of process control. A
variety of HVAC systems are simultaneously installed in each FRP, so their baseline and improved performance can be
experimentally characterized on a common basis by rotating operation from one system to the next and exposing each
system to a range of operating conditions. This approach enables a large number of physical cases to be tested on the FRPs
without major renovations to the test buildings. Each FRP will utilize several hundred channels of measured data to achieve
performance characterization, FDD verification, and other experimental objectives. In the first year, the first temporary
baseline ‘test buildings’ will be designed, erected, instrumented, and commissioned. ORNL will then collaborate with
remote industry partners for R&D using the light commercial building FRPs. In subsequent years, the baseline test building
$6+5$(

will be modified to test a series of ‘retrofits’, or replaced with new ‘test building’.
The single story FRP will be a conventional pre-engineered metal building structure. The building envelope
manufacturers’ industry association partner in this project is interested in determining the most cost effective methods to
improve the energy efficiency of both existing and new pre-engineered metal buildings. The partner wants to demonstrate
that this type of building structure can continue to meet and exceed the minimum building performance criteria as defined
in ASHRAE 90.1 and continue to compete in a cost effective manner compared to conventional building structures.
The envelope components that will be assessed in the 2-story FRP represent a significant portion of existing 20-year-
old light commercial buildings. Forty-one percent of these structures have brick, stucco or stone cladding; and 29% have
either single-ply or build-up roofing (EIA 2003, Table B6). The building envelope research team will develop the test plan
to characterize the energy performance of the baseline structure, design and install the instrumentation package, oversee the
data acquisition system, collect data, and share the data with industry partners. The researchers, in consultation with
industry partners for wall construction, will develop the wall retrofit and its test protocol. Likewise industry partners for
roof construction will participate in the formation of a low-sloped roof retrofit and the experimental plan to evaluate its
performance.
Equipment research is another important aspect of this project. For the 1-story FRP, the baseline metal test building
will be a single space (one HVAC zone) with two HVAC system types installed in parallel. Preliminary selections include a
packaged rooftop unit (RTU) with Direct-Expansion (DX) cooling and natural gas heating, and a gas engine-driven heat
pump packaged rooftop unit. For the 2-story FRP, the baseline brick-walled building with low-sloped roof will be a multi-
zone HVAC application (e.g., five 2-story zones, the 4 perimeter orientations and 1 core zone). Preliminary selections
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
include a packaged Variable Air Volume (VAV) RTU serving shutoff VAV boxes in each zone, multiple water-source heat
pumps (WSHPs), and a multi-split variable refrigerant flow (VRF) system. Both the SWHP system and the VRF system will
share a dedicated outdoor air system (DOAS) that will deliver tempered fresh air to each zone. Final HVAC system
determinations for both platforms will be made in consultation with the sponsor and industry partners. Figures 2 and 3
represent the sketches of possible test buildings with different building envelope and HVAC retrofit measures applied.
BIM to BEM Work Flow

Generally, modelers spend a large amount of time creating the building definition for the energy analysis software,
even though the building definition was previously created in some version of CAD or BIM software. Building geometry
design in CAD constitutes the major portions of overall BIM. To avoid redundant effort, various energy modeling software
vendors are striving to create a reliable mechanism to import the building geometry data from CAD/BIM software. In
addition to time saved in geometry creation, the additional benefits could be reduced cost of modeling and reduced errors.
Industry Foundation Classes (IFC), gbXml and Extensive Markup Language (XML) are the major formats being used for
BIM to BEM import by BEM software. However, in real practice, using BIM to BEM translation often requires extensive
troubleshooting procedures; so many modelers often prefer to use their traditional modeling method, i.e. manually
developing the building definition for the simulation model.
This FRP project aims to evaluate the benefits and limitations of the direct connection mechanisms between BIM and
BEM software. The overall work flow demonstrating interoperability between BIM and BEM software will be
demonstrated in the context of a real building. The interoperability will also be evaluated using existing building simulation
software and their included translation/import/export mechanisms.
Design workflows used in light commercial building architecture and engineering will also be evaluated. Working with
the FRP design team, the research team will use, evaluate, and document the model interoperability (i.e. import, export, and
translation) of several 2D and 3D design tools, HVAC system engineering tools, and tools for equipment selection and
system sizing. A set of test cases that will inform DOE's interoperability efforts going forward will be developed. These test
cases will be used to produce a detailed interoperability evaluation, including gap analysis, of a number of important
software products. The research team will work with key industry partners to improve the functionality, robustness and
200
ASHRAE Transactions
Figure 2 Sketches of 2 story FRP with different building envelope retrofit measures.
Figure 3 Sketches of 2 story FRP with different building HVAC retrofit measures.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

fidelity of BIM and BEM data exchange in existing software products. Autodesk, Bentley, and DesignBuilder are all active
partners in this project.
In parallel with design and construction, the research team will develop energy models of the baseline test buildings,
based on a BIM model developed by an Architectural and Engineering design firm and ORNL.
Calibration and Validation of Building Energy Models

With regard to energy modeling research, the modeling of the FRPs, M&V and further calibration of models will be
investigated. A base case model for the initial test buildings using the envelope and equipment details will be created in
EnergyPlus and the selected list of simulated parameters will be compared with the measured data collected at 15 minute
intervals. Software such as AECOsim, OpenStudio, DesignBuilder, Simergy, and eQuest with user friendly graphical user
interfaces will be used to model the FRPs. Findings will be used to determine the shortcomings of current data exchange
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
mechanisms between BIM and energy modeling products. An onsite weather station will provide the actual weather data
used in the simulations and will be prepared in the widely used TMY3 format. Based on current calibration methods, a
methodology will be developed to calibrate the base model. The following list of parameters will be compared with the
measured data (15 min interval):
1. Whole building electrical energy consumption (kWh), demand (kW), gas consumption (kBtu) and profile (monthly,
daily, hourly)
2. End use energy consumption (kWh, kBtu)
3. Zone air temperatures (°F), relative humidity (RH)
4. Heat flux through various building envelope elements (walls, windows, roof, floor (underground slab)) (Btu/hr-ft2)
5. Zone transmitted solar (W)
6. HVAC equipment power (kWh), fan electric power(kWh), pump electric power(kWh), boiler gas consumption
(kBtu), compressor electric power (kWh), gas heat exchanger fuel consumption (kBtu)
7. HVAC supply temperature (varied by HVAC type) (°F)
8. Infiltration rate based on blower door test and tracer gas method (ACH or CFM)
9. Illumination (lux level) map
The calibration methodologies developed during the first year will be refined in the subsequent years, as buildings are
modified with more efficient technologies. A series of sensitivity and parametric studies will be carried out to achieve this
goal. The sensitivity analysis will include scaling of model to include various building geometries, sizes, building types,
different climatic zones.
The whole building simulation validation, based on a comparison of measured and simulated values, will also help to
identify the source of errors, such as initial assumptions or algorithms at the component level. These individual aspects
could include ground temperatures, thermostat control based on operative temperature, daylight distribution in the zones
(illuminance maps), natural ventilation (airflow network model) and hybrid ventilation, and recent HVAC types including
localized cooling, etc. These tasks will be coordinated with the EnergyPlus development team, and the findings from this
research will be considered in a future version of EnergyPlus. The simulations, measurements and subsequent validation
from both buildings will help in generalizing the model validation techniques using larger sets of measured data.
SUMMARY
This paper introduces a project which offers the unique capability of exposing ‘test buildings’ to natural weather
conditions for purposes of R&D leading to system/building-level advanced energy efficiency solutions for new and retrofit
applications. The test buildings will provide the opportunity to prove solutions in a low risk environment so they can be
202
ASHRAE Transactions
deployed with speed and scale in actual commercial buildings. Two of the research aspects, BIM and BEM, were described
in detail. The BIM demonstration will explore the benefit of a direct connection between BIM and energy analysis software.
With regard to energy modeling research, the modeling of the FRPs, M&V and further calibration of models will be
investigated. For the first year, the research program will focus on: EnergyPlus modeling using BIM inputs to simulate
energy use, construction of the flexible research platform, deployment of the instrumentation and digital acquisition system,
data collection and analysis/comparison between model results and measurement, and validation analysis efforts for Energy
Plus. Following the acquisition of data for the baseline building, the program will then focus on retrofit construction
document preparation, additional energy modeling, retrofitting the building shell, data collection on the retrofitted platform
and analysis of the data. It is anticipated there will only be time to complete the initial baseline construction and analysis
and planning for execution of the retrofit case during FY12.
This work could then serve as reference material for the building modeling community and building owners. The
calibration capability envisaged in this project will serve to generate confidence in the use of energy modeling tools and
their predicted energy savings for small commercial buildings. In the long term, accurate modeling and M&V will result in
increased energy savings, better data to assess financing for retrofits, as well as more accurate emissions predictions.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
ACKNOWLEGEMENTS
Funding for this project was provided by field work proposal under the Department of Energy Building Technology
Activity Number BT0201000. Oak Ridge National Laboratory is managed by UT-Battelle, LLC, for the U.S. Dept. of
Energy under contract DE-AC05-00OR22725. This manuscript has been authored by UT-Battelle, LLC, under Contract
Number DE-AC05-00OR22725 with the U.S. Department of Energy. The United States Government retains and the
publisher, by accepting the article for publication, acknowledges that the United States Government retains a non-exclusive,
paid-up, irrevocable, world-wide license to publish or reproduce the published form of this manuscript, or allow others to
do so, for United States Government purposes.
DISCLAIMERS
This paper was prepared as an account of work sponsored by an agency of the United States Government. Neither
the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied,
or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus,
product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any
specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise, does not necessarily
constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof.
The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States
Government or any agency thereof.
REFERENCES
ASHRAE, 2005. ASHRAE Handbook-Fundamentals. Atlanta: American Society of Heating Refrigeration and Air
Belzer DB., 2009. Energy Efficiency Potential in Existing Commercial Buildings: Review of Selected Recent Studies.
PNNL-18337, Pacific Northwest National Laboratory, Richland, WA.
Bentley, 2011. AECOsim, Interdisciplinary Building Design, Analysis, and Simulation, http://www.bentley.com/en-
US/Products/AECOsim+Energy+Simulator/.
DesignBuilder, 2011. DesignBuilder building simulation [online]. Gloucestershire, UK: DesignBuilder Software Ltd.:
http://www.designbuilder.co.uk/.
Eastman, Chuck, Teicholz, Paul, Sacks, Rafael and Liston, Kathleen (2008). BIM Handbook: A Guide to Building
Information Modeling for Owners, Managers, Designers, Wiley. ISBN 978-0-470-18528-5.
$6+5$(

Energy Information Administration (EIA). 2003. 2Commercial Building Energy Consumption and Expenditures (CBECS),
Public Use Data, Microādata files on EIA website:
http://www.eia.doe.gov/emeu/cbecs/cbecs2003/public_use_2003/cbecs_pudata2003.html.
EnergyPlus, 2011. EnergyPlus Energy Simulation Software and supplemental documentations. http://www.energyplus.gov,
US Department of Energy.
eQuest, 2011. The Quick Energy Simulation Tool, http://doe2.com/equest/.
Judkoff, R., D. Wortman, B. O’Doherty, and J. Burch., 2008. A Methodology for Validating Building Energy Analysis
Simulations, National Renewable Energy Laboratory Technical Report NREL/TP-550-42059, Boulder, CO.
National Renewable Energy Laboratory (NREL), 2011. OpenStudio - a cross-platform collection of software tools to
support whole building energy modeling using EnergyPlus and advanced daylight analysis using Radiance,
http://openstudio.nrel.gov/, National Renewable Energy Laboratory, Boulder, CO.
Simergy, 2011. Simergy-Graphical User Interface for EnergyPlus, http://gundog.lbl.gov/projects/gui.
U.S. Department of Energy (DOE), 2009. Buildings Energy Data Book. Prepared for U.S. Department of Energy Office of
Energy Efficiency and Renewable Energy by D&R International, Ltd. Silver Spring, MD.
http://buildingsdatabook.eren.doe.gov/.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
204
ASHRAE Transactions
SA-12-C026
8VLQJ(QHUJ\0RGHOVWR9HULI\8WLOLW\
&RQVXPSWLRQ

-DUHG$+LJJLQV3(&(0 6WHSKHQ')RVWHU(,7 -5\DQ%DLOH\(,7
$6+5$(0HPEHU $6+5$($VVRFLDWH0HPEHU
$%675$&7
:LWKXWLOLW\FRVWVSURMHFWHGWRLQFUHDVHLQWKHQHDUIXWXUHPDQ\GHYHORSHUVDQGSODQQHUVDUHEHJLQQLQJWRYLHZ
HQHUJ\FRQVHUYDWLRQPHDVXUHVPRUHVHULRXVO\(QHUJ\PRGHOHUVDUHJRLQJWREHFRQVXOWHGPRUHIUHTXHQWO\LQRUGHU
WRGHWHUPLQHWKHIHDVLELOLW\RILPSOHPHQWLQJWKHVHPHDVXUHVLQWRIXWXUHSURMHFWV+RZHYHUEHIRUHMXVWLILHGGHFLVLRQV
FDQ EH PDGH UHJDUGLQJ LQFRUSRUDWLQJ HQHUJ\ FRQVHUYDWLRQ PHDVXUHV D GHWHUPLQDWLRQ PXVW EH PDGH RQ KRZ
DFFXUDWHFXUUHQWHQHUJ\PRGHOLQJSUHGLFDWLRQVDUHEDVHGRQDFWXDOPHDVXUHGGDWD7KLVSDSHUZLOODGGUHVVDVWXG\
SHUIRUPHGRQIRXUIDFLOLWLHVFRPSDULQJSURMHFWHGHQHUJ\XVHYHUVXVPHDVXUHGGDWD7KHIDFLOLWLHVHQFRPSDVVDZLGH
UDQJH RI EXLOGLQJ W\SH +9$& V\VWHPV FOLPDWH UHJLRQV DQG RSHUDWLRQ VFKHGXOHV 7KH GDWD SURYLGHG ZLOO DVVLVW
GHYHORSHUVDQGSODQQHUVZLWKXQGHUVWDQGLQJWKHEHQHILWVRIHQHUJ\PRGHOLQJDQGRIIHULGHDVWRKHOSIXUWKHUUHGXFH
GLVFUHSDQFLHVEHWZHHQPRGHOHGDQGPHWHUHGGDWD

,1752'8&7,21

,QWKHSDVWHQHUJ\PRGHOLQJKDVEHHQXVHGSULPDULO\IRUOLIHF\FOHFRVWDQDO\VLVWRGHWHUPLQHWKHSD\EDFNRI
YDULRXV EXLOGLQJ FRPSRQHQWV VXFK DV PHFKDQLFDO V\VWHPV LQFUHDVHG HQYHORSH LQVXODWLRQ DQG KLJK SHUIRUPDQFH
JOD]LQJ+RZHYHUZLWKWKHH[SORVLYHLQWHUHVWLQWKHJUHHQEXLOGLQJLQGXVWU\WKHGHPDQGIRUHQHUJ\PRGHOLQJKDV
LQFUHDVHG GUDPDWLFDOO\ :LWK SURJUDPV VXFK DV WKH /HDGHUVKLS LQ (QHUJ\ DQG (QYLURQPHQWDO 'HVLJQ /(('
HQHUJ\PRGHOLQJKDVEHFRPHDUHJXODUO\XVHGVHUYLFHIRUPDQ\HQJLQHHULQJILUPV7KHIHGHUDOJRYHUQPHQWUHTXLUHV
DQ HQHUJ\ PRGHOLQJ DQDO\VLV IRU DOO QHZ FRQVWUXFWLRQ DQG PDMRU UHQRYDWLRQV EDVHG RQ WKH UHTXLUHPHQWV VHW LQ WKH
(QHUJ\3ROLF\$FWRI6HYHUDOVWDWHJRYHUQPHQWVKDYHDOVREHJXQWRUHTXLUHHQHUJ\PRGHOLQJDQDO\VLVIRUVWDWH
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
IXQGHGIDFLOLWLHVDQGXQLYHUVLWLHV)RUWKHVHSURJUDPVDQHQHUJ\PRGHOLVFUHDWHGWRGHWHUPLQHWKHDPRXQWRIHQHUJ\
FRVW VDYLQJV EHWZHHQ D SUHGHWHUPLQHG EDVHOLQH EXLOGLQJ DQG WKH SURSRVHG EXLOGLQJ DOWHUQDWLYHV 7KH EDVHOLQH
EXLOGLQJ DOWHUQDWLYH LV FUHDWHG EDVHG RQ $SSHQGL[ * RI D SDUWLFXODU YHUVLRQ RI $6+5$( 6WDQGDUG DQG WKH
SURSRVHGEXLOGLQJDOWHUQDWLYHLVEDVHGRQWKHDFWXDOGHVLJQRIWKHQHZEXLOGLQJ
+RZHYHURQFHWKHHQHUJ\PRGHODQDO\VLVKDVEHHQSHUIRUPHGDQGWKHUHVXOWVKDYHEHHQDFFHSWHGVHYHUDORI
WKHVH IDFLOLWLHV GR QRW KDYH D PHDVXUHPHQW DQG YHULILFDWLRQ SURJUDP LQ SODFH WR GHWHUPLQH LI WKH EXLOGLQJ LV
SHUIRUPLQJ DFFRUGLQJ WR WKH HQHUJ\ PRGHO UHVXOWV 7KH EXLOGLQJ RZQHUV WKDW KDYH WDNHQ WKH WLPH WR PHDVXUH WKH
DFWXDO EXLOGLQJ SHUIRUPDQFH DJDLQVW WKH SUHGLFWHG EXLOGLQJ SHUIRUPDQFH KDYH IRXQG RXW WKDW WKHLU EXLOGLQJV UDUHO\
PDWFK XS WR WKH SUHGLFWHG SHUIRUPDQFH 6HYHUDO FDVHV KDYH EHHQ IRXQG WR SHUIRUP ZRUVH WKDQ WKH SUHGLFWLRQV
KRZHYHUWKHUHKDYHEHHQLQVWDQFHVZKHQWKHDFWXDOPHDVXUHPHQWVSHUIRUPPXFKEHWWHUWKDQWKHSUHGLFWHGUHVXOWV
:LWKWKHJURZLQJJUHHQEXLOGLQJLQLWLDWLYHVDQGWKHQDWLRQZLGHSXVKWRZDUGQHW]HURHQHUJ\EXLOGLQJVWKHHQHUJ\
PRGHOLQJVHFWRUZLOOVHHDJURZLQJGHPDQGIRUPRUHDFFXUDWHPRGHOLQJUHVXOWV
-DUHG$+LJJLQV3(&(0LVDFRQVXOWLQJHQJLQHHUZLWK3DUNKLOO6PLWK &RRSHU,QF/XEERFN7H[DV6WHSKHQ')RVWHU(,7
LVDPHFKDQLFDOGHVLJQHQJLQHHUZLWK3DUNKLOO6PLWK &RRSHU,QF/XEERFN7H[DV-5\DQ%DLOH\(,7LVDPHFKDQLFDOGHVLJQ
HQJLQHHUZLWK3DUNKLOO6PLWK &RRSHU,QF/XEERFN7H[DV
$6+5$(


$1$/<6,60(7+2'2/2*<

7KLVSDSHUZLOOFRPSDUHHQHUJ\PRGHOLQJUHVXOWVWRDFWXDOPHWHUHGXWLOLW\GDWDIRUIRXUIDFLOLWLHV7KHHQHUJ\
PRGHOV ZHUH VHW XS EDVHG RQ DFWXDO EXLOGLQJ HQYHORSH FRQVWUXFWLRQ FRPSRQHQWV OLJKWLQJ SRZHU GHQVLWLHV
PHFKDQLFDO V\VWHP HIILFLHQFLHV ZDWHU KHDWLQJ HIILFLHQFLHV DQG PLVFHOODQHRXV QDWXUDO JDV DQG HOHFWULFDO ORDGV
2SHUDWLQJDQGXWLOL]DWLRQVFKHGXOHVZHUHEDVHGRQIDFLOLW\RSHUDWLQJFDOHQGDUVDQGVWDQGDUGZRUNKRXUV(DFKLQLWLDO
PRGHOZDVFRQVWUXFWHGDQGWKHUHVXOWVZHUHDQDO\]HGWRGHWHUPLQHHDFKV\VWHPFRPSRQHQW¶VHQHUJ\XVH7KHPRGHO
FRPSRQHQWVRIOLJKWLQJ PLVFHOODQHRXVORDGVDQG+9$&V\VWHPRSHUDWLRQVHWEDFNWHPSHUDWXUHV ZHUHWKHQWXQHG
EDVHGRQWKHDFWXDOXWLOLW\GDWDFRPSDULVRQVWRDFFXUDWHO\PDWFKWKHDFWXDOXWLOLW\FRQVXPSWLRQ7KHSHUFHQWDJHRI
WRWDOHQHUJ\XVHRIHDFKYDULDEOHZDVFRPSDUHGWRLQGXVWU\VWDQGDUGVLQRUGHUWRGHWHUPLQHZKLFKYDULDEOHUHTXLUHG
WXQLQJLQRUGHUWRDFFXUDWHO\PDWFKWKHXWLOLW\FRQVXPSWLRQ
(DFK RI WKH IDFLOLWLHV XVHG IRU DQDO\VLV ZHUH HLWKHU GHVLJQHG DV QHZ FRQVWUXFWLRQ RU PDMRU UHQRYDWLRQ RU
UHFHLYHG IDFLOLW\ DVVHVVPHQW DQG PDVWHU SODQQLQJ E\ 3DUNKLOO 6PLWK &RRSHU ,QF RI/XEERFN 7H[DV 7ZR RI
WKHVHIDFLOLWLHVDUH/(('FHUWLILHGEXLOGLQJVDQGDOOIDFLOLWLHVKDYHJRQHWKURXJKXWLOLW\EHQFKPDUNLQJ$PL[WXUH
RI HOHPHQWDU\ DQG PLGGOH VFKRRO IDFLOLWLHV DQG DQ RIILFH EXLOGLQJ ZHUH XVHG IRU WKH DQDO\VLV 7KH IRXU VHOHFWHG
IDFLOLWLHVDUHVXPPDUL]HGLQ7DEOHEHORZDGGLWLRQDOGHWDLOHGGHVFULSWLRQVZLOOEHSURYLGHGLQWKHGLVFXVVLRQDQG
DQDO\VLVVHFWLRQRIWKHUHSRUW

7DEOH&DVH6WXG\'HVFULSWLRQV
&DVH
%XLOGLQJ'HVFULSWLRQ 6L]HIW 6L]HP /RFDWLRQ
1R
&DVH (OHPHQWDU\6FKRROVWWK*UDGH %RUJHU7H[DV
&DVH (OHPHQWDU\6FKRRO3UH.WK*UDGH (O3DVR7H[DV
&DVH 0LGGOH6FKRRO*\PQDVLXP 5DGFOLII.HQWXFN\
&DVH (QJLQHHULQJ$UFKLWHFWXUH2IILFH /XEERFN7H[DV

0RGHOV ZHUH FRQVWUXFWHG EDVHG RQ FRQVWUXFWLRQ GRFXPHQWV FRQVXOWLQJ ZLWK WKH GHVLJQ WHDPV DQG PHHWLQJ
ZLWKIDFLOLW\PDQDJHUVIRUHDFKIDFLOLW\)XOO\HDUZHDWKHUILOHVFRQWDLQLQJKRXUVRIZHDWKHUGDWDZHUHXVHGIRU
HDFK ORFDWLRQ WKH VXPPHU GHVLJQ FRROLQJ GU\ EXOE DQG ZHW EXOE WHPSHUDWXUHV ZHUH XVHG DV ZHOO DV WKH
ZLQWHUKHDWLQJGHVLJQGU\EXOEWHPSHUDWXUH:HDWKHUILOHW\SHVXVHGIRUWKHDQDO\VLVZHUH70<70<RU70<
VHOHFWLRQRIWKHZHDWKHUILOHVZHUHEDVHGRQWKHDFWXDOFOLPDWHORFDWLRQZLWKWKHH[FHSWLRQRI%RUJHU7H[DVZKLFK
XVHG $PDULOOR 7H[DV DV WKH UHIHUHQFHG ZHDWKHU ZKLFK LV DSSUR[LPDWHO\ PLOHV NLORPHWHUV VRXWKZHVW RI
%RUJHU$VXPPDU\RIWKHIDFLOLW\EXLOGLQJHQYHORSHYDOXHVDQGPHFKDQLFDOV\VWHPVDUHVKRZQLQ7DEOHEHORZ

7DEOH$&DVH6WXG\%XLOGLQJ(QYHORSHDQG0HFKDQLFDO6\VWHP'HVFULSWLRQV,38QLWV
:DOO 5RRI *OD]LQJ
&DVH1R 6+*& 0HFKDQLFDO6\VWHP
89DOXH 89DOXH 89DOXH
&DVH 3DFNDJHG5RRIWRS$LU&RQGLWLRQLQJ8QLWVZLWK*DV+HDW
&DVH 3LSH&HQWUDO6\VWHPZLWK9DULDEOH$LU9ROXPH8QLWV
&DVH 6SOLW6\VWHP&RROLQJZLWK*DV)LUHG%RLOHU+HDWLQJ
&DVH 9DULDEOH9ROXPH7HUPLQDO6\VWHPZLWK3DFNDJHG5RRIWRS8QLWV

8YDOXHXQLWVDUH%78K)IW
6RODUKHDWJDLQFRHIILFLHQW6+*&UHIHUVWRWKHSHUFHQWDJHRIOLJKWWUDQVPLWWHGWKURXJKDQRSHQLQJ
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
206
ASHRAE Transactions
7DEOH%&DVH6WXG\%XLOGLQJ(QYHORSHDQG0HFKDQLFDO6\VWHP'HVFULSWLRQV6,8QLWV
:DOO 5RRI *OD]LQJ
&DVH1R 6+*& 0HFKDQLFDO6\VWHP
89DOXH 89DOXH 89DOXH
&DVH 3DFNDJHG5RRIWRS$LU&RQGLWLRQLQJ8QLWVZLWK*DV+HDW
&DVH 3LSH&HQWUDO6\VWHPZLWK9DULDEOH$LU9ROXPH8QLWV
&DVH 6SOLW6\VWHP&RROLQJZLWK*DV)LUHG%RLOHU+HDWLQJ
&DVH 9DULDEOH9ROXPH7HUPLQDO6\VWHPZLWK3DFNDJHG5RRIWRS8QLWV

8YDOXHXQLWVDUH:&P
6RODUKHDWJDLQFRHIILFLHQW6+*&UHIHUVWRWKHSHUFHQWDJHRIOLJKWWUDQVPLWWHGWKURXJKDQRSHQLQJ

7KHVH PRGHOV ZHUHQRWFRQVWUXFWHGEDVHGRQ $SSHQGL[ * RI $6+5$(6WDQGDUG7KHREMHFWLYHRI WKLV
SDSHULVWRGHWHUPLQHLIDQHQHUJ\PRGHOFDQEHXVHGWRYHULI\DFWXDOEXLOGLQJXWLOLW\SHUIRUPDQFHGDWDDFFXUDWHO\
+RZHYHUPDQ\RIWKHUHFRPPHQGDWLRQVRI$SSHQGL[*ZHUHXVHGWRGHYHORSWKHLQLWLDOPRGHOVIRUWKHVWXG\

',6&866,21$1'$1$/<6,6

&DVH6WXG\(OHPHQWDU\6FKRROLQ%RUJHU7H[DV

$QHZHOHPHQWDU\VFKRROZDVFRQVWUXFWHGLQ%RUJHU7;LQ7KHIWPIDFLOLW\FRQVLVWV
RI FODVVURRPV DGPLQLVWUDWLRQ RIILFHV DQ LQIRUPDWLRQ FHQWHU J\PQDVLXP FDIHWHULD DQG FRRNLQJ IDFLOLWLHV 7KH
EXLOGLQJ HQYHORSH FRQVLVWV RI VWHHO IUDPHG ZDOOV ZLWK D EULFN H[WHULRU D ORZ SLWFKHG PRGLILHG ELWXPHQ URRI DQG
GRXEOHSDQHGORZHFRDWHGJOD]LQJ7KHPHFKDQLFDOV\VWHPXWLOL]HGLV6((5SDFNDJHGURRIWRSDLUFRQGLWLRQLQJ
XQLWVZLWKQDWXUDOJDVIXUQDFHV7KHPHFKDQLFDOXQLWVDOVRFRQWDLQFDUERQGLR[LGHHFRQRPL]HUV7KHPDMRULW\RIWKH
IDFLOLW\ XVHV WKUHH ODPS UHFHVVHG OLJKW IL[WXUHV WKDW FRQWDLQ 7 IOXRUHVFHQW ODPSV ZLWK HDFK URRP FRQWDLQLQJ GXDO
VZLWFKLQJ2YHUDOOOLJKWLQJSRZHUGHQVLW\IRUWKHEXLOGLQJLV:IW:P7KHPLVFHOODQHRXVHOHFWULFDO
ORDGLVDVVXPHGDW:IW :P7ZREWXKN:GRPHVWLF KRW ZDWHUERLOHUVSURYLGHKRW
ZDWHUIRUWKHNLWFKHQDQGVHYHUDOVPDOOFDSDFLW\HOHFWULFZDWHUKHDWHUVSURYLGHGRPHVWLFKRWZDWHUIRUWKHUHPDLQGHU
RIWKHVFKRRO$QHQHUJ\PDQDJHPHQWFRQWUROV\VWHPZDVLQVWDOOHGWKDWSURYLGHVVXSHUYLVRU\FRQWURORIWKH+9$&
V\VWHPDQGOLJKWLQJV\VWHPZLWKRFFXSDQF\RYHUULGHDOORZHG7KHHOHFWULFLW\DQGQDWXUDOJDVXWLOLW\FRPSDULVRQV
DUH VKRZQ LQ WKH ILJXUH EHORZ 7KH HOHFWULFDO FRQVXPSWLRQ WRWDOV IRU WKH PHWHUHG DQG PRGHOHG VFHQDULRV ZHUH
N:K DQG N:K UHVSHFWLYHO\ 7KH QDWXUDO JDV FRQVXPSWLRQ WRWDOV IRU PHWHUHG DQG PRGHOHG
VFHQDULRVZHUH&&)PDQG&&)PUHVSHFWLYHO\7KHHQHUJ\XVHLQGH[(8,IRU
WKLVIDFLOLW\IRUWKHPHWHUHGDQGPRGHOHGVFHQDULRVZDVN%78IW\UN-P\UDQGN%78IW\U
N-P\UUHVSHFWLYHO\

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

(OHFWULFDO&RQVXPSWLRQ&RPSDULVRQ

(OHFWULF&RQVXPSWLRQ

N:K

0HWHUHG

0RGHOHG

-DQXDU\
)HEUXDU\
0D\
-XO\
-XQH
2FWREHU
0DUFK
$SULO
6HSWHPEHU
'HFHPEHU
1RYHPEHU
$XJXVW

1DWXUDO*DV&RQVXPSWLRQ&RPSDULVRQ

1DWXUDO*DV&RQVXPSWLRQ

&&)
0HWHUHG

0RGHOHG

-DQXDU\
)HEUXDU\
0DUFK
0D\
-XO\
2FWREHU
$SULO
-XQH
6HSWHPEHU
'HFHPEHU
$XJXVW
1RYHPEHU
)LJXUH&DVH6WXG\8WLOLW\&RPSDULVRQV
7KHFOLPDWHUHJLRQIRUWKLVIDFLOLW\H[SHULHQFHGDPXFKFROGHUZLQWHUWKDQDYHUDJHGXULQJWKHXWLOLW\GDWDUDQJH
SURYLGHG$GGLWLRQDOO\WKHIDFLOLW\LVORFDWHGIXUWKHUQRUWKWKDQWKHZHDWKHUGDWDORFDWLRQXVHGDQGGRHVH[SHULHQFH
VOLJKWO\FROGHUWHPSHUDWXUHVWKDQWKHZHDWKHUGDWDORFDWLRQ7KH$XJXVWDQG6HSWHPEHUPRQWKVIRUWKHXWLOLW\GDWD
UDQJH ZHUH DOVR XQVHDVRQDEO\ FRROHU ZKLFK FRQWULEXWHV WR WKH PRGHOHG GDWD VKRZLQJ D KLJKHU HOHFWULFDO
FRQVXPSWLRQGXULQJWKRVH PRQWKV6XPPHUVFKRRO ZDVDVVXPHGWREHLQVHVVLRQ ZLWK EUHDNIDVWDQGOXQFK PHDOV
EHLQJ VHUYHG KRZHYHU EDVHG RQ WKH DFWXDO PHWHUHG JDV XVDJH LW DSSHDUV WKDW FRRNLQJ DQG NLWFKHQ GRPHVWLF KRW
ZDWHUSURGXFWLRQZHUHQRWXWLOL]HG

&DVH6WXG\(OHPHQWDU\6FKRROLQ(O3DVR7H[DV

$QHZHOHPHQWDU\VFKRROZDVFRQVWUXFWHGLQ(O3DVR7;LQ7KHIWPIDFLOLW\FRQVLVWV
RI FODVVURRPV DGPLQLVWUDWLRQ RIILFHV DQ LQIRUPDWLRQ FHQWHU J\PQDVLXP FDIHWHULD DQG FRRNLQJ IDFLOLWLHV 7KH
EXLOGLQJHQYHORSHFRQVLVWVRIPDVRQU\ZDOOVZLWKDVWRQHH[WHULRUDORZSLWFKHGPRGLILHGELWXPHQURRIDQGGRXEOH
SDQHG ORZH FRDWHG JOD]LQJ ZLWK ILQQHG H[WHULRU RYHUKDQJV LQVWDOOHG DERYH WKH JOD]LQJ 7KH PHFKDQLFDO V\VWHP
XWLOL]HG IRU WKH FODVVURRP DQG DGPLQLVWUDWLRQ DUHDV ZDV D IRXUSLSH FHQWUDO V\VWHP ZLWK WZR KLJK HIILFLHQF\ DLU
FRROHGFKLOOHUVDQGQDWXUDOJDVILUHGERLOHUV7KHDLUVLGHV\VWHPXVHGYDULDEOHYROXPHDLUKDQGOLQJXQLWVGHOLYHULQJ
DLU WR YDULDEOH DLU YROXPH ER[HV ZLWK KRW ZDWHU UHKHDW FRLOV 7KH NLWFKHQ FDIHWHULD DQG J\PQDVLXP DUHDV XVHG
HYDSRUDWLYH FRROLQJ DQG D KRW ZDWHU KHDWLQJ V\VWHP 7KH PDMRULW\ RI WKH IDFLOLW\ XVHV IRXU ODPS UHFHVVHG OLJKW
IL[WXUHVWKDWFRQWDLQ7IOXRUHVFHQWODPSVZLWKHDFKURRPFRQWDLQLQJGXDOVZLWFKLQJ2YHUDOOOLJKWLQJSRZHUGHQVLW\
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
208
ASHRAE Transactions
IRU WKH EXLOGLQJ LV :IW :P 7KH PLVFHOODQHRXV HOHFWULFDO ORDG LV DVVXPHG DW :IW :P
7ZREWXKN:GRPHVWLFKRWZDWHUERLOHUVSURYLGHKRWZDWHUIRUWKHVFKRRODQGNLWFKHQ$QHQHUJ\
PDQDJHPHQWFRQWUROV\VWHP ZDVLQVWDOOHGWKDWSURYLGHV VXSHUYLVRU\FRQWURORI WKH+9$&V\VWHP7KHHOHFWULFLW\
DQGQDWXUDOJDVXWLOLW\FRPSDULVRQVDUHVKRZQLQWKHILJXUHEHORZ7KHHOHFWULFDOFRQVXPSWLRQWRWDOVIRUWKHPHWHUHG
DQGPRGHOHGVFHQDULRVZHUHN:KDQGN:KUHVSHFWLYHO\7KHQDWXUDOJDVFRQVXPSWLRQWRWDOVIRU
PHWHUHGDQGPRGHOHGVFHQDULRVZHUH&&)PDQG&&)PUHVSHFWLYHO\7KH(8,
IRU WKLV IDFLOLW\ IRU WKH PHWHUHG DQG PRGHOHG VFHQDULRV ZHUH N%78IW\U N-P\U DQG
N%78IW\UN-P\UUHVSHFWLYHO\

ϭϬϬϬϬϬ
(OHFWULF&RQVXPSWLRQ
ϴϬϬϬϬ
ϲϬϬϬϬ
N:K
ϰϬϬϬϬ 0HWHUHG
ϮϬϬϬϬ 0RGHOHG
Ϭ
0D\
2FWREHU
-DQXDU\
0DUFK
)HEUXDU\
-XO\
$SULO
-XQH
6HSWHPEHU
'HFHPEHU
$XJXVW
1RYHPEHU

ϲϬϬϬ
ϱϬϬϬ
ϰϬϬϬ
ϯϬϬϬ
ϮϬϬϬ 0HWHUHG
&&)
ϭϬϬϬ
0RGHOHG
Ϭ
-DQXDU\
)HEUXDU\
0DUFK
0D\
-XO\
-XQH
2FWREHU
$SULO
6HSWHPEHU
'HFHPEHU
1RYHPEHU
$XJXVW

7KHFOLPDWHUHJLRQIRUWKLVIDFLOLW\W\SLFDOO\KDVDYHU\PLOGZLQWHU+RZHYHUGXULQJ)HEUXDU\RIWKHPHWHUHG
XWLOLW\ GDWD SURYLGHG WKH WHPSHUDWXUHV ZHUH WZHQW\ILYH GHJUHHV )DKUHQKHLW EHORZ QRUPDO IRU D WZR ZHHN SHULRG
7KLVH[SODLQVWKHVKDUSVSLNHLQQDWXUDOJDVFRQVXPSWLRQGXULQJWKH)HEUXDU\PRQWK$GGLWLRQDOO\GXULQJ0DUFK
WKH PRQWKO\ WHPSHUDWXUHV ZHUH VOLJKWO\ ORZHU WKDQ DYHUDJH ZKLFK H[SODLQV WKH ORZHU HOHFWULFDO FRQVXPSWLRQ DQG
KLJKHUQDWXUDOJDVFRQVXPSWLRQ,WZDVDVVXPHGWKDWVXPPHUVFKRROGLGQRWWDNHSODFHDQGWKDWWKHVXPPHUVHWEDFN
WHPSHUDWXUH ZDVDW)&KRZHYHUEDVHGRQWKH HOHFWULFDOXWLOLW\FRQVXPSWLRQLWDSSHDUVWKDWDVXPPHU
VFKRROVHVVLRQPD\KDYHWDNHQSODFH

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

&DVH6WXG\0LGGOH6FKRRO*\PQDVLXPLQ5DGFOLII.HQWXFN\

$QHZ PLGGOHVFKRROJ\PQDVLXP ZDVFRQVWUXFWHGLQ 5DGFOLII.<LQ7KHIW PIDFLOLW\
FRQVLVWVRIDJ\PQDVLXPFRXUWDUHDFRDFK¶VRIILFHVDQGORFNHUURRPV7KHEXLOGLQJHQYHORSHFRQVLVWVRIPDVRQU\
ZDOOVZLWKDEULFNH[WHULRUSLWFKHGVWDQGLQJVHDPPHWDOURRIDQGKLJKHIILFLHQF\JOD]LQJ7KHPHFKDQLFDOV\VWHP
XWLOL]HGIRUWKHIDFLOLW\LVDWRQN:VSOLWV\VWHPDLUKDQGOLQJXQLWFRXSOHGZLWKDEWXKN:
QDWXUDO JDV ILUHG ERLOHU /LJKWLQJ LQ WKH J\PQDVLXP LV KLJK ED\ IL[WXUHV ZLWK IRXU 7 +2 IOXRUHVFHQW ODPSV SHU
IL[WXUH 7KH DGMDFHQW VSDFHV XVH UHFHVVHG OLJKW IL[WXUHV ZLWK IRXU 7 IOXRUHVFHQW ODPSV SHU IL[WXUH 7KH RYHUDOO
OLJKWLQJ SRZHU GHQVLW\ DVVXPSWLRQ IRU WKH EXLOGLQJ LV :IW :P 7KH PLVFHOODQHRXV HOHFWULFDO ORDG LV
DVVXPHGDW:IW:PLQWKHRIILFHDUHDRQO\7ZREWXKN:GRPHVWLFKRWZDWHUKHDWHUV
SURYLGH KRW ZDWHU IRU WKH ORFNHU URRP UHVWURRPV DQG VKRZHUV &RQWURO IRU WKH +9$& V\VWHP LV SURYLGHG E\ D
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
VLQJOH SURJUDPPDEOH WKHUPRVWDW DQG OLJKWLQJ LV HTXLSSHG ZLWK RFFXSDQF\ VHQVRUV 7KH HOHFWULFLW\ DQG QDWXUDO JDV
XWLOLW\FRPSDULVRQVDUHVKRZQLQWKHILJXUHEHORZ7KHHOHFWULFDOFRQVXPSWLRQWRWDOVIRUWKHPHWHUHGDQGPRGHOHG
VFHQDULRVZHUHN:KDQGN:KUHVSHFWLYHO\7KHQDWXUDOJDVFRQVXPSWLRQWRWDOVIRUPHWHUHGDQG
PRGHOHGVFHQDULRVZHUH&&)PDQG&&)PUHVSHFWLYHO\7KH(8,IRUWKLVIDFLOLW\
IRUWKHPHWHUHGDQGPRGHOHGVFHQDULRVZHUHN%78IW\UN-P\UDQGN%78IW\U
N-P\UUHVSHFWLYHO\
ϯϬϬϬϬ
(OHFWULF&RQVXPSWLRQ
ϮϱϬϬϬ
ϮϬϬϬϬ
N:K
ϭϱϬϬϬ
0HWHUHG
ϭϬϬϬϬ
0RGHOHG
ϱϬϬϬ
Ϭ
-DQXDU\
)HEUXDU\
0D\
-XO\
2FWREHU
0DUFK
$SULO
-XQH
6HSWHPEHU
'HFHPEHU
$XJXVW
1RYHPEHU
ϭϮϬϬ
ϭϬϬϬ
ϴϬϬ
ϲϬϬ
ϰϬϬ 0HWHUHG
&&)
ϮϬϬ
0RGHOHG
Ϭ
)HEUXDU\
0D\
-XO\
2FWREHU
-DQXDU\
0DUFK
$SULO
-XQH
6HSWHPEHU
'HFHPEHU
$XJXVW
1RYHPEHU
210
ASHRAE Transactions
'XULQJ 1RYHPEHU RI WKH XWLOLW\ GDWD SURYLGHG WKLV FOLPDWH UHJLRQ H[SHULHQFHG DQ XQXVXDOO\ ZDUP PRQWK
ZKLFKFDQEHVHHQLQWKHDEQRUPDOHOHFWULFDOFRQVXPSWLRQVSLNHDQGWKHODFNRIQDWXUDOJDVXVDJH+RZHYHUWKLVKDV
SURPSWHGWKHRZQHURIWKHIDFLOLW\WRGRVRPHDGGLWLRQDOLQYHVWLJDWLRQLQWRWKHHOHFWULFDOPHWHUUHDGLQJVDVWKHXVDJH
IRUWKHPRQWKLVDOPRVWGRXEOHWKDWRIWKHXVDJHVHHQLQWKH'HFHPEHUDQG-DQXDU\PRQWKV$GGLWLRQDOO\WKHVSDFH
KHDWLQJHTXLSPHQWLVDJHGSDVWLWVXVHIXOOLIHDQGKDVEHHQKDYLQJVHYHUDORSHUDWLQJLVVXHVZKLFKFRXOGFRQWULEXWHWR
WKHODUJHPHWHUHGQDWXUDOJDVFRQVXPSWLRQVKRZQLQWKHILJXUHDERYH

&DVH6WXG\(QJLQHHULQJ$UFKLWHFWXUH2IILFHLQ/XEERFN7H[DV

$QHZRIILFHIRUDQHQJLQHHULQJDUFKLWHFWXUHILUPZDVFRQVWUXFWHGLQ/XEERFN7;LQ$QDGGLWLRQWRWKH
RIILFH ZDV FRQVWUXFWHG LQ 7KH FRPELQHG IW P IDFLOLW\ FRQVLVWV RI D VWDII RIILFHV FRPPRQ
ZRUNDUHDVFRQIHUHQFHURRPVDQGDEUHDNURRP7KHEXLOGLQJHQYHORSHFRQVLVWVRIVWHHOIUDPHGZDOOVZLWKDEULFN
H[WHULRUDORZSLWFKHG732PHPEUDQHURRIRQWKHILUVWSRUWLRQRIWKHEXLOGLQJDQGDYHJHWDWLYHURRIRQWKHDGGLWLRQ
DQG GRXEOH SDQHG ORZH FRDWHG JOD]LQJ 7KH PHFKDQLFDO V\VWHP XWLOL]HG LV KLJK HIILFLHQF\ SDFNDJHG URRIWRS DLU
FRQGLWLRQLQJ XQLWV ZLWK QDWXUDO JDV IXUQDFHV 7KH PHFKDQLFDO XQLWV DOVR FRQWDLQ FDUERQ GLR[LGH HFRQRPL]HUV DQG
GHOLYHU DLU WR YDULDEOH YROXPH WHUPLQDOV 7KH PDMRULW\ RI WKH RIILFHV XVH WZR ODPS UHFHVVHG OLJKW IL[WXUHV WKDW
FRQWDLQ 7 IOXRUHVFHQW ODPSV 7KH FRPPRQ ZRUN DUHDV FRQWDLQ SHQGDQW PRXQWHG OLJKWLQJ ZLWK D GD\ OLJKWLQJ
FRQWUROV\VWHPWRUHGXFHWKHDPRXQWRIOLJKWLQJSRZHU7KHRYHUDOOOLJKWLQJSRZHUGHQVLW\IRUWKHEXLOGLQJLV
:IW :P7KHPLVFHOODQHRXVHOHFWULFDOORDGLVDVVXPHGDW:IW :P$QHQHUJ\PDQDJHPHQW
FRQWUROV\VWHPZDVLQVWDOOHGWKDWSURYLGHVVXSHUYLVRU\FRQWURORIWKH+9$&V\VWHP7KHHOHFWULFLW\DQGQDWXUDOJDV
XWLOLW\FRPSDULVRQVDUHVKRZQLQWKHILJXUHEHORZ7KHHOHFWULFDOFRQVXPSWLRQWRWDOVIRUWKHPHWHUHGDQGPRGHOHG
VFHQDULRVZHUHN:KDQGN:KUHVSHFWLYHO\7KHQDWXUDOJDVFRQVXPSWLRQWRWDOVIRUPHWHUHGDQG
PRGHOHGVFHQDULRVZHUH&&)PDQG&&)PUHVSHFWLYHO\7KH(8,IRUWKLVIDFLOLW\
IRUWKHPHWHUHGDQGPRGHOHGVFHQDULRVZHUHN%78IW\UN-P\UDQGN%78IW\U
N-P\UUHVSHFWLYHO\

ϯϱϬϬϬ
(OHFWULF&RQVXPSWLRQ
ϯϬϬϬϬ
ϮϱϬϬϬ
ϮϬϬϬϬ
N:K
ϭϱϬϬϬ
0HWHUHG
ϭϬϬϬϬ
ϱϬϬϬ 0RGHOHG
Ϭ
-DQXDU\
)HEUXDU\
0D\
-XO\
-XQH
2FWREHU
0DUFK
$SULO
6HSWHPEHU
'HFHPEHU
1RYHPEHU
$XJXVW

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

ϭϬϬϬ
ϴϬϬ
ϲϬϬ
&&)
ϰϬϬ 0HWHUHG
ϮϬϬ 0RGHOHG
Ϭ
-DQXDU\
)HEUXDU\
0DUFK
0D\
-XO\
$SULO
2FWREHU
-XQH
6HSWHPEHU
'HFHPEHU
1RYHPEHU
$XJXVW
7KHUHLVDQDQRPDO\ VKRZQLQWKHVXPPHU PRQWKV IRUWKHQDWXUDO JDV XVDJH $OOGRPHVWLF ZDWHU KHDWLQJLV

SURYLGHGE\HOHFWULFZDWHUKHDWHUVDQGQRDGGLWLRQDOJDVILUHGDSSOLDQFHVH[LVWRWKHUWKDQWKHVSDFHKHDWLQJV\VWHP
:KLOH FRPSDULQJ WKH GDWD DQ LQYHVWLJDWLRQ ZDV SXW LQ SODFH WR GHWHUPLQH ZK\ QDWXUDO JDV FRQVXPSWLRQ KDV EHHQ
VKRZLQJ XS GXULQJ WKH VXPPHU PRQWKV ,W ZDV GLVFRYHUHG WKDW WZR RI WKH QDWXUDO JDV SUHVVXUH UHJXODWRUV IRU WKH
SDFNDJHGURRIWRSXQLWVKDGEHHQOHDNLQJWKURXJKRXWWKH\HDUFDXVLQJWKHDQRPDO\VKRZQLQWKHILJXUHDERYH

&21&/86,216

$V VHHQ LQ WKH ILJXUHV DERYH WKH HOHFWULFLW\ XVDJH LV YHU\ FRPSDUDEOH EHWZHHQ WKH PRGHOHG DQG PHWHUHG
VFHQDULRV 7KH QDWXUDO JDV FRQVXPSWLRQ IRU WKH PRGHOHG VFHQDULR IROORZV D PRUH XQLIRUP IORZ WKDQ WKH DFWXDO
PHWHUHGGDWD7KLVLVEHOLHYHGWREHGXHWRWKHPRGHOLQJSURJUDPUHTXLULQJOHVVKHDWLQJIXHOWKDQLVDFWXDOO\XVHGLQ
WKHHTXLSPHQW7KHSURJUDPDOVRPD\GHWHUPLQHWKDWDORZHUOHDYLQJDLUWHPSHUDWXUHLVUHTXLUHGIRUKHDWLQJDVSDFH
DQGDGMXVWWKHIXHOLQSXWDFFRUGLQJO\
2QHRIWKHPDMRUDGYDQWDJHVWRHQHUJ\PRGHOVLVWKDWPRGHOVSURYLGHDEUHDNGRZQRIHQHUJ\XVDJHLQYDULRXV
DUHDV RI WKH EXLOGLQJ VXFK DV OLJKWLQJ PLVFHOODQHRXV SRZHU +9$& V\VWHP FRPSRQHQWV LH VXSSO\ IDQV KHDW
UHMHFWLRQHTXLSPHQWHWFDQGVHYHUDORWKHUDVSHFWV,IDIDFLOLW\KDVLQVWDOOHGDVXEPHWHULQJV\VWHPWKHEXLOGLQJ
RZQHUFRXOGWUDFNWKHDFWXDOXVDJHFRPSDUHGWRWKHSUHGLFWHGXVDJHDQGPDNHDGMXVWPHQWVLIUHTXLUHGWRLQGLYLGXDO
EXLOGLQJ FRPSRQHQWV LI WKH XVDJH LV KLJKHU WKDQ H[SHFWHG 8QIRUWXQDWHO\ WKH IDFLOLWLHV LQ WKLV VWXG\ ZHUH QRW
HTXLSSHGZLWKDVXEPHWHULQJV\VWHP
(QHUJ\PRGHOVPD\DOVRKHOSWURXEOHVKRRWRSHUDWLQJSUREOHPV ZLWKHTXLSPHQW$VVHHQLQ&DVH6WXG\LW
ZDVGLVFRYHUHGWKDWWKHUHZHUHVRPHQDWXUDOJDVOHDNVLQWKHV\VWHP%\NQRZLQJWKDWWKHUHVKRXOGQRWKDYHEHHQ
DQ\ QDWXUDOJDV XVDJHWKHEXLOGLQJRZQHU ZDVDEOHWRGHWHUPLQHWKDWWKHUH ZDVDSUREOHP$GGLWLRQDOO\LQ&DVH
6WXG\WKHEXLOGLQJRZQHUGHFLGHGWRORRNIXUWKHULQWRWKH1RYHPEHUPRQWKHOHFWULFDOXVDJHVLQFHLWZDVXQXVXDOO\
KLJK7KHUHPD\KDYHEHHQDPHWHUUHDGLQJSUREOHPWKDWFDXVHGDPXFKKLJKHUXVDJHRQWKHXWLOLW\ELOOWKDQZDV
VHHQDQ\RWKHUPRQWK
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
212
ASHRAE Transactions
,W DSSHDUV WKDW WKH HQHUJ\ PRGHOV DUH DFFXUDWH IRU GHWHUPLQLQJ WKH HQHUJ\ XVDJH IRU D W\SLFDO \HDU %\
SHUIRUPLQJ WKLV DQDO\VLV EHIRUH FRQVWUXFWLQJ D IDFLOLW\ DQG KDYLQJ WKH DQDO\VLV DYDLODEOH WKH EXLOGLQJ RZQHU RU
GHYHORSHUZLOOEHDEOHWRWUDFNWKHHQHUJ\XVDJHDQGFRPSDUHWKHGLIIHUHQFHVWKURXJKRXWWKHOLIHRIWKHIDFLOLW\7KLV
ZLOODOVRKHOSGHWHUPLQHPDMRUGLVFUHSDQFLHVDVWKH\RFFXUVRWKDWDFWLRQWRUHVROYHWKHLVVXHPD\EHWDNHQTXLFNHU
$GGLWLRQDO VXE PHWHULQJ FDSDELOLWLHV ZLOO IXUWKHU HQKDQFH WKH DELOLW\ WR DGMXVW WKH HQHUJ\ XVDJH IRU SDUWLFXODU
EXLOGLQJ V\VWHPV DQG KHOS QRW RQO\ EXLOGLQJ RZQHUV DQG GHYHORSHUV EXW DOVR HQHUJ\ PRGHOLQJ SURIHVVLRQDOV
XQGHUVWDQGZKHUHIXUWKHUDWWHQWLRQRQEXLOGLQJHQHUJ\XVDJHLVUHTXLUHG

$&.12:/('*(0(176

7KHFDVHVWXGLHVIRUWKLVSDSHUZHUHEDVHGRQEXLOGLQJVRZQHGE\WKH%RUJHU,QGHSHQGHQW6FKRRO'LVWULFW(O
3DVR,QGHSHQGHQW6FKRRO'LVWULFW)RUW.QR[&RPPXQLW\6FKRROVDQG3DUNKLOO6PLWK &RRSHU,QF
120(1&/$785(
+9$& KHDWLQJYHQWLODWLQJDQGDLUFRQGLWLRQLQJ

/((' /HDGHUVKLSLQ(QHUJ\DQG(QYLURQPHQWDO'HVLJQ
6((5 VHDVRQDOHQHUJ\HIILFLHQF\UDWLR
N:K NLORZDWWKRXU
&&) KXQGUHGFXELFIHHW
(8, (QHUJ\8VH,QGH[
%78 %ULWLVK7KHUPDO8QLW
N%78 %ULWLVK7KHUPDO8QLWV
N- .LOR-RXOH
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---

IW VTXDUHIHHW
P
VTXDUHPHWHUV

P FXELFPHWHUV
\U \HDU
+2 KLJKRXWSXW
732 WKHUPRSODVWLFSRO\ROHILQ

5()(5(1&(6

$16,$6+5$($6+5$(6WDQGDUG(QHUJ\6WDQGDUGIRU%XLOGLQJVH[FHSW/RZ5LVH5HVLGHQWLDO
%XLOGLQJV$WODQWD$PHULFDQ6RFLHW\RI+HDWLQJ5HIULJHUDWLRQDQG$LU&RQGLWLRQLQJ(QJLQHHUV,QF
$6+5$($6+5$(+DQGERRN)XQGDPHQWDOV$WODQWD$PHULFDQ6RFLHW\RI+HDWLQJ5HIULJHUDWLRQDQG$LU
&RQGLWLRQLQJ(QJLQHHUV,QF
'2(³'2(EDVHG%XLOGLQJ(QHUJ\8VHDQG&RVW$QDO\VLV6RIWZDUH´ZZZGRHFRP
7KH7UDQH&RPSDQ\7UDFHY
7KH7UDQH&RPSDQ\7UDFH%XLOGLQJ(QHUJ\DQG(FRQRPLF$QDO\VLV9HUVLRQ8VHUV0DQXDO
$6+5$(

SA-12-C027
Investigating the Performance of Two

Types of Solar Domestic Water Heating
(SDWH) Systems with Drain Water Heat
Recovery through Computer Simulation
and Experimental Analysis
Kamyar Tanha Alan S. Fung Wey H. Leong

Student Member ASHRAE ASHRAE Member ASHRAE Member
ABSTRACT
The objective of this paper is to investigate the performance of two different types of Solar Domestic Hot Water (SDHW) systems , one with a gas
boiler as the auxiliary heating source and the other with an electric tank as the auxiliary heating source, and to investigate the contributions of
different components of the systems in lowering the annual energy demand and GHG emissions for producing 225 litres (59.4 gal) per day of
domestic hot water. The systems are for the twin Archetype Houses located in Vaughan, Ontario, Canada. The annual energy consumption of the
gas boiler and electric tank from the experimental analysis were to be 3718.7 kWh (12.69 MBtu) and 2778 kWh (9.48 MBtu), respectively.
The TRNSYS models of the two systems were also created and validated with the experimental results. Using the TRNSYS models for a typical
Canadian household in Toronto, Ontario, the annual energy consumption by the gas boiler and electric backup tank were concluded to be 3636.9
kWh (12.41 MBtu) and 2681.5 kWh (9.15 MBtu), respectively. The annual thermal energy output by the flat plate and evacuated tube solar
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
thermal collectors were found to 2010.8 kWh (6.86 MBtu) and 1390.6 kWh (4.74 MBtu), respectively. The annual heat recovery by the
Drain Water Heat recovery (DWHR) units were concluded to be 792.8 kWh (2.7 MBtu).
INTRODUCTION
Domestic water heating (DWH) is estimated to be the second largest energy end-use for Canadian households,
exceeded only by space heating which accounts for about 18% of each household energy consumption (NRCan, 2009).
Statistics show that although there has been an approximate decrease of 20% in per household annual energy used to heat
water since 1990, the overall energy required for this purpose has increased by about 6% from 243.0 PJ (230.3 GBtu) to
257.9 PJ (244.4E GBtu), which is due to the increase in the number of housings (NRCan, 2009). Decreasing energy
requirements and conserving energy, alongside with exploring renewable sources of energy has become more important as
the concerns for limited energy resources in the near future continues to grow, especially in cold climate countries. This
paper focuses on two hybrid Solar Domestic Hot Water (SDHW) systems with a drain water heat recovery (DWHR)
system in two semi-detached houses located in Vaughan, Ontario, Canada. The first house, designated as House-A, has a
Kamyar Tanha is a MASC candidate, Alan S. Fung is an associate professor, and Wey H. Leong is an associate professor in the
Department of Mechanical and Industrial Engineering, Ryerson University, Toronto, ON, Canada.
214
ASHRAE Transactions
gas boiler as the auxiliary heating source and the second house, House-B, is equipped with an electric backup tank as the
auxiliary heating source for DHW production. The DWHR systems are for recovering part of the energy contained in drain
water coming from different sources. Both main pipe and coil of the DWHR heat exchanger are made from copper to
enhance the heat transfer between the two flows. The solar thermal collector of House-A is a flat plate collector and the
solar thermal collector of House-B is a heat pipe based evacuated tube collector type with 20 evacuated tubes.
The “Sustainable Archetype Houses” project has been implemented by Toronto and Region Conservation Authority
(TRCA) along with the Building Industry and Land Development (BILD) Association with the aim of demonstrating
different sustainable housing technologies in the near and medium term future. Figure 1 shows an overview picture of the
twin archetype houses. House A is the semi-detached house located on the left side of Figure 1.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 1 Overview of TRCA twin Archetype houses
A comprehensive energy monitoring system has been implemented in the houses, which is useful in investigating the
effectiveness and efficiency of the mechanical systems within the houses (Zhang et al., 2011). It should also be added that
the two solar systems are active, indirect systems which are closed from the atmosphere and the circulating fluid for both
solar loops are a mixture of 60% distilled water and 40% propylene glycol.
SIMULATION
Figure 2 displays the schematics of the two SDHW models. Both solar collectors are south facing with an inclined
angle of 25°. The two models presented in Figure 2 are simulated in TRNSYS to investigate the overall performances of the
two systems. Standard components from TRNSYS library are used except for the DWHR unit for which a new model was
created. As seen in Figure 2, both systems have been configured in a way that the DWHRs only preheat the water going to
hot water taps, in other words, the water going to the solar preheat tanks is first preheated by the DWHR units. It should
also be noted that the role of the tempering valves in the two systems is to make sure that the supplied hot water
temperature to the services is kept below 48°C (118°F) to prevent the scalding effects of the hot water with higher
temperatures.
$6+5$(


(a) (b)
Figure 2 Schematic view of the two SDHW systems: a) SDHW system of House-A and b) SDHW system of House-B
TRNSYS Models
The components used within the two models are as followed:
Solar Thermal Collectors: The main component of the flat plate collector is the copper absorber that is coated with
Sol-titanium which ensures a high absorption of solar radiation and low emission of thermal radiation. The evacuated tube
collector consists of 20 evacuated tubes and an absorber with sol-titanium coating is an integral part of the tubes to ensure
high absorption of solar radiation. The gross and aperture areas of the flat plate collector are 2.51 m2 (27 ft2) and 2.33 m2
(25.1 ft2) and the gross and aperture areas of the evacuated tube collector are 2.88 m2 (31 ft2) and 2.11 m2 (22.7 ft2),
respectively. The flow rate at test conditions for both solar loop pumps is 0.02 kg/s-m2 (0.004 lb/s-ft2), same as the actual
flow rate by the two pumps. The operational efficiency equation used for the model is:
ܶ௠ െ ܶ௔ ሺܶ௠ െ ܶ௔ ሻଶ
ߟ௖௢௟ ൌ ߟ଴ െ ݇ଵ ൬ ൰ െ ݇ଶ ቈ ቉ሺͳሻ
‫ܫ‬௧ ‫ܫ‬௧
Where Tm is the collector’s inlet and outlet average temperature, Ta is the ambient temperature and I is the solar
irradiation. ߟ଴ is the optical efficiency and ݇ଵ and ݇ଶ are thermal loss coefficients. From manufacturer’s datasheets, ߟ଴ , ݇ଵ
and ݇ଶ values for the flat plate collector are 81 %, 3.48 W/m2K (0.61 Btu/hr.ft2 °F)and 0.0164 W/m2K2 (0.0016 Btu/hr.ft2
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
° 2
F ), respectively. Similarly, these values are 83.8 %, 1.18 W/m2K (0.21 Btu/hr.ft2 °F) and 0.0066 W/m2K2 (0.0006
Btu/hr.ft2 °F2), for the evacuated tube collector. It should be noted that these values have been determined by the
manufacturer in accordance with the EN 12975 standard. The thermal efficiency of the collector can also be obtained as
shown in Equation (2):
216
ASHRAE Transactions
݉ሶ‫ܥ‬௣ ൫ܶ௢ǡ஼௢௟ െ ܶ௜ǡ௖௢௟ ൯
ߟ௖௢௟ ൌ ሺʹሻ
‫ܣ‬௖௢௟ Ǥ ‫ܫ‬௧
where ݉ሶ is the mass flow rate, ‫ܥ‬௣ is the fluid specific heat, ܶை and ܶ௜ outlet and inlet temperature of fluid through the
collector and Acol is the collector area.
Solar Preheat Tank(s): The tanks are assumed to be stratified tanks with a volume of 300 litres (79.5 gal) and with two
immersed heat exchangers. The lower heat exchangers for both tanks are used for the solar loops. The upper heat
exchanger of the tank in House-A is connected to the gas boiler and the upper heat exchanger in House-B can be
connected to an auxiliary heating source like the desuperheater loop of Ground Source Heat Pump/Co-generation unit.
The stored water in the tanks gets warmed up with heat exchange from antifreeze fluid in the heat exchanger. The model
used for the tanks is Type 60t with two immersed heat exchangers for the one in House-A and one immersed heat
exchanger for the other one, since this paper’s focus is only on the performance of the SDHW system with the electric tank
as the backup. The heat exchanger coils are 0.025 m (1 in) in diameter and the upper heat exchangers areas are 0.9 m2 (9.7
ft2) with 6 litres (1.6 gal) of liquid content and the lower ones have an area of 1.5 m2 (16.1 ft2) with 10 litres (3.8 gal) of
liquid content. Both tanks have no heating elements.
Auxiliary Tank: The tank is assumed to be a stratified tank with a volume of 184 litres (48.6 gal). The tank is modelled
by assuming that it consists of eight fully mixed equal volume segments with an equal height of 0.15 m (0.49 ft). Each
segment has an assigned node. The type used to model the tank is Type 4a with two electric resistance heating elements.
The first element is located in the eighth node with node one being the top most node with a set point temperature of 52°C
(125.6°F) and a dead band of 5°C (41°F) and with a maximum heating rate of 3000 W (10.24 kBtu/hr). The second element
is located in the fourth node with a set point temperature of 60°C (140°F)and a dead band of 5°C (41°F) and with a
maximum heating rate of 3000 W (10.24 kBtu/hr). The tank loss coefficient is set to be 1.7 kJ/hr-m2-K (0.08 Btu/hr.ft2 °R)
from experimental analysis.
Gas Boiler: The model used for this component is auxiliary heater Type 6. The maximum heating rate is set to 84000
kJ/hr (79.6 kBtu/hr). From the collected data, the average efficiency of 55.5% is used for the heater.
Solar Loop Circulation Pump(s): The pumps are used to circulate the water and propylene glycol solution within the
solar loops. The model used for the pumps is Type 110 which is a variable speed pump that is able to maintain any outlet
mass flow rate between zero and a rated value. The rated flow rate for both pumps is 120 kg/hr (264.5 lb/hr) and the rated
power is 50 W (170.6 Btu/hr). A differential controller with hysteresis (Type 2b) is used to activate the pumps. The pumps
start to run when the temperature difference between the solar preheat tanks sensors, located inside the lower heat
exchanger coils and close to the outlet ports of the tanks, and the solar collectors’ outlets reache 6.7°C (44.1°F) and stop
when the temperature difference is less than 4.5°C (40.1°F). There is also a high limit cut-off to stop the pumps when solar
preheat tank outlet temperatures exceed 90°C (194°F).
DWHR Units: The performance and the effectiveness of both units have been obtained through a series of
experiments performed at the site and the average effectiveness for different flow rates has been defined. The main pipe is
7.6 cm (3 in) in diameter and 91 cm (36 in) in length and the wrapping tubes are ½ inch in diameter. A new model has been
created for the two DWHRs. The DWHRs were assumed to be counter flow heat exchangers. The current model is based
on the work done by Zaloum et al. (2007). According to their report, different DWHR units can be characterized by the
Number of heat Transfer Units (NTU) vs. flow rate curve due to the fact that the units flow rates, inlet and outlet
temperatures are the most important variables in determining the performance of the units. The correlation is in the format
shown in Equation (3):
ܷܰܶ ൌ ‫ܣ‬ଵ Ǥ ݉ሶି஺మ ሺ͵ሻ
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

From the actual data collected from the DWHR units at the site, the two coefficients for the NTU vs. flow rate
correlation formula for different flow rates were determined to be:
݉ሶ = Drain water mass flow rate (lit/min)

A1= 3.166
A2= 0.66
Once the curve for the NTU is obtained, the unit’s effectiveness and the heat transfer can be obtained.
Hot Water Draw Profile

No permanent occupant lives in the twin houses due to the nature of the project The daily hot water draw profile
used in this study is the minutely water draw profile based on the Annex 42 schedule by the International Energy Agency
(Knight et al., 2007). The daily water draw is 225 litres (59.4 gal), as specified in several studies on the daily hot water draw
in North America, with the end use temperature of 43°C (109.4°F). The temperature drop from the shower heads/faucets
to the DWHR’s inlet was found to be approximately 4°C (39.2°F). The daily hot water consumption for simultaneous water
draw sources is assumed to be about 160 litres (42.3 gal) (Hendron & Burch, 2007). It is also accepted that about 40% of
the daily hot water usage is from showers (Jordan & Vajen, 2000). The actual daily hot water draw patterns in minutes is
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
shown in Figure 3. Occuring flow rates at different times are all constant.
7 26.50
6 22.71
FlowRate(Lit/min)
5 18.93
FlowRate(gpm)
4 15.14
3 11.36
2 7.57
1 3.79
0 0.00
0 120 240 360 480 600 720 840 960 1080 1200 1320 1440
Time(min)
Figure 3 High Resolution Daily Hot Water Draw Pattern
The Archetype Houses water source are from local wells in the Kortright Centre. The average monthly water mains
temperatures were measured at the site. The solar irradiance on the collectors’ surface was measured by the pyranometer
with the inclination angle of 25°.
RESULTS ANALYSIS
The performance analysis of the two SDHW systems and their components were performed in two testing periods,
one in summer and the other in winter. Once the heat generations / recoveries by the two systems during the two testing
periods were obtained, these values can be extrapolated to a whole year performance of the two systems. The daily thermal
energy outputs of the collectors during a typical year performance can be derived, using the average daily solar radiations,
218
ASHRAE Transactions
obtained from TRNSYS 16. The monthly natural gas usage of gas boiler in House-A and electricity usage in House-B, and
the monthly heat recoveries by the two DWHRs can be achieved using the average monthly water mains temperatures. The
average efficiency of the boiler, with respect to the thermal energy supplied to the DHW tank, was found to be
approximately 56%. The average efficiency of the electric backup tank, obtained from the two testing periods, was found to
be approximately 79%. From the experimental data, the annual heat recovery by the two DWHRs was shown to be 788.5
kWh (2690.5 kBtu). The annual thermal energy output by the flat plate solar thermal collector and energy consumption of
the mini gas boiler in House-A were concluded to be 2038 kWh (6.95 MBtu) and 3718.7 kWh (12.69 MBtu), respectively.
Similarly, the annual thermal energy output by the evacuated tube solar thermal collector and energy consumption of the
electric backup tank in House-B were concluded to be 1383 kWh (4.72 MBtu) and 2778 kWh (9.5 MBtu), respectively.
The performances of the two TRNSYS models must be validated first, and it is only then that they can be used to
predict the energy consumption / producing of the SDHW systems and their components. The models were simulated
using the average measured monthly water mains temperatures and the weather data file of metropolitan Toronto. The
thermal energy outputs from the two solar thermal collectors in the TRNSYS models were compared with the experimental
results, and the models’ thermal energy outputs prediction were concluded to be quite close to the experimental results.
Figure 4 displays the flat plate solar thermal collector’s thermal energy outputs from TRNSYS model and experimental
results.
13
12
11
10
HeatGeneration(kWh)
9
8
7
6
5
4
3
ExperimentalResults
2
TRNSYSModelResults
1
0
0 100 200 300 400 500 600 700
AverageDaily SolarRadiation (W/m 2 )

Figure 4 Comparison of experimental and TRNSYS model’s thermal energy output of the flat plate solar
thermal collector of House-A
The annual heat recovery by the DWHR units from TRNSYS models were calculated to be 824.9 kWh (2.8 MBtu)
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
which showed a 4.6% difference from the experimental approach. The annual energy consumption by the gas boiler and
thermal energy ouput by the flat plate solar thermal collector from House-A SDHW system’s TRNSYS model were found
to be 2012.5 kWh (6.9 MBtu) and 3690.5 kWh (12.6 MBtu), respectively, which showed 1.3% and 0.8% difference from the
experimental data. Similarly, the annual energy consumption by the electric backup tank and thermal energy ouput by the
evacuated tube solar thermal collector for the second TRNSYS model were found to be 1400.8 kWh (4.78 MBtu) and
2713.5 kWh (9.26 MBtu) respectively, which showed 1.3% and 2.3% difference from the experimental data.
The two validated models can now be used to investigate the performance of such systems for typical households in
other locations, with similar daily hot water consumption. For this purpose, related TRNSYS weather data files and the
$6+5$(

average monthly water mains temperatures should be implemented in the models. In this study, the performances of the
two SDHW systems were simulated for a typical household in Toronto, Ontario. For this case, the annual heat recovery by
the DWHR units was obtained to be 792.8 kWh (2.7 MBtu). The annual thermal energy output by the flat plate and
evacuated tube solar thermal collectors were found to 2010.8 kWh (6.9 MBtu) and 1390.6 kWh (4.74 MBtu), respectively.
The annual energy consumption by the gas boiler and electric backup tank were also found to be 3636.9 kWh (12.41 MBtu)
and 2681.5 kWh (9.15 MBtu), respectively. Using the respective average efficiencies of 56% and 79% for the gas boiler and
electric backup tank, the annual operational cost / savings, for providing 225 lit/day of DHW with the end use water
temperature of 43°C from different components of the two SDWH systems were also calculated. Using the 2010 average
natural gas rate of ¢45.75 per m3 (Statistics Canada, 2011) for the residential sector in the province Ontario, the annual
energy cost was calculated to be $161.7. The annual cost savings by the flat plate solar thermal collector and DWHR were
also calculated to be $161 and $63.5, respectively. With the 2009 GHG emission factor per kilogram equivalent CO2 for
natural gas of 1.879 kg (4.14 lb) (Environment Canada, 2011), the annual GHG emissions reductions by the flat plate
collector and DWHR were also concluded to be 661.7 kg (1459 lb) and 260.9 kg (575.2 lb), respectively.
The electricity cost for the SDHW system of House-B was also calculated to be $364.7, based on the 2009 average
electricity rate of ¢13.6 per kWh (London Economics LLD, 2011) for the residential sector in Ontario. The annual cost
savings by the evacuated tube solar thermal collector and DWHR were concluded to be $240 and $136.8, respectively. With
the 2009 GHG emission factor per kilogram equivalent CO2 for electricity of 0.16 kg (0.35 lb) (Environment Canada,
2011), the annual GHG emissions reductions by the evacuated tube collector and DWHR were also shown to be 282.4 kg
(622.6 lb) and 161 kg (355 lb), respectively.
CONCLUSION
This study has investigated the performance of two different SDHW systems of the twin Archetype Houses located at
Kortright Centre in Vaughan, Ontario, Canada, with the daily hot water draw of 225 litres (59.4 gal) at 43°C (109.4°F)
through experimental analysis and TRNSYS simulations. The first system consists of a flat plate solar thermal collector and
a gas boiler with DWHR. The second system consists of an evacuated tube solar thermal collector with an electric tank and
DWHR. From experimental analysis, the annual energy consumption of the gas boiler and electric tank were found to be
3718.7 kWh (12.69 MBtu) and 2778 kWh (9.48 MBtu), respectively. Simulation results also showed that the addition of
DWHR and flat plate collector for House-A and DWHR and evacuated tube collector for House-B would result in 1777
kWh (6.1 MBtu) and 1738 kWh (5.9 MBtu) of energy savings, from the two base case models without these components.
The results from TRNSYS models were validated with the experimental results. The models were then used to simulate the
performances of such SDHW systems in the city of Toronto. The annual energy consumption by the gas boiler and electric
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
backup tank were shown to be 3636.9 kWh (12.4 MBtu) and 2681.5 kWh (9.15 MBtu). The annual heat recovery by the
DWHR units were concluded to be 792.8 kWh (2.7 MBtu). Moreover, the annual thermal energy output by the flat plate
and evacuated tube solar thermal collectors were found to 2010.8 kWh (6.9 MBtu) and 1390.6 kWh (4.74 MBtu), which
showed that the flat plate collector was capable of providing about 31% more heat. The annual cost savings by the flat plate
solar thermal collector and DWHR in House-A were shown to be $161 and $63.5, and the GHG emission reductions for
the two components were also concluded to be 661.7 kg (1459 lb)and 260.9 kg (575.2 lb), respectively. Similarly, the annual
cost savings by the evacuated tube solar thermal collector and DWHR in House-B were shown to be $240 and $136.8, and
the GHG emission reductions for these two were shown to be 282.4 kg (622.6 lb) and 161 kg (355 lb), respectively.
ACKNOWLEDGEMENT
The authors would like to thank Regional Municipality of Peel, Regional Municipality of York, City of Toronto,
Building Industry and Land Development (BILD) Association, Toronto and Region Conservation Authority (TRCA),
MITACS/Accelerate Ontario, Reliance Home Comfort and Union Gas for their financial support of this project. The
220
ASHRAE Transactions
authors also express their gratitude to Mr. Dahai Zhang, Mr. David Nixon, Mr. Warren Yates, Professor Leo Salemi, and
Mr. Derrik Ashby for their tremendous help in implementing the DAQ system at the site.
NOMENCLATURE
‫ܣ‬௖௢௟ - Collector area (m2)

‫ܥ‬௣ - Specific heat (kJ/kg.K)
ߝ - Heat exchanger effectiveness (%)
‫ܫ‬௧ - Solar incident radiation on tilted surface (W/m2)
݉ሶ – Solar loop flow rate (kg/hr.m2)
NTU – Number of Thermal Units
Q – Heat transfer rate (kJ/hr)
ܶ௠ - Collector mean inlet and outlet temperature (°C)
ܶ௔ - Ambient air temperature (°C)
ܶ௜ǡ௖௢௟ - Collector inlet temperature (°C)
ܶ௢ǡ௖௢௟ - Collector outlet temperature (°C)
ߟ଴ - Collector optical efficiency (%)
ߟ௖௢௟ - Collector efficiency (%)
REFERENCES
Environment Canada. (2011). Fuel Combustion. Retrieved November 2011, from www.ec.gc.ca.
Gill, G., & Fung, A. (2011). Energy and Environmental Analysis of Residential Hot Water Systems; A Study for Ontario,
Canada. ASHRAE Transactions , 117, Part2, 506-520.
Hendron, R., & Burch, J. (2007). Development of Standardized Domestic Hot Water Event Schedules for Residential
Buildings. Energy Sustainability. Long Beach, CA.
Jordan, U., & Vajen, k. (2000). Influence of the DHW Load Profile on the Fractional Energy Savings: A case Study of a
Solar Combi-System with TRNSYS Simulations. Solar Energy , 60, 197-208.
Knight, I., Kreutzer, N., Manning, M., Swinton, M., & Ribberink, H. (2007). European and Canadian non-HVAC Electric and
DHW Load Profiles for Use in Simulating the Performance of Residential Cogeneration Systems. International Energy Agency (IEA
Annex 42).
London Economics LLD. (2011, March). Independent Power Producers Society of Alberta. Retrieved November 2011, from
www.ippsa.com.
NRCan. (2009, September). Energy Efficiency Trends in Canada: 1990 to 2007. Retrieved March 2011, from Natural Resources
Canada: http://oee.nrcan.gc.ca/publications/statistics/trends09/pdf/trends.pdf
Parker, D. (2003). Research highlights from a large scale residential monitoring study in a hot climate. Energy and Buildings 35
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
, 863-876.
Perlman, M., & Mills, B. (1985). Development of Residential Hot Water Use Patterns. ASHRAE Transactions , 657-679.
Statistics Canada. (2011). Energy Statistics Hanbook, 2nd Quarter 2011. Statistics Canada.
Stevenson, D. (1983). Residential Hot Water Use Patterns (Volume I). Report for the Canadian Electrical Association Research
and Development, Montreal, QC.
Tanha, K., Fung, A., & Leong, W. (2011). Investigating the Performance of a Hybrid Solar Domestic Hot Water System
Through Computer Simulation and Data Analysis. Buiding Simulation. Sydney, Australia.
Viessmann Ltd. (2010). Vitotech technical guide. Retrieved April 12, 2011, from www.viessmann.com.
Zaloum, C., Lafrance, M., & Gusdorf, J. (2007). Drain Water Heat Recovery Characterization and Modeling. Ottawa: Sustainable
Buildings and Communities, Natural Resources Canada.
Zhang, D., Barua, R., & Fung, A. (2011). TRCA-BILD Archetype Sustainable House - Overview of Monitoring System and
Preliminary Results for Mechanical Systems. ASHRAE Transactions , 117, Part 2, 597-612.
$6+5$(

SA-12-C028
Decision Making for HVAC&R System

Selection for a Typical Office Building in the
UK
Mehdi Shahrestani Runming Yao, PhD Geoffrey K Cook, PhD

Student Member ASHRAE Member ASHRAE
ABSTRACT
Demands for thermal comfort, better indoor air quality together with lower environmental impacts have had ascending trends in the last decade. In
many circumstances, these demands could not be fully covered through the soft approach of bioclimatic design like optimization of the building
orientation and internal layout. This is mostly because of the dense urban environment and the internal energy loads of the building. In such cases,
heating, ventilation, air-conditioning and refrigeration (HVAC&R) systems make a key contribution to fulfilling the requirements of the indoor
environment. Therefore, the most appropriate HVAC&R system must be selected. In this study, a robust decision making approach for
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
HVAC&R system selection is proposed. Technical performance, economic and environmental impacts of 36 permutations of primary and
secondary systems are taken into account to choose the most appropriate HVAC&R system for a case study office building. The building is a
representative of the dominant form of office buildings in the UK. Dynamic performance evaluation of HVAC&R alternatives using the
TRNSYS package together with life-cycle energy cost analysis is undertaken to provide a reliable basis for decision making. The six scenarios
considered in this paper broadly cover the decision makers' attitudes on HVAC&R system selection. These scenarios are analyzed through an
Analytical Hierarchy Process (AHP), and sensitivity analyses which prove the robustness of the final decision for each scenario. One of the
significant outcomes of this study reveals that the compound heating, cooling and power system (CCHP), despite the high level of energy demand
and also a higher initial investment, is one of the top-ranked alternatives due to the lower energy cost and C02 emissions. The proposed approach
and the results of this study could be used by researchers and designers, especially in the early stages of a design process where all the involved bodies
commonly face a lack of time, information and the tools necessary for the evaluation of a variety of HVAC&R systems.
INTRODUCTION
Building energy consumption accounts for more than 38% of the UK energy used in 2009. Besides, HVAC&R
systems demand more than 60% of the building energy consumption in the UK (DECC, 2010). This significant energy
demand together with the ascending trend in utilizing HVAC&R systems (BSRIA, 2008) underline the importance of
selecting the most appropriate HVAC&R system during the design process.
HVAC&R system selection is conducted in the early stages of a design process (Phillips, 2008). Decisions in these
early stages significantly influence the total building energy performance and establish up to 90% of the life time occupants'
satisfaction and building running costs (Elovitz, 2002). Therefore, "HVAC&R designers are responsible for considering
various systems and recommend one or two that will meet the project requirements"(ASHRAE, 2008). However, the lack
of tools to investigate a variety of HVAC&R systems and the paucity of systematic selection methods appear to result in to
Mehdi Shahrestani is a PhD student, Runming Yao and Geoffrey K Cook are readers in the School of Construction Management and
Engineering, University of Reading, UK.
222
ASHRAE Transactions
a basic rule of thumb approach to HAVC&R system selection (Ellis and Mathews, 2002; Langmaid, 2004). This,
consequently, might result in missing some promising alternatives (Maor and Reddy, 2004).
The open literature reveals two dominant approaches for HVAC&R system selection: 1-Decision making based on
the data gathered from simulation studies 2-Decision making based on data collected from large and small-scale surveys. A
key investigation using the first approach was conducted by Avgelis and Papadopoulos (2009). Dynamic performance
evaluation of HVAC&R systems along with multi-criteria decision making have been utilized to choose the most
appropriate system. Simulation of the alternatives using the TRNSYS package along with conducting a formal decision
making process using the ElectreIII method are the undeniable strengths of their study. Unfortunately, the investigated
systems are limited to few alternatives. This deficiency was addressed through the development of an expert model for
HVAC&R system selection in order to automatically synthesize a complete set of possible alternatives (Maor and Reddy,
2004). Despite covering a broad range of alternatives, life-cycle cost was the only criteria used to rank the alternatives.
In the second approach, various studies based on large and small-scale surveys have been undertaken. Shams et
al.(1994) developed a knowledge-based model for HVAC&R system selection. Using the same approach, a prototype
knowledge-based model (ESCHER) was developed to assist decision makers in HVAC&R system selection (Fazio and
Bedard, 1989). Even though capturing the experts’ knowledge is the rationale behind the development of the ESCHER
model, Shams et al. (1994) demonstrated that in some cases there are many conflicts not only between decision making
attitudes of different experts but also between experts and the open literature. These studies were based on capturing the
decision makers' attitudes on system selection without analyzing the decision making criteria for system selection. Other
researchers attempted to capture HVAC&R system evaluation criteria as well as decision maker's preference from on-site
surveys to permit formal decision making (Wang et al., 2009). An advanced Fuzzy decision making method, which has been
used by Wang et al. (2009) considers few alternatives and the surveyed results were not clearly described. It seems that the
main attempt was to introduce a new application for the advance decision making method rather than to mitigate the
complexity of the decision making process for HVAC&R system selection. A review of the existing building energy
benchmarks developed using the on-site survey of buildings reveals that the performance of HVAC&R systems are not
addressed in detail (Brigges et al., 1992; ECG-19, 2000; CIBSE-TM46, 2008; Torcellini et al., 2008). The investigations
show that the survey results do not provide sufficient details to make a clear distinction between surveyed HVAC&R
systems. Moreover, in many cases, the root cause of an unexpected performance of a system could not be identified. These
two deficiencies have been coarsely overcome by introducing a typical HVAC&R system within energy benchmarks
(Brigges et al., 1992; ECG-19, 2000; CIBSE-TM46, 2008; Torcellini et al., 2008).
RESEARCH DESIGN
In this study, the first approach of decision making for HVAC&R system selection is adopted. The decision is based
on the data gathered from simulation of the variety of HVAC&R systems for a case study office building. The second
approach to decision making is not adopted due to several reasons. These include: 1-The difficulty in capturing the decision
makers' attitudes on system selection and the uncertainty in relation to compliance with the open literature (Shams et al.,
1994) 2-The difficulty in identifying and capturing the detailed performance of different components of HVAC&R systems
in existing buildings (Brigges et al., 1992) 3-There is a concern about the dependency of the performance of HVAC&R
systems on building specifications (Korolija et al., 2011).
Among the variety of HVAC&R system simulation tools, Transient System Simulation (TRNSYS) has been selected
due to several advantages including thermal comfort analysis (Klein et al., 2009). A formal decision making process is
conducted using the outcome of the simulations. In relation to the nature of HVAC&R system selection that is performed
in a design and construction environment rather than in a research study, a user-friendly decision making method is always
more appropriate. The user-friendly structure of the Analytic Hierarchy Process (AHP) method along with the
acknowledged qualities of this method, especially in engineering applications (Yang et al., 2010) are the main reasons for
adopting the AHP for decision making in this study (Saaty, 1990).
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

BUILDING SPECIFICATIONS
Daylit cellular form office buildings account for more than 67% of office buildings in England and Wales (Steadman
et al., 2000). Therefore, a daylit cellular office building has been selected for this case study (Figure 1). The glazed area is
assumed to be 20% of the floor area (Gakovic, 2000) and the thermal conductance of walls, roof and windows are
respectively assumed as 0.28, 0.18 and 1.8 W/m2.K (0.049, 0.032 and 0.32 Btu/hr.ft2.°F) (HM Government, 2010a). The
infiltration rate is set to 0.3 air changes per hour (CIBSE, 2006). Also, the ventilation rate is assumed to be 10
litres/sec.person (21.2 cfm/person) (HM Government, 2010b). Occupancy density of the prototypical building is
considered about 10 m2/person (107.6 ft2/person) (CIBSE, 2006). A power load of 15 W/m2 (4.75 Btu/hr.ft2 ) is taken
into account for lighting energy loads (SLL, 2009). Finally, the electrical equipment load is assumed to be 200 W/person
(682 Btu/hr. person ) (CIBSE, 2006). The building is assumed to be in use only during weekdays from 8 am to 6 pm. The
indoor temperature is set to 23°C (73.4 °F) in cooling mode and 22°C (71.6 °F) during heating mode along with 45%
relative humidity (CIBSE, 2006). The models of the building and HVAC&R systems are created in TRNSYS and simulated
using the London-Gatwick weather data file (Klein et al., 2009).
40
5
12
2
Figure 1 Isometric model and typical plan of the case study office building. (All dimensions in meters)
HVAC&R SYSTEMS SPECIFICATIONS

In this study, 36 permutations of three primary and 12 secondary systems have been investigated. Tables 1 and 2
illustrate both the primary and secondary systems respectively.
Table 1. Primary systems.
No. Heating and cooling parts
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
1 Gas boiler with reciprocating air cooled chiller

2 Gas boiler with absorption chiller and cooling tower
3 Compound heat and power (CHP) with absorption chiller and cooling tower (CCHP)
Table 2. Secondary systems.

No. Part 1 Part 2 Part 3
1 Constant air volume (CAV) - -
2 Variable air volume (VAV) - -
3 Constant air volume (CAV) Heat recovery (thermal wheel) -
4 Variable air volume (VAV) Heat recovery (thermal wheel) -
5 Constant air volume (CAV) Economizer -
6 Variable air volume (VAV) Economizer -
7 Constant air volume (CAV) - Reheat coil
8 Variable air volume (VAV) - Reheat coil
9 Constant air volume (CAV) Heat recovery (thermal wheel) Reheat coil
10 Variable air volume (VAV) Heat recovery (thermal wheel) Reheat coil
11 Constant air volume (CAV) Economizer Reheat coil
12 Variable air volume (VAV) Economizer Reheat coil
224
ASHRAE Transactions
It is assumed that all primary systems are installed in a plant room close to the building and each floor has a separate
air- handling unit. The energy efficiency of HVAC&R system components are defined based on the actual manufacturer
information within the acceptable range recommended by the recent standards (ASHRAE, 2010).
AHP MULTI-CRITERIA DECISION MAKING

The AHP is a multi-criteria decision making method which is formed based on a pair-wise comparison of decision
criteria and alternatives in a hierarchy structure. The hierarchy structure provides an overview of the inherent relation
between goal, criteria and alternatives. The AHP method is implemented in practice through five principal steps (Saaty,
1990): 1-structuring the decision hierarchy 2-collecting data and pair-wise comparison 3-checking the consistency of
material judgments 4-applying the weighting method to calculate both criteria and alternatives weights 5-aggregating the
weights and ranking the alternatives. Figure 2 illustrates the hierarchy structure of the decision making process for
HVAC&R system selection used in this study.
Figure 2 Hierarchy structure of the decision making for HVAC&R system selection.
Matrix D, represents the pair-wise comparison of criteria (C) and matrix E, represents the pair-wise comparison of
alternatives (A) with respect to each criterion when there are 'n' criteria and alternatives.
⎡ C1/C1 C1/C 2 ... C1/C n ⎤ ⎡ A 1/A 1 A 1/A 2 ... A 1/A n ⎤
⎢ ⎥ ⎢ A /A A 2 /A 2 ... A 2/A n ⎥
⎢C 2/C1 C 2 /C 2 ... C 2 /C n ⎥ ⎢ 2 1 ⎥
D = ⎢ . . . ⎥ E=⎢ . . . ⎥ (1)
⎢ . ... ⎥ ⎢ ... ⎥
⎢ . . ⎥ . . .
⎢ ⎥
⎣⎢ C n /C1 C n /C 2 ... C n /C n ⎦⎥ ⎣⎢ A n /A 1 A n /A 2 ... A n /A n ⎦⎥
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Pair-wise comparison of criteria and alternatives are measured based on a fundamental nine-point comparison scale, 1:
equal importance, 3: moderate importance, 5: strong importance, 7: very strong importance, 9: absolute importance (Saaty,
1990). Then, subjective priority vector (weights of the criteria) and objective priority vectors (weights of the alternatives) are
calculated using the Eigenvector Method (Saaty, 2003). The general eigenvector is obtained by perturbation of the following
consistent formulation where, W is a priority vector (eigenvector) and the eigenvalue of the matrix E is represented by λmax .
⎡ A 1/A 1 A 1/A 2 ... A 1/A n ⎤ ⎡ w1 ⎤ ⎡ w1 ⎤
⎢ ⎥ ⎢ ⎥ ⎢ ⎥
⎢ A 2/A 1 A 2 /A 2 ... A 2/A n ⎥ ⎢w 2 ⎥ ⎢w 2 ⎥
E.W = ⎢ . . . ⎥ . ⎢ . ⎥ = λ max . ⎢ . ⎥ (2)
⎢ . ... ⎥ ⎢ ⎥ ⎢ ⎥
⎢ . . ⎥ ⎢ . ⎥ ⎢ . ⎥
⎢⎣ A n /A 1 A n /A 2 ... A n /A n ⎥⎦ ⎢⎣ w n ⎥⎦ ⎢⎣ w n ⎥⎦
This approach has to be followed to determine the objective priority vectors with respect to each criterion and also to
calculate the subjective priority vector (Saaty, 1990). After developing the comparison matrix and compiling both subjective
and objective priority vectors (weights), the consistency of the comparison matrix is calculated. The AHP uses a method to
evaluate the consistency of a comparison matrix by introducing a Consistency Index (C.I.) and a Consistency Ratio (C.R.).
$6+5$(

The C.I. of a comparison matrix is given by C.I.=( λmax-n)/(n-1) and C.R. is obtained from the ratio of C.I. and the Random
Index (R.I.) which were introduced by Saaty (1990).
The approach of the criteria evaluation for each alternative is introduced in this part. First, energy consumption of
HVAC&R systems is evaluated using the outcome of simultaneous simulation of HVAC&R systems and the building. It is
defined according to the following equation where there are 'j' energy-consuming processes within a HVAC&R system:
TEC = ∑1j EC j (3)
Also, the energy-related CO2 emissions of HVAC&R systems are calculated by considering CO2 emission factors of
natural gas and electricity, from National Grid equal to 0.19 and 0.55 kg-CO2/kW.h (0.42 and 1.21 lb-CO2/kW.h)
respectively (CIBSE-TM46, 2008). It should be remembered that, the main goal of utilizing HVAC&R systems is to
provide acceptable indoor environmental quality. The users' satisfaction is measured through the consideration of thermal
comfort and indoor air quality (BS/EN:15251, 2007). In this study, an index for deviation from set-point temperature (ITC)
is taken into account to measure the level of thermal comfort. With the same approach, an index for indoor air quality
(IIAQ) is adopted to measure the level of indoor air quality according to the following equations where there are 'i' spaces
which are studied in 'h' hours:
∑1i ∑1h ( DSTi , h × X i , h )
ITC = (4)
∑1i ∑1h X i , h
∑1i ∑1h ( PPM i , h × X i , h )
IIAQ = (5)
∑1i ∑1h X i , h
Moreover, life-cycle energy costs of the alternatives are calculated based on the net present value (NPV) of money
using the following equation (6). In this study, a 20-year life is assumed for HVAC&R systems along with 3.5% discount
rate (Churcher, 2008). Also, the initial costs of the main equipment for each alternative are estimated based on the recent
cost references (Langdon, 2011).
C
NPV = (6)
(1 + R )N
DISCUSSION AND RESULTS

To broadly cover the variety of decision makers' attitudes, this study considered six scenarios for subjective priority
vectors (weights of the criteria) which are shown in Table 3. The consistency ratio of these priority vectors are also shown
in this table. These values are within the acceptable level of 0-0.10 (Saaty, 1990).
After pair-wise comparison of the alternatives with respect to each criterion, the objective priority vectors (weights of
the alternatives) are calculated using the Eigenvector method (EM). These values are shown in the left side of Table 4
(columns 4 to 9). Multiply combination of the subjective priority vectors (Table 3) and the objective priority vectors (Table
4, columns 4 to 9) culminates in the final ranking of the alternatives under different scenarios. These results are
demonstrated in the right side of Table 4 (columns 10 to 15).
Table 3. Weights of the criteria (subjective priority vectors) under six different scenarios.
Scenario Energy CO2 ITC IIAQ Life-cycle Initial Cost Consistency
No. consumption emissions energy Ratio
cost
1 0.50 0.10 0.10 0.10 0.10 0.10 0.020
2 0.10 0.50 0.10 0.10 0.10 0.10 0.015
3 0.10 0.10 0.10 0.10 0.50 0.10 0.020
4 0.10 0.10 0.10 0.10 0.10 0.50 0.023
5 0.25 0.05 0.30 0.30 0.05 0.05 0.075
6 0.05 0.25 0.30 0.30 0.05 0.05 0.068
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
226
ASHRAE Transactions
Table 4. Weights (objective priority vector) and final ranks of the alternatives for different scenarios.
HVAC&R Alternatives priority vectors with respect to different criteria Final alternatives Ranks
system based on scenarios No. 1-6
No P S Energy CO2Life-cycle Initial
ITC IIAQ 1 2 3 4 5 6
consumption emissions
energy Cost
cost
1 1 1 0.0308 0.0206 0.0222 0.0289 0.0201 0.0505 11 22 23 1 18 31
2 1 2 0.0337 0.0250 0.0239 0.0267 0.0252 0.0363 8 17 17 5 17 28
3 1 3 0.0381 0.0241 0.0230 0.0289 0.0231 0.0350 3 18 18 6 11 26
4 1 4 0.0416 0.0297 0.0297 0.0266 0.0296 0.0276 1 7 11 13 2 10
5 1 5 0.0316 0.0215 0.0226 0.0289 0.0211 0.0411 12 23 25 4 19 32
6 1 6 0.0304 0.0260 0.0256 0.0267 0.0264 0.0312 6 16 16 12 15 25
7 1 7 0.0309 0.0169 0.0332 0.0289 0.0157 0.0418 13 32 32 3 7 19
8 1 8 0.0338 0.0218 0.0337 0.0267 0.0211 0.0316 10 21 24 11 6 16
9 1 9 0.0375 0.0204 0.0335 0.0285 0.0188 0.0306 4 26 26 14 3 13
10 1 10 0.0419 0.0268 0.0338 0.0267 0.0259 0.0248 2 12 14 17 1 8
11 1 11 0.0315 0.0174 0.0334 0.0289 0.0161 0.0352 14 33 34 9 8 20
12 1 12 0.0344 0.0226 0.0335 0.0267 0.0219 0.0277 9 20 21 19 5 14
13 2 1 0.0262 0.0188 0.0222 0.0289 0.0187 0.0459 24 31 31 2 28 36
14 2 2 0.0290 0.0229 0.0239 0.0267 0.0237 0.0339 19 25 20 10 25 34
15 2 3 0.0315 0.0217 0.0230 0.0289 0.0213 0.0327 15 27 27 15 20 33
16 2 4 0.0346 0.0266 0.0295 0.0266 0.0272 0.0261 5 15 15 21 9 15
17 2 5 0.0292 0.0202 0.0226 0.0289 0.0199 0.0380 20 30 30 7 24 35
18 2 6 0.0308 0.0233 0.0256 0.0267 0.0237 0.0294 16 24 22 20 21 30
19 2 7 0.0257 0.0157 0.0321 0.0289 0.0149 0.0386 26 36 36 8 16 29
20 2 8 0.0290 0.0202 0.0337 0.0267 0.0199 0.0298 21 29 28 18 13 22
21 2 9 0.0311 0.0186 0.0322 0.0289 0.0176 0.0289 18 34 33 22 10 23
22 2 10 0.0348 0.0243 0.0337 0.0267 0.0240 0.0236 7 19 19 24 4 12
23 2 11 0.0291 0.0168 0.0321 0.0289 0.0158 0.0329 22 35 35 16 14 27
24 2 12 0.0310 0.0206 0.0336 0.0267 0.0201 0.0262 17 28 29 23 12 21
25 3 1 0.0165 0.0322 0.0217 0.0289 0.0350 0.0200 36 13 12 30 36 24
26 3 2 0.0170 0.0477 0.0232 0.0266 0.0438 0.0173 28 3 4 26 34 3
27 3 3 0.0198 0.0346 0.0222 0.0289 0.0392 0.0170 30 8 7 35 33 17
28 3 4 0.0204 0.0527 0.0240 0.0266 0.0503 0.0150 23 1 1 25 27 1
29 3 5 0.0188 0.0331 0.0219 0.0289 0.0375 0.0187 34 10 9 31 35 18
30 3 6 0.0187 0.0453 0.0235 0.0266 0.0447 0.0160 27 4 3 28 32 5
31 3 7 0.0158 0.0307 0.0308 0.0289 0.0334 0.0185 35 14 13 32 29 11
32 3 8 0.0162 0.0445 0.0263 0.0267 0.0412 0.0162 32 5 6 29 31 4
33 3 9 0.0190 0.0331 0.0306 0.0289 0.0375 0.0159 31 9 8 36 22 7
34 3 10 0.0195 0.0493 0.0263 0.0267 0.0474 0.0142 25 2 2 27 26 2
35 3 11 0.0180 0.0318 0.0310 0.0289 0.0359 0.0170 33 11 10 34 23 9
36 3 12 0.0179 0.0426 0.0263 0.0267 0.0422 0.0151 29 6 5 33 30 6
Note: 'P' and 'S' represent the associated number of primary system (Table1) and secondary system (Table2) respectively.
The final ranking of the variety of HVAC&R systems (Table 4) reveals that, when the energy consumption is the
dominant criterion (first scenario, Table 3), the reciprocating chiller with gas boiler linked into a VAV air distribution
equipped with heat recovery is the best alternative (Table 4). The general trend of the highly ranked alternatives is towards
the reciprocation chiller with gas boiler in the primary part and VAV systems in the secondary part. Moreover, when the
energy-related CO2 emissions are the dominant criterion (second scenario, Table 3), the CCHP linked to a VAV system
equipped with heat recovery is seen as the best option (Table 4). Highly ranked alternatives under this scenario are
predicted for the systems linked to a CCHP unit. This trend is also observed for the third scenario (dominance of life-cycle
energy cost) (Table 3). When the initial cost is the major concern (fourth scenario, Table 3), the reciprocating chiller with
boiler linked to a CAV distribution system has the highest ranking (Table 4). Under this scenario, the highly ranked
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

alternatives mostly belong to the systems with reciprocating chillers and gas boilers in the primary part and a CAV
distribution system in the secondary part (Table 4). In the fifth scenario (Table 3) when the thermal comfort and indoor air
quality have the highest priority followed by a moderate priority for energy consumption, the reciprocating chiller and boiler
linked to a VAV system equipped with heat recovery and reheat coil is the top ranked alternative (Table 4). Finally, in the
sixth scenario when the indoor air quality and thermal comfort have the highest priority followed by a moderate priority for
energy-related CO2 emissions (Table 3), the CCHP linked into a VAV air distribution system equipped with heat recovery
has the highest rank (Table 4). In general, the highly ranked alternatives under this scenario belong to the systems linked to
a CCHP unit. To examine the robustness of the results, a series of sensitivity analyses is performed. The dominant criteria
weights are altered between -20% and +20% by taking into account the constant relative weight for the rest of the criteria.
The outcomes of the analyses show that the order of the first five highly ranked alternatives is not changed. This highlights
the robustness of the HVAC&R system selection approach adopted in this study.
CONCLUSION
This study proposes a robust decision making approach for HVAC&R system selection. Simultaneous dynamic
simulation of HVAC&R systems within a prototypical office building has produced a reliable set of information about
performance characteristics of a variety of HVAC&R systems. This information along with the associated cost estimations
form the basis of the decision making process. Providing six scenarios to represent a variety of different weighted criteria
made this study applicable to decision makers with different attitudes. The ranking of the alternatives under the six
scenarios shows that the CCHP system, despite the high level of energy demand and also a higher initial investment is
highly likely to be selected as a highly ranked HVAC&R system when the energy-related CO2 emissions are of importance.
Finally, the sensitivity analyses reveal that for all these six scenarios, the order of the first five highly ranked alternatives is
not changed. The proposed approach and the results of this study could be used by researchers and designers, especially in
the early stages of a design process when all those involved face a lack of time, information and the tools for evaluation of a
variety of alternatives in HVAC&R system selection.
ACKNOWLEDGMENTS
Mehdi Shahrestani would like to acknowledge the School of Construction Management and Engineering of the
University of Reading for partially funding his doctorate study.
NOMENCLATURE
TEC = Total energy consumption (kW.hr or Btu)
EC = Energy consumption of each process within a HVAC&R system (kW.hr or Btu)
DST = Deviation from the set-point temperature (K or °F)
ITC = Index for deviation from the set-point temperature (K or °F)
IIAQ = Index for indoor air quality (ppm)
PPM = Part per million CO2 concentration (ppm)
NPV = Net present value (£)
X = Number of people (-)
C = Annual cost (£/year)
R = Discount rate (-)
N = Year of occurrence (-)
Subscripts
i,h = space 'i' at hour 'h'
j = energy consuming process 'j' within HVAC&R systems
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
228
ASHRAE Transactions
REFERENCES
ASHRAE. 2008. ASHRAE handbook-HVAC systems and equipment. Atlanta: American Society of Heating, Refrigeration
and Air-conditioning Engineers, Inc.
ASHRAE. 2010. ANSI/ASHRAE/IESNA Standard-90.1, Energy standard for buildings except low-rise residential
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
buildings. Atlanta: American Society of Heating Refrigerating and Air-conditioning Engineers, Inc.
Avgelis, A. and Papadopoulos, A.M. 2009. Application of multi-criteria analysis in designing HVAC systems. Energy &
Buildings, 41: 774-780.
Brigges, R.S., Crawley, D.B. and Schliesing, J.S. 1992. Energy requirements of office buildings, vol-1. Battelle: Pacific
Northwest Laboratory for Gas Research Institute.
BS/EN:15251. 2007. Indoor environmental input parameters for design and assessment of energy performance of buildings
addressing indoor air quality, thermal environment, lighting and acoustics. London: British/European Standard.
BSRIA. 2008. World market for air conditioning study. Bracknell: Building Services Research and Information Association.
Churcher, D. 2008. Whole-life costing analysis. Bracknell: Building Services Research and Information Association.
CIBSE-TM46. 2008. Energy benchmark. London: Chartered Institution of Building Services Engineers.
CIBSE. 2006. CIBSE guide A: Environment design. London: Chartered Institution of Building Services Engineers.
DECC. 2010. Digest of UK energy statistic. London: Department of Energy and Climate Change.
ECG-19. 2000. Energy consumption guide 19: Energy use in offices. London: The Government's Energy Efficiency Best
Practice programme.
Ellis, M.W. and Mathews, E.H. 2002. Needs and trends in building and HVAC system design tools. Building and
Environment, 37(5): 461-470.
Elovitz, D.M. 2002. Selecting the right HVAC system. ASHARE Journal, 44(1): 24-30.
Fazio, P. and Bedard, C. 1989. Knowledge-based system approach to building envelope design. Computer aided design,
21(8): 519-527.
Gakovic, B. 2000. Areas and types of glazing and other opening in the nondomestic building stock. Environmental and
planning B: Planning and design, 27: 667-694.
HM Government. 2010a. Building regulations-conservation of fuel and power in new buildings other than dwelling: Part
L2A. London: Her Majesty’s Stationery Office.
HM Government. 2010b. Building regulations-ventilation: Part F. London: Her Majesty’s Stationery Office.
Klein, S., Beckman, W., Mitchell, J., Duffie, J., Duffie, N., Freeman, T., Mitchell, J. and Braun, J. 2009. TRNSYS 17: A
Transient System Simulation Program. Wisconsin: Solar Energy Laboratory, University of Wisconsin.
Korolija, I., Marjanovic-Halburd, L., Zhang, Y. and Hanby, V. 2011. Influence of building parameters and HVAC systems
coupling on building energy performance. Energy and Buildings, 43(6): 1247-1253.
Langdon, D. 2011. Mechanical and electrical services price book. London: Taylor & Francis.
Langmaid, J. 2004. Choosing building services. Bracknell: Building Services Research and Information Association.
Maor, I.H. and Reddy, T.A. 2004. Development and Implementation of HVAC-KBCD: A knowledge-based expert system
for conceptual design of HVAC&R systems-Part 1: general framework. ASHRAE Transactions, 110(2): 232-257.
Phillips, R. 2008. Plan of work: Multi-Disciplinary Services. London: Royal Institute of British Architects.
Saaty, T.L. 1990. How to make a decision: the analytic hierarchy process. European journal of operational research, 48: 9-
26.
Saaty, T.L. 2003. Decision-making with the AHP: Why is the principal eigenvector necessary. European journal of
operational research, 145: 85-91.
Shams, H., Maxwell, G.M., Nelson, R.M. and Leonard, C. 1994. Development of knowledge-based system for the selection
of HVAC system types for small buildings. ASHRAE Transactions, 100(1): 203-17.
SLL. 2009. Code for lighting CD-ROM. London: Society of Light and Lighting.
Steadman, P., Bruhnes, H.R., Holtier, S., Gakovic, B., Rickaby, P.A. and Brown, F.E. 2000. A classification of built forms.
Environmental and planning B: Planning and design, 27: 73-91.
Torcellini, P., Deru, M., Griffith, B., Benne, K., Halverson, M., Winiarski, D. and Crawley, D. 2008. DOE Commercial
building benchmark models. California: National Renewable Energy Laboratory.
Wang, J.J., Jing, Y.Y. and Zhang, C.F. 2009. Fuzzy multi-criteria evaluation model of HVAC schemes in optimal
combination weighting method. Building services engineers, 30(4): 287-304.
Yang, Y., Li, B. and Yao, R. 2010. A method of identifying and weighting indicators of energy efficiency assessment in
Chinese residential buildings. Energy Policy, 38: 7687-97.
$6+5$(

SA-12-C029
Laboratory Evaluation of Aftermarket Boiler

Control System
Patricia Rowley Paul Glanville, PE

Associate Member ASHRAE
ABSTRACT
Aftermarket boiler control systems offer a potential option for energy efficiency retrofits of commercial gas-fired boilers. These devices are installed
on existing boilers with no interruption to boiler operation. The systems monitor boiler operation and adjust boiler cycling rates to reduce annual
fuel consumption. Aftermarket boiler control systems are seeing initial market and technical success and thus, are candidates for utility-sponsored
energy efficiency programs.
Through a third-party laboratory evaluation of an aftermarket boiler control system, preliminary data reported from this study quantify the fuel
savings under a range of heating demand profiles. A laboratory test stand is developed to deliver a variable heating load to the test boiler in order
to simulate a typical building installation. The Department of Energy (DOE) Energy Plus model, using the standard reference building model of
a secondary school, is used to create a 24-hour heating demand profile for typical commercial buildings.
Boiler operation and energy consumption is monitored under both steady-state heating loads and 24-hour simulated heating demand profiles. Fuel
consumption is measured with and without the aftermarket control system to determine the reduction in burner cycles, fuel savings, and system
temperatures. Under most conditions, the aftermarket control system reduces boiler cycling and fuel consumption as compared to the baseline tests.
The maximum fuel savings occurs during periods of peak heating demand. However, the magnitude of fuel savings varies depending on the control
setpoints of the boiler, specifically the aquastat differential (or deadband). This paper summarizes preliminary data from this laboratory
evaluation of an aftermarket boiler control system within a range of heating loads to aid in the identification of applications where it may provide
the most fuel savings.
INTRODUCTION
Gas-fired boilers provide heating for over 30% of commercial heated floor space in the U.S. and have median
lifetimes of 25-35 years.1,2 A retrofit solution that can provide energy savings and reduce CO2 emissions would be a cost-
effective option for existing commercial boilers relative to a replacement upgrade. Aftermarket boiler control systems are
designed to be installed on existing boilers with minor interruption to boiler operation. By measuring the temperature at
the boiler return and supply, or the boiler return only, these systems monitor heating loads at the boiler and adjust boiler
cycling through the aquastat controls to prevent unnecessary and inefficient operation. The control device evaluated in this
report incorporated an algorithm to evaluate prior demand patterns, current demand ramp up, and other factors to delay
firing during periods of perceived low demand and release it to fire during significant heating demands. Aftermarket boiler
control systems are seeing initial market and technical success and are now candidates for utility-sponsored energy
efficiency programs. Commercially available aftermarket boiler control systems have reported energy savings up to 30% in
Patricia Rowley is a Senior Engineer at the Gas Technology Institute, Des Plaines, IL. Paul Glanville, PE is a Principal Engineer at
the Gas Technology Institute, Des Plaines, IL
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
230
ASHRAE Transactions
some cases, while other field evaluations of these devices have shown marginal reductions in gas consumption.3,4
The goal of this study is to evaluate an aftermarket boiler controller system in a controlled laboratory setting to
evaluate its performance relative to baseline operation during typical building heating load patterns. The laboratory test
stand also provides a means to assess the variability of these gains relative to load profile and temperature settings in order
to determine the parameters for optimal performance and identify the best applications for maximum energy savings
relative to climate and building type.
LABORATORY TEST SETUP

An automated laboratory test stand was developed to evaluate the performance of a hot water boiler with aftermarket
controls and to quantify the range of energy savings that could potentially be achieved in field installations. The laboratory
test stand was used for steady state testing and to simulate the heating demand from a typical building. The test stand,
shown in Figure 1, consists of a 195 MBH (57 kW) residential boiler installed in a primary/secondary configuration. While
the single-stage residential sized hot water boiler is a fraction of the size of typical boilers this device is installed on, the test
stand was operated in a manner representative of typical installations, in agreement with the controls manufacturer . The
boiler circulator pump and the secondary loop pump operate continuously throughout the tests at approximately equal flow
rates, equivalent to a hydronic system simple circuit configuration.
Figure 1 Laboratory test stand to evaluate aftermarket boiler control systems was automated to simulate 24-hour heating load profiles.
$6+5$(

--`,`,``,,`,`,``
The laboratory test rig was constructed to reproduce a given heating load pattern from the perspective of the test
boiler. The secondary loop includes a 60 gallon (227 L) storage tank to simulate the total pipe volume in a typical hydronic
installation based on recommendations from the manufacturer. The heating load at the boiler is generated by a cooling
water loop circulating through a heat exchanger in the secondary loop. The cooling loop incorporates an electric chiller,
storage, and a variable speed pump. A second heat exchanger added to the cooling loop provides the higher heating loads
required for the winter heating demand profile and pipe heat loss. A real-time data acquisition system collects natural gas
pressure, gas volume flow rates, water flow rates, water and gas temperatures every second during steady state testing and
every ten seconds during 24-hour tests. Water temperatures are recorded at the boiler supply/return and throughout the
secondary loop. The boiler heat output is calculated based on the temperature difference at the boiler supply and return,
and the flow rate of the boiler circulator pump.
The heating load at the boiler was controlled by the flow rate in the cooling loop using a variable voltage controller on
the pump. Data collected during calibration test runs was used to determine the relationship between the input voltage to
the pump controller and the corresponding heat demand measured at the boiler. Automating the pump controls in the
cooling loop allows the test stand to simulate variations in heating load at the boiler to match a given 24-hr heating load
profile of typically building.
ENERGY PLUS MODELING OF HEATING LOAD
Heating profiles were determined using Energy Plus models based on DOE reference buildings for a secondary
school, post 1980, using climate data for Chicago using TMY3 data5. The Energy Plus model boiler capacity was based
on a scale factor of 1.2 peak load and the 24 hour secondary school heating load was scaled for the test setup using
the ratio of the reference building boiler capacity to the capacity of the test boiler
Twenty-four hour heating profiles were selected based on average outdoor temperatures. A winter heating
profile, scaled for the test setup, is shown in Figure 2. During the 24-hour period, the default building energy
management system turns on the HVAC for 1 minute every 20 minutes during the morning (hours 1-6) and evening (hours
21-24) while the building is unoccupied. At 6:00 am the HVAC is turned on, creating a rapid increase in demand which
then reduces to a slightly lower level during the remainder of the day. The HVAC system returns to the night cycle settings
at 21:00.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
180,000
160,000
Heating Demand - Scaled for Test Setup (Btu/h)
140,000
120,000
100,000
80,000
60,000
40,000
20,000
-
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Time (hrs)
Figure 2. Scaled 24-hour input heating profile for a winter day based on Energy Plus models for DOE Reference buildings for a Chicago
Secondary School, post 1980.
232
ASHRAE Transactions
The aftermarket boiler controller was provided and installed in the test stand by the manufacturer. The boiler control
system consists of a small electronics box which cuts into the boiler aquastat wire to control the boiler. Two surface-
mounted temperature sensors are installed on the primary loop tubing at the boiler supply and return. Installation was
completed in less than 2 hours with no disruption to the heating system. The manufacturer also participated in a design
review of the test stand and testing protocol.
EXPERIMENTAL RESULTS
Test Matrix
The test matrix for this project is shown in Table 1. Thirty minute steady-state tests and 24-hr heating profiles based
on Energy Plus models were conducted using the test stand with the gas-fired boiler. Baseline tests were conducted with
the aftermarket controller powered off. For both the baseline and controller test runs, the boiler aquastat setpoint was
175F (79C) with a differential of 5F (3C), corresponding to a high temperature limit of 175F (79C) and a low limit of 170F
(77C). In addition, a third test run was conducted for each test condition to compare the performance of the aftermarket
controller with simply increasing the boiler differential from 5F (3C) to 15F (8C), corresponding to an aquastat low limit of
160F (71C). In actual practice, boiler differentials can be higher or lower than 15F (8C) depending on the installation.
Based on the manufacturer specifications for this aftermarket controller, it is only effective on boilers with aquastat
differentials less than 15F (8C).
Table 1. Proposed Boiler Control Test Matrix

DOE Reference Heating Load Test Condition Aquastat Aquastat Test
Building Setpoint Differential Duration
(oF/oC) (oFoC) (hr)
Steady State Low Heating Demand Baseline 175/79 5/3 0.5
Testing (120 MBH [35 kW]) Aftermarket Controller 175/79 5/3 0.5
Baseline with 15F (8C) Differential 175/79 15/8 0.5
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
High Heating Demand Baseline 175/79 5/3 0.5
(159 MBH [47 kW]) Aftermarket Controller 175/79 5/3 0.5
Baseline with 15F (8C) Differential 175/79 15/8 0.5
Secondary Average Winter Baseline 175/79 5/3 24
School, Chicago, Aftermarket Controller 175/79 5/3 24
Post-1980 Baseline with 15F (8C) Differential 175/79 15/8 24
Shoulder Season Baseline 175/79 5/3 24
Aftermarket Controller 175/79 5/3 24
Baseline with 15F (8C) Differential 175/79 15/8 24
Steady State Testing

A summary of the steady state test results are shown in Table 2. The aftermarket controller reduced the number of
boiler cycles by almost half, as compared to baseline tests. Fuel consumption was also reduced during both high and low
heating demand by 9.0% and 3.4%, respectively. In baseline tests, the boiler cycles approximately same amount under low
or high demand (10.6 vs. 11 cycles), but since the total amount of natural gas consumed is significantly less with a lower
heating demand, this reduces the opportunity for natural gas savings at lower heating demand.
During steady state tests, increasing the boiler aquastat differential from 5F (3C) to 15F (8C), i.e. changing the boiler
lower limit from 170F (77C) to 160F (71C), also reduced the number of cycles and natural gas consumption. During the 30
minute test, the aftermarket controller reduced the number of cycles from 11 to 6, while the 15F (8C) differential reduced
the cycles to 7. Similarly, increasing the differential to 15F reduced fuel consumption by 7.7% and 5.4% at high and low
demand, respectively, compared to the energy savings produced by the aftermarket controller of 9.0% and 3.4%.
$6+5$(

Table 2. Summary of 30-Minute Steady State Testing
Temp. Gas Usage Reductio No. of Secondary Temp.
Differential (cu. ft./ cu. m.) n in Gas Cycles Loop Decrease
(oF/oC) Usage Temp (oF/oC)
(oF/oC)
High Heating Demand (159 MBH [47 kW])
Baseline 5/3 56.1/1.6 11.0 167.5/75.3
Aftermarket Controller 5/3 51.1/1.5 9.0% 6.0 162.6/72.6 4.8/2.7
Baseline with 15F (8C) Differential 15/8 51.8/1.5 7.7% 7.0 162.8/72.7 4.6/2.6
Low Heating Demand (110 MBH [35 kW])
Baseline 5/3 26.5/0.8 10.5 169.8/76.6
Aftermarket Controller 5/3 24.6/0.7 3.4% 4.0 160.2/71.2 9.6/5.3
Baseline with 15F (8C) Differential 15/8 23.5/0.7 5.4% 5.0 162.2/72.3 7.6/4.2
Another factor to consider is the lower secondary loop temperatures resulting from both the aftermarket controller
and the increased aquastat differential. The aftermarket controller reduced the secondary loop average temperature by 4.8F
(2.7C) at high demand and 9.6F (5.3C) at low demand, compared to the baseline. Increasing the boiler aquastat differential
to 15F lowered the secondary loop temperature at low and high demand by 4.6F (2.6C) and 7.6F (4.2C), respectively.
Lower secondary loop temperatures may potentially affect building room temperatures or occupants comfort, but comfort
issues were not evaluated within this test scope.
24-hour Heating Profiles Boiler Heat Demand without Controls - Baseline

Average Winter Day, Chicago, Secondary School
To simulate actual operating conditions, the aftermarket
Thermostat controller
Setpoint: 175F was evaluated using a typical winter heating profile
with 5F Differential
based on DOE reference building models of a secondary school in a Chicago climate using TMY3 data. Figure 3 displays
200
the input heating profile in red. The actual boiler output during baseline testing is shown in blue as the boiler cycles on/off
190
180
to meet the demand. The boiler output with the aftermarket controller is show in Figure 4.
170
160
150
140
130
Heat Demand at Boiler (MBH)
120
110
100
90
80
70
60
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
50
40
30
20
10
0
0 60 120 180 240 300 360 420 480 540 600 660 720 780 840 900 960 1020 1080 1140 1200 1260 1320 1380 1440
Time (minutes)
Figure 3. Test boiler heat output compared to the input 24-hour heating profile for baseline boiler test without aftermarket controller.
234
ASHRAE Transactions
Boiler Heat Demand with M2G
Average Winter Day, Chicago, Secondary School
Thermostat Setpoint: 175F with 5F Differential
170
160
150
140
130
120
110
Heat Demand at Boiler (MBH)
100
90
80
70
60
50
40
30
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
20
10
0
0 60 120 180 240 300 360 420 480 540 600 660 720 780 840 900 960 1020 1080 1140 1200 1260 1320 1380 1440
Time (minutes)
Figure 4. Test boiler heat output with aftermarket controller compared to the input 24-hr heating profile
Table 3 contains a summary of 24-hr winter heating profile test results. Reductions in natural gas consumption
produced by the aftermarket controller ranged from 4.8% to 11.5% compared to the baseline case. As seen in the steady
state tests, the aftermarket controller produced the highest energy savings during occupied hours with higher heating load
than the unoccupied hours. In addition, the number of cycles was reduced by 57% compared to the baseline. Increasing
the aquastat differential from 5F (3C) to 15F (8C) also reduced the number of boiler cycles by 45% and gas consumption by
4.2%.
Table 3. Heating Profile Evaluation

Gas Usage Reduction in Number of Cycles Reduction in
(therms/MJ) Gas Usage Cycles
Series 1
Baseline 10.63/1,121 380
Aftermarket Controller 9.4/992 11.5% 164 56.8%
Series 2
Baseline 11.96/1,262 388
Aftermarket Controller 11.38/1,200 4.8% 184 52.6%
Baseline with Increased Differential 11.46/1,209 4.2% 212 45.4%
CONCLUSION
The laboratory boiler test stand developed for this study provides a useful and flexible tool to simulate different
buildings and climate zones in order to evaluate boiler controls and performance. This test method can be used to identify
building applications and climates where an aftermarket controller can provide the most energy savings. The researchers
will continue to evaluate these aftermarket control devices using further variations in simulated building type and climate
zone.
x Aftermarket controller can reduce energy consumption up to 11% for a 24 hr period and reduce boiler cycling by up to
57%.
x The largest energy savings by the aftermarket controller occurred during high demand, most likely due to the higher
energy use resulting in a greater opportunity for reducing gas consumption.
x In laboratory tests, the boiler aquastat differential had a significant impact on the energy use and cycling of the boiler.
$6+5$(

x The benefit of the aftermarket controller is very dependent on the existing boiler setpoints. It will not provide
significant savings for boilers with a differential of 15F (8C) or more (in agreement with manufacturer specifications).
This may explain some of the conflicting reports about the field performance of these units.
x Both the aftermarket controller and a larger aquastat differential decreased the test stand secondary loop temperatures
by as much as 10F (6C) below baseline testing. Field testing is needed to understand the effect of lower temperatures
on building comfort and condensation issues.
x Although these tests have demonstrated the potential for fuel savings by increasing the aquastat differential, i.e.,
widening the primary loop temperature band, there are potential risks associated increasing the variance that may
compromise the integrity of the boiler. Depending on the boiler setpoint temperature, increasing the variance of the
primary loop may increase the potential for condensation within the flue and increase the magnitude of thermal cycling
the heat exchanger.
REFERENCES
1 Energy Information Administration, 1995 Commercial Buildings Energy Consumption Survey.

2 Building Energy Data Book, Table 5.3.9 Major Commercial HVAC Equipment Lifetimes and Ages,
http://buildingsdatabook.eren.doe.gov/docs/xls_pdf/5.3.9.pdf.
3 Manufacturer’s website, Greffen Systems, M2G Product Specifications, http://greffensys.com/products.php, January
2012.
4 Cove, G., Hammer, J. and Butcher, T., A Technology Demonstration and Validation Project for Intellidyne Energy
Saving Controls, New York State Energy Research and Development Authority (NYSERDA).
5 Wilcox, S. and W. Marion. 2008. User's Manual for TMY3 Data Sets, NREL/TP-581-43156. April 2008. Golden,
Colorado: National Renewable Energy Laboratory.

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---

SA-12-C030
Using Steady State Predictions to Improve

the Transient Response of a Water to Air
Heat Exchanger
David A. Hodgson, PhD Peter M. Young, PhD Charles W. Anderson, PhD

ASHRAE Member
Douglas C. Hittle, PhD William S. Duff, PhD Daniel B. Olsen, PhD

ASHRAE Member
ABSTRACT

!

!

"

"

!

#
$

%
INTRODUCTION

!" #$ %&&&' $

( ('

(

) $

) *

David A. Hodgson is an assistant professor in the Department of Mechanical Engineering, Union College, Schenectady, New York. Peter M.
Young is a professor in the Department of Electrical and Computer Engineering, Colorado State University (CSU), Fort Collins, Colorado.
Charles W. Anderson is a professor in the Department of Computer Science, CSU. Douglas C. Hittle is an emeritus professor, William S. Duff
is a professor, and Daniel B. Olsen is an assistant professor in the Department of Mechanical Engineering, CSU.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

)

"

ERROR MODULATED STATE SETTING CONTROL

)

+ )

,
(
)

" )

) (

*
Background
- - .//%0 - .//%0 .//%' + -
%&&1' %&&2' )
-

$

"

-

)

)

)

)

(

(

THE CONTROL LAW

34 % .

5($ 5$ 5

$ '
'$ '

'$ 6 '
(

5'$ 67
)
4

)

0 $

'85( '96 ' %'
6:85 '95; '; 676' .'
5 "
(
5 "$
< 7

(

'
) 5
*

+
$ 5

$ 5
)

- =$
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---

>
$

238
ASHRAE Transactions

?

5@5:;67 6A;$ " 6A
" $ ;67 6A;

+$ 5

)

(
(

"

./%%'
THE EXPERIMENTAL SYSTEM

) *
$ !3(3) $
!3(3)

> B

% )

,
$

$

4
'$ $
$
$
4

$

!)-' )

C)DEF C+ %&&&' C

- .//%0 .//%0 .//20 ./%%'
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
& '( )

!

*

!

Air Side
? )

)
$6+5$(


)

) *
$

, $
$ *
$ *

Water Side

)
(C

?

)

)

)

)

, *
*

Control Configuration
) !3(3)

*

C>?'

)
+

$ $

$ ) $
+
$ )$

= .
+

) $ '$ $ ) $
$ ) )

)

$
+
(

$

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
"
& +( $

,-#-! * %

PRELIMINARY INVESTIGATION OF THE REPEAT HVAC SYSTEM

E

$
!3(3) )

$
240
ASHRAE Transactions
(
$
-

%/G H/G='

H/G %..G='
System Gain
= I

)

+

I/J HJ %HJ
2/J
KHJ &HJ ) /.1G /1I G='

) //.KG //H G='

) K2
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
& .( !

!

/

/

Nominal Controller
) (
) L M

?$ .//&'
I
+

, 5( 8..J:G %.J:G='$ 5 8//2HJ:G //1. J: G='
Steady State Model

3
4

" " $
"

$6+5$(

4
(

"
4

" )

$

$
$
)
%H%

)
$
$
$ " N%$%O
4
= 1 ) 4 1.J )

$ 67$

& 0( )

#

ERROR MODULATED STATE SETTING CONTROL APPLIED TO THE HVAC REPEAT SYSTEM

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
!3(3) - 5 -

=

5 ) 5

, /$ /%5$ /.5$ /15$ /&5' )

5
= H $
5
) + H

) 4

'

) % "
5 58/

(
5 !C> (

)

$
242
ASHRAE Transactions

)

5 !C>
+ 4

%$ 5 &HJ

!C>

& 1( )

,-#-! !

!

! $

& 232

Table 1. Results of Error Modulated State Setting Control Tests

( I& /K% G %HG =' /H& G %%G =' M:
/% IK /HP G %/G =' /IH G /PIG =' &&KJ
/. I. /1P G /KIG =' /IP G /PHG =' &&&J
/1 H% /PP G %.G =' /1I G /22G =' &%1J
/& 1/ /H& G %%G =' /I2 G /P2G =' &2HJ
CONCLUSION
3

!3(3)

)

(

0 4
=

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(


+$ $ 5$

NOMENCLATURE
5( 8 (
5 8
5 8 3

) 8 )
) 8 )
) 8 )

) 8 )

) 8 )

)8 )

8 =
8 =
)8 -

8
8
8
8
6 8 >

8 > Q
'
8
67 8 (

"
6A 8

"
REFERENCES
$ $ $ -$ 5"$ -$ !C 5
$ %&&2 >
$ +
(

$

3$ %% 1',1.%&
$ CD$ .//% CC?
$ $
' $ C> )$
> B$ = $ $ .//%
$ CD$ E$ C!$ R$ (C$ $ -$ $ $ )$ S$ - $ .//2

$ $
'
$ 3 E $
I& .',%IP%12
- $ $ .//% M M+ (

$ C> )$
> B$ = $ $ .//%
- $ $ $ -$ R$ (C$ $ $ CD $ .//% M + (

(
DC$ HK2%
$ >S$ D -$ %&&1 M

(
I 333

$ %K1&H1
$ -$ $ $ R$ (C$ - $ $ CD $ .//% (,

(' + MM'
$ T/%/IH
$ -$ ./%%

$
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
(U$ BC - (

) C+
$ %&&& C)DE HI$ M
+$ C

?$ 5$ .//& C 3$ (
$
!" $ $ R #$ %&&& -
E 3
$ ) (: C

(
$ >$ B 5
244
ASHRAE Transactions
SA-12-C031
A High-Efficiency, Reduced-Emissions,
Combustion Control System for
Commercial and Industrial Boilers
Guido Poncia, PhD Christoph Haugstetter David Liscinsky
Junqiang Fan, PhD Joseph Mantese, PhD Meredith Colket, PhD
Heidi Hollick Ricardo Garvey Yan Zhang, PhD
Kenneth Swanson Michael McCarron Duffy Parlett
ABSTRACT
A cost effective solution to the problem of improving boiler efficiency and reducing emissions by means of a novel combustion control system and a
sensor package is described, with applications for hot water or steam generation plants and capacity larger than 10 MMbtu/h. The system
continuously monitors flue gas emissions and reduces them by means of regulation of the burner’s inlet fuel valve and air damper. The new burner
management system incorporates a novel control algorithm, low cost sensors to monitor exhaust composition, and a user friendly tool for
visualization of boiler performance. The controller continuously maintains the optimum proportion of fuel and air feeding the burner in order to
reduce inefficiencies arising from excess air content while preventing unsafe operation arising from incomplete combustion. This new system is an
evolution of commercially available "O 2 trim" solutions and contrasts with legacy systems with preset mechanical linkage to fix the ratio of fuel to
air. The new control algorithm accounts for variation of environmental conditions, e.g., inlet air temperature and humidity, and system
degradation, e.g., linkage position drift. Use of this feedback control system is anticipated to enable increases in year averaged efficiencies
approximately 5%. Hence, significant annual fuel savings can be expected to defray costs related to the new system. The new system design and
data obtained during a demonstration of the system at the US Army's Watervliet Arsenal in Watervliet, NY will be presented.
INTRODUCTION
Combustion of fossil fuels is still by far the most utilized technology for generating hot water and steam in industrial
and commercial applications. Environmentally cleaner and renewable energy technologies lack the flexibility and availability
required for most near term applications. High efficiency, low emission combustion is therefore considered the most viable
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
approach to reduce fuel cost and mitigate undesired environmental effects. In the United States there are approximately
163,000 industrial and commercial boiler systems delivering steam for industrial processes, space heating and hot water.
Boilers with capacity larger than 10 MMbtu/h account for 28% of the total, and provide 85% of the overall US boiler
Guido Poncia is a Project Leader, David Liscinsky, Junqiang Fan, and Heidi Hollick are Staff Research Engineers, Yan Zhang is a Senior
Research Engineer, Ricardo Garvey is a Test Engineer, Joseph Mantese and Meredith Colket are Fellows, all at United Technologies
Research Center, East Hartford, CT.
Chris Haugstetter is a Staff Research Engineer at Hamilton Sundstrand, Windsor Locks, CT.
Ken Swanson is Director of Engineering, Mike McCarron, and Duffy Parlett are Staff Engineers, all at Fireye, Derry NH.
$6+5$(

capacity. 93% of all such systems are more than 10 years old (ONRL, 2005) and typically operate at efficiencies between 70-
80% (Harrold, 1999). Under pressures from rising fuel costs and increasingly stringent policies limiting the emissions of
polluting gases and overall carbon footprint, boiler owners are looking at cost effective ways to improve legacy systems.
Three possible paths to renovation are among the currently available ones: (1) replacement with new boilers (either
condensing boilers allowing efficiencies above 90% or non-condensing ones with improved heat exchanger, burner and
control system); (2) replacement of the burner for better air/fuel mixing and combustion or (3) replacement/adoption of
state of the art combustion control systems (Showers, 2009). While the first and second paths lead to the highest efficiency
gains, they are capital investment intensive with payback in multiple years (Durkin, 2006), and often require significant
infrastructural changes which further add to cost. In the case of condensing units, higher efficiency is gained at the price of
reduced fuel flexibility and potential durability issues associated with corrosion. An upgrade of the combustion control
system is a more cost effective solution (Eoff, 2008) often generating payback in less than one year (Wright et al., 2001) due
to lower first cost and significant recurring fuel savings associated with more efficient boiler operation.
The State of the Art (SoA) approach to upgrading the combustion control systems consists of substituting the
mechanical linkage between the air inlet damper and the fuel inlet valve with a digital controller acting on electromechanical
positioning servomechanisms (Eoff, 2008 and Alleman, 2010). The controller sets the opening of fuel and air inlets at all
working conditions (firing range) of the boiler as imposed by the installer during a commissioning phase. In addition to this
so-called parallel positioning controller, an O 2 trim function ensures that the oxygen concentration measured in the exhaust
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
gases is kept at a pre-set low value (generally 4%, depending on the burner installed), thus allowing efficient operation under
all boiler working conditions. While the current technology can ensure efficiencies around 80%, it has the following
shortcomings, which prevent reaching the highest possible efficiency gains through combustion control:
x Flue gas O 2 concentration cannot be further reduced because of safety concerns associated with incomplete
combustion. For this reason, efficiency is not increased further, limiting gains to ~5%.
x Continuous emission monitoring is unavailable, preventing minimum emission operation and real time verification
compliance with air permits.
x Commercially available O 2 sensor stack probes are expensive (~$10K sensor installed cost), thus decreasing the
economic attractiveness of the retrofit.
Hence there is a need for a safe, low cost, robust approach that can be easily retrofitted into legacy boiler systems,
with continuous optimization of air/fuel controls to attain maximum efficiency while monitoring and controlling operation
to meet local emission regulations. Satisfying this need would reduce fuel consumption and carbon footprint in older boiler
systems enabling them to be operated at the highest efficiencies possible through tight closed-loop control while
maintaining low CO and NO x emissions.
This paper describes how this need was addressed with the introduction of a new combustion control solution which
utilizes information from a continuous flue gas monitoring system made with low cost commercial grade sensors to regulate
the burner’s inlet fuel valve and air damper achieving maximum combustion efficiency. The paper is organized as follows:
First, an introduction to conventional boiler controls will be presented followed by the description of the novel technology
solution. A demonstration on a legacy, gas fired single burner 25 MMbtu/h boiler located at the Watervliet Arsenal central
steam plant will also be discussed, including the definition of standard performance metrics and a preliminary evaluation of
boiler performance. Final remarks will address the next steps leading to completion of the technology demonstration.
B OILER EFFICIENCY CONTROL PRACTICE AND CHALLENGES

Boiler fuel efficiency can be controlled by setting the correct proportion of fuel and air feeding the burner, and
depends on the unburned fuel, the inlet and outlet temperature of the gases, and the oxygen content of the exhaust (BSI,
1987). Boiler efficiency decreases as the air/fuel ratio increases; this change is accompanied by increase in the exhaust
oxygen concentration. In contrast, very low air/fuel ratios results in incomplete combustion and potentially unsafe
conditions manifested by a sharp increase in exhaust CO concentration. In legacy systems the fuel-air ratio is maintained by
246
ASHRAE Transactions
a mechanical linkage, while state of the art (SoA) solutions are based on electronic parallel servomechanism positioning and
O 2 trim technology. The unavailability of information on flue gas composition and relatively imprecise positioning of air
and fuel opening require linkage systems to be set to operate often with 8%-10% excess O 2 (DoE, 2006 and WSUEP,
2003) to guarantee an adequate safety margin (Eoff, 2008). Part load operation, variable environmental conditions, system
drift, and linkage hysteresis over time cause performance degradation towards either more inefficient operation or
potentially unsafe conditions (Figure 1). For this reason, legacy boilers have efficiencies of about 75%. In SoA solutions,
improved positioning and O 2 concentration measurement enable the reduction of safety margins to a typical value of 4%
and reduction of variability due to environmental factors and degradation, allowing efficiencies around 80%. The boiler and
burner characteristics affect the efficiency curve and the region of safe operation and consequently the achievable efficiency.
Desired
operating
region
Figure 1 – Trade off between safe operation and efficiency via modulation of fuel/air ratio. Information on stack CO enables detection of
best performance operation.
The ideal control solution that maximizes combustion efficiency without compromising boiler safety would need to
identify the fuel/air ratio corresponding to the lowest possible oxygen concentration where all available fuel is spent and
operate the boiler at that ratio, adapting to boiler demand changes and variations of the surrounding environment.
Proximity to that operating point can be achieved by sensing the concentration of carbon monoxide sharply increasing with
the onset of incomplete combustion. The proposed algorithm leverages therefore the feedback measurement of CO in
addition to O 2 and adjusts the position of the air servomechanism to maximize efficiency gains (a CO/O 2 trim control
approach). In turn, the control solution is able to detect unsafe operation via direct CO monitoring, while improving part
load performance, and adapts to degradation, changing conditions, and fuel properties.
THE CO/O 2 TRIM COMBUSTION CONTROL SYSTEM

Figure 2 shows a block diagram of the proposed control system. A new efficiency control algorithm is introduced to
replace the existing O 2 trim control on an existing platform of a commercially available boiler control system. The
efficiency algorithm communicates with air/fuel positioning controls to dispatch optimal settings for the air and fuel servo-
mechanisms actuating the air damper and natural gas supply valve. Information on the concentrations of O 2 , CO, and NO x
are provided to the controller by a continuous emissions measurement unit which leverages low cost gas sensor technology
utilized in the automotive industry. The system also includes a Graphical User Interface based on a commercially available
building management system, who communicates via MODBUS to the boiler controller. The GUI is able to report all
operating information about the boiler, including energy performance and continuous monitoring of polluting emissions to
the boiler operator. The solution enabled quick turnaround from concept to demonstration prototype, as existing hardware
components were fully leveraged to implement the new control functionality.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

Figure 2 – Block diagram of the proposed system, comprising of a new efficiency algorithm, low cost sensors for flue gas characterization,
and a graphical user interface to visualize efficiency gains.
The Controller
As anticipated, an existing control platform was used to implement the new CO/O 2 trim algorithm. This approach
maintained and leveraged the existing features that guarantee the boiler safe operation, including the flame safeguard
system, interlocks, and alarm management. In addition, the algorithm was engineered so that the system would fall back to
O 2 trim operation in case of failure of the CO measurement sensor of the control system. The algorithm control scheme is
illustrated in Figure 3. A positioning control selects the positions of air and fuel servos based on information on the boiler
desired firing rate. The O 2 trim control corrects the air position by ensuring that the O 2 concentration in the stack tracks a
desired reference value, also dependent on the firing rate and set at boiler commissioning. The proposed control algorithm
adds to the standard O 2 trim a correction module which uses the CO concentration measurement to prevent unsafe fuel
rich conditions from being reached. In this way, the boiler can be operated as close as possible to the point of maximum
efficiency. As the correction module is separate from the other control function modules, it can be disabled without
interrupting the operation of the standard O 2 trim algorithm.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 3 – Flow diagram of the proposed algorithm. The CO concentration measurement is used to correct the O 2 trim algorithm.
Operation of the new control algorithm was tested following a two-step approach. First, rapid prototyping was used
to verify the algorithm functionality. A virtual representation of the controller was implemented in LabView to drive the
actual air and fuel servomechanisms on an experimental boiler. A lab grade gas analyzer was used to collect information on
oxygen and carbon monoxide concentrations. Through rapid prototyping, quick changes and refinements of the algorithm
were possible, leading to the final version. The final algorithm was then coded for implementation on the control product
platform and tested to ensure correct and safe operation. Figure 4 illustrates the operation of the new control algorithm.
For a prescribed firing rate, the controller reduces the oxygen levels gradually until a rapid increase of the CO concentration
248
ASHRAE Transactions
is observed. At that point, O 2 levels are increased to avoid the accumulation of unspent fuel.
O2
CO
Figure 4 – Effects of air trim. Oxygen levels are adjusted to enable operation closer to the stoichiometric point. A spike in CO
concentration forces the system to step back to higher oxygen concentration levels.
The Gas Sensing Device

Selection of low cost but accurate, reliable and durable sensing devices to continuously measure the concentrations of
O 2 , NO x , and CO in the boiler exhaust is critical to enable reliable monitoring and feedback control of combustion for
optimal boiler efficiency and minimum environmental impact. In addition, packaging and integration of sensors into the
system must be carefully carried out to ensure measurement reliability in a harsh environment characterized by high
temperature and, in case of oil fired systems, of soot formation in the exhaust. Finally, sensor characteristics and placement
must be selected to ensure adequate response time for the controller to react promptly to changing settings and
environmental effects. While response times of 5 seconds are required for O 2 measurement systems to allow the controller
to promptly act, the sensor response time for NO x and CO could be extended more than 30s as CO-based corrections of
the O 2 reference value are typically performed within a timescale of minutes.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 5 – Linearity and accuracy of the CO sensor.
The adopted solution for continuous gas monitoring includes:

Gas sampling probe and analysis package. Effective gas extraction, conditioning, and feeding into the gas sensors
is obtained by using a commercially available gas probe enabling in-situ measurements of O 2 and NO x and extraction of
exhaust gas. The analysis package comprises of devices for reduction of water content, elimination of soot and sulfur which
could contaminate the CO sensor, and apparatus to maintain stable environmental conditions. Provisions are made to
$6+5$(

enable installation of the sensor analysis package in proximity of the stack to minimize gas transport time.
Gas sensors. Measurements of O 2 and NO x are obtained by means of automotive grade yttria-stabilized zirconia
sensors. This solution enables to sensibly reduce the cost of today’s commercially available in situ O 2 sensing systems for
boilers, while meeting requirements for precision, stability, and reliability. Extensive testing on boiler enabled to confirm the
viability of the approach. Electrochemical sensors for CO monitoring were selected because of their low cost and
robustness. Low cost also enabled to implement redundant CO measurements for improved reliability. The sensor has a
dynamic response of about 20 seconds, which was considered satisfactory for control operation.
Figure 5 illustrates the linearity and accuracy of the CO measurements obtained with the gas sensing device.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Deviations from known calibration points are less than 7% over the entire operating range.
THE DEMONSTRATION
The demonstration of the CO/O 2 trim algorithm is being conducted on a 30 year old 25MMbtu/h dual fuel steam
boiler which is part of the boiler plant of Watervliet Arsenal in Watervliet NY. The boiler is a water-wall unit which is used
during plant startup and in periods of peak demand. This auxiliary boiler is fitted with a new burner and uses linkage-based
positioning of fuel and air opening, which is typical of the boilers (in the <50MMBTU/h size range) of that age found at
US Army sites. The boiler is ideal for this type of demonstration as:
It is representative of typical installations in the US.
Since the boiler is used for startup and emergency, it can be taken offline for installation purposes without jeopardizing
the operation of the whole plant.
The boiler can be operated independently on actual demand enabling characterization over the entire operational range.
Figure 6 – Boiler instrumentation to enable performance characterization.
The demonstration has been conducted in three steps, i.e. (I) the installation of a new sensing and monitoring system
and characterization of baseline with legacy control, (II) the installation of the control platform and characterization of SoA
O 2 trim control, and (III) software update to CO/O 2 trim control and characterization. Accordingly, the boiler was
instrumented with metering devices to measure boiler properties such as airside inlet and outlet temperatures, water side
flow rate and inlet temperatures, steam flow, pressure and temperature to allow the precise quantification of boiler fuel to
steam efficiency. Direct measurement of sufficiently accurate air flow measurements turned out to be impractical at the
demonstration site and was replaced with an estimate based on fuel flow and exhaust O 2 measurements instead. Flue gas
composition was also measured including oxygen, NO x , and CO to monitor combustion characteristics as well as
emissions. Laboratory grade instruments or other suitable instrumentation was used for gas species measurements as part of
the boiler monitoring system. Boiler process variables i.e., water/steam pressure were also measured. Sensors were
connected via MODBUS to a data collection unit connected to a data acquisition unit also including the proposed graphical
250
ASHRAE Transactions
user interface. Figure 6 illustrates a schematic of the boiler system and location of all sensors.
Once the baseline characterization was completed, the legacy controller was removed and updated with the control
system inclusive of new servomechanisms and an O 2 measurement probe. Critical operating variables of the controller were
included in the data acquisition system via MODBUS. At the end of SoA control characterization tests, the software of the
controller was updated with the new algorithm and a new multi-sensing device was installed replacing the O 2 measurement
probe. Data collection during the three phases has been occurring by operating the boiler over a period of time of about
two months for each control configuration (baseline, O 2 trim, and CO/O 2 trim) by operating the boiler at 5 different
operating conditions for at least 1 hour, ranging from the “low fire” position to 100% capacity.
Performance Metrics
Comparison among the three control configurations was based on standard performance metrics to evaluate the
ability of a given control strategy to enable energy savings. Ultimately, a boiler operator is interested in continuously
achieving operating cost savings in face of the upfront investment in the new control technology. Savings come from lower
fuel utilization associated with increased boiler efficiency. Two efficiency metrics are defined:
Combustion Efficiency. A measurement of performance of the combustion process independent of other factors
contributing to the overall boiler efficiency (BS 845-2, 1987). Combustion efficiency is an index of the completeness of
combustion or the burner’s ability to burn fuel which is quantified by measurement of the release of usable thermal energy
to the boiler. Combustion efficiency is computed as:
(1)
where K 1g is a constant parameter dependent on the carbon content of the fuel, T net (=T flue -T in ) is the temperature
difference of the inlet air, T in and the outlet flue gas temperature, T flue , %O 2 is the oxygen content of the exhaust, K 3 is the
weight conversion of the fuel to gas, and K 2 is the theoretical CO 2 content. The second term in (1) represents the dry
losses; the third term represents the wet losses due to heat in the moisture in the stack exhaust.
Fuel-to-steam efficiency. The ratio of the useful energy available in the steam to the fuel energy required to produce
it (ASME PTC 4-2008). Although several factors contribute to fuel-to-steam efficiency (e.g. boiler and burner type and age,
and operating conditions), the focus has been on efficiency improvements associated with the update to new boiler control
technology. This metric was calculated from available instrumentation data, assuming steam saturation conditions, as:
(2)
where and are steam and fuel mass flows, p is the steam pressure, h s is the steam saturation enthalpy, H f the higher
heating value of fuel. Natural gas heating values used were obtained from the supplier as daily averages for the region,
confirmed through sampling of as delivered gas on several different days. Fuel-to-steam efficiency equals the combustion
efficiency minus heat losses through radiation and convection.
PRELIMINARY RESULTS
A series of experiments were conducted on the demonstration boiler at Watervliet Arsenal to assess performance
during operation at different load levels (25%, 50%, 75% and full load). Measurements were collected by averaging data
samples during steady state operation. In addition, 24 hour operation at a fixed firing rate was collected to measure
variability of data induced by varying exogenous conditions (e.g. inlet air temperature, humidity, and fuel heating value). At
least three samples were collected for a given load condition. Overall, acceptable repeatability of test results was established.
Figure 7 reports preliminary comparisons between baseline and O 2 trim control. It can be observed that combustion
efficiency gains of 2-3% are observed consistently across the firing range. Boiler operation with the O 2 trim control results
in more complete combustion and reduction of stack losses. Combustion efficiency gains result in overall improvement of
fuel to steam efficiency of 2-4% with higher gains at higher firing rate operating points. Differences associated with firing
rate are likely to be related to the change in effectiveness of the water tube heat exchanger rather than phenomena involving
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

the combustion process. Indeed, the function of the O 2 trim controller is to improve combustion efficiency independent of
the boiler configuration. Its installation on a legacy boiler results in fuel to steam efficiency gains and associated fuel
savings, but the magnitude of those gains will depend on the boiler configuration, as well as its operation.
Figure 7 – Performance metrics computed for baseline and O2 trim control
CONCLUSIONS AND NEXT STEPS

A combustion efficiency control system update is a rapid and inexpensive way to improve a boiler’s combustion
process with positive impact on efficiency gains and ultimately fuel savings and reduced carbon emissions. A new
combustion control solution utilizing feedback information of O 2 and CO concentrations in the flue gases via a low cost,
robust sensing apparatus has been presented in this paper, together with a systematic approach for quantification of boiler
performance. New control solutions and quantification methods are being demonstrated on a legacy steam boiler at
Watervliet Arsenal, enabling accurate computation of fundamental performance metrics. Preliminary results show that the
upgrade to a state of the art O 2 trim control enables combustion efficiency gains of 2-3% across the operating range of the
boiler and fuel to steam efficiencies of 2 to 4% dependent on boiler operation. The performance of the next generation
O 2 /CO control solution is expected to provide additional efficiency gains. Its performance relative to baseline and O 2 trim
control will be quantified during the first quarter of 2012.
ACKNOWLEDGEMENTS
The authors would like to acknowledge the financial support of the Environmental Security Technology Certification
Program (ESTCP) of the U.S. Department of Defense and the support of all staff at the Watervliet Arsenal central boiler
plant with installation, test, and demonstration activities.
REFERENCES
Alleman, J. (2010). Opportunities in Industrial Boiler Efficiency. Northwest Food Processors Association, Jun 2010.
ASME PTC 4-2008 (2009). Fired Steam Generators – Performance Test Codes. American Society of Mechanical Engineers.
BS 845-2. (1987). Methods for assessing thermal performance of boilers for steam, hot water and high temperature heat
transfer fluids: a comprehensive procedure. British Standards Institution.
Durkin, T.H. (2006). Boiler System Efficiency. ASHRAE Journal, Jul 2006, p. 51-57
Eoff, D. (2008). Understanding Fuel Savings In the Boiler Room . ASHRAE Journal, Dec 2008, p. 38-43
Harrold, D. (1999, November 1). Incomplete Combustion Burns Money. Control Engineering.
ORNL. (2005). Characterization of the US Industrial Commercial Boiler Population. Energy and Environmental Analysis.
Showers, G. (2009). Industrial-Boiler Replacement. HPAC Engineering, Jan 2009.
Wright, A., Hart, F., Russell, C., and Jaber, D.. (2001). Best Practices Steam Resources and Tools: “Old” News is “New”
News!. In Steam Digest.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,
252
ASHRAE Transactions
SA-12-C032
+\EULG6RODU7KHUPDODQG*URXQG6RXUFH
+HDW3XPS6\VWHP
)DU]LQ05DG3(QJ$ODQ6)XQJ3(QJ3K':H\+/HRQJ3(QJ3K'
$6+5$(0HPEHU$6+5$(0HPEHU1RWDQ$6+5$(0HPEHU
$%675$&7
7KLVGRFXPHQWGHPRQVWUDWHVWKDWK\EULGJURXQGVRXUFHKHDWSXPSV\VWHP*+63FRPELQHGZLWKVRODUWKHUPDOFROOHFWRUVLVDIHDVLEOHFKRLFH
IRUVSDFHFRQGLWLRQLQJIRUKHDWLQJGRPLQDWHGKRXVHV,WZDVVKRZQWKDWWKHVRODUWKHUPDOHQHUJ\VWRUDJHLQWKHJURXQGFRXOGUHGXFHDODUJHDPRXQW
RIJURXQGORRSKHDWH[FKDQJHU*+;OHQJWK&RPELQLQJWKUHHVRODUWKHUPDOFROOHFWRUVZLWKWRWDODUHDRIP>IW@WRWKH*6+3V\VWHP
ZLOOUHGXFH*+;OHQJWKE\IURPP>IW@WRP>IW@7KHV\VWHPPDOIXQFWLRQLQJLQWKHFRROLQJVHDVRQZDVDOVRGHWHFWHGDQG
RSWLRQVIRUIL[LQJWKHSUREOHPZHUHSUHVHQWHG6HQVLWLYLW\DQDO\VLVZDVFDUULHGRXWIRUGLIIHUHQWFLWLHVRI&DQDGDDQGUHVXOWVZHUHGHPRQVWUDWHGWKDW
9DQFRXYHUZLWKPLOGHVWFOLPDWHFRPSDUHGWRRWKHUFLWLHVZDVWKHEHVWFDQGLGDWHIRUWKHSURSRVHGVRODUK\EULG*6+3V\VWHPZLWKPP
>IWIW@*+;OHQJWKUHGXFWLRQWRVRODUFROOHFWRUDUHDUDWLR2YHUDOOV\VWHPHFRQRPLFYLDELOLW\ZDVDOVRHYDOXDWHGXVLQJD\HDUOLIHF\FOHFRVW
DQDO\VLV7KHDQDO\VLVVKRZHGWKDWWKHUHLVVPDOOHFRQRPLFEHQHILWLQFRPSDULQJWRFRQYHQWLRQDO*6+37KHQHWSUHVHQWYDOXHRIWKHSURSRVHG
K\EULGV\VWHPDQGFRQYHQWLRQDO*6+3V\VWHPZHUHHVWLPDWHGWREHDQGUHVSHFWLYHO\
,QDGGLWLRQDVHQVLWLYLW\DQDO\VLVKDVEHHQFDUULHGRXWWRGHWHUPLQHWKHUHODWLRQVKLSEHWZHHQWKHVRODUFROOHFWRUDUHDDQGWKHJURXQGORRSKHDW
H[FKDQJHU*+; OHQJWK ,W ZDV VKRZQ WKDW WKH UDWLR RI *+; OHQJWKUHGXFWLRQ WRVRODU SDQHO DUHD RI PP >IWIW@ UHQGHULQJ WKH
RSWLPXPUDWLRZKLFKFRUUHVSRQGHGWRP>IW@*+;OHQJWKUHGXFWLRQZLWKP>IW@VRODUFROOHFWRUVDUHD
,1752'8&7,21
*HRWKHUPDODSSOLFDWLRQVIRUEXLOGLQJVDUHPRVWO\OLPLWHGWRIXOOGHSHQGHQFHRQJURXQGVRLOWHPSHUDWXUHIRURI
WKH KHDWLQJ DQG FRROLQJ HQHUJ\ $OWKRXJK WKHUH DUH WKH DGYDQWDJHV RI ORZ HQHUJ\ DQG PDLQWHQDQFH FRVWV LQ IDYRU RI WKLV
DSSURDFKVSDFHOLPLWDWLRQVDQGKLJKLQLWLDOFRVWVPD\UHVWULFWDIXOOJHRWKHUPDOLQVWDOODWLRQ5HVWULFWLYHUHJXODWLRQVVXFKDV
PDQGDWLQJDPLQLPXPERUHKROHVL]HJURXWLQJPDWHULDOVZDJHUDWHVDQGKHDWH[FKDQJHPHWKRGJHQHUDOO\LQFUHDVHWKHFRVW
RIVXFKV\VWHP7KHLQLWLDOFRVWPD\SXWWKHSURMHFWDERYHWKHEXGJHWDQGLQVRPHFDVHVWKHGULOOLQJFRQGLWLRQVPD\SUHYHQW
WKHXVHRIDODUJHFRQYHQWLRQDOFORVHGORRSERUHKROHILHOG
,QPDQ\EXLOGLQJVDQQXDOO\WKHDPRXQWVRIKHDWH[WUDFWHGIURPDQGLQMHFWHGLQWRWKHJURXQGDUHQRWEDODQFHG7KH
YHUWLFDOFORVHGORRSLVDFRPPRQW\SHRIHDUWKFRXSOLQJPRVWO\XVHGLQEXLOGLQJVWKDWKDYHOLPLWHGODQGDUHDV,QGHVLJQLQJ
WKLVW\SHRIV\VWHPFRQVLGHUDWLRQPXVWEHPDGHRIWKHWKHUPDOUHVSRQVHRIWKHJURXQGWKURXJKRXWWKHH[SHFWHGSURMHFWOLIH
LH \HDUV $Q DQQXDO LPEDODQFH LQ JURXQG ORDG ZLOO OHDG WR ORZHU KHDW SXPS HQWHULQJ IOXLG WHPSHUDWXUHV ()7 LQ
KHDWLQJGRPLQDWHG EXLOGLQJV RU KLJKHU KHDW SXPS ()7V LQ FRROLQJGRPLQDWHG EXLOGLQJV WR D SRLQW ZKHUH HTXLSPHQW
FDSDFLW\PD\EHFRPSURPLVHGLIWKHJURXQGORRSKHDWH[FKDQJHULVQRWORQJHQRXJK7KLVLPEDODQFHUHTXLUHVHLWKHUDYHU\
ODUJH JURXQG ORRSKHDW H[FKDQJHU RU VRPHPHFKDQLVP IRU DVVLVWLQJ WKH V\VWHPE\ VXSSOHPHQWLQJ GHILFLW KHDW RU UHMHFWLQJ
)DU]LQ05DGLVDUHVHDUFKIHOORZDW&OHDQ(QHUJ\'HYHORSPHQWV0LVVLVVDXJD2QWDULR$ODQ6)XQJDQG:H\+/HRQJDUHDVVRFLDWH
SURIHVVRUVLQWKH'HSDUWPHQWRI0HFKDQLFDODQG,QGXVWULDO(QJLQHHULQJ5\HUVRQ8QLYHUVLW\7RURQWR2QWDULR
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

H[FHVVKHDW%HFDXVHWKHFRVWRILQVWDOOLQJDYHU\ODUJHJURXQGORRSKHDWH[FKDQJHUPD\EHH[FHVVLYHWKHUHDUHDQXPEHURI
DOWHUQDWH ZD\V WR DVVLVW D *6+3 7KHVH LQFOXGH VRODU FROOHFWRU ZKLFK LQMHFWV DGGLWLRQDO KHDW LQWR WKH JURXQG IRU KHDWLQJ
GRPLQDWHG EXLOGLQJV DQG FRROLQJ WRZHU ZKLFK UHMHFWV H[FHVV KHDW LQWR WKH DWPRVSKHUH IRU FRROLQJ GRPLQDWHG EXLOGLQJV
6\VWHPVWKDWLQFRUSRUDWHERWKDJURXQGKHDWH[FKDQJHUDQGDQDERYHJURXQGKHDWH[FKDQJHUDUHFRPPRQO\UHIHUUHGWRDV
K\EULG*6+3,QK\EULGV\VWHPVWKHSHDNKHDWSXPS()7IURP\HDUWR\HDUVKRXOGEHDSSUR[LPDWHO\HTXDO,QWKLVVWXG\
WKHV\VWHPXWLOL]HVDVRODUWKHUPDOFROOHFWRUDVDQDERYHJURXQGKHDWH[FKDQJHUFDOOHGDVRODUDVVLVWHGJURXQGVRXUFHKHDW
SXPS6$*6+3
7KHSXUSRVHRIWKLVVWXG\ZDVWRHYDOXDWHWKHSHUIRUPDQFHDQGYLDELOLW\RIK\EULGJHRWKHUPDOKHDWSXPSV\VWHPVZLWK
VRODUWKHUPDOFROOHFWRUVIRUUHVLGHQWLDOKRXVHDSSOLFDWLRQVLQ&DQDGD7KHPDLQREMHFWLYHZDVWRSHUIRUPDV\VWHPVLPXODWLRQ
DSSURDFKWRDVVHVVWKHIHDVLELOLW\RIWKLVNLQGRIK\EULGV\VWHPLQKHDWLQJGRPLQDWHGEXLOGLQJV$QDFWXDOUHVLGHQWLDOEXLOGLQJ
ZDVPRGHOHGDQGWKHUHVXOWVFRPSDUHGWRDFWXDOGDWDFROOHFWHGE\PRQLWRULQJWKHUHODWHGRSHUDWLRQRIHTXLSPHQWVWKURXJK
VSHFLILF PRQWKV ,W ZRXOG EH LGHDO LI WKLV VWXG\ DWWUDFWV WKH LQWHUHVW RI UHVHDUFKHUV DQG FRQWUDFWRUV DQG SURYLGHV YDOXDEOH
LQIRUPDWLRQIRUGHVLJQLQJDQGLQVWDOOLQJWKLVNLQGRIK\EULGV\VWHPLQKHDWLQJGRPLQDWHGEXLOGLQJVLQ&DQDGD
+286(02'(/'(6&5,37,21
7KHKRXVHVHOHFWHGIRUWKHSURSRVHGVWXG\LVORFDWHGLQWKH&LW\RI0LOWRQ2QWDULR7KHKRXVHZDVRQHRIWZRHQHUJ\
HIILFLHQWGHPRQVWUDWLRQKRXVHVEXLOWE\DORFDOEXLOGHULQ,WLVDGHWDFKHGWZRVWRUH\EXLOGLQJKDYLQJIWP
RIKHDWHGDUHDLQFOXGLQJWKHEDVHPHQW
$VSHUEXLOGHUVSHFLILFDWLRQVWKHKRXVHWHPSHUDWXUHLVVHWDW&>)@DQG&>)@LQWKHKHDWLQJDQGFRROLQJ
SHULRGVUHVSHFWLYHO\$LUOHDNDJHDW3D>LQ+J@LV$&+OV$FRQWLQXRXVYHQWLODWLRQRI$&+OV
WKURXJKKHDWUHFRYHU\YHQWLODWLRQV\VWHP+59LVDOVRFRQVLGHUHG7KHVHQVLEOHLQWHUQDOKHDWJDLQIURPRFFXSDQWVLVVHWWR
EHN:KGD\>N%WXGD\@7KHRFFXSDQF\RIWKHKRXVHLVWZRDGXOWVDQGWZRFKLOGUHQIRURIWKHWLPHZLWKDKRW
ZDWHU FRQVXPSWLRQ RI OLWHUVGD\ 7KH EDVH ORDGV DUH FRQVLGHUHG WR EH N:KGD\ LQFOXGLQJ LQWHULRU DQG H[WHULRU
OLJKWLQJDSSOLDQFHVXVHDQGRWKHUV751%XLOGDFRPSRQHQWRIWKH7516<6VLPXODWLRQVRIWZDUHZDVXVHGWRJHQHUDWHWKH
KRXVHORDGSURILOH
7KHFOLPDWHRI7RURQWR2QWDULRZKLFKLVDERXWNP>PL@HDVWRI0LOWRQZDVFKRVHQIRUWKLVVWXG\
6$*6+302'(/'(6&57,37,21
7KHJURXQGORRSKHDWH[FKDQJHU*+;V\VWHPFRQVLVWVRIIRXUYHUWLFDOFORVHGORRSFLUFXLWVMRLQHGLQSDUDOOHO(DFK
ERUHKROH KDV P > LQ@ GLDPHWHU DQG P >IW@ OHQJWK 7KH\ DUH ORFDWHG P >IW@ DSDUW IURP HDFK RWKHU LQ WKH
EDFN\DUGDQGPHUJHLQDP>IW@EHORZJUDGHDUHD)LJXUHVKRZVWKLVDUUDQJHPHQW
)LJXUHVKRZVDVFKHPDWLFRIWKHVRODUDVVLVWHGJURXQGVRXUFHKHDWSXPS6$*6+3V\VWHP
7KH*+;ORRSLVFRQQHFWHGLQSDUDOOHOWRWKHVRODUWKHUPDOFROOHFWRUV7KHVRODUFROOHFWRUVUHFHLYHDSHUFHQWDJHRIWKH
WRWDOIORZIURPWKHJURXQGORRSH[FKDQJHU7ZRFLUFXODWLRQSXPSVIRUPSDUWRIWKHKHDWSXPSV\VWHPDQGWKH\DUHORFDWHG
XSVWUHDP DQG GRZQVWUHDP RI WKH *+; IORZ $ VROHQRLG YDOYH DQG D FRQWURO YDOYH FRQWURO WKH IORZ UDWH WR WKH VRODU
FROOHFWRUV
7KH 7516<6 PRGHOLQJ HQYLURQPHQW VWXGLR ZDV XVHG WR FRQVWUXFW WKH V\VWHP XVLQJ VWDQGDUG DQG QRQVWDQGDUG
FRPSRQHQWPRGHOV7KHFRPSRQHQWPRGHOVXVHGDUHDVEHORZ
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
+HDW3XPS0RGHO
*URXQG/RRS+HDW([FKDQJHU0RGHO
6RODU&ROOHFWRU0RGHO
:DWHU7DQN0RGHO
,Q)ORRU5DGLDQW+HDWLQJ0RGHO
254
ASHRAE Transactions
*UD\:DWHU+HDW5HFRYHU\0RGHO
9HQWLODWLRQ0RGHO
)ORZ&RQWURODQG3XPS&RPSRQHQW0RGHO
+HDW3XPS(TXLSPHQW&RQWURODQG6FKHGXOLQJ

)LJXUH*URXQGORRSERUHKROHOD\RXW
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
)LJXUH6FKHPDWLFGLDJUDPRI6$*6+3V\VWHPFRQILJXUDWLRQ
$6+5$(

',6&866,21$1'5(68/76
2SWLPXP)ORZWR6RODU7KHUPDO&ROOHFWRUV
,QWKHDVEXLOWV\VWHPWKHSHUFHQWDJHRIWKHJURXQGORRSIORZUDWHGLYHUWWRWKHVRODUFROOHFWRUVZDVVWXGLHGXVLQJWKUHH
GLIIHUHQWIORZTXDQWLWLHV)URPWKHHTXLSPHQWPDQXIDFWXUHUWKHPD[LPXPDFFHSWDEOHIORZUDWHVWRWKHKHDWSXPSDQGWKH
WKUHH VRODUSDQHOV ZHUH GHWHUPLQHGWREH NJKU >JSP@ DQG NJKU >JSP@ UHVSHFWLYHO\7KLVPHDQV WKDWWKH
WRWDOPD[LPXPIORZUDWHLQWKHJURXQGORRSH[FKDQJHUZDVWKHVXPRIWKHWZRLHNJKU>JSP@7KHPD[LPXP
IORZGLYHUVLRQIURPWKHJURXQGORRSZDVGHWHUPLQHGWREHRIWKHWRWDOIORZ7KLVVWXG\VKRZHGWKDWE\LQFUHDVLQJWKH
IOXLGIORZWRWKHVRODUSDQHOVIURP]HURWRRIWKHKHDWSXPSIORZUDWHWKHRYHUDOOV\VWHPHQHUJ\FRQVXPSWLRQLQWKH
KHDWLQJPRGHGHFUHDVHVZKHUHDVLQWKHFRROLQJPRGHWKHV\VWHPHQHUJ\FRQVXPSWLRQLQFUHDVHV7KLVWUHQGLVLQIDYRURI
WKHV\VWHPSHUIRUPDQFHLQWKHKHDWLQJVHDVRQDQGDJDLQVWLWLQWKHFRROLQJVHDVRQ7KLVLQGLFDWHVWKDWIRU7RURQWRZHDWKHU
FRQGLWLRQV D UHVLGHQWLDO KRXVH ZLWK D K\EULG *6+3 V\VWHP FDQ EHQHILW IURP KLJKHU IORZ GLYHUVLRQ WR VRODU WKHUPDO
FROOHFWRUVLQWKHKHDWLQJVHDVRQKRZHYHUWKHUHYHUVHLVWUXHLQWKHFRROLQJVHDVRQ
6HQVLWLYLW\$QDO\VLVRIWKH*URXQG7KHUPDO&RQGXFWLYLW\
7KHUPDOFRQGXFWLYLW\LVRQHRIWKHSK\VLFDOSURSHUWLHVQHHGHGIRUVL]LQJ*6+3$VWKHH[DFWVRLOWKHUPDOFRQGXFWLYLW\
IRU WKH VLWHRI WKHKRXVHZDV XQNQRZQ WKHKRXVHZDV PRGHOHGZLWK GLIIHUHQW VRLOWKHUPDOFRQGXFWLYLWLHVDQGWKHHIIHFWV
UHFRUGHG &KDQJLQJ WKH VRLO WKHUPDO FRQGXFWLYLW\ OHDGV WR D FKDQJH LQ WKH HQWHULQJ IOXLG WHPSHUDWXUH ()7 WR WKH KHDW
SXPS7KLVLVDYHU\LPSRUWDQWSDUDPHWHUIRUWKHHIILFLHQF\RIWKHKHDWSXPS7KHKHDWSXPSRIWKHKRXVHLVGHVLJQHGWR
ZRUN EHWZHHQ &>)@ DQG &>)@ $QDO\VLV ZDV FDUULHG RXW ZLWK IRXU GLIIHUHQW VRLO W\SHV LQ WKH UDQJH RI
:P.>%WXKU)IW@WR:P.>%WXKU)IW@7KHUHVXOWVKRZHGWKDWLQWKHKHDWLQJVHDVRQWKHORZHUWKHUPDO
FRQGXFWLYLW\ OHDGV WR D ORZHU ()7 WR WKH KHDW SXPS 6RLO W\SHV ZLWK D WKHUPDO FRQGXFWLYLW\ ORZHU WKDQ
:P.>%WXKU)IW@OHDGWRDQ()7ORZHUWKDQ&DQGFDXVHDPDOIXQFWLRQRIWKHKHDWSXPS2QWKHRWKHUKDQGD
KLJKHU VRLO WKHUPDO FRQGXFWLYLW\ OHDGV WR KLJKHU ()7 LQ KHDWLQJ VHDVRQ DQG PRUH HIILFLHQW KHDW SXPS RSHUDWLRQ ,Q
FRQFOXVLRQEDVHGRQRYHUDOOUHJLRQDOVRLOW\SHVRLOZLWK:P.>%WXKU)IW@LVDUHDVRQDEOHVHOHFWLRQDQGWKLVZDV
WKHYDOXHXVHGLQWKHUHVWRIWKHVWXGLHV+DYLQJEHWWHUKHDWWUDQVIHULQWKHYLFLQLW\RIWKHJURXQGKHDWH[FKDQJHUORRSLVLGHDO
IRUWKHV\VWHPZKHUHDVKDYLQJORZHUWKHUPDOFRQGXFWLYLW\LQWKHEDFNILOOYROXPHZRXOGEHJRRGIRURYHUDOOWKHUPDOVWRUDJH
WKHUHODWLRQVKLSEHWZHHQVRODUFROOHFWRUDUHDDQGJURXQGORRSOHQJWKZDVRQHRIWKHLPSRUWDQWDVSHFWVRIWKLVVWXG\
6RODU&ROOHFWRU$UHDDQG*URXQG/RRS+HDW([FKDQJHU/HQJWK5HODWLRQ
,Q WKH KHDWLQJ VHDVRQ WKH UHVXOWV RI WKLV VWXG\ VKRZHG WKDW IRU WKLV VSHFLILF KRXVH ORFDWHG LQ 7RURQWR UHJLRQ WKUHH
VRODUSDQHOVLQWKHV\VWHPKHOSHGWRUHGXFHWKHWRWDOJURXQGORRSKHDWH[FKDQJHU*/+(OHQJWKE\FRPSDUHGWRWKH
V\VWHPWKDWRQO\KDGDFRQYHQWLRQDOKHDWSXPS,QFUHDVLQJWKHQXPEHURIVRODUSDQHOVIURPWKUHHWRVL[GLGQRWGRXEOHWKH
*/+(OHQJWKUHGXFWLRQZKHUHDVLWVWUHQGZRXOGEHLQWKHUDQJHRIWRDIWHUWKDW$QRSWLPXPUHODWLRQVKLSRIWKUHH
VRODU SDQHOV ZLWK WKH UHGXFHG */+( OHQJWK RI )LJXUH VKRZV SHUFHQWDJH JURXQG ORRS OHQJWK UHGXFWLRQ YHUVXV
QXPEHURIVRODUSDQHOV,QWKHFRROLQJVHDVRQDGGLQJVRODUSDQHOVWRWKHV\VWHPZRXOGKDYHDQHJDWLYHLPSDFWRQWKHV\VWHP
SHUIRUPDQFH DQG DQ LQFUHDVH LQ WKH KHDW SXPS HQHUJ\ FRQVXPSWLRQ 7KHUHIRUH LQ KHDWLQJ GRPLQDWHG SODFHV ZKHUH WKH
FRROLQJVHDVRQLVVKRUWWKLVLQFUHPHQWDOHQHUJ\FRQVXPSWLRQIRUVSDFHFRROLQJZRXOGQRWEHVLJQLILFDQWFRPSDUHGWRWKH
SRWHQWLDO VDYLQJV LQ WKH KHDWLQJ VHDVRQ )LJXUH VKRZ HQWHULQJ IOXLG WHPSHUDWXUHV ()7V WR WKH KHDW SXPS LQ D W\SLFDO
KHDWLQJDQGFRROLQJVHDVRQUHVSHFWLYHO\
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
6\VWHP&RVW$QDO\VLV
$\HDUOLIHF\FOHDQDO\VLVRIWKHV\VWHPVKRZHGRQO\VPDOOHFRQRPLFEHQHILWIRUWKHK\EULGV\VWHPFRPSDUHGWRWKH
V\VWHP ZLWK RQO\ D *6+3 7KLV ZDV GXH WR WKH ORZ ERUHKROH GULOOLQJ FRVW RI P>IW@ $W WKH WLPH RI VWXG\ WKH
256
ASHRAE Transactions
ERUHKROH GULOOLQJ FRVWV ZHUH HVWLPDWHG WR EH LQ WKH UDQJH RI P>IW@ WR P>IW@ IRU GLIIHUHQW JURXQG
FRQGLWLRQV+RZHYHUIRUWKHFDVHRIKLJKHUGULOOLQJFRVWVWKHHFRQRPLFEHQHILWVZRXOGEHFRQVLGHUDEOHEHFDXVHRIWKH
UHGXFWLRQRI*/+(OHQJWKGXHWRWKHWKUHHVRODUFROOHFWRUV
)LJXUH*URXQGORRSOHQJWKUHGXFWLRQSHUFHQWDJHYHUVXVQXPEHURIVRODUSDQHOV
)LJXUH$QQXDO()7WRWKH+3IRU6$*6+3DQG*6+3V\VWHPV
7DEOHVKRZVWKH1HW3UHVHQW9DOXH139RIK\EULGVRODUJURXQGVRXUFHKHDWSXPSV\VWHPFRQVLGHULQJERUHKROH

FRVW RI P>IW@ VRODUFROOHFWRUVFRVW RI P >IW@HOHFWULFLW\FRVW RI N:KDQG LQWHUHVW UDWHRI
'XULQJWKHV\VWHPOLIHF\FOHRI\HDUVWKH6$*6+3V\VWHPHQHUJ\FRQVXPSWLRQLQFUHDVHVVOLJKWO\\HDURYHU\HDU7KLV
HIIHFWFRUUHVSRQGVWRWKHUHGXFWLRQRID&>)@LQWKHPLQLPXP()7WRWKH+3IURPWKHILUVWWRWK\HDU,QWKHFDVH
RI*6+3V\VWHPZLWKRXWVRODUFROOHFWRUVWKHUHZLOOEHKLJKHUHQHUJ\FRQVXPSWLRQGXHWRDQHDU&>)@UHGXFWLRQRI
PLQLPXP()7WRWKHKHDWSXPSIURPWKHILUVWWRWK\HDU7KLVHIIHFWZDVQRWFRQVLGHUHGLQWKLVVWXG\DVLWZDVEH\RQG
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

WKHVFRSHRIWKLVZRUN
7DEOH1HWSUHVHQWYDOXHRIK\EULGVRODUJURXQGVRXUFHKHDWSXPSV\VWHP

6RODU&ROOHFWRU */+( &RVW$QDO\VLV
1R $UHD 7RWDO/HQJWK ,QLWLDO&RVW 2SHUDWLRQ&RVW
RI
3DQ $QQXDO 3UHVHQW 1HW3UHVHQW
HOV P>IW@ P>IW@ 6RODU&RVW */+(DQG+3 &RVW 9DOXH 9DOXH

>@ >@
>@ >@
>@ >@
>@ >@
>@ >@
>@ >@

2SWLPXPEDODQFHEHWZHHQWKH*/+(VL]HDQGVRODUFROOHFWRUVL]H
)LHOG6WXG\DQG9HULILFDWLRQ
)RU WKLV VWXG\ WKHUH ZDV OLPLWHG ILHOG GDWD DYDLODEOH WR YDOLGDWH WKH VLPXODWLRQ UHVXOWV )RU KHDW SXPS HQHUJ\
FRQVXPSWLRQ WKHUH ZHUH RQO\ GD\V RI GDWD LQ WKH KHDWLQJ VHDVRQ DQG QR FRQVLGHUDEOH GDWD LQ WKH FRROLQJ VHDVRQ
+RZHYHUWKHUHZDVDOPRVWHLJKWPRQWKVRIGDWDDYDLODEOHIRUWKH()7WRWKHKHDWSXPSLQ7KHFRPSDULVRQRIWKH
VLPXODWLRQUHVXOWVZLWKWKHILHOGGDWDVKRZHGDWRGHYLDWLRQLQHQHUJ\FRQVXPSWLRQV7KHVRXUFHRIWKLVGHYLDWLRQ
ZDV SDUWO\ GXH WR WKH ZHDWKHU GDWD 70< XVHG LQ WKH VLPXODWLRQ %\ DGMXVWLQJ WKH VLPXODWLRQ UHVXOWV ZLWK WKH DFWXDO
ZHDWKHUGDWDGHULYHGGHJUHHGD\VIRUIRU7RURQWRWKLVGHYLDWLRQZDVUHGXFHGWRWR
6ROXWLRQVWRWKH3UREOHPRIWKH$VEXLOW6\VWHP
7KH H[LVWLQJ V\VWHP KDG SUREOHPV IXQFWLRQLQJ SURSHUO\ LQ WKH FRROLQJ VHDVRQ 7KH DQDO\VLV UHVXOWV VKRZHG WKDW WKH
V\VWHP ZDV QRW SURSHUO\ VL]HG IRU WKH FRROLQJ VHDVRQ DV WKH ()7V WR WKH KHDW SXPS ZHUH H[FHHGLQJ WKH DOORZDEOH ()7
GHILQHG E\ WKH PDQXIDFWXUHU 7KLV KDSSHQHG IURP -XQH WR $XJXVW DOPRVW WKH HQWLUH FRROLQJ VHDVRQ 6LPXODWHG VROXWLRQV
LQFOXGHDVWRSSLQJWKHIORZWRWKHVRODUSDQHOLQFRROLQJVHDVRQEVHOHFWLQJDKHDWSXPSZLWKDPRGLILHGFRROLQJFDSDFLW\
DQGVSHFLILFDWLRQDQGFLQFUHDVLQJWKH*/+(OHQJWK$OOWKUHHVROXWLRQVDUHDSSOLFDEOHZLWKFDVHDEHLQJWKHVLPSOHVWDQG
FRVW OHDVW WR LPSOHPHQW 7KLV SUREOHP FRXOG KDYHEHHQ SUHYHQWDEOH LI WKH ERUHKROH OHQJWKVZHUH LQFUHDVHG E\ 7KLV
VROXWLRQFRXOGEHMXVWLILHGDVWKHV\VWHPZRXOGSHUIRUPEHWWHULQWKHKHDWLQJVHDVRQLQVSLWHRIH[WUDERUHKROHFRVW)LJXUH
VKRZVWKHH[LVWLQJKHDWSXPS()7OLPLWVLQV\VWHPZLWKWKUHHVRODUSDQHOV,WVKRZVWKDWLQDOPRVWHQWLUHFRROLQJVHDVRQ
WKH()7VZLOOIDOODERYHWKHKHDWSXPSFDSDFLW\OLPLW&>)@DQGV\VWHPVWRSVIXQFWLRQLQJ
6\VWHP9LDELOLW\LQ'LIIHUHQW&LWLHVRI&DQDGD
%\FRQVLGHULQJWKHVDPHKRXVHFKDUDFWHULVWLFVWKHHIIHFWVRIGLIIHUHQWFOLPDWHVLQ&DQDGDZHUHLQYHVWLJDWHG)RUWKLV
SXUSRVH VL[ &DQDGLDQ FLWLHV LQ GLIIHUHQW UHJLRQV ZHUH VWXGLHG 7KH UHVXOWV DUH WDEXODWHG LQ 7DEOH ,W LV VKRZQ WKDW LQ
GLIIHUHQWFLWLHVLQJHQHUDODVWKHUDWLRRIDQQXDOJURXQGKHDWLQJORDGWRDQQXDOJURXQGFRROLQJORDGRIWKHKRXVHGHFUHDVHV
WKHUHGXFHG*/+(OHQJWKWRVRODUFROOHFWRUDUHDUDWLRGHFUHDVHV7KLVUDWLRZDVPP>IWIW@IRU9DQFRXYHUZLWK
DQDQQXDOKHDWLQJORDGWRDQQXDOFRROLQJORDGUDWLRRI)RU(GPRQWRQZLWKDQQXDOKHDWLQJORDGWRFRROLQJORDGUDWLRRI
UHVXOWHG LQ UHGXFHG */+( OHQJWK WR VRODU FROOHFWRU DUHD UDWLR RI PP >IWIW@ $ KLJKHU UHGXFWLRQ UDWLR
9DQFRXYHU LQGLFDWHV EHWWHU YLDELOLW\ RI WKH K\EULG *6+3 V\VWHP 7KLV ZRXOG QRW EH DQ DEVROXWH FRQFOXVLRQ DV RWKHU
SDUDPHWHUVVXFKDVUDWLRRIKHDWLQJGHJUHHGD\VWRFRROLQJGHJUHHGD\VLQHDFKFLW\DOVRDIIHFWVWKHFRQFOXVLRQ
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
258
ASHRAE Transactions
&21&/86,21
,QWKLVVWXG\RYHUDOOV\VWHPYLDELOLW\ZDVHYDOXDWHGDQGH[LVWLQJV\VWHPSUREOHPVZHUHGHWHFWHGWKURXJKWKHGHGLFDWHG
PRGHOLQJDQGVLPXODWLRQRIWKHLQVWDOOHGVRODUDVVLVWHG*6+3V\VWHP
9LDELOLW\RI6\VWHP
7KHUHVXOWRIWKLVVWXG\KDYHVKRZQWKDWWKHK\EULG*6+3V\VWHPFRPELQHGZLWKVRODUWKHUPDOFROOHFWRUVFRXOGEHD
IHDVLEOHFKRLFHIRUVSDFHFRQGLWLRQLQJIRUKHDWLQJGRPLQDWHGKRXVHV)RUWKHKRXVHLQWKLVVWXG\WKHVHDVRQDOVRODUWKHUPDO
HQHUJ\VWRUDJHLQWKHJURXQGLQWKHK\EULGV\VWHPZDVVXIILFLHQWWRRIIVHWODUJHDPRXQWRI*/+(OHQJWKWKDWZRXOGKDYH
EHHQUHTXLUHGLQFRQYHQWLRQDO*6+3V\VWHPV7KHHFRQRPLFEHQHILWRIVXFKV\VWHPGHSHQGVRQFOLPDWHDVZHOODVERUHKROH
GULOOLQJFRVW
6\VWHP6LPXODWLRQ$SSURDFK
7KLV VWXG\ GHPRQVWUDWHV WKH YDOXH LQ XWLOL]LQJ D V\VWHP VLPXODWLRQ DSSURDFK WR HYDOXDWLQJ DOWHUQDWLYHV LQ FRPSOH[
V\VWHPV7KHKRXUO\WLPHVWHSVLPXODWLRQIRUWKHLPSOHPHQWDWLRQRIFRPSOH[FRQWURODQGRSHUDWLRQVWUDWHJLHVHQDEOHGWKH
DVVHVVPHQW RI WUDQVLHQW V\VWHP UHVSRQVHV 7KLV VWXG\ ZLOO EH IXUWKHU HQKDQFHG E\ H[DPLQLQJ DQG DQDO\]LQJ GLIIHUHQW
FRQILJXUDWLRQ DQG FRQWURO VWUDWHJLHV WKH LQWHUDFWLRQ RI GLIIHUHQW FRPSRQHQWV DQG SRWHQWLDO EHQHILWV LQ EURDGHU
JHRJUDSKLFDODUHDV
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
)LJXUH([LVWLQJKHDWSXPS()7OLPLWVV\VWHPZLWKWKUHHVRODUSDQHOV
$6+5$(

7DEOH6\VWHPSHUIRUPDQFHLQGLIIHUHQWFLWLHV
&LW\ 6RODU&ROOHFWRU *URXQG/RRS+HDW([FKDQJHU $QQXDO6\VWHP(QHUJ\ 5HGXFHG

$UHD /HQJWK %KROHV 0 LQ()7 6SDFH+HDWLQJ &RROLQJ /HQJWK&RO

RI3DQHO P >IW @ P>IW@ OHQJWKP>IW@ &>)@ +HDWLQJ0 ->N%WX@ &RROLQJ0->N%WX@ PP >IWIW @
9DQFRXYHU
>@ >@ >@ >@ >@
>@ >@ >@ >@ >@ >@ >@

7RURQWR
>@ >@ >@ >@ >@
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
>@ >@ >@ >@ >@ >@ >@
0RQWUHDO
>@ >@ >@ >@ >@
>@ >@ >@ >@ >@ >@ >@

2WWDZD
>@ >@ >@ >@ >@
>@ >@ >@ >@ >@ >@ >@

+DOLID[
>@ >@ >@ >@ >@
>@ >@ >@ >@ >@ >@ >@

(GPRQWRQ
>@ >@ >@ >@ >@
>@ >@ >@ >@ >@ >@ >@

5()(5(1&(6
.DYDQDXJK63´$QDO\VLVDQG'HYHORSPHQWRID'HVLJQ0HWKRGIRU+\EULG*HRWKHUPDO+HDW3XPSVµ'UDIW8QLYHUVLW\
RI$ODEDPD7XVFDORRVD0DUFK
6LQJK-%´$GYDQWDJHRI8VLQJWKH+\EULG*HRWKHUPDO2SWLRQµ/HWWHUWRFRQWUDFWRU-XQH
%RVH-(6PLWK0'DQG6SLWOHU-'´$GYDQFHVLQ*URXQG6RXUFH+HDW3XPS6\VWHP$Q,QWHUQDWLRQDO2YHUYLHZµ
3URFHHGLQJVRIWKHVHYHQWK,QWHUQDWLRQDO(QHUJ\$JHQF\+HDW3XPS&RQIHUHQFH%HLMLQJ0D\
751EXLOGDQG0XOWL]RQH%XLOGLQJPRGHOLQJ7516<6PDQXDO9ROXPH
(VVHQWLDO,QQRYDWLRQ7HFKQRORJ\&RUS(,*HR([FKDQJHV\VWHP6XUUH\%&&DQDGDKWWSZZZHLWHFKFRUSFRP
7516<60DQXDO7(66OLEUDU\9HU
$6+5$( 6WDQGDUG ´0HWKRGV RI WHVWLQJ WR GHWHUPLQH WKH SHUIRUPDQFH RI VRODU FROOHFWRUVµ $6+5$( $WODQWD

7RURQWR+\GUR(OHFWULF5HVLGHQWLDOUDWHVKWWSZZZWRURQWRK\GURFRPHOHFWULFV\VWHPUHVLGHQWLDOUDWHVLQGH[FIP
%DQNRI&DQDGD$YHUDJHORDQUDWHVKWWSZZZEDQNRIFDQDGDFD
5DG)DU]LQ´9LDELOLW\RIK\EULGJURXQGVRXUFHKHDWSXPSV\VWHPZLWKVRODUWKHUPDOFROOHFWRUVµ5\HUVRQ8QLYHUVLW\0DVWHU
WKHVLV7RURQWR-XO\
5DG))XQJ$6/HRQJ:´&RPELQHG6RODU7KHUPDODQG*URXQG6RXUFH+HDW3XPS6\VWHPµWK,%36$&RQIHUHQFH
*ODVJRZ8.-XO\
:DQJ(Q\X )XQJ$ODQ6 4L&KHQJ\LQJDQG/HRQJ:H\+3HUIRUPDQFHSUHGLFWLRQRIDK\EULGVRODUJURXQGVRXUFH
KHDWSXPSV\VWHP(QHUJ\DQG%XLOGLQJV,Q3UHVVGRLMHQEXLOG
260
ASHRAE Transactions
SA-12-C033
Zero Refrigeration for Data Centres in the

USA
Robert Tozer PhD CEng Sophia Flucker CEng

MASHRAE
ABSTRACT
Data centre energy consumption has grown significantly in recent years; cooling energy forms a substantial part of this and presents a significant
opportunity for efficiency improvements. With best practice air management in place, it is possible to reduce airflow and increase supply air
temperatures. This can be achieved through separation of hot and cold air streams by cold aisle, hot aisle or rack exhaust containment systems.
Once server air inlet temperatures are within a narrow range, close to the supply temperature from cooling units, it is possible to increase
temperature set points with the confidence that IT equipment receives air at an acceptable temperature. This allows energy savings to be realised
through more efficient refrigeration cycle operation and increased opportunities to benefit from free cooling; in many cases it may be possible to
remove refrigeration altogether.
There are several different methods for free cooling using economizer cycles in data centres, which can be used instead of / in conjunction with
traditional refrigeration for full or partial free cooling. These include direct and indirect air side and water side free cooling systems. The typical
overall approach (difference between ambient wet bulb condition and data hall supply air temperature) is given for each of the free cooling methods,
together with the projected maximum design ambient temperatures in different locations and the resulting maximum data hall supply air
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
temperatures.
The results indicate the extent to which zero refrigeration is possible for each type of free cooling system throughout the USA, when supplying air
into the data hall within the ASHRAE recommended and allowable ranges. Zero refrigeration solutions result in a significant saving on capital
cost due to the reduction in mechanical plant and associated supporting electrical infrastructure i.e. switchboards, generators, power distribution etc.
In many US climates 100% free cooling (zero refrigeration) is possible and significant operational and capital cost savings can be realised.
INTRODUCTION
In recent years, the industry has started to focus on how to operate more efficient data centres. The main drivers are:
• Regulatory requirements
• Market pressures around Corporate Social Responsibility (CSR)
• Increasing energy costs
Data centre energy consumption has increased with the growing demand for IT services. Consumers and
governments are pressurising companies to reduce their environmental impact, which has, along with energy price
increases, driven data centre operators to review how efficiently they run their data centre facilities. US data centre energy
use was predicted to double to more than 120 billion kWh from 2006 to 2011 (US EPA Energy Star 2007). The
Robert Tozer is a Visiting Fellow at London South Bank University and Managing Director at Operational Intelligence Ltd. Sophia Flucker is a
Director at Operational Intelligence Ltd.
$6+5$(

Department of Energy called upon companies to improve data centre energy efficiency by 10% in 2 years and 30% in 5
years (US DoE 2009). ‘Green’ issues are subject to more media attention and sustainability achievements often feature in
companies’ annual reports and marketing material. Many operators market themselves using their Power Usage
Effectiveness (PUE) performance and LEED certification.
Operators face cost pressures, particularly during a time of economic uncertainty. Limited budgets may result in new
build projects being placed on hold and require the continuing operation of legacy facilities. Rising energy costs also
increase operational spend, for example, a 1MW IT load facility, with total load of 2MW and utility rate of $0.10 / kWh has
an annual energy bill of $1,752,000. If the rate increases to $0.14 / kWh, this represents an additional annual cost of
$700,800.
The growth in the market has led to a skills shortage which presents a risk to the reliability of data centre operations.
This can be managed through staff training and experience and also design simplification.
DESIGN PARADIGM CHANGE
In recent years important improvements to data centre designs have led to improved energy efficiency with typical
PUE = 2+ (Data Centre Infrastructure Efficiency, DCiE = 0.5 or lower) for legacy facilities being replaced with new
designs with PUE = 1.2 or lower now possible (DCiE = 0.83 or higher). Most of the energy reduction has been in the
cooling systems, through:
• Best Practice Air Management: minimising bypass and recirculation air in the data hall by physically separating hot
and cold air streams (through containment) and controlling on cooling unit supply air temperature to deliver a
narrow range of temperatures at IT equipment inlet.
• Reduced cooling unit fan speeds / air volumes through the use of variable frequency drives.
• Warmer / wider environmental conditions: increasing operating temperatures to deliver air at the server inlet in
line with IT hardware vendor specifications; 18 – 27°C (64.4°F – 80.6°F) recommended and higher allowable,
reducing refrigeration needs and increasing free cooling opportunity.
• Use of free cooling (economizer) designs
By minimising the cooling system energy consumption, data centre energy efficiency is increased and capital,
operational and reliability costs, i.e. total cost of ownership, minimised.
Best Practice Air Management
The objective of putting best practice air management in place is to allow airflow to be reduced and temperatures to
be increased. The concept of flooding the data hall with cold air to prevent overheating and ensure reliability proves
ineffective as the supplied air is not directed to the heat loads and the return air not separated, resulting in local areas of
high temperatures supplied to IT equipment. These cooling problems are often misunderstood; Air Performance Metrics
allow operators to quantify the effectiveness of cooling in the data centre (Tozer et al 2009). Often air is oversupplied,
resulting in wastage of fan energy and bypassing of excess air; ideally the air volume should be reduced to just what is
sufficient for cooling and slight pressurisation of the cold air stream, where contained. By using variable speed drives the
air volume delivered can be varied in line with demand, e.g. as additional / higher density IT equipment is installed /
workload increases. Using variable speed fans produces particularly significant energy savings at part load; many data halls
operate at part load for a significant proportion (if not all) of their life, with redundant plant increasing the installed cooling
capacity.
When server air inlet temperatures are controlled such that they are close to the raised floor supply temperature,
temperature set points can be increased with the confidence that IT equipment receives air within the appropriate range.
Higher temperatures result in energy savings through more efficient refrigeration cycle operation and increased
opportunities to operate with free cooling and may permit refrigeration to be removed altogether (Flucker, Tozer 2011).
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
262
ASHRAE Transactions
Once hot spots have been discovered, where IT equipment receives higher temperature air, a typical solution is to
decrease the cooling unit set point. This does not deal with the root cause of the problem and contributes to additional
energy wastage; it makes the overcooling issue worse.
Hot and cold air streams can be separated by using containment which minimises bypass and recirculation and the
range of temperatures supplied to the IT equipment. There are various containment types:
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---

Figure 1 – Cold aisle containment, Semi-containment, Hot aisle containment, Direct rack containment
Cold aisle containment is commonly used, where the cold aisle is closed with a roof and doors and the rest of the data
hall is the same temperature as the hot aisle. Semi-containment is a variation on this, where curtains are fitted above the
cold aisle, blocking the air path from the hot aisle. This works well as a retrofit option, particularly where there are cabinets
of different heights. Hot aisle containment involves ducting the air from the hot aisle back to the cooling units, usually by
way of a ceiling plenum, with the rest of the data hall is the same temperature as the cold aisle. This may only be possible in
a new build as it requires coordination with other overhead services, such as cable trunking. Direct rack containment
employs a chimney at the back of the cabinets (deeper than normal) to duct hot air back to the cooling units. This method
keeps the hot air outside of the room and may become more widely adopted as hot aisle temperatures increase due to
increasing IT equipment exhaust temperatures and air supply temperatures. The impact on fire systems should be
considered, i.e. placement of detectors / suppression heads, flame retardant properties of materials used.
Controlling on cooling unit return air temperature, even with the measures above in place can still result in a range of
temperatures delivered to the IT equipment. This is due to the non-uniform nature of load distribution in most data halls;
each cooling unit will deliver a delta T proportional to its load and thus supply a different temperature (Tozer 2009).
Changing the cooling unit temperature control strategy to supply air control allows this range to be minimised.
These best practices and others can be found in the EU Code of Conduct for Data Centres Best Practice Guidelines
(European Commission 2011).
Environmental Conditions
ASHRAE Technical Committee 9.9 (Mission Critical Facilities, Technology Spaces and Electronic Equipment) has
defined the range of acceptable conditions for IT equipment in conjunction with the hardware manufacturers. The
recommended envelope has been expanded (ASHRAE 2008) and further allowable ranges defined (ASHRAE 2011); the
ranges for class 1- 4 equipment during operation are summarised in the table below and indicated on the psychrometric
charts in Figure 2.
Table 1 – Summary table of environmental conditions at IT equipment inlet published by ASHRAE
! !

%,& %,& %,& %,&
Recommended
A1 to A4 18 to 27 64.4 to 80.6 5.5°C (41.9°F) DP to 60% RH 15 59
Allowable
A1 15 to 32 59 to 89.6 20-80% RH 17 62.6
A2 10 to 35 50 to 95 20-80% RH 21 69.8
A3 5 to 40 41 to 104 -12°C (10.4°F) DP & 8% RH to 85% RH 24 75.2
A4 5 to 45 41 to 113 -12°C (10.4°F) DP & 8% RH to 90% RH 24 75.2
$6+5$(

These increased temperatures allow the possibility for data centres to benefit from free cooling (or economized
cooling) throughout the world.
The Miami and San Francisco climate envelopes are shown shaded in Figure 2, with the inner zones indicating the
occurrence of more than 90% of annual hours. This data suggests that ambient air in San Francisco does not exceed the
allowable maximum temperature for class A2 and rarely for class A1 and hence free cooling should be exploited.

Figure 2 - Miami and San Francisco climate envelopes
Free Cooling
There are several different methods for free cooling in data centres, which can be used in the place of / in
complement to traditional refrigeration for full or partial free cooling:
Direct air free cooling Ambient air is filtered, humidified and refrigerated where necessary and brought into the
data hall. The cooling system (and electrical plant) needs to be sized for the worst case maximum refrigeration load.
Indirect air free cooling Ambient air and data hall internal air is simultaneously and separately passed through a
heat exchanger for heat transfer. Adiabatic humidification on the external air stream allows evaporative cooling at hot dry
ambient conditions. Any refrigeration capacity required is minimal (supplementary only), allowing reduced sizing of
mechanical and electrical plant and therefore capital cost saving.
Quasi-indirect air free cooling (thermal wheel) Similar to indirect air free cooling except constant volume of fresh
air brought in to data hall which requires treatment (more than required for pressurisation).
Water side free cooling Heat is rejected to ambient without the use of a refrigeration cycle, for example using
cooling towers or dry coolers. This is often an easier retrofit option compared with air side free cooling designs as this
method usually requires modification to external plant only and is typically less demanding in terms of plant footprint.
Figure 3 – Direct air side free cooling, Indirect air side free cooling, Quasi-indirect air side free cooling, Water side free cooling
Adiabatic or evaporative cooling can be used effectively to reduce energy consumption as part of these solutions to reduce
refrigeration requirements, especially to treat hot dry ambient air; it requires significantly less energy than refrigeration but
will increase water consumption. Adiabatic cooling is governed by wet bulb rather than dry bulb ambient temperature and
in some climates may allow thousands of additional hours of free cooling.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
264
ASHRAE Transactions
Figure 4 - Twb / Tdb diagrams for Phoenix AZ & San Francisco CA
ENERGY EFFICIENCY
Based on the weather data for different locations, the total number of refrigeration hours can be modelled, and hence
the total energy consumption of the whole data centre facility for a given IT load. The most widely used and recognised
metric for data centre efficiency is PUE (The Green Grid 2007). Guidelines have been published to clarify how to collect
and report data (Data Centre Metrics Coordinated Taskforce 2011). Very low PUEs can be achieved by minimising the
energy used for cooling by using the techniques described in this paper. An analysis was made to understand the
correlation between PUE and a range of psychrometric parameters, including dry bulb temperature, wet bulb temperature,
enthalpy, dew point temperature; averages, maximums, standard deviations. Average wet bulb temperature was found to
exhibit the best correlation; the results are shown below for different cooling systems, assuming an electrical PUEe of 1.14
(DCiEe = 0.88).
Figure 5 – Modelled PUE with direct and indirect air side free cooling systems versus average wet bulb temperature
Figure 6 - Modelled PUE with indirect and semi-indirect air side and water side free cooling systems versus average wet bulb temperature
ZERO REFRIGERATION
In many climates, 100% free cooling or zero refrigeration is possible. This reduces not only the energy cost and the
operating cost associated with maintenance but also the capital cost by around 30%, as removing refrigeration plant reduces
the maximum load allowing the electrical infrastructure to be reduced.
The typical approach (difference between ambient wet bulb condition and data hall supply air temperature) is given
for each of the free cooling methods in the table below, along with the statistical maximum temperatures based on design
ambient conditions in different locations and the resulting maximum data hall supply air temperatures. These maximums
are derived from historical data and do not take into account heat island effects of cities or global warming; they are also
lower than the extreme conditions, which would provide a very conservative design specification, given the limited decrease
in reliability caused by short periods operating at higher temperatures (ASHRAE 2011) and the trend for increased
$6+5$(
--`,`,``,,`,`,```,

environmental condition tolerances of IT hardware.
Figure 7 – Maximum temperature derivations
Supply air temperatures exceeding the recommended and allowable ranges are indicated in blue and red respectively.
Table 2 - Free cooling temperatures
! %,&
" ! ! "
),%+#(,& "+, '(,%('#*,&
%'(#*,&
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
%,& %,& %,& %,& %,& %,& %,& %,&
Acapulco 28.4 83.2 32.4 90.4 35.4 95.8 40.4 104.8
Billings 19.8 67.7 23.8 74.9 26.8 80.3 31.8 89.3
Charlotte 25.6 78.1 29.6 85.3 32.6 90.7 37.6 99.7
Chicago 26.3 79.4 30.3 86.6 33.3 92.0 38.3 101.0
Denver 18.6 65.4 22.6 72.6 25.6 78.0 30.6 87.0
Detroit 25.8 78.4 29.8 85.6 32.8 91.0 37.8 100.0
Jacksonville 27.1 80.8 31.1 88.0 34.1 93.4 39.1 102.4
Las Vegas 23.3 73.9 27.3 81.1 30.3 86.5 35.3 95.5
Los Angeles 21.7 71.1 25.7 78.3 28.7 83.7 33.7 92.7
Memphis 27.3 81.1 31.3 88.3 34.3 93.7 39.3 102.7
Mexico City 17.0 62.7 21.0 69.9 24.0 75.3 29.0 84.3
Miami 27.2 81.0 31.2 88.2 34.2 93.6 39.2 102.6
Minneapolis 25.9 78.6 29.9 85.8 32.9 91.2 37.9 100.2
Montreal 24.1 75.4 28.1 82.6 31.1 88.0 36.1 97.0
New Orleans 27.4 81.3 31.4 88.5 34.4 93.9 39.4 102.9
New York 25.6 78.1 29.6 85.3 32.6 90.7 37.6 99.7
Phoenix 24.9 76.9 28.9 84.1 31.9 89.5 36.9 98.5
Portland 21.5 70.7 25.5 77.9 28.5 83.3 33.5 92.3
San Antonio 26.0 78.8 30.0 86.0 33.0 91.4 38.0 100.4
San Francisco 19.3 66.7 23.3 73.9 26.3 79.3 31.3 88.3
Seattle 20.1 68.1 24.1 75.3 27.1 80.7 32.1 89.7
Tulsa 26.7 80.1 30.7 87.3 33.7 92.7 38.7 101.7
Winnipeg 23.8 74.8 27.8 82.0 30.8 87.4 35.8 96.4
266
ASHRAE Transactions

Figure 8 - Requirements for refrigeration using indirect air side free cooling (4K / 7.2°F approach), water side free cooling (7K / 12.6°F
approach), water side free cooling (12K / 21.6°F approach). Recommended range achieved shown with round green marker, allowable range
round blue marker with dot, refrigeration required shown with red pin
The results indicate that using indirect air side free cooling, temperatures can be supplied to the data hall within the
ASHRAE A1 allowable range for all of the USA locations analysed and for many of the western cities, temperatures are
within the recommended range; zero refrigeration is possible throughout the USA with this method.
With a high efficiency water side free cooling system (with larger heat exchange area and footprint), in some USA
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
locations zero refrigeration is possible whilst remaining in the recommended range, however in warmer climates supply
temperatures increase into the allowable range.
With a legacy water side free cooling system, refrigeration will be required in warmer climates to keep supply
temperatures within the allowable range, although not in some cooler USA climates.
THE FUTURE
The trend for increasing energy prices and pressure to reduce costs and improve environmental performance should
continue to drive operators towards exploiting energy saving opportunities in data centre cooling systems. Data centre
densities and temperatures look likely to continue to rise. This opens up opportunities to increase zero refrigeration which
is already available in many locations. It also allows simpler cooling designs to be used which are more reliable and help
operators to reduce data centre operating and capital costs.
$6+5$(

Mexico Guadalajara 20.5 68.8
APPENDIX
Mexico Tijuana 24.1 75.3
Table 3 – Statistically derived maximum wet bulb Mexico Monterrey 26.9 80.5
temperatures Mexico Merida 28.8 83.8
$! ! Mexico Mazatlan 28.1 82.6
°C °F Mexico Veracruz 28.4 83.1
Alabama Birmingham 26.2 79.2 Mexico Tampico 29.3 84.8
Alaska Anchorage 16.4 61.5 Mississippi Jackson 27.2 80.9
Arizona Tucson 23.0 73.3 Missouri Kansas City 27.1 80.7
Arkansas Little Rock 27.3 81.2 Nebraska Omaha 25.9 78.6
California Los Angeles 21.7 71.1 New Hampshire Manchester 25.0 77.1
Canada Calgary 18.3 64.9 New Jersey Newark 26.1 79.0
Canada Medicine Hat 19.8 67.6 New Mexico Albuquerque 19.0 66.1
Canada Vancouver 19.7 67.4 North Dakota Fargo 25.0 77.0
Canada Toronto 24.4 76.0 Ohio Cleveland 25.4 77.7
Colorado Colorado Springs 17.9 64.2 Pennsylvania Philadelphia 26.3 79.3
Cuba Havana 28.1 82.6 Rhode Island Providence 25.5 77.8
Delaware Wilmington 26.2 79.1 South Carolina Columbia 26.2 79.2
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Dist. of Columbia Washington 26.4 79.5 South Dakota Sioux Falls 26.1 78.9
Georgia Atlanta 25.6 78.1 Tennessee Nashville 26.0 78.8
Hawaii Honolulu 25.5 78.0 Texas Austin 26.7 80.0
Idaho Boise 19.8 67.6 Texas Houston 27.0 80.7
Illinois Chicago 26.3 79.4 Utah Salt Lake City 20.2 68.3
Indiana Indianapolis 26.4 79.5 Vermont Burlington 24.3 75.7
Iowa Des Moines 26.4 79.6 Virginia Richmond 26.4 79.5
Kansas Wichita 25.7 78.3 West Virginia Charleston 25.6 78.1
Kentucky Louisville 26.6 79.9 Wisconsin Milwaukee 25.9 78.7
Maine Portland 24.2 75.6 Wyoming Cheyenne 17.8 64.0
Maryland Baltimore 26.2 79.1
Massachusetts Boston 25.3 77.6
REFERENCES
ASHRAE, Environmental Guidelines for Datacom Equipment -Expanding the Recommended Environmental Envelope,
2008.
ASHRAE, Thermal Guidelines for Data Processing Environments – Expanded Data Center Classes and Usage Guidance,
2011.
Data Centre Metrics Coordinated Taskforce, Recommendations for Measuring and Reporting Overall Data Center
Efficiency, 2011.
European Commission, Best Practices for the EU Code of Conduct on Data Centres, version 3, 2011.
Flucker, Tozer, Data Centre Cooling Air Performance Metrics, CIBSE Technical Symposium, 2011
The Green Grid, Green Grid Data Center Power Efficiency Metrics: PUE and DCiE, 2007.
Tozer, R, Salim, M, Kurkjian, C, Air Management Metrics in Data Centres, ASHRAE Transactions, vol. 115, part 1,
Chicago meeting, CH-09-009, 2009.
Tozer, Global Data Centre Energy Strategy, DatacenterDynamics, 2009.
US Department of Energy, U.S. Data Centers Save Energy Now, 2009.
US Environmental Protection Agency Energy Star Program, Report to Congress on Server and Data Center Energy
Efficiency Public Law 109-431, 2007.
268
ASHRAE Transactions
SA-12-C034
The Challenge of Low Energy Design in

Low Energy Cost States
Stephen W. Duda, PE
Member ASHRAE
ABSTRACT
ASHRAE is strongly emphasizing the need for sustainable or low-energy design practice through programs, papers,
standards, the Advanced Energy Design Guide series, and the Net-Zero Energy Building initiative. The author of this paper
does not dispute the urgent need for energy-conscious design. However, low-energy mechanical equipment and systems are
sometimes more costly to purchase and install than their more traditional counterparts. In many areas of the Midwestern
United States, low utility rate structures are available, in some cases as low as $0.04 per kilowatt-hour for electricity. These
utility rates tempt building owners and decision-makers to forego more expensive but lower-energy HVAC options due to
lack of an attractive return on investment. This paper explores the conundrum experienced by responsible engineers in
attempting low-energy design where very low utility costs drive financial decisions in the opposite direction. Included are
examples from ASHRAE’s “50% Advanced Energy Design Guide for Small to Medium Office Buildings.”
INTRODUCTION
A review of the Technical Program listing at any recent ASHRAE Annual or Winter Conference will show an
abundance of sessions dedicated to design and engineering of low-energy consuming buildings and systems, with some
sessions even touting net-zero energy buildings. As a regular attendee at these conferences, this author has eagerly listened
and then sought opportunities to apply those sustainable design concepts in practice. However, in portions of the Midwestern
United States, very low electric utility rates tempt building owners and decision-makers to forego more expensive but lower-
energy HVAC options due to lack of an attractive return on investment. As far back as 2006, this author has lamented the
difficulty of generating interest in green or sustainable design in some regions of the country, if cost-benefit or payback
periods are an important consideration [ASHRAE 2006]. Responsible and creative engineers continue to implement low-
energy design where possible, but this paper highlights challenges posed by very low utility costs driving financial decisions
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
in the opposite direction.
Utility Rate Survey

SSPC 90.1 or Standing Standards Project Committee 90.1 (i.e., the committee responsible for the development of
ASHRAE Standard 90.1) [ASHRAE 2007] evaluates possible energy conservation measures by using a utility rate of
$0.0939 per kilowatt-hour (kWh) as the basis for life cycle and payback in their analyses [Taylor/McGuire 2008]. Electric
utility rates in certain parts of the United States, particularly the east and west coasts, often exceed $0.10/kWh. For example,
Stephen W. Duda, PE is Assistant Director of Mechanical Engineering for Ross & Baruzzini, Inc. in St. Louis, MO and is an ASHRAE-Certified
Building Energy Assessment Professional. He is immediate past Chair of Technical Committee 9.01 and is now Section Head 8 for TAC.
$6+5$(

an October 2008 ASHRAE Journal article [ibid] analyzed pipe sizing options based on an electric utility consumption rate of
$0.15/kWh in their hometown of Oakland, California.
On the contrary, electric utility rates in low-utility cost States such as this author’s home of Missouri, are well below
these benchmarks. Examples include the following, taken from utility company websites under large general power
commercial classifications.
The electric consumption rate in Columbia, Missouri offered by the dominant service provider is $0.055/kWh for large
general service, as found via their website [http://www.gocolumbiamo.com/WaterandLight/Rates/lg-ele.php].
The city electric utility in Springfield, Missouri charges only $0.0291/kWh [http://www.cityutilities.net/pricing/pricing-
electric-largegeneralpower.pdf] plus a variable fuel surcharge, which brings the overall net consumption rate up to about
$0.04/kWh as of December 2011. Similarly, a rate of $0.0411/kWh is available in Little Rock, Arkansas when summing base
rate plus fuel surcharge [http://www.entergy-arkansas.com/content/price/tariffs/eai_lgs.pdf and http://www.entergy-
arkansas.com/content/price/fuel/fa_eai.pdf]. Tulsa, Oklahoma offers fuel-adjusted electric consumption rates very similar to
that of Springfield, MO [https://www.psoklahoma.com/account/bills/rates/RatesAndTariffs.aspx].
It is also important to accurately account for any tiered consumption charges imposed by the electric utility. By tiered
consumption charges, we mean a billing structure that charges (for example) $0.08 per kWh for the first 100 kWh per month,
$0.06 per kWh for the next 300 kWh per month, $0.04 per kWh for all kWh beyond 400 kWh per month, or similar. When
assessing a potential energy-conservation measure but assuming that all other energy consuming devices in the building are
fixed, one must select the last tier or tiers of billing rate that will actually be applicable. In the above example, if it is known
that the building does or will use more than 400 kWh per month for fixed loads such as lights, computers, and processes, then
one must model the potential energy-conservation measure at the $0.04/kWh rate, because that is the tier that will realize any
differences in consumption.
With this in mind, the dominant electric utility company in St. Louis, MO applies a multi-tier rate for general service,
with $0.045/kWh charged in last tier [http://ameren.com/sites/aue/Rates/Documents/umbe34rt3M.pdf]. An almost identical
situation exists in Omaha, NE at $0.043/kWh in last tier [http://ww3.oppd.com/rates/OppdRateManual.pdf#nameddest=231].
The reason for lower rates in these regions varies from political and legislative factors, to the fact that plentiful
availability of coal and natural gas locally for generation keeps transportation costs to a minimum.
The above survey spans six utility companies in three Climate Zones (3A, 4A, 5A) in four central States, with rates
spanning from $0.04 to $0.055/kWh. For the financial examples in this paper, an electric consumption rate of $0.05 per kWh
will be applied for the purpose of illustration based on this survey.
Demand charges should also be considered in any thorough energy analysis, but that would add a complexity to this
paper not important to this paper’s thesis. A brief survey of the above utilities shows that demand charges are also lower in
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
the Midwestern USA by a magnitude similar to the difference in consumption charges.
AEDG Overview
The Advanced Energy Design Guide (AEDG) series was developed by ASHRAE to provide recommendations to
achieve specific energy savings when compared to the minimum code requirements of ASHRAE Standard 90.1. To do so, the
AEDG committee modeled prototypical buildings representative of construction for their class, and two sets of hourly
simulations were run for each prototype. The first set meets the minimum requirements of Standard 90.1, and the second uses
the recommendations in the AEDG. Each set of prototypes was simulated in eight climate zones further divided into moist
and dry regions, for a total of 16 climate locations. Energy-savings ideas that did not offer a reasonable return on investment
in the above simulations were not included in the AEDG; or were recommended only in climate zones where they did offer a
reasonable rate of return.
The AEDG does not disclose what energy rates were used in the simulations, but since $0.0939/kWh for electricity is
used by the Standard 90.1 committee for their work, it is assumed that the AEDG uses the same rate.
The chief limitation of the AEDG series is that it based in new construction, whereas existing buildings use most of our
energy. Gordon Holness, 2009-10 ASHRAE President, emphasized in his presidential theme “Sustain our Future by
270
ASHRAE Transactions
Rebuilding our Past” that the greatest opportunity to change our energy future is in existing buildings. He said that 75% to
85% of all buildings that will exist in 2030 are already in existence, and that if every new building used zero energy, it would
only make a small reduction in global energy consumption by 2030 [Holness 2009]. And in a weaker construction economy
experienced in the Midwestern USA in 2010-11, less new construction and greater renovation construction has been the
norm.
Existing buildings pose a far greater challenge in terms of sustainable design on a payback or cost-benefit basis. Even
with low electric utility rates, it is still possible to calculate a reasonable payback period for the purchase of a high-efficiency,
higher-cost building chiller versus a modest-efficiency, lower-cost building chiller. After all, in new construction even lesser-
efficiency options have a first cost. The projected long-term energy savings of the higher-efficiency option needs only to
offset the differential first cost. However, it is a far greater challenge to justify the voluntary replacement of a 15-year old
chiller that is still in very good functioning condition. An energy analysis in a retrofit situation often compares a high-
efficiency, high first cost purchase to no first cost at all – the status quo – continuing operation with existing equipment that
was already paid for. Depending on the building owner’s accounting or depreciation method, the projected long-term energy
savings in this case may need to offset the entire first cost.
EXAMPLES
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
This portion of the conference paper explores some specific examples of energy conservation strategies suggested by
ASHRAE’s 50% Advanced Energy Design Guide for Small to Medium Office Buildings [ASHRAE 2011a]. Each example
demonstrates the approximate payback period of the measure if using the Oakland, CA electric utility consumption rate of
$0.15/kWh, the “average” national electric utility consumption rate of $0.0939/kWh, and a rate more typical of that found in
low-cost States of $0.05/kWh. First, a brief review of the financial modeling methodology is necessary.
Financial Modeling Methodology

In order to specify or recommend an energy conservation measure on behalf of a building owner, the consulting
engineer should consider the initial cost (purchase and installation), the annual energy cost, the annual maintenance cost, and
a time-rate cost of money (i.e., interest rate, discount rate, or opportunity cost rate).
Some engineers and decision-makers use a Simple Payback analysis. In Simple Payback, both inflation and the time-
value of money are neglected. For example, if Machine A costs $2000 more than Machine B, but Machine B uses $500 worth
of energy per year more than Machine A does, then Machine A is said to have a Simple Payback of four years. Assuming
both machines are intended to last beyond four years, Machine A would be the better long-term selection.
For a short time horizon, life cycle cost results may not differ significantly from a simple payback approach. However,
building HVAC systems are intended to be long-life purchases. Over a longer life span, inflation and the time-value of
money are very significant components. In the age of high computing power and easy-to-use spreadsheet software programs,
a life cycle cost analysis accounting for these factors is recommended in lieu of a Simple Payback method.
Several factors are considered in determining the life cycle cost or payback period of each studied energy conservation
measure, including purchase costs, maintenance costs, and yearly energy consumption. The life cycle approach is valuable
because it quantifies these factors into one lump sum, present-day equivalent dollar amount for easy comparison of
alternatives.
The energy cost to be expended each year of the life cycle period is converted to present-day dollars using the Uniform
Series Present Worth Factor (USPWF) at a specified interest rate and a specified annual energy cost inflation [ASHRAE
2011b]. In other words, we determine exactly how much money would have to be deposited today in an interest-bearing
investment to fund the annual energy bills and just run out of money at the end of the life cycle period.
ሺଵା୧ሻ౤ ିଵ
ൌ (1)
୧ሺଵା୧ሻ౤
$6+5$(

where AP is the periodic recurring payment, i is the Interest Rate per period (months, years, etc.) and n is the
number of periods in units consistent with i.
The above equation is easy to program into popular spreadsheet software programs for repeated use. As one example, at
6.5% yield, an initial investment of about $4,156 would fund five annual energy bills of $1,000 each. Even if the energy cost
increases at a rate of 3% per year, an initial investment of $4,682 would be sufficient. The life cycle cost is therefore first cost
plus present-day equivalent energy cost plus present-day equivalent maintenance and other operating costs, summed over the
life cycle study period. With this method, an alternative system that may cost somewhat more to purchase, but requires lower
annual energy costs, can be studied. This way, an ‘apples-to-apples’ single-number comparison can be made between each
alternative.
Throughout this paper, an annual interest rate of 6.5% and an annual energy inflation rate of 3% are applied.
Example 1: Oversized Ductwork

Suggestion HV20 from the 50% AEDG–SMO [ASHRAE 2011a], applicable in all climate zones, is related to ductwork
design. One specific recommendation therein is to size all supply air duct for a friction rate not exceeding 0.08-inches water
column (20 Pa) per 100 linear feet (30 m) of duct run. This recommendation appears to apply to all supply air ducts, whether
they are mains, trunks, or sub-branches. The engineering office of this author happened to be modeling a 12,000 cfm (5,660
L/s) VAV supply air system for an actual medium-size office building in Climate Zone 4A as this conference paper was
being drafted, so we decided to put HV20 to the test. This HV20 recommendation is evaluated using the following modeling:
x Conventional sizing criteria of 2000 fpm (10 m/s) maximum velocity and 0.15”/100ft (1.25 Pa/m) maximum
friction, as would often be applied by this office.
x An aggressive sizing criterion of 0.08”/100ft (0.67 Pa/m) maximum friction in accordance with the AEDG.
x $7.46 per pound ($16.45/kg) installed cost of ductwork [based on Means 2011].
x A combined fan-motor efficiency of 65%.
x A duct height limitation of 16-inches (400 mm) inside dimension due to ceiling clearance available.
This analysis will assume that all duct friction losses are bad. A valid argument can be made that friction in ductwork
operating in heating mode is energy-neutral; because friction manifests itself as heat, and heat is desired in heating mode.
(Even in economizer cooling mode, anytime the outdoor temperature is lower than the desired supply air temperature, heat
from duct friction only changes the proportion of outdoor air to return air without energy penalty.) That fact will be ignored,
but bear in mind it would worsen any projected payback period of oversized ducts if it were considered, especially in winter-
dominated climates.
Space does not allow the full duct cost and energy model to be reproduced herein, but the analysis can be summarized
as follows: The full network of ductwork in the conventional system weighs 13,520 pounds (6,133 kg) and costs $100,859.
The same network of ductwork when re-figured under the aggressive criterion weighs 14,850 pounds (6,735 kg) and costs
$110,781.
Using proprietary commercial energy modeling software, the fan energy loss in the conventional ducts due to friction
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
on an annual basis is projected to be 25,309 kWh. The model uses an annual estimated loading and VAV part load curve
based on projected occupancy schedules, weather conditions, and so forth, for this specific building. (Even if the model is
inexact, it still allows a fair comparison of two alternatives.) Using the same building with same operating characteristics, the
more aggressive duct sizing criterion would have a duct energy loss due to friction of 15,818 kWh per year.
A Uniform Series Present Worth method of Equation (1) is applied at three electric utility rates, an annual interest rate
of 6.5% and an annual energy inflation rate of 3%, to determine the projected payback of the more aggressive ductwork
sizing criterion. At $0.15/kWh, the aggressive approach pays back its initial increased investment in 8 years. At
$0.0939/kWh, the aggressive approach pays back in 14 years. At $0.05/kWh, the aggressive approach does not offer a
payback within the first 25 years, and the analysis was discontinued at 25 years since paybacks beyond that point are rarely of
interest to building owners.
272
ASHRAE Transactions
Therefore, in a low-utility cost region such as those cited previously, it may be very difficult to justify this particular
strategy on a financial basis.
Example 2: Air-cooled Chiller Efficiency

Suggestion HV14 from the 50% AEDG–SMO [ASHRAE 2011a], applicable in all climate zones, is related to premium-
efficiency mechanical equipment, specifically the air-cooled chiller. A chiller with full-load energy use of 1.20 kW/ton (COP
of 2.93) or better and an integrated part-load value (IPLV) of 0.96 kW/ton (COP of 3.66) or higher, according to AHRI rating
methods, is recommended. We will compare this to a “conventional” air-cooled chiller using 1.30 kW/ton (COP of 2.71) and
an IPLV of 1.15 kW/ton (COP of 3.05). Again, the engineering office of this author happened to be modeling a 150-ton air-
cooled chiller for an actual medium-size office building in Climate Zone 4A as this conference paper was being drafted, so
we decided to put HV14 to the test. We obtained two chiller product selections from several chiller manufacturers, each
providing a “standard” selection near the conventional efficiency described above, and a “premium” selection meeting the
requirements of HV14 in AEDG-SMO.
Again, space does not allow the full chiller energy model to be reproduced herein, but the analysis can be summarized
as follows: The average of the “standard” chiller selections costs $101,000. The average of the “premium” chiller selections
is $117,000. Using the same proprietary commercial energy modeling software and annual energy model described in
Example 1, the energy consumed by the “standard” chiller on an annual basis is projected to be 296,010 kWh. Using the
same building with same operating characteristics, the “premium” chiller selection would have an energy use of 247,104
kWh per year.
A Uniform Series Present Worth method of Equation (1) is applied at three electric utility rates, an annual interest rate
of 6.5% and an annual energy inflation rate of 3%, to determine the projected payback of the “premium” chiller versus the
“standard” selection. At $0.15/kWh, the “premium” chiller pays back its initial increased investment in 3 years. At
$0.0939/kWh, the “premium” chiller pays back in 4 years. At $0.05/kWh, the “premium” chiller pays back in 8 years. While
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
8 years is often considered a reasonable return on investment for a piece of equipment that is expected to last much longer
than that, anecdotally a significant number of building owners and financial decision-makers will not accept measures with
paybacks longer than 5 years.
Example 3: Chiller Replacement, Existing Construction

The point was made earlier in this paper of the importance of energy retrofits in existing buildings. While even low
energy-cost States can show a reasonable payback period for the premium chiller in Example 2 above, it is a far greater
challenge to justify the voluntary replacement of an older, less efficient chiller that is still in very good functioning condition.
Let us further imagine for illustrative purposes that the energy consumption of this older chiller, considering both its original
efficiency and some degradation over time, is 1.40 kW/ton (COP of 2.51) with an IPLV of 1.30 kW/ton (COP of 2.71). We
now run the exact same energy analysis for the same building as in Example 2, but consider the first cost of the “standard”
chiller to be zero because it already exists and was already paid for. In this case, the differential in energy cost of the
“premium” chiller must justify its entire purchase and installation cost.
At $0.15/kWh, the high-efficiency replacement chiller pays back its investment in 11 years. At $0.0939/kWh, the
replacement chiller pays back in 19 years. At $0.05/kWh, the replacement chiller does not offer even a 25-year payback, so
the building owner will likely decide to keep the older, less efficient chiller for its entire useful life span.
Example 4: Any Generic Option with an 8-Year Payback

Many more examples could be run, but the thesis of this paper is not to criticize the AEDG nor Standard 90.1. The goal
is to point out that any conservation measure with a higher first cost will have a longer payback period where utility rates are
low. In fact, simple math can demonstrate this challenge even without a specific energy conservation measure being studied.
As a hypothetical generic example, take any generic premium-efficiency piece of equipment that costs $110,000 up front and
$6+5$(

uses 50,000 kWh per year, versus $100,000 first cost and 65,000 kWh per year for the conventional system. These
hypothetical numbers were specifically chosen because they result in an 8-year payback period at the $0.0939/kWh electric
consumption rate favored by the AEDG and SSPC 90.1, again using an annual interest rate of 6.5% and an annual energy
inflation rate of 3%. That same situation when changed to a $0.05/kWh rate yields a payback period of 16 years; a time
horizon that is unattractive to many building owners and financial decision-makers.
CONCLUSIONS
We have seen that electric utility rates in some portions of the Midwestern USA are only half of the rate assumed by
SSPC 90.1 in their analyses, and are only one-third of the rate found in the highest rate areas of the country. This fact renders
some energy-conservation measures unattractive from a financial payback or cost-benefit basis, as seen in the examples
above. The practice of touting short payback periods when based upon electric rates of $0.09/kWh or higher, whether in
technical seminars or commercial sales presentations, has little meaning to the consulting engineer practicing where rates are
half that figure.
Clearly, consulting and energy engineers in low utility cost regions need additional tools from organizations like
ASHRAE in order to encourage low-energy design in other ways. Those tools might include more emphasis on consequential
benefits of sustainable design, more emphasis on integrated design, mandatory standards, or even legislative approaches to
energy conservation.
The AEDG series hints at the consequential benefits of sustainable design in an appeal to building or business owners:
“Energy-efficient building offers a great possibility to enhance the working environment, including
indoor air quality, thermal comfort, and natural lighting. Studies show an enhanced working environment
can increase worker productivity, improve staff retention, increase corporate brand awareness, have
significant rental and sales prices premiums, and may contribute directly to corporate social responsibility
through reduced greenhouse gas emissions.” [ASHRAE 2011a]
This might be the greatest area where ASHRAE could provide additional tools to its member engineers struggling with
low-energy design in low energy cost States: Further hard science and concrete mathematics to demonstrate the correlation
between sustainable design and the owner’s bottom line using factors other than direct energy cost savings.
Integrated design is a valid approach to sustainability in the new construction market. When designing a new building
on a greenfield site, implementation many architectural green building features (better insulation, green roofs, solar shading
and screening strategies, optimum orientation, and so forth) can sometimes result in down-sizing of HVAC equipment, with
no net cost add to the project. The AEDG-SMO includes a full chapter on Integrated Design. But again with reference to the
theme of Presidential Member Holness theme, this is far less possible in the renovation market.
The final tools that may be effective involve legislative ideas: A so-called carbon tax, significant tax policy incentives
for energy-conservation measures, mandatory energy codes and their increased enforcement, or similar approaches. Even
increased enforcement of mandatory energy codes may not have a large impact, since these are usually applied only to new
construction or major renovations, and not to existing building owners who are considering voluntary energy retrofits. This
author is not necessarily advocating a political or legislative approach to energy-conscious design, but this conference paper
would not be complete without at least a passing mention of those possibilities.
NOMENCLATURE
AEDG = Advanced Energy Design Guide
AEDG-SMO = Advanced Energy Design Guide for Small to Medium Office Buildings (see References)
AP = Annual Payment
COP = Coefficient of Performance (energy removed divided by work energy in consistent units)
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
274
ASHRAE Transactions
i = Interest Rate per time period (i and n must be expressed in consistent time periods)
IPLV = Integrated Part-Load Value as defined by AHRI Standard 550/590.
kW = kilowatt [both an I-P and an SI unit]
kWh = kilowatt-hour [both an I-P and an SI unit]
USPWF = Uniform Series Present Worth Factor
VAV = Variable Air Volume
Superscripts
n = number of study periods (e.g., number of months or years)
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Symbols
$ = Dollars (United States currency)
REFERENCES
ANSI/ASHRAE. 2007. Standard 90.1, Energy Standard for Buildings Except Low-Rise Residential Buildings. Atlanta:
American Society of Heating Refrigeration and Air Conditioning Engineers, Inc.
ASHRAE. 2006. Energy Efficiency is Buzz for Expo Visitors, Exhibitors. ASHRAE Journal, Vol. 48, No. 3, pp. 8. Atlanta:
ASHRAE. 2011a. 50% Advanced Energy Design Guide for Small to Medium Office Buildings. Atlanta: American Society of
Heating Refrigeration and Air Conditioning Engineers, Inc.
ASHRAE. 2011b. Handbook of HVAC Applications, Chapter 37. Atlanta: American Society of Heating Refrigeration and Air
Holness, Gordon. 2009. Sustaining Our Future by Rebuilding Our Past. ASHRAE Journal, Vol. 51, No. 8, pp. 16. Atlanta:
Taylor, S.T. and McGuire, M. 2008. Sizing Pipe Using Life-Cycle Costs. ASHRAE Journal, Vol. 50, No. 10, pp. 24. Atlanta:
Means. 2011. RSMeans Mechanical Cost Data, 34th Edition. Norwell, MA: Reed Construction Data.
$6+5$(

SA-12-C035
Targeting a Net-Zero Energy Student

Center: Part 1 - Thermal and Electrical Load
Minimization
Trevor Caldwell Kelton Friedrich Brad Gregus

Student Member ASHRAE Student Member ASHRAE Student Member ASHRAE
Ryan Verschuere Jordan Anderson Aaron Murenbeeld
Kaveh Arfaei Mahsa Boroumand David S. Wilkinson, PE
Student Member ASHRAE Student Member ASHRAE
James S. Cotton, PE
Member ASHRAE
ABSTRACT
In the pursuit of sustainability, a change to the standard building design process, which normally starts with an
architect followed by a civil engineer then lastly a mechanical engineer designing the HVAC and energy system, has been
suggested. The proposed design process will have a mechanical engineer design the thermal envelop, HVAC and related
energy systems first so that operational energy efficiency was the first goal achieved in the design. Next, the civil engineer
would do analysis on the structural components, and determine the most sustainable construction materials to be used in
the building. Finally, the architect would have to work within the constraints set by the mechanical and civil engineer
when designing the layout of the building. With this new approach, the goal of sustainability would be entrenched in the
design process and not an afterthought.
To start implementing the proposed design process, work on the first step of the new building design process was
started on the proposed Engineering Center for Experiential Learning (ExCEL) to be located at McMaster University in
Southern Ontario. Student teams were given the task of designing the energy related components of the building to see if it
was possible to reach a net-zero energy building. The work was split into two phases, Part 1 was to determine the
building’s thermal and electrical loads and reduce these through conservation measures. The second stage, discussed in the
subsequent Part 2 paper, was to analyze the various site generation and HVAC systems that could meet the loads
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
determined in the current paper.

The student team looked at thermal envelop factors such as insulating properties, air tightness and building
orientation to utilize passive solar heating and day lighting. Solar and internal gains were considered, and the design tried
to minimize solar gains in the summer and reduce internal electrical gains year round with high efficiency lighting and
Caldwell, Friedrich, Gregus, Verschuere, Anderson, Murenbeeld, Arfaei and Boroumand were Engineering Students at McMaster University,
Hamilton, Ontario. David S. Wilkinson is the Dean of Engineering, McMaster University, Hamilton, Ontario. James S. Cotton is an Associate
Professor in the Department of Mechanical Engineering, McMaster University, Hamilton, Ontario.
276
ASHRAE Transactions
equipment. Infiltration loads were reduced by increasing air tightness of the building. Ventilation loads were minimized
by a combination of heat recovery ventilators.
With the various conservation measures implemented, a daily energy balance was done to find the annual electrical,
heating and cooling loads of the building. With the combination of many methods to reduce building loads and site
generation and thermal load meeting system combinations, presented in the Part 2 paper, it was shown that a net-zero
energy institutional building is possible.
INTRODUCTION
A new engineering student center, named the Engineering Center for Experiential Learning (ExCEL), has
been proposed to provide a shop space for engineering extracurricular teams and offices for student clubs, such as a
sustainable building management club, for the university students. The center analyzed is a two story, 12,000 ft2
(1,115 m2) building. To keep in line with the hands on learning approach, the building’s energy and HVAC system
design will be student led and incorporate teams from various engineering disciplines. These student teams would
be the first to start the proposed new building design process, focusing on the energy systems of the building.
Immediately the student teams came to consensus on using a performance based design with the ultimate goal of
achieving a net zero energy building.
As there are many definitions of net zero the team began by setting a target of: “the amount of energy
produced from renewable resources on site that is exported off site, is greater than or equal to the amount of energy
drawn from the grid over the course of a year.” Early in the design process it was determined that a high
performance building design, both thermally and electrically, would need to be used to reduce the amount of site
generation required to reach the goal. From this the building design was split into two stages. One was analysis on
the methods to reduce the thermal and electrical loads of the building. The second, which is discussed in a
subsequent paper Part 2, is analysis of the systems to meet the loads of the building. To determine the thermal and
electrical loads, the building was assessed on a daily energy balance approach, and then compared with a building
energy simulation program. With yearly energy requirement analysis of the high performance building design
completed the amount of site energy generation required could be determined to allow for load meeting systems to
be compared.
ELECTRICAL ENERGY DEMAND
The daily electrical energy needs of the ExCEL building were analyzed in terms of varying monthly usage
scenarios. From a typical school year different energy use schedules were evaluated in terms of weekday, weekend,
and holiday usage. Also a May to June case was evaluated to represent the lower number of people present on
campus, followed by an August case when campus utilization is at its lowest. The estimated plug load energy use
per day can be seen in Table 1.
Table 1. Plug Load Energy Use

Power for Weekdays Holidays Weekends May to July August
single (kWh/day) (kWh/day) (kWh/day) (kWh/day) (kWh/day)
component
(W)
Laptops 65 14.6 7.5 7.3 2.5 1.6
Computers 300 72.0 7.2 36.0 7.2 0.0
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

LCD 120 14.4 1.4 7.2 1.4 0.0
Projector 1000 7.9 0.2 7.9 1.9 0.0
Microwave 1000 1.9 0.2 1.9 1.9 0.0
Vacuum 1500 0.7 0.4 0.7 0.7 0.7
Garage Door 1000 0.2 0.2 0.2 0.2 0.0
Bathroom 150 21.6 21.6 21.6 21.6 21.6
Fan
The lighting was assumed to be provided by T8 Florescent bulbs with an 80 lumens/watt efficiency. The
amount of time these lights were on during a day was also estimated for the various loading seasons throughout the
year following the same method used to determine the plug load. The circulation fans were predicted to be 3.4 hp
(2.5 kW) in size and on continuously to provide a constant ventilation rate. Other equipment pumps were
accounted for, but their use was predicted to be more intermittent. The shop portion of the building was treated
with a separate usage schedule for 2.7 hp (2 kW) of power it was assumed to draw at a maximum. The estimated
monthly energy usage for each of the electrical loads can be seen in Figure 1. The summation of the monthly
electrical usage for each type of load would give the total for the month. Adding all months up, the total annual
electrical load was estimated to be 317E6 Btu (93 MWh). This gives an electrical power density of 26.6
kBtu/ft2/year (83 kWh/m2/year).
Electrical load conservation measures were selecting high efficiency lighting and fans as well as choosing
energy star appliances.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 1 Estimated monthly electrical load based on expected usage of ExCEL building.
THERMAL ENERGY DEMAND
Before the peak and annual thermal audit could be conducted, the building parameters that affect the energy
balance had to be determined. To reduce energy losses by conduction, the most thermally efficient construction
materials were chosen for the thermal envelop based on available options present in the Heating, Refrigeration and
278
ASHRAE Transactions
Air Conditioning Institute of Canada’s handbook (1993).
To allow visibility of the learning activities inside and outside the ExCEL building the east and south faces of
the building were expected to be completely covered in windows. The windows were specified as triple glazed glass
with a 1/2” air spacing, an emissivity of 0.2 and a thermal conductivity of 0.28 Btu/h•°F•ft2 (1.6 W/m2•ÝC).
They will be mounted tightly utilizing low profile thermally efficient metal framing. The other two faces of the
building were solid walls consisting of an 8” solid concrete studded wall with R13 insulation, R8 Sheathing, an
inside finish, and an effective thermal conductivity of 0.032 Btu/h•°F•ft2 (0.18 W/m2•ÝC). The roof of the
building will consist of 8” light concrete with R50 insulation above deck and a suspended ceiling below. The
effective thermal conductivity is 0.0067 Btu/h•°F•ft2 (0.038 W/m2•ÝC). The basement walls and floor consist of
concrete with R19 insulation. The basement floor utilized a heat loss factor of 1.1 Btu/h•ft2 (3.4 W/m2).
The building infiltration was analyzed and specified that the quality of construction for this building was
tight. As the building will utilize a forced air ventilation system, the building will operate at a positive pressure.
The building was specified to have 3 double doors, a single door, and an overhead door. Infiltration through the
doors was found to be 1730 ft3/min (49 m3/min) (ASHRAE 2009). Heat loss through the doors was estimated to
be reduced by 60% by utilizing an efficient air curtain (Daikin 2010). Finally the ventilation system was analyzed.
For a building with 60 occupants, all require 20 ft3/min (0.6 m3/min) (ASHRAE 2009), this equates to 1200
ft3/min (34 m3/min) overall ventilation. A paint room for contaminated air handling was also used, it was 20x20x8
ft (6x6x2.4 m) room, and was predicted to have 6 air changes per hour (ASHRAE 2009). Utilizing a 50% efficient
heat recovery system, this heat loss for ventilating this room can be reduced.
In an effort to reduce the cooling load while still utilizing passive heating in the winter, shading on the
windows was analyzed for a number of overhang lengths. Solar radiation was analyzed with the solar admittance
method (ASHRAE 1997) in regards to building profile, time of year, and shading overhang length and was found to
be 682E6 Btu (200 MWh) on an unshaded building per year. After analyzing various overhang lengths, it was
observed that the overhang length to window height ratio of 50% provided the reduced summer loading effect
while maintaining the winter solar load, as shown in Figure 2. Therefore it was chosen as the most suitable option
with the total annual solar radiation being reduced to 426E6 Btu (125 MWh).
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

Figure 2 Total monthly solar heat gain varying with shading.
Peak Thermal Requirements
To allow for HVAC component maximum capacity sizing, peak winter heating load and peak summer
cooling load calculations were performed. The peak heating and cooling loads on the ExCEL building were
analyzed in regards to average daily temperature. The hourly temperature averages were collected from
Environment Canada’s Climate website (2011) and were averaged to get daily averages, which are used as the base
temperature in the calculations for peak heating/cooling loads.
Calculation of the peak winter heating load was done by analysis of sources of thermal loss in the building
design. The building was assumed to be unoccupied with no electrical gains. Night time was the selected as there is
no solar gain during this time. Using the previously mentioned construction constraints, the peak thermal losses
for the coldest day of the year were calculated and can be seen in Table 2.
Table 2. Peak Heating Thermal Losses

Losses
E6 Btu/h (kW)
Transmission 59 (17.3)
Infiltration 16 (4.8)
Office Ventilation 13.7 (13.7)
Shop Ventilation 47 (3.7)
Total Sensible Heat Loss 135 (39.5)
The peak summer cooling load included an analysis of the same thermal sources as the peak heating
requirements with the addition of internal and solar gains as shown in Table 3. Utilizing the shading with 50%
overhang length to window height ratio, solar gains amount to approximately 28% of the total peak thermal gains
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
280
ASHRAE Transactions
of the building in the summer.
Table 3. Peak Cooling Thermal Gains

Losses
E6 Btu/h (kW)
Glass 8.5 (2.5)
Walls 1.7 (0.5)
Roof 0.44 (0.13)
Infiltration 24 (7.1)
Ventilation 21 (6.3)
Occupants 29 (8.4)
Electrical 75 (21.9)
Thermal Gains (without 160 (46.9)
Solar)
Solar Thermal Gains 61 (17.9)
(Overhang with 50% length
to height ratio)
Total Thermal Gains 221 (64.8)
Monthly Heating and Cooling Loads
The monthly heating and cooling loads on the ExCEL building were analyzed in regards to average daily
temperature, identical to the calculations used to determine the peak thermal requirements. Thermal gains and
losses were added up daily to determine the amount of conditioning load required. This time scale will not account
for the hourly changes between heating and cooling requirements, and will assume they will cancel each other out.
These loads were evaluated monthly and show the breakdown of heating and cooling loads experienced by the
building. Figure 3 indicates heating gains such as passive solar loading, heating by occupancy, and heating by
electrical. Figure 4 shows heat loss through the roof, walls, basement, glass, ventilation and infiltration. Thus it
was calculated that the total required yearly cooling and heating load respectively was 112E6 Btu (32.7MWh) and
298E6 Btu (87.4MWh).
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

Figure 3 Monthly average thermal gain.
With the annual energy consumption determined as well as the peak heating and cooling power required, a
comparison between the utilization of the heating and cooling systems can be done. Given that the peak cooling
power required is higher than the peak heating requirements but the annual energy for cooling is less than heating,
the installed capacity for cooling will be underutilized compared to the installed capacity for heating. This is the
case for many buildings in climate zones that have higher heating degree days than cooling. From this utilization
analysis a compromise was suggested to install a smaller cooling system that will cost less initially and get more
utilization and allow the temperature to rise in the conditioned space on the few hottest days of the year. The
discomfort on a few days of the year was considered less important than the improvement in sustainability due to
the higher utilization of the smaller cooling system.
Figure 4 Monthly average thermal loss.
To summarize the thermal and electrical energy conservation methods compared to constructing a building
which was code compliant in Ontario, Canada (ASHRAE 90.1 2004), Table 4 was developed.
Table 4. Comparison of Energy Efficient Measures

Conservation Measure Code Compliance Proposed for ExCEL Building
Thermal Conductivity
Walls 0.048 Btu/h•°F•ft2 (0.28 0.032 Btu/h•°F•ft2 (0.18
W/m2•ÝC) W/m2•ÝC)
Windows 0.30 Btu/h•°F•ft2 (1.74 0.28 Btu/h•°F•ft2 (1.6 W/m2•ÝC)
W/m2•ÝC)
Roof 0.026 Btu/h•°F•ft2 (0.15 0.0067 Btu/h•°F•ft2 (0.038
Ground Floor 0.67 Btu/h•°F•ft2 (3.79 0.051 Btu/h•°F•ft2 (0.29
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Air Tightness 1 ACH or Greater 0.4 ACH
282
ASHRAE Transactions
Passive Solar Design Not Considered Window orientation, solar heating,
day lighting and shading considered
Heat Recovery Ventilator Not Used Cross flow HRV (50% efficient)
Electrical Load Reduction Not Considered High efficiency lighting and
equipment
The daily energy balance calculation done by hand to determine the annual thermal and electrical loads as well
as the peak loads was compared with an energy simulation program, ECOTECT. With the same assumptions used
in the energy simulation program, a close agreement between the two methods shows that the numerous
summation calculations in the by hand method were done correctly.
CONCLUSION
As step one in new the proposed sustainable building design process, conservation methods and technologies
were suggested to reduce the electrical and thermal loads of ExCEL building. Utilizing ASHRAE recommended
energy gains and losses calculations and an energy balance was performed on the proposed building. From the
hottest and coldest day of the year analysis the peak loads were estimated for sizing the max capacity of heating and
cooling systems. An annual energy balance based on a daily time step was done to determine the annual electrical
and thermal loads of the building. From this estimated annual load it would now be possible to compare
combinations of HVAC and site generation systems to see if a net-zero energy building is possible with this level of
energy conservation implemented in the ExCEL building. The discussion of the various HVAC and site generation
systems considered and if they are able to achieve a net-zero energy building can be seen in Part 2.
ACKNOWLEDGMENTS
The authors wish to thank the ASHRAE Hamilton Chapter for their advice, suggestions and support.
REFERENCES
ASHRAE. 2009. ASHRAE Handbook-Fundamentals. Atlanta: American Society of Heating Refrigeration and Air
ASHRAE Standard 90.1. 2004. Energy Standard for Buildings Except Low-Rise Residential Buildings. Atlanta:
American Society of Heating, Air-Conditioning and Refrigeration Engineers, Inc.
Daikin. 2010. Unique air curtain solution from Daikin & Biddle provides energy and CO2 savings. Daikin Inc.
Environment Canada. 2010. National Climate Data and Information Archive.
http://climate.weatheroffice.gc.ca/climateData/canada_e.html
HRAIC. 1993. HRAIC Handbook- Small Commercial Heat Gain and Heat Loss Calculations Manual. Heating,
Refrigeration and Air Conditioning Institute of Canada.
$6+5$( --`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---

SA-12-C036
Targeting a Net-Zero Energy Student

Center: Part 2 – Systems to Meet Building
Loads
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Kelton Friedrich Trevor Caldwell Brad Gregus
Ryan Verschuere Jordan Anderson Aaron Murenbeeld
Kaveh Arfaei Mahsa Boroumand David S. Wilkinson, PE
Student Member ASHRAE Student Member ASHRAE
James S. Cotton, PE
Member ASHRAE
ABSTRACT
A new sustainability focused approach to the building design process has been proposed where first mechanical
engineers design the building energy systems, next the civil engineers determine the most sustainable construction
materials and finally the architect works within these constraints to come up with the building layout. The first step in the
proposed design process was started on a building to be located at McMaster University in Southern Ontario, the proposed
Engineering Center for Experiential Learning (ExCEL). To meet the Center’s objectives, which is to provide hands on
learning experience, student teams worked on designing the energy related components of the building to see if it was
possible to reach a net-zero energy building. Split into two stages, the design first focused on determining the building’s
thermal and electrical loads and reducing these through conservation measures, as discussed in the Part 1 paper. The
second phase, which is the discussion of this paper, was to analyze the various site generation and HVAC systems that
could meet the predetermined building loads.
To achieve the goal of net-zero energy, site generation from renewable resources has to be larger than all the building
loads. For generation of electrical energy on site the renewable resources of wind, geothermal and solar were considered. It
was found that solar photovoltaic system was the only practical form of electrical generation. With a system selected for the
electrical loads, the thermal loads of the building had to be met by an HVAC system. Besides a conventional natural gas
furnace and roof mounted cooling system, a geothermal heat pump system was analyzed. It was found that with a high
performance geothermal system and a novel continuous surface solar array, it was possible to meet all the buildings
electrical and thermal loads with electricity generated on site over the course of a year, allowing for the net-zero energy
designation.
Friedrich, Caldwell, Gregus, Verschuere, Anderson, Murenbeeld, Arfaei and Boroumand were Engineering Students at McMaster University,
Hamilton, Ontario. David S. Wilkinson is the Dean of Engineering, McMaster University, Hamilton, Ontario. James S. Cotton is an Associate
Professor in the Department of Mechanical Engineering, McMaster University, Hamilton, Ontario.
284
ASHRAE Transactions
In addition, a tri-generation system consisting of a micro natural gas turbine with heat recovery on the exhaust and
an absorption chiller system was considered for meeting the electrical, heating and cooling loads of the building. Due to
the use of non-renewable natural gas, the system could not achieve the net-zero energy goal. However, the system allowed
for a more sustainable use of natural gas due to the higher efficiently of the natural gas consumed over conventional
natural gas power plants. Therefore the tri-generation was considered a comparable alternative for showcasing sustainable
technologies within the building for the students and general public to learn from.
INTRODUCTION
A new engineering student center, named the Engineering Center for Experiential Learning (ExCEL), has
been proposed to provide a shop space for engineering extracurricular teams and offices for student clubs, such as a
sustainable building management club, for the university students. The center analyzed is a two story, 12,000 ft2
(1,115 m2) building. To keep in line with the hands on learning approach, the building’s energy and HVAC system
design will be student led and incorporate teams from various engineering disciplines. Immediately the student
teams came to consensus on using a performance based design with the ultimate goal of achieving a net zero energy
building.
As there are many definitions of net zero the team began by setting a target of: “the amount of energy
produced from renewable resources on site that is exported off site, is greater than or equal to the amount of energy
drawn from the grid over the course of a year.” The first stage of the design process to come up with the most
efficient building design, both thermally and electrically to reduce the amount of site generation required to reach
the goal. The energy conservation methods employed and the determination of the buildings annual loads can be
seen in the preceding paper Part 1. With the loading found it is the aim if this paper to perform a comparative
analysis of the various site energy generation and HVAC systems that could meet the predetermined loads of the
building. The systems comparison will be done on a triple bottom line analysis, where economic, environmental
and social concerns will be addressed. Given the premise of the building is to aid in experiential learning, the ability
of the load meeting system to be a teaching tool for sustainable technologies to educate and inspire future students
is the primary social criterion.
From the previous analysis the thermal loads are larger than the electrical load. There are various systems that
can meet the thermal loads, these systems can use different forms of input energy and have different efficiencies in
converting the input energy into useful energy. To provide a base comparison, a conventional HVAC system will
be analyzed. To utilize the higher efficiencies of a heat pump, a geothermal ground loop with a heat pump will be
considered for providing heating and cooling in the building. In addition, a tri-generation system which includes a
natural gas turbine with exhaust heat recovery, an absorption chiller and thermal storage would be considered due
to the increased utilization of the potential energy in natural gas compared to conventional systems.
To achieve a net-zero designation, some form of energy generation is required on site to meet the building’s
loads. Renewable energy that is incident on the building site is the least disputed energy generation source to allow
the net-zero energy designation and used to define net-zero energy for this project. For the location of the ExCEL
building, geothermal energy, solar energy and wind energy are the potential renewable sources. From the lack of
consistently high winds speeds and ice throw safely hazards, wind generation was not considered feasible. Due to
the drilling costs to go deep enough to make electrical generation from geothermal possible, it is not possible to
generate electricity from geothermal at the site. However, meeting thermal loads is still possible with geothermal.
This left only solar energy as the renewable energy source to generate electricity with. Solar thermal would also be
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

able to help meet the hot water thermal needs of the building. Due to hot water only being needed for hand
washing in the building, the hot water thermal needs were neglected in this analysis.
SOLAR PHOTOVOLTAIC ENERGY
Analysis of solar photovoltaic panels of the roof of the building was done first as this form of site generation
could meet the electrical loads of the building regardless of the HVAC system used to meet the thermal loads of the
building. If the annual electrical energy generation from this method was greater than the required annual electrical
load of the building, the excess electrical energy could be used to power the systems used to meet the thermal loads
of the building. If the site generation was greater than both the required annual electrical loads of the building and
the required load of the system used to meet the thermal loads, then a net-zero designation would be possible. Since
the solar panel system only generates electricity and does not meet the thermal loads directly, separate analysis of it
and the load meeting method are required determine if the net-zero designation is possible.
The building energy feasibility software RETScreen (2004) was utilized to design the photo-voltaic (PV) array.
The design objective is to maximize the generated electrical energy annually from the building site and was done by
maximizing area available to place solar panels, using a high efficiency product and spacing panels to reduce
shading.
It was determined using RETScreen (2004) that a photovoltaic panel tilt angle of 37° will maximize the
energy available annually to a single panel. The panels would face directly south optimizing the azimuth angle
throughout the year. In the traditional horizontal flat roof layout, a decision was made to space panels so that
minimum shading would occur throughout the year. Consequently, the solar noon angle for the shortest day of the
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
year was used to determine the spacing between panels. This spacing would ensure that the panels will undergo a
minimal amount of shading annually, while keeping them close enough to maximize the number of panels that
could be placed on the rooftop. It was imperative that a high-efficiency panel was used to maximize annual energy
available, hence a 14.9% efficiency module was chosen. Additionally, a local vendor was chosen to reduce the CO2
emissions in transport of the panels to the university campus. In total, using the entire 6000 ft2 (557 m2) of the
rooftop, and placing a row of panels on the south-facing sunshade, 2649 ft2 (246 m2) of optimally angled panels
could be implemented. This array would result in 188E6 Btu (55 MWh) of direct current electrical energy being
available to the student center annually. Utilizing a 90% efficient inverter system to convert the direct current to
the standard alternating current used in buildings, 171E6 Btu (50 MWh) of electrical energy annually could be
provided to the ExCEL building for standard use.
The electrical generation from the standard solar farm lay out of the PV panels on the roof was unable to
generate more electrical energy in a year than was consumed by the high efficiency building, 318E6 Btu (93 MWh).
From this analysis alone it was not possible to achieve a net-zero energy building with these assumptions. Rather
than purchase higher efficiency but more expensive per watt of power generation PV panels, a change to the PV
array layout was proposed. This was to have a continuous panel array at a close to optimal angle for year round
electrical production. This greatly changes the roof of the building, but without needing space between rows of
panels to prevent shading and allowing the array to shade the area on the north side of the building a significant
yearly increase in electrical production is possible. With no spacing between the PV cells it was estimated that in
covering the same roof space as the solar farm layout, with the same type of panels and inverter gave 478E6 Btu
286
ASHRAE Transactions
(140 MWh) of annual electrical production. Since this was greater than the estimated building electrical load it was
possible to achieve a net-zero energy building with this array configuration.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
LOAD MEETING: CONVENTIONAL SYSTEM
With the thermal loads reduced by various high performance conservation methods, meeting this load can still
be done by non-high performance conventional systems. Conventional methods would be building housed furnaces
and chillers. Utilizing a condensing natural gas furnace that has an efficiency of 90% and a roof mounted chiller
with a COP of 2 in the summer cooling season the energy consumption by this method can be calculated. The
expected energy consumption with this method can be seen in Table 1. Due to the amount of natural gas
consumed in the conventional HVAC case it was not possible to achieve net-zero energy even with the
conservation measures.
Table 1. Annual Building Energy Consumption with the Conventional System

Load Building Requirements Consumption to Meet Tonnes of CO2
E6 Btu (MWh) Load
E6 Btu (MWh)
Electricity 317 (93) 317 (93) 21.2
Heating 300 (88) 334 (98)* 17.6
Cooling 113 (33) 58 (17) 3.8
Total 730 (214) 709 (208) 42.6
*Input energy source is natural gas
LOAD MEETING: GEOTHERMAL HEAT PUMP SYSTEM
A vertical closed loop, ground source heat pump system was evaluated for providing the buildings heating
requirements. Analysis was performed using RETScreen (2004) with results verified using ASHRAE models
(Kavanaugh 2003). The system was designed to meet peak heating demands at a COP of 4 with a 117 °F (47 °C)
temperature differential between the load side exit fluid and the fluid incoming from the geothermal source, and
additionally, to balance the energy exchanged with the ground on a yearly basis to ensure sustainability of the
resource. RETScreen was utilized first in order to compare to the ASHRAE model. It was found that RETScreen
had difficulties modeling high efficiency buildings and subsequently the loads in the summer were significantly
larger in RETScreen than in the aforementioned loading profile shown in this report. Thus, the ASHRAE method
for sizing geothermal borefields for feasibility studies suggested by Michel Bernier (2010) was used to verify the
results. Utilizing RETScreen’s loading profile, the ASHRAE model agrees within 6% of the length predicted by
RETScreen. However, when utilizing the loading from Part 1, the ASHRAE model gives a 33% larger borefield
than the RETScreen model gave for the loading profile RETScreen came up with. The main difference between the
RETScreen model and the model proposed by Bernier is that the model proposed by Bernier takes into account
long term thermal imbalances and when using the loading outlined previously in this report, the long term thermal
imbalance rises to 187E6 Btu (55MWh). Thus, when the yearly imbalance is forced to zero, the model agrees within
3% of the RETScreen model.
$6+5$(

Once the geothermal system was sized, the annual average COP of the system needed to be estimated. In the
interest of high efficiency it was assumed the heating working fluid temperature would not be much higher than
the required conditioned space temperature. This is to reduce the temperature gradient the heat pump has to
overcome thereby improving the COP. Also it was assumed that in the summer, to provide cooling, the heat pump
would not be used, and the fluid would just be circulated in the heat exchangers, as the ground temperature is the
desired cooling fluid temperature. From this it was assumed that a COP of 4 for winter heating and a COP of 20
for summer cooling would be possible. With these assumptions the expected energy consumption with this method
can be determined and seen in Table 2.
Table 2. Annual Building Energy Consumption with the Geothemal System

E6 Btu (MWh) Load
E6 Btu (MWh)
Electricity 317 (93) 317 (93) 21.2
Heating 300 (88) 75 (22) 4.9
Cooling 113 (33) 7 (2) 0.5
Total 730 (214) 399 (117) 26.6*
* Assuming all electricity is provided at Ontario grid average greenhouse gas emissions per kWh (OPA 2010)
When comparing on a triple bottom line basis, the geothermal load meeting method can be shown to be both
environmentally and socially feasible, however it is not cost effective. Comparing the geothermal system and a
conventional system it can be seen that geothermal system emits 60% of greenhouse gas emissions that the
conventional system does with the electricity for both systems being provided by the Ontario power grid.
Financially, a geothermal system requires a larger initial investment compared to a conventional system. While
maintenance and input energy costs tend to be lower for a geothermal system as compared to a conventional
system, the energy cost of electricity is on the rise relative to the predicted cost of natural gas which may not allow
the geothermal system owner to recoup the higher installation costs in a reasonable payback period.
From this analysis with the unconventional continuous solar panel array is it possible to achieve a net-zero
energy building as only electrical energy is required as input energy for the geothermal system giving a total annual
requirement of 399E6 Btu (117 MWh) which is less than 478E6 Btu (140 MWh) of electrical energy that could be
produced by the array with the inverter system. With all the electrical power being provided by the solar array,
this system would have no net annual greenhouse gas emissions while meeting the thermal loads. None of the
systems that use natural gas are able to be greenhouse gas emissions free, giving the geothermal system a significant
environmental advantage. With this ability to showcase a net-zero energy building with no greenhouse gas
emissions during operation makes the geothermal system attractive from a social standpoint as it would be a great
learning tool for teaching future students about sustainability.
LOAD MEETING: TRI-GENERATION SYSTEM
To allow more flexibility in the electrical generation, a tri-generation system installed in the building was
considered. Tri-generation is a micro gas turbine with heat recovery system and absorption chilling system to be
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`
288
ASHRAE Transactions
used as a combined source of heating and cooling as well as electrical power. The natural gas will be combusted in
the natural gas turbine first, driving a generator to generate electricity. The hot exhaust gas that exits the natural gas
turbine will go through a heat exchanger which will be similar to the one in a natural gas furnace, allowing the
exhaust gas to heat the building in winter. In summer the heat from the exhaust gas of the natural gas turbine
would be the required heat input into an absorption chiller, allowing cooling of the building. To optimize
efficiency, the micro gas turbine will be heating led, as generating electricity with no heat required will end up
sending potential energy in the exhaust gas in to the atmosphere like standard generation plants. After a review of
available micro gas turbines it was suggested to select a natural gas turbine that is 25% efficient at converting the
potential energy in natural gas to electrical energy with the capability of producing 102E6 Btu/h (30kW) of
electrical power and 184E6 Btu/h (53.9 kW) of useful thermal power when operating at full load (Capstone
Turbine Corporation 2010). This is greater than the peak heating requirements of the building of 135E6 Btu/h
(39.5 kW), so the micro gas turbine would not be required to run continually on the coldest day of the year. To
provide the cooling needs of the building, an absorption cooling system was studied. The suggested absorption
system with a heat input of 171E6 Btu/h (50.2kW) would provide 120E6 Btu/h (35.2 kW) of cooling (Yazaki
Energy Systems, Inc. 2010) This level of cooling is less than the peak cooling required, 221E6 Btu/h (64.8 kW), so
either a larger micro gas turbine is required, a supplementary cooling unit is added or the building is warmer for a
few days of the year. While meeting just the thermal loads of the building the tri-generation was able to produce
270E6 Btu (79MWh) of electrical energy per year. The annual energy consumption was calculated and can be seen
in Table 3.
Table 3. Annual Building Energy Consumption with the Tri-generation System

E6 Btu (MWh) Load
E6 Btu (MWh)
Electricity 317 (93) 48 (14) 3.2
Heating 300 (88)
Cooling 113 (33)
Natural Gas 1000 (293) 52.7
Total 730 (214) 1048 (307) 56.0
When comparing on a triple bottom line basis, the tri-generation system is not financially or environmentally
feasible, but can provide social benefits. Compared to the conventional load meeting system, the tri-generation
system would increase the carbon footprint of the building by approximately 31%. However, if the tri-generation
was implemented in a grid system where all the electrical power was provided by coal and natural gas plants, the
tri-generation system would in this case be more environmentally friendly as it would result in less greenhouse gas
than a conventional system. The tri-generation system is not able to outperform the conventional system from an
environmental point of view where the ExCEL building is located due to the large amount of electricity supplied to
the grid by low greenhouse gas generation. Only with thermal storage allowing for time of use control would the
tri-generation system emit less greenhouse gases than the conventional where the ExCEL building is located. A tri-
generation system would also require additional space and more complex equipment with more complex
$6+5$( --`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---

maintenance, making it more expensive to maintain than a conventional system. The input energy cost of the tri-
generation system would be cheaper annually than a conventional system as all the energy comes from
considerably less expensive natural gas than pulling, more expensive per unit of energy, electricity from the grid.
However, the energy savings would not be high enough to allow for an effective payback period on the extra initial
cost of the tri-generation system, so subsidies would be required to make the system financially feasible. Much like
the geothermal system, the tri-generation system would be beneficially socially as it would provide a learning tool
to educate future students about the sustainability of using what is waste heat to conventional natural gas power
plant for useful purposes.
Even though electrical energy is produced on site, the ability to call this a net-zero building is questionable.
This is because non-renewable natural gas transported into the site is used as the energy source. However,
compared to the conventional system this is an improvement, as high quantity of energy in the form of electricity
is produced from the potential in natural gas. The tri-generation system is more thermally efficient than a
commercial natural gas turbine generator, as most of the energy in the exhaust gas is utilized. So although a tri-
generation system does not allow for a net-zero designation, it allows a more sustainable use of natural gas than
conventional systems. Along with the improvement of the overall utilization of the potential energy in natural gas,
the tri-generation system gives more flexibility in the generation of electricity than the solar panel array. If the all
electricity generated can be sold back to the grid, the economics of both the solar panel array and the tri-generation
system will not affect each other, so they can both be implemented in the building at the same time.
THERMAL ENERGY STORAGE
The purpose of thermal storage would be to control the time of day energy is either pulled from the grid in
the conventional and geothermal case or supplied to the grid in the tri-generation case. The motivation for this
control is time of use pricing and reduction in carbon foot print. Due to the large amount of low greenhouse gas
emitting nuclear and hydro plants providing power to the electrical grid the ExCEL building would draw from,
there is the potential to draw power from these base load providers. When electrical power demand is greater than
what these base load providers can give, fast responding but high greenhouse gas emitting coal and natural gas
plants are started up that increase the price of electricity. Thermal storage allows for only running the conventional
and geothermal systems when electrical demand is low and the price is cheaper due to base load power plants
providing it. Thermal storage would allow for running the tri-generation system only when other natural gas plants
are running at peak electrical demand times thereby utilizing the higher total thermal efficiency of the tri-
generation system over the conventional natural gas plants.
CONCLUSION
Neither the geothermal heat pump system nor the tri-generation system had enough energy cost saving
relative to their initial cost to be financially beneficial when compared to a conventional load meeting system. On
the other hand, both the systems provided more social benefits, particularly from an experiential learning
perspective, than the conventional system. Finally, since the geothermal system with a solar array can allow the
building to be net-zero energy with no greenhouse gas emissions, it clearly came out as the most environmentally
--`,`,``,,`,`,```,`,,``,`
290
ASHRAE Transactions
friendly.
To provide the most benefits a duel load meeting system is proposed. In this case both the geothermal and tri-
generation methods are housed in the ExCEL building. However, only the solar array and the geothermal system
would be used to meet the loads of the ExCEL building. As mentioned before the geothermal heat pump combined
with the high efficiency building and a large enough solar array could achieve a net-zero energy building. Instead of
meeting the loads of the ExCEL building, the tri-generation system would just be meeting campus loads of steam
and chilled water. In this way all the experiential learning of the tri-generation system can still be a part of the
ExCEL building as well as using natural gas in a more sustainable way than conventional natural gas power plants
do. The additional advantage of this duel showcase is the tri-generation system does not have to be sized to meet
the loads of the ExCEL building, it can be bigger or smaller depending on available space and initial funding.
However, the biggest barrier, that of the initial cost, would be significantly higher due to purchasing both systems.
With the geothermal and tri-generation methods meeting different loads, their pay back periods will not affect one
another, so the economics of each method will not be affected.
ACKNOWLEDGMENTS
The authors wish to thank the ASHRAE Hamilton Chapter for their advice, suggestions and support.
REFERENCES
Bernier, M., Marchino D. and Phillipe, M. 2010. Vertical Geothermal Borefields. ASHRAE Journal 52(7) pp. 20–
28.
Capstone Turbine Corporation. 2010. C30 Natural Gas Data Sheet. www.capstoneturbine.com:
http://docs.capstoneturbine.com/home.asp
Ontario Power Authority. 2010. Supply Mix Background Reports.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
RETScreen International. 2004. RETScreen International Clean Energy Decision Support Center.
Steven Kavanaugh, W. M. 2003. ASHRAE Handbook: Geothermal Energy. Atlanta: American Society of Heating,
Ventillation and Air Conditioning Engineers.
Yazaki Energy Systems, Inc. 2010. Water Fired SingleǦEffect Chillers and ChillerǦHeaters.
http://yazakienergy.com/waterfiredapplication.htm
$6+5$(

SA-12-C037
Contaminant Removal Effectiveness

of Displacement Ventilation Systems
During Heating Season; Summary
Results From Three Field Studies
Boualem Ouazia, Ph.D. Iain Macdonald, Ph.D. Michel Tardif, PE
Alexandra Thompson, Ph.D. Daniel Booth

ABSTRACT
A series of field studies on the performance of displacement ventilation (DV) were carried out in
schools across Canada during the heating season over the last three years. Indoor air quality was evalu-
ated using the IAQ index Contaminant Removal Effectiveness (CRE) based on carbon dioxide concentra-
tions for three schools equipped with DV systems. Results from the three field studies have shown values of
CRE to be higher than 0.7 (ventilation effectiveness specified in ASHRAE 62.1-2007 when DV systems are
used for heating, with ceiling return). The measured CRE in the three schools was often greater than 1.0
even in the middle of the heating season. The results show that as long as the supply air is cooler than the
room air, DV system works well in achieving good contaminant stratification and ventilation effectiveness
in the space when combined with a perimeter heating system.
INTRODUCTION
Generally, there are two types of mechanical ventilation system applied to commercial buildings: mix-
ing ventilation and stratified ventilation. With mixing ventilation the conditioned air is supplied from ceil-
ing outlets at a relatively high velocity. The air is discharged at a high velocity to provide a well mixed
room air condition. This air distribution pattern causes contaminated room air to mix with the supply air.
Most commercial buildings in Canada and the United States use this type of overhead distribution system.
Stratified ventilation, including displacement ventilation (DV) and underfloor air distribution (UFAD), is a
means of providing cool supply air directly to the vicinity of occupants in a space. The conditioned air is
supplied near the floor level at low velocity and falls towards the floor due to gravity and spreads across the
room until it comes into contact with heat sources. The cool supply air slowly rises as it picks up heat from
occupants and equipment. The warm polluted air rises towards the ceiling where it is exhausted from the
space. This vertical airflow pattern near each occupant often referred to as a thermal plume, makes it less
likely that pollutant will spread. The DV system provides effective ventilation, since the supply air is deliv-
ered directly to each occupant. Stratified ventilation systems, are methods of delivering space conditioning
in offices and other commercial buildings that are increasingly being considered a serious alternative to
conventional ceiling-based air distribution systems because of the potential significant benefits that it can
provide (Bauman, 2003). With the appropriate design and application, stratified systems can improve in-
door air quality (REHVA, 2002; Bauman, 2003 and Chen, 2003, Chen and Glicksman, 2003; Bauman and
Daly, 2003) and reduce energy use (Hu et al., 1999; Im et al., 2005). Many previous studies (Lin et al.,
2005; Kobayashi and Chen, 2003; and Yuan et al., 1998) indicate that the design parameters, such as selec-
tion of diffusers, supply flow and temperature, perimeter heating and control/operation scheme, have sig-
nificant impact on the ventilation effectiveness and energy performance of stratified air distribution sys-
tems.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Boualem Ouazia, Iain Macdonald, Alexandra Thompson are research officers and Daniel Booth is a technical offi-
cer at the National Research Council Canada, Ottawa, Ontario. Michel Tardif is a manager at Natural Resources Can-
ada.
292
ASHRAE Transactions
DV is a good application for facilities where IAQ is a major concern. Densely occupied spaces with
open floor plans, such as classrooms, restaurants, and theatres are excellent applications. DV has also been
successfully used in open spaces with high ceilings, such as airport terminals, fitness centers, atria, and
casinos. Educational facilities and classrooms in particular, benefit from DV because they require signifi-
cant amounts of outdoor air to maintain acceptable IAQ (Bauman, 2003).
The approach presented in ASHRAE Standards 113-2005 and 129-1997 is often used in the literature
as the reference method for the assessment of air distribution in rooms. The IAQ performance of the air
distribution system is often assessed by using air distribution effectiveness or contaminant removal effec-
tiveness (CRE). CRE characterizes the ability of a system to remove air-borne contaminants.
The latest edition of Standard 62 (ANSI/ASHRAE Standard 62.1-2007) allows some adjustment in
ventilation rates based on the air distribution effectiveness (Ez) of the air distribution system, a feature that
may give credit to UFAD and DV systems. Mixing-type air distribution systems result in a well mixed
space at best, characterized by Ez of 1.0, as determined in accordance with ASHRAE Standard 129-1997.
By definition mixing-type systems cannot provide Ez higher than 1, which corresponds to improved air
change effectiveness at the breathing level in the space. Displacement ventilation systems are known to
provide improved air distribution effectiveness in the occupied zone. In ASHRAE Standard 62.1-2007, the
following default values for air distribution effectiveness are recommended for different air distribution
system configurations and modes of operation:
x 1.2 for displacement ventilation system in cooling mode,
x 0.7 for displacement ventilation system in heating mode,
x 1.0 for an overhead system in cooling mode, and
x 0.8 for an overhead system in heating mode.
The parametric study of ASHRAE Research Project 1373 (Jiang and Chen (RP-1373 2008)) proposed
various values for air distribution effectiveness associated with different systems and operating conditions.
In cooling mode, the CRE with UFAD and DV systems is 1.05 – 1.35 for offices and classrooms. In heat-
ing mode, the CRE for the DV and UFAD systems is between 0.7 – 1.0.
Even though stratified systems are currently in place in North American buildings and in a situation
where systems are being designed and installed at an increasingly rapid pace, a full understanding of their
performance has yet to be achieved. There is some data collected for stratified system operating in cooling
season, but there is no data reported on the performance of such systems during the heating season. Cana-
dian climate is characterized by long and cold winter, which requires the ventilation systems to operate in
both heating and cooling modes. Further on site research (rather than experiment conducted in controlled
environment chamber or simulation work) is required to measure the actual performance of DV with AHU
in heating mode and assess the impact of a perimeter heating system on the performance of a DV system.
In this paper, results from three field studies conducted in schools equipped with displacement ventilation
systems are summarized. The results presented in this paper are limited to the performance of DV related to
indoor air quality.
STUDY BUILDINGS
School 1, located in Calmar south west of the city of Edmonton (Figure 1), is the first school in Alberta
built to be LEED (Leadership in Energy & Environmental Design) Canada-NC 1.0 accredited. School 2 is
located in Scarborough, a residential suburb area of Toronto (Figure 1) and is an award winning school
recognized for its sustainable energy design and is the first “LEED Silver” certified elementary school in
Canada. School 3 is located in Pierrefonds, a residential area in the north-west of Montreal Island (Figure
1) and is also a LEED certified school, recognized for its sustainable energy design.
School 1
The school is a one-story building with a rectangular plan oriented along the East-West axis, with
classrooms located in the east and west wings (see Figure 2). The school’s key sustainability features rele-
vant to indoor environmental quality include: operable windows, displacement ventilation, passive solar
control, and daylighting in all occupied spaces. Daylighting is provided by south-facing upper sealed rib-
bon windows and lower peripheral operable windows. Users have access to light switches, lower peripheral
operable windows but not to thermostat control.
The ventilation system is 100% fresh air (100% OA) with a heat recovery wheel) for preheating and
pre-cooling of the incoming air. The heat recovery wheel is switched on whenever supply fans are operat-
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

ing and there is a requirement for pre-conditioning the outside air using exhaust air. The air is conditioned
through heating/cooling coil and then delivered to the space (classrooms and computer room) through
sidewall diffusers at floor level and exhausted through return grilles located at ceiling level. In heating sea-
son, supplemental heat is delivered to the spaces through hot water shelf radiators located at ceiling level
over the windows on the perimeter.
Figure 1. Field Studies Location.
Figure 2. Plan of school 1 - monitored classrooms (numbered rooms).

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
School 2
The school is a two-storey “H” shaped building oriented along the East-West axis as shown in Figure 3
and has a gross floor area of 5,554 m2 (59,783 ft2). The school has a hybrid ventilation system integrating
DV and natural ventilation (operable windows). The school’s key sustainability features relevant to indoor
environmental quality include: operable windows, DV, passive solar control, and daylighting in all occu-
pied spaces. Users have access to light switches, On/Off ventilation switch and thermostat. Perimeter heat-
ing is provided by a hydronic radiant floor system. Mechanical air supply in classrooms is delivered
through sidewall diffusers located at floor level and exhausted through returns located in the ceilings.
The school is equipped with VAV DV system which consists of two Air Handling Units: AHU 1 for
the kindergarten and second floor north and AHU 2 for first floor and second floor south. The air is condi-
tioned (through a heating coil or cooling coil) and then delivered to the space (classrooms) through sidewall
diffusers at floor level. Each classroom has five diffusers installed in the interior wall at floor level. Three
returns are installed at the ceiling close to the exterior wall. Supplemental heat delivery to classrooms is
achieved through a hydronic radiant floor heating system located on the perimeter. The temperature of the
water in the hydronic system is controlled at a moderate temperature. Heat is then radiated and convected
from the floor to the room and occupants.
School 3
The school is a two-storey building with two attached sections with separate outdoor air intake. The
old section was built in 1969 and equipped with an overhead (mixing) ventilation system. The new section
294
ASHRAE Transactions
completed in 2007 and is equipped with a VAV DV system and a hydronic floor radiant heating system.
The new section of the school is a LEED certified building, recognized for its sustainable energy design.
Plan of the school is shown in Figure 4. The school has a gross floor area of 13,700 m 2 (147,466 ft2) and
has around 1,450 students from kindergarten to grade 1.
3 4
2

Mechanical air supply in classrooms located in the new section is delivered by DV through corner dif-
fusers located at floor level with the exhaust air grilles located in the ceilings; in the old section, air is de-
livered through ceiling diffusers with the exhaust air grilles located in the ceilings. Mechanical air supply in
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Gymnasium, located in the new section, is delivered by DV through sidewall diffusers located at floor level
with exhaust air grilles located in the ceiling. The secondary heat delivery to the two classrooms and gym-
nasium in the new section is achieved through a hydronic radiant floor heating system located on the pe-
rimeter.
New section
PROCEDURE
Measurement Schedule
The field studies took place in April 21st, 2009 for school 1, February 10 th 2010 for school 2 and March
th
10 2011 for school 3. In school 1, measurements were taken at three to four locations over sections of
floor close to exterior wall (windows), in the centre and close to interior wall. In school 2, measurement
locations were over sections of floor without hydronic radiant heating (close to interior wall), over floor
with hydronic radiant heating (close to exterior wall) and at the border between the two sections of floor
(centre of the room). In school 3, measurements were taken at only one central location in each classroom,
using an instrumented main pole. The room orientation, diffusers type and number, occupancy, number of
measurement locations and outdoor temperature for that day are shown in Tables 1, 2 and 3. The typical
diffusers installed in the study buildings are shown in Figure 5.
$6+5$(

Table 1. School 1 – Monitored spaces.
Classroom 1 2 3 4
Orientation North South South South
Diffuser type (#) Wall (3) Wall (3) Wall (4) Wall (3)
No. of pupils 17 12 31 15
No. of measurement locations 4 4 3 3
Supply temperature - °C (°F) 20.5 (68.9) 21.7 (71.0) 22.8 (73.0) 23.7 (74.7)
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Outdoor temperature - °C (°F) 9.7 (45.5) 14.9 (58.8) 19.1 (66.4) 21.2 (70)
Table 2. School 2 - Monitored Spaces.
Classroom 1 2 3 4
Orientation North South-West South South-East
Diffuser type (#) Wall (5) Wall (4) Wall (4) Wall (5)
No. of pupils 28 33 24 20
No. of measurement locations 3 3 3 3
Supply temperature - °C (°F) 19.0 (66.2) 18.0 (64.4) 18.8 (65.8) 18.5 (65.3)
Outdoor temperature - °C (°F) -7.4 (17.7) -7.1 (19.2) -5.4 (22.3) -4.5 (23.9)
Table 3. School 3 - Monitored Spaces.
Classroom 1 2 3 (gymnasium)
Orientation North Centre South
Diffuser type (#) Corner (4) Corner (3) Wall (15)
No. of pupils 27 17 25
No. of measurement locations 1 1 1
Supply temperature - °C (°F) 20.5 (68.9) 19.0 (66.2) 19.0 (66.2)
Outdoor temperature - °C (°F) 0.8 (33.4) 1.0 (33.8) 2.0 (35.6)
School 1
School 2
School 3 School 3 - Gymnasium
Figure 5. Diffusers type.
Measurement Procedure
The performance assessment of the DV systems were conducted during one day and consisted of a
short monitoring period, three periods of three minutes each, for each location in the monitored spaces. The
classrooms were monitored during occupied periods (classrooms 1 and 2 in the morning, 3 and 4 in the
afternoon). Measurements were carried out in compliance with ASHRAE Standards 113-2005, 129-1997
and 55-2010. Instrumented main pole with anemometers at 0.1, 0.6, 1.1 and 1.7 m (0.3. 2.0, 3.6, 5.6 ft),
RTDs at 0.1, 0.6, 1.1, 1.7 and 2.2 m (0.3, 2.0, 3.6, 5.6 and 7.2 ft), CO2 sensors at 1.1 and 1.7 m (3.6 and 5.6
ft), was used for detailed spatial and vertical characterization of the indoor environment. A return pole was
used to measure the air temperature and CO2 level at the return grilles in the rooms and a reference pole
was used to measure the reference air temperature and relative humidity at a height of 1.1 m (0.3 ft) above
the floor in the room. The sensors chosen for this study meet the accuracy requirements set by ASHRAE
standard 113-2005. The CO2 sensor accuracy is ±30 ppm + 2% of the reading, temperature sensor (RTD)
accuracy is ±0.1°C and accuracy of anemometer is summation of ±1% full scale and ±1.5% of the reading.
ANALYSIS
The performance of DV systems in the three schools were assessed using three criteria: (1) thermal
comfort – vertical air temperature difference and draft ratio and (2) IAQ – CRE. Due to limited space, only
CRE values are presented and discussed in this paper.
296
ASHRAE Transactions
Both air change effectiveness or contaminant removal effectiveness are frequently used to characterize
ventilation effectiveness. Air change effectiveness) can be represented by the mean age of air:
ሺெ௘௔௡஺௚௘஺௜௥ሻ೐ೣ೓ೌೠೞ೟
‫ܧ‬௭ ൌ ሺெ௘௔௡஺௚௘஺௜௥ሻ (1)
್ೝ೐ೌ೟೓೔೙೒೥೚೙೐
For a perfectly mixed system the air distribution effectiveness is 1. DV usually has an average lower
mean age of air in the breathing zone, an indication that supply air is more efficiently distributed to occu-
pants.
When used to describe effectiveness at removing airborne contaminants, it is defined by:
஼೐ ି஼ೞ
‫ ܧܴܥ‬ൌ (2)
஼೚೥ ି஼ೞ
Where Cs, Ce, Coz are the CO2 concentrations in the supply air, exhaust air, and occupied zone, respec-
tively.
In this section CRE values are presented to address indoor air quality. CO 2 concentrations were meas-
ured at two heights, breathing height of seated and standing person (1.1 m (3.6 ft) and 1.7 m (5.6 ft)), dif-
fuser (floor level) and return grille (ceiling level), respectively. CRE indicates the efficiency of the
ventilation process in terms of controlling the occupants’ exposure to an indoor contaminant emitted at
locations spatially distributed within aclassroom. The measured CO2 concentrations were used for CRE
calculation, which should be representative of the CRE for other occupant-generated pollutants. Since CO2
measurements were made at the breathing height of seated and standing adults, for each measurement loca-
tion within a classrooms two corresponding values of local CRE were calculated, one based on CO 2 meas-
urement at seated breathing height and one based on CO2 measurements at standing breathing height – both
calculations also used CO2 data from the supply diffuser and return grille.
The field measurements at school 1 were carried out under spring outdoor conditions with the school
HVAC system functioning in a heating mode and the measured air supply temperature at diffusers varied
between 20.5°C (68.9°F) and 23.7°C (74.7°F). The field measurements at school 2 were carried out under
winter outdoor conditions with the school HVAC system functioning in a heating mode and the measured
air supply temperature at diffusers varied between 18°C (64.4°F) and 23°C (73.4°F). The field measure-
ments at school 3 were carried out under winter outdoor conditions with school HVAC functioning in a
heating mode and measured air supply temperature at diffusers varied between 19.0°C (66.2°F) and 20.5°C
(68.9°F).
The average CRE based on CO2 measurements at the breathing height of seated adults 1.1 m (3.6 ft),
standing adult 1.7 m (5.6 ft) and at the nearby supply diffuser and return grille are presented for both
schools in Figure 6, Figure 7 and Figure 8. CRE at the three schools were better than what has been re-
ported for DV systems in heating mode (Jiang et al. 2008) with the exception of classroom 4 in school 1
where the measurement was taken in the afternoon with outdoor temperature at 21.1°C (70°F). Classroom 4
in school 1 was supplied with 100% outdoor air at 23.7 ºC (74.7°F), which resulted in poor ventilation ef-
fectiveness in the occupied zone (CRE<1).
The accuracy in CRE (absolute values) is presented in Figure 9, which are comparable to the uncer-
tainty estimated by ASHRAE Standard 129-1997 on measuring air-change effectiveness in field-based
measurements of ±0.16.
Figure 10 plots the CRE values for two heights, breathing heights of seated (1.1 m/3.6 ft) and standing
person (1.7 m/5.6 ft), in each monitored school with respect to the difference between supply air tempera-
ture and the mean value of temperature measurements at several locations/heights. The results show that
when the supply temperature air is cooler than the room air, the DV system in a heating mode was able to
achieve satisfactory contaminant removal effectiveness in the space, represented by CRE generally greater
than 0.90.
Comparing the IAQ index CRE results to values previously reported in the literature, it is clear that the
CRE values measured in theses field studies were better than what has been reported for DV systems in
heating mode (Jiang et al. 2008). As a result of their parametric study, ASHRAE Research Project 1373
proposes various values for air distribution effectiveness. In heating mode with air supplied at 28 to 30°C
(82.4°F to 86.0°F), the CRE for the DV and UFAD systems was 0.70 – 1.0 for indoor spaces. The results of
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
their simulation study for classrooms showed that the DV system did not perform as well as during cooling
mode. Field studies in heating season with supplied air as low as 18°C (64.4°F) coupled with supplemental
$6+5$(

perimeter heating showed CRE values often higher than 1 which do not support CRE ranges proposed by
Jiang et al. in heating mode.
2.0 2.0
1.8 1.1 m 1.8 1.1 m
1.6 1.7 m 1.6 1.7 m
1.4 1.4
1.2 1.2
CRE (-)
CRE (-)
1.0 1.0
0.8 0.8
0.6 0.6
0.4 0.4
0.2 0.2
0.0 0.0
1 2 3 4 1 (1) 1 (2) 2 (1) 2 (2) 3
Classroom
Classroom
Figure 6. Classroom average CRE at school 1.
Figure 8. Classroom average CRE at school 3.
2.0
0.50
1.8 1.1 m 0.45
1.6 1.7 m
0.40
1.4
0.35
1.2
CRE (-)
0.30
Error (-)
1.0
0.25
0.8
0.20 school 1 - 1.1 m
0.6 school 1 - 1.7 m
0.4 0.15 school 2 - 1.1 m
0.10 school 2 - 1.7 m
0.2 school 3 - 1.1 m
0.0 0.05 school 3 - 1.7 m
1 2 3 4
0.00
Classroom 0.0 0.3 0.6 0.9 1.2 1.5 1.8 2.1 2.4 2.7 3.0
Figure 7. Classroom average CRE at school 2. CRE (-)
Figure 9. Accuracy in CRE.
The current field studies results showed clearly how the supply air temperature has a major impact on
the CRE resulting from a DV system. As shown in Figure 10, when the supply air is cooler than the room
air and the DV system is coupled with a perimeter heating system, the DV system in heating mode was able
to achieve satisfactory IAQ in the occupied zone characterized by CRE values higher than 0.9.
CONCLUSIONS
These three field studies have shown that displacement ventilation can provide satisfactory indoor air
quality in the occupied zone during the heating season. This was achieved by heating the space using both
the ventilation system and the low temperature heating systems, hydronic shelf radiators and hydronic floor
heating, as perimeter heating system.
The air quality index (CRE) was found to be higher than what was previously reported for DV system
in heating mode. These results contradict previous findings on DV performance, which states that the CRE
index is below 1 for DV in heating mode and that the DV system does not perform as well as during cool-
ing mode. These three field studies where the displacement ventilation systems were supplying air to the
spaces at fairly low temperatures and coupled with a perimeter heating systems showed good performance
of a displacement ventilation system in heating season.
In these studies, CRE was often higher than unity and at worst very close to unity, which is the value
typically observed with conventional overhead air supply systems. The field studies show an opportunity to
save energy by reducing the rate of outdoor air supply rate because of the high value of contaminant re-
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
298
ASHRAE Transactions
moval effectiveness associated with DV systems in heating season. The CRE range for DV systems with
low supply air temperature and supplemental heating system should be 1.0 to 1.2.
2.0
1.8
1.6
1.4 school 1 (1.1 m)
1.2 school 1 (1.7 m)
CRE (-)
1.0 school 2 (1.1 m)
0.8 school 2 (1.7 m)
0.6 school 3 (1.1 m)
0.4 school 3 (1.7 m)
0.2
0.0
-2 -1 0 1 2 3 4 5
T air- Tsupply (K)
Figure 10. CRE versus temperature difference between the supply and room air.
ACKNOWLEDGEMENTS
This project was supported by funding provided by both the federal government department of Natural
Resource Canada and by the National Research Council Canada.
REFERENCES
ASHRAE Standard 62.1 (2007) Ventilation for acceptable indoor air quality. Atlanta: American Society of
Heating Refrigerating and Air-conditioning Engineers, Atlanta, Inc.
ASHRAE Standard 55 (2010) “Thermal environmental conditions for human occupancy” American Soci-
ety of Heating, Refrigerating and Air-Conditioning Engineers, Inc.
ASHRAE Standard 113 (2005) “Method of testing for room air diffusion” American Society of Heating,
Refrigerating and Air-Conditioning Engineers, Inc.
ASHRAE Standard 129 (1997) (RA 2002) “Measuring air change effectiveness” American Society of
Heating, Refrigerating and Air-Conditioning Engineers, Inc.
Bauman F S (2003) “Underfloor Air Distribution (UFAD) Design Guide” American Society of Heating,
Refrigerating and Air-Conditioning Engineers, Inc.
Chen Q and Glicksman L (2003) “System Performance Evaluation and Design Guidelines for Displace-
ment Ventilation” American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc.
Hu, S., Q. Chen and L. Glicksman (1999). Comparison of energy consumption between displacement and
mixing ventilation systems for different US buildings and climates. ASHRAE Transactions 105(2),
pp. 453-464.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Im, P., A. Cho, and J. Haberl (2005). Literature review on under floor air distribution (UFAD) system.
Technical Report of the Energy Systems Laboratory, Texas A&M University System, USA.
Jiang Z. and Chen Q (2008) “Air distribution effectiveness with stratified air distribution systems” ASH-
RAE Research Project – 1373.
Kobayashi, N., and Q. Chen (2003). Floor Supply Displacement Ventilation in a Small Office. Indoor and
Built Environment. 12(4), pp. 281-291.
LEED Canada-NC 1.0 (2004) “LEED Canada Reference Guide” Canada Green Building Council.
Lin, Z., T.T. Chow, K.F. Fong, Q. Wang, and Y.Li (2005) Comparison of performance of displacement and
mixing ventilations. Part I: thermal comfort. International Journal of Refrigeration 28, pp. 276-287.
Lin, Z., T.T. Chow, K.F. Fong, C.F. Tsang, and Q. Wang (2005) Comparison of performance of displace-
ment and mixing ventilations. Part II: IAQ. International Journal of Refrigeration 28, pp. 288-305.
Skistad, H., Mundt, E., Nielsen, P.V., Hagstrom, K., Railio, J. (2002) “Displacement ventilation in non-
industrial premises” REHVA, Federation of European Heating and Air-conditioning Associations.
Yuan, X., Q. Chen, and L. Glicksman (1998). Critical review of displacement ventilation. ASHRAR Trans-
actions. 104(1A), pp. 78-89.
$6+5$(

SA-12-C038
6XSSO\$LU&2%DVHG'HPDQG&RQWUROOHG
9HQWLODWLRQIRU0XOWL=RQH+9$&6\VWHPV

1DELO1DVVLI3K'3(
$6+5$(0HPEHU
$%675$&7
7KH SDSHU GLVFXVVHV D QHZ YHQWLODWLRQ FRQWURO VWUDWHJ\ EDVHG RQ VXSSO\DLU &2 FRQFHQWUDWLRQ IRU PXOWL]RQH +9$&
V\VWHPV7KHVWUDWHJ\RIIHUVDJUHDWRSSRUWXQLW\WRUHGXFHHQHUJ\FRQVXPSWLRQLQ+9$&V\VWHPVZKLOHSURYLGLQJWKHUHTXLUHG
YHQWLODWLRQDFFRUGLQJWR$6+5$(VWDQGDUG7KHVWUDWHJ\UHTXLUHVD3,RU3,'FRQWUROOHUDORQJZLWKD&2VHQVRU
ORFDWHG LQ WKH PDLQ VXSSO\ DLU GXFW WR FRQWURO WKH VXSSO\ DLU FRQFHQWUDWLRQ E\ DGMXVWLQJ WKH RXWGRRU DLU GDPSHU ,W DOVR
LQFOXGHVDFRQWURODOJRULWKPWRG\QDPLFDOO\UHVHWWKH&2FRQFHQWUDWLRQVHWSRLQWRQDSURSHUWLPHLQWHUYDOEDVLVDQGXVLQJ
WKHDFWXDO]RQHDLUIORZUDWHV7KHVWUDWHJ\LVHYDOXDWHGE\VLPXODWLRQVRQPXOWL]RQHRIILFHEXLOGLQJFRQGLWLRQHGE\DW\SLFDO
9$9V\VWHPDQGXQGHUYDULRXV86$ORFDWLRQV7KHUHVXOWVVKRZWKDWDVLJQLILFDQWVDYLQJFRXOGEHREWDLQHGE\LPSOHPHQWLQJ
WKHSURSRVHGVWUDWHJ\DVFRPSDUHGWRWKHGHVLJQRFFXSDQF\$6+5$(6WDQGDUGSURFHGXUHDQGWKLVVDYLQJGHSHQGV
PDLQO\RQDFWXDORFFXSDQF\SURILOHDQGEXLOGLQJORFDWLRQV
,1752'8&7,21
6HYHUDOYHQWLODWLRQFRQWUROVWUDWHJLHVDUHSURSRVHGIRU+9$&V\VWHPV&DUSHQWHU:DQJDQG;XDQG1DVVLI
HWDOLQRUGHUWRLPSURYHWKHTXDOLW\RIDLULQVLGHEXLOGLQJVDQGUHGXFHWKHDVVRFLDWHGHQHUJ\XVH7KHVHPHWKRGVZKLFK
KDYHEHHQGLVFXVVHGLQPDQ\SXEOLFDWLRQVPD\RUPD\QRWVDWLVI\WKHQHZUHTXLUHPHQWVRI$6+5$($6+5$(
7KH &2EDVHG GHPDQGFRQWUROOHG YHQWLODWLRQ '&9 LV RQH RI WKH VWUDWHJLHV WKDW FRXOG ORZHU HQHUJ\ XVH E\
UHGXFLQJRYHUYHQWLODWLRQRIEXLOGLQJV$ODODZLDQG.UDUWL7D\ORU6WDQNH&RQJUDGDFDQG.XOLF/XHW
DO:KHQVXFKVWUDWHJ\LVDSSOLHGE\GHWHFWLQJWKH &2FRQFHQWUDWLRQLQUHWXUQDLULW PD\UHVXOWLQSRRU DLUTXDOLW\
LQVLGHFHUWDLQ]RQHVLQDPXOWL]RQHV\VWHP,QDGGLWLRQPRVW'&9VWUDWHJLHVDUHEDVHGRQIORZUDWHSHUSHUVRQ$6+5$(
6WDQGDUG WKURXJK DQG WKRVH PD\ QRW FRPSO\ ZLWK WKH QHZ YHQWLODWLRQ UHTXLUHPHQWV RI $6+5$( 6WDQGDUG
WKURXJKZKLFKUHTXLUHVWZRYHQWLODWLRQUDWHVRQHLQWHQGHGWRGLOXWHWKHFRQWDPLQDQWVJHQHUDWHGE\RFFXSDQWV
DQGRWKHUIRUEXLOGLQJUHODWHGVRXUFHV%HFDXVHRIWKRVHWZRGLIIHUHQWYHQWLODWLRQUDWHVWKHUHTXLUHGVSDFH&2FRQFHQWUDWLRQ
RUWKHLQGRRURXWGRRUGLIIHUHQFHLVQRORQJHUFRQVWDQWDVLWZDVLQ$6+5$(6WDQGDUGWKURXJK6WDQNH
0XUSK\1DVVLI7KHUHTXLUHGVSDFH&2FRQFHQWUDWLRQYDULHVZLWKWKHRFFXSDQWVPDNLQJDQ\&2EDVHG'&9
VWUDWHJ\ KDUG WR DSSO\ DQG FRPSO\ H[DFWO\ ZLWK WKH UHFRPPHQGDWLRQV RI WKH 6WDQGDUG $OWKRXJK WKH VWDQGDUG
SHUPLWVDSSO\LQJ&2EDVHG'&9VWUDWHJ\RUDQ\G\QDPLFYHQWLODWLRQUHVHWVWUDWHJ\FRQVLGHUDEOHFKDOOHQJHVZLWKPXOWL]RQH
+9$&V\VWHPVDULVH,QDGGLWLRQDVLWLVFRVWO\DQGGLIILFXOWEXWQRWLPSRVVLEOHWRHVWLPDWHDFFXUDWHO\WKHDFWXDORFFXSDQWV
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
LQHDFKVSDFHWKHVWDQGDUGSURFHGXUHLVPRVWO\EDVHGRQWKHGHVLJQRFFXSDQF\SURILOHOHDGLQJWRRYHUYHQWLODWHWKHVSDFHV
\LHOGLQJOHVVWKDQGHVLJQRFFXSDQWVDQGFRQVHTXHQWO\ZDVWHRIHQHUJ\$VDUHVXOWRIWKHVHFKDOOHQJHVDQDOWHUQDWLYHVXSSO\
*UHHQVERUR1&
300
ASHRAE Transactions
DLU&2EDVHG'&9VWUDWHJ\6$'&9LVSURSRVHGWRPDLQWDLQWKH&2FRQFHQWUDWLRQLQVXSSO\DLUDORZHQRXJKWRPHHWWKH
YHQWLODWLRQUHTXLUHPHQWVLQDOO]RQHVDQGLPSURYHHQHUJ\HIILFLHQF\$OWKRXJKWKHSURSRVHGVWUDWHJ\XVHVGHVLJQRFFXSDQF\
$6+5$(SURFHGXUHWRGHWHUPLQHWKHVXSSO\DLU&2VHSRLQWWKHVWUDWHJ\LVVWLOODEOHWRIROORZVRPHZKDWWKHDFWXDO
RFFXSDQF\ SDWWHUQ GXH WR WKH ORZHU &2 UHWXUQ DLU FRQFHQWUDWLRQ ZLWK RIIGHVLJQ RFFXSDQF\ 7KLV VXSSO\ DLU &2EDVHG
YHQWLODWLRQ VWUDWHJ\ ZDV SURSRVHG E\ 1DVVLI EDVHG RQ WKH FULWLFDO ]RQHV DQG YHQWLODWLRQ HIILFLHQF\ ,Q WKLV SDSHU DQ
LPSURYHPHQWZLOOEHGRQHRQWKHUHVHWDOJRULWKPRIWKHVXSSO\DLU&2FRQFHQWUDWLRQVHWSRLQWE\XVLQJWKHGHVLJQRFFXSDQF\
$6+5$(6WDQGDUGPXOWL]RQHYHQWLODWLRQSURFHGXUH
352326('675$7(*<
,Q PXOWL]RQH +9$& V\VWHPV WKH VXSSO\ DLU WHPSHUDWXUH FRQWURO PDLQWDLQV WKH VXSSO\ DLU FROG HQRXJK WR PHHW WKH
WKHUPDOORDGLQVSDFHV6LPLODUO\WKHVXSSO\DLU&2FRQWURODVSURSRVHGLQWKLVSDSHUPDLQWDLQVWKH&2FRQFHQWUDWLRQLQ
VXSSO\DLUORZHQRXJKWRPHHWWKHYHQWLODWLRQORDGLQVSDFHV,QWKLVDVSHFWWKHSURSRVHGVWUDWHJ\FDOOHG6$'&9UHTXLUHVD
ORFDO3,RU3,'FRQWUROOHUDORQJZLWKD&2VHQVRUORFDWHGLQWKHVXSSO\DLUGXFWDVVKRZQLQ)LJXUH7KHFRQWUROOHURXWSXW
LVGHWHUPLQHGE\FRPSDULQJWKHPHDVXUHG&2FRQFHQWUDWLRQLQWKHVXSSO\DLUZLWKLWVVHWSRLQWWRPRGXODWHWKHRXWGRRUDLU
GDPSHUDFFRUGLQJO\7KHVXSSO\DLU&2FRQFHQWUDWLRQVHWSRLQWFRXOGEHDOZD\VIL[HGEDVHGRQGHVLJQFRQGLWLRQVLQDZD\WR
PHHW WKH SHDN YHQWLODWLRQ ORDG EXW WKLV ZRXOG EH LQHIILFLHQW 7R WUDFN WKH DFWXDO DLUIORZ UDWHV DQG RFFXSDQF\ LQ VSDFHV D
FRQWURODOJRULWKPLVQHHGHGWRG\QDPLFDOO\UHVHWWKHVXSSO\DLU&2FRQFHQWUDWLRQVHWSRLQWRQDSURSHUWLPHLQWHUYDOEDVLV
HJHDFKPLQXWH7KHDOJRULWKPXVHVDFWXDO]RQHDLUIORZUDWHUHDGLQJVWKDWDUHPRVWO\DYDLODEOHLQ''&V\VWHPEXWZLWK
WKHGHVLJQRFFXSDQF\SURILOHDVZHDVVXPHGWKDWWKHLQIRUPDWLRQRQWKHDFWXDORFFXSDQF\LVQRWDYDLODEOH'LIIHUHQWPHWKRGV
FRXOGEHXVHGWRGHYHORSVXFKDOJRULWKPZLWKWKHJRDORISURYLGLQJYHQWLODWLRQUDWHDVSRVVLEOHFORVHWRWKDWUHFRPPHQGHGE\
WKH$6+5$(6WDQGDUG,QHDUO\SXEOLFDWLRQ1DVVLIWKHDOJRULWKPZDVSURSRVHGEDVHGRQWKHFULWLFDO]RQH
DQGYHQWLODWLRQHIILFLHQF\,QWKLVSDSHUWKHDOJRULWKPZLOOEHEDVHGVLPSO\RQWKH³PLQLPXPYHQWLODWLRQUDWH9RW´FDOFXODWHG
E\WKH$6+5$(VWDQGDUGPXOWL]RQHYHQWLODWLRQSURFHGXUHDVGLVFXVVHGEHORZ
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
7KH PDVV EDODQFH HTXDWLRQ WR SUHGLFW WKH GLIIHUHQFH EHWZHHQ UHWXUQ DLU &2 FRQFHQWUDWLRQ &U DQG RXWGRRU &2
FRQFHQWUDWLRQ&RDWVWHDG\VWDWHFRQGLWLRQV
ܰ௕
ܸ௢௧ ൌ ሺͳሻ
ሺ‫ܥ‬௥ െ ‫ܥ‬௢ ሻ
8VLQJWKH&2FRQFHQWUDWLRQLQVXSSO\DLU&VDQGV\VWHPVXSSO\DLUIORZ9SVWKHVWHDG\VWDWHPDVVEDODQFHEHFRPHV
ܰ௕
ܸ௣௦ ൌ ሺʹሻ
ሺ‫ܥ‬௥ െ ‫ܥ‬௦ ሻ

7RSURYLGHWKHPLQLPXPV\VWHPRXWGRRUDLUIORZ9RWUHTXLUHGE\WKHVWDQGDUGWKH&2FRQFHQWUDWLRQLQVXSSO\DLU&VLV
JLYHQXVLQJE\(TXDWLRQDQG
ͳ ͳ
‫ܥ‬௦ ൌ ‫ܥ‬௢ ൅ ܰ௕ ቆ െ ቇሺ͵ሻ
ܸ௢௧ ܸ௣௦

7KH1ELVWKH &2 JHQHUDWLRQUDWH1E &î3EDQGLWLVD IXQFWLRQRISHRSOHQXPEHU3E LQHQWLUHVSDFHVFRYHUHGE\D
SDUWLFXODUV\VWHP7KH&LVDFRQVWDQWYDOXHUHODWHGWRWKHRFFXSDQF\DFWLYLWLHVOHYHOGLHWKHDOWKDQGHWF
(TXDWLRQUHSUHVHQWVWKHUHTXLUHGYDOXHRIWKHVXSSO\DLU&2FRQFHQWUDWLRQVHWSRLQWWRSURYLGHWKHPLQLPXPV\VWHP
RXWGRRU DLU IORZ 9RW GHWHUPLQHG E\ WKH PXOWL]RQH YHQWLODWLRQ SURFHGXUH RI $6+5$( VWDQGDUG FRQVLGHULQJ WKH
V\VWHPYHQWLODWLRQHIILFLHQF\8QGHUVWHDG\VWDWHFRQGLWLRQVDQGGHVLJQRFFXSDQF\SURILOHWKHSURSRVHGVWUDWHJ\6$'&9
E\PDLQWDLQLQJWKHVXSSO\DLU&2FRQFHQWUDWLRQDVFDOFXODWHGE\(TXDWLRQSURYLGHVH[DFWDPRXQWRIYHQWLODWLRQUDWH9RW
UHTXLUHGE\WKHVWDQGDUG+RZHYHUXQGHURIIGHVLJQRFFXSDQF\FRQGLWLRQVDQGDVLWLVKDUGLQSUDFWLFHWRHVWLPDWHWKHDFWXDO
RFFXSDQF\LQHDFKVSDFHWKHVWDQGDUGSURFHGXUHLVXVXDOO\DSSOLHGXVLQJWKHGHVLJQRFFXSDQF\DQGWKHUHE\LWZLOOUHVXOWLQ
RYHUYHQWLODWLRQLQFHUWDLQVSDFHV\LHOGLQJOHVVWKDQGHVLJQRFFXSDQWV&RQYHUVHO\DOWKRXJKWKHSURSRVHGVWUDWHJ\XVHVGHVLJQ
RFFXSDQF\ GHVLJQRFFXSDQF\ $6+5$( VWDQGDUG SURFHGXUH WR GHWHUPLQH WKH &2 VXSSO\ VHW SRLQW WKH UHWXUQ &2
$6+5$(

FRQFHQWUDWLRQDQGWKHQYHQWLODWLRQUDWHZLOOEHOHVVWKDQGHVLJQYDOXH+LJKHU³XQXVHGÝHQWLODWLRQDLUIURPRYHUYHQWLODWHG
VSDFHV SURGXFHV ORZHU UHWXUQ DLU &2 FRQFHQWUDWLRQ DQG FRQVHTXHQWO\ LQWURGXFHV OHVV DPRXQW RI IUHVK DLU WR PDLQWDLQ D
VSHFLILF VXSSO\ DLU &2 VHW SRLQW 7KLV LV LPSRUWDQW DGYDQWDJH RI &2EDVHG FRQWURO FRPSDULQJ WR WKH XVH RI $6+5$(
VWDQGDUGFDOFXODWLRQV2WKHUDGYDQWDJHLVE\XVLQJDVLPSOH3,FRQWUROOHUWRPRGXODWHWKHRXWGRRUDLUGDPSHUDQGWKHQ
DYRLGLQJWKHXVHRIDLUIORZPHWHUORFDWHGDWRXWGRRUDLUGXFW7KHPDLQLVVXHRIWKH6$'&9LVWKHSRVVLELOLW\RIUHFLUFXODWLQJ
DSRUWLRQRIWKHEXLOGLQJUHODWHGYHQWLODWLRQDVWKHUHWXUQDLUFRQWDLQVDPL[WXUHRIERWKRFFXSDQF\DQGEXLOGLQJYHQWLODWLRQ
UDWHV7RDYRLGWKLVSUREOHPDVZLOOGLVFXVVHGODWHUDPLQLPXPGDPSHUSRVLWLRQOLPLWVKRXOGEHHVWDEOLVKHGWRSURYLGHIUHVK
DLUQROHVVWKDQIRUH[DPSOHRIGHVLJQRFFXSDQWV7KLVZLOOHQVXUHVXSSO\LQJYHQWLODWLRQUDWHKLJKHUWKDQWKHUHTXLUHG
EXLOGLQJUHODWHGRQH
ZĞƚƵƌŶĨĂŶ
ĂŵƉĞƌƐ
DŽƚŽƌ
ϭ Ϯ ŝ
&ŝůƚĞƌĂŶĚĐŽŽůŝŶŐĂŶĚ
ŚĞĂƚŝŶŐĐŽŝůƐ ^ƵƉƉůǇĨĂŶ s
Ěǌϭ sĚǌϮ sĚǌŝ

ŽŶƚƌŽůƐŝŐŶĂů >ŽĐĂůW/ŽŶƚƌŽůůĞƌ KϮ^ƵƉƉůǇĂŝƌĐŽŶĐĞŶƚƌĂƚŝŽŶ
KϮ^ƵƉƉůǇĂŝƌ
^ƵƉĞƌǀŝƐŽƌǇĐŽŶƚƌŽůůĞƌƋƵĂƚŝŽŶϯ ŽŶĞĂŝƌĨůŽǁƌĂƚĞ
ƐĞƚƉŽŝŶƚ
sƚŽĚĞƚĞƌŵŝŶĞĚďĂƐĞĚŽŶĚĞƐŝŐŶŽĐĐƵƉĂŶĐǇ
ƌĞĂĚŝŶŐƐsĚǌŝ
ƵƐŝŶŐ^,Z^ƚĂŶĚĂƌĚƉƌŽĐĞĚƵƌĞ
)LJXUH 6FKHPDWLFRIVXSSO\DLU&2FRQFHQWUDWLRQFRQWUROVWUDWHJ\ZLWKDW\SLFDO9$9V\VWHP
5(68/76
$RQHVWRU\RIILFHEXLOGLQJLVXVHGIRUWKHVLPXODWLRQVDQGDQDO\VLV,WLVDUHFWDQJOHIRRWSULQWVKDSHZLWKIORRUDUHDRI
IWPLQFOXGLQJILYH]RQHVHDVWZHVWQRUWKVRXWKDQGFRUHZLWKWKHIROORZLQJVSHFLILFDWLRQVDDUHDVRI
IWIWIWIWDQGIWPPPPDQGPESRSXODWLRQRI
DQGDQGFGHVLJQDLUIORZUDWHVRIFIPFIPFIPFIPFIP/V/V
/V/V/VUHVSHFWLYHO\7KHVLPXODWLRQVZHUHPDGHZLWKWKHIROORZLQJDVVXPSWLRQV7KH]RQHDLU
GLVWULEXWLRQHIIHFWLYHQHVV(]LVVHOHFWHGWREHHTXDOWKH&2JHQHUDWLRQUDWH1RIWKHRFFXSDQWVLVHTXDOWRFIP
/VRI&2SHUSHUVRQWKHRXWGRRUDLU&2FRQFHQWUDWLRQLV330WKH5S FIPSHUVRQDQG5D FIPIW5S
/VSHUVRQ DQG 5D /VP DUH XVHG LQ PLQLPXP EUHDWKLQJ ]RQH RXWGRRU DLU HTXDWLRQ )RU DQDO\VLV SXUSRVH
YDULRXVRSHUDWLRQFRQGLWLRQVSRLQWVDUHUDQGRPO\JHQHUDWHGYDU\LQJWKH]RQHRFFXSDQF\IURPWRRIDGHVLJQ
YDOXHDQGWKH]RQHDLUIORZUDWHVIURPWRRIGHVLJQYDOXH7KHQWKHYHQWLODWLRQUDWHVUHVXOWHGIURPWKHSURSRVHG
VWUDWHJ\6$'&9DUHFRPSDUHGZLWKWKRVHFDOFXODWHGE\$6+5$(6WDQGDUGSURFHGXUHLQWZRZD\VDXVLQJDOZD\V
WKH GHVLJQ RFFXSDQF\ DQG E XVLQJ WKH DFWXDO RFFXSDQF\ ,Q ERWK FDVHV WKH DFWXDO DLUIORZ UDWHV DUH XVHG )LUVW LQ WKH
IROORZLQJVHFWLRQLWLVDVVXPHGWKDWWKHDLUIORZUDWHDQGSRSXODWLRQLQHDFK]RQHYDU\VLPLODUO\VXFKDVRQH]RQHV\VWHPEXW
ZLWK WKH UHVXOWHG FRQVWDQW YHQWLODWLRQ HIILFLHQF\ RI 7KH VLPXODWHG DLU YHQWLODWLRQ UDWHV DUH LOOXVWUDWHG LQ )LJXUH DQG
)LJXUH2QO\KXQGUHGRSHUDWLQJSRLQWVDUHVKRZQIRUWKHFODULW\,QWKH)LJXUHDIRUGLVFXVVLRQSXUSRVHWKHDUHDLVVHWWR
EH]HURRU5DVHWWREH]HURDQGWKHYHQWLODWLRQUDWHLVWKHQRQO\IXQFWLRQRIRFFXSDQWVDVLWZDVLQWKHROGYHUVLRQRIWKH
VWDQGDUG $VWKH6$'&9LV EDVHGRQ &2FRQFHQWUDWLRQ DVDQLQGLFDWLRQRIRFFXSDQF\WKH YHQWLODWLRQUDWHVFDOFXODWHGE\
DFWXDORFFXSDQF\ VWDQGDUG SURFHGXUH DQG E\ 6$'&9 DUH TXLWH VLPLODU $OWKRXJK WKH 6$'&9 LV EDVHG RQ WKH GHVLJQ
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
302
ASHRAE Transactions
RFFXSDQF\ WKH VWUDWHJ\ IROORZV VRPHZKDW WKH DFWXDO RFFXSDQF\ DQG SHUIRUPV VLPLODU WR WKH DFWXDORFFXSDQF\ VWDQGDUG
SURFHGXUH QRW VLPLODU WR WKH GHVLJQRFFXSDQF\ VWDQGDUG SURFHGXUH ,Q )LJXUH E WKH DUHD LV FRQVLGHUHG RU 5D DV
UHFRPPHQGHGE\WKHUHFHQWYHUVLRQRIVWDQGDUG7KHYHQWLODWLRQUDWHGHWHUPLQHGE\WKHDFWXDORFFXSDQF\VWDQGDUGSURFHGXUH
GURSV IRUP FIP WR FIP /V WR /V ZKHQ WKH SRSXODWLRQ FKDQJHV IURP GHVLJQ WR QR SHRSOH LQ
VSDFHV2EYLRXVO\WKHEXLOGLQJVRXUFHYHQWLODWLRQUDWHLVHTXDOWRFIP/V7KHYHQWLODWLRQUDWHGHWHUPLQHGE\
WKH GHVLJQRFFXSDQF\ VWDQGDUG SURFHGXUH LV DOZD\V FRQVWDQW DW FIP /V 7KH 6$'&9 WUDFNV WKH
SRSXODWLRQFKDQJHVHYHQLWLVEDVHGRQWKHGHVLJQRFFXSDQF\7KHYHQWLODWLRQUDWHGURSVTXLFNO\ZLWKWKHSRSXODWLRQDQGLW
JRHVWR]HURLIWKHUHLVQRORZHUOLPLWHVWDEOLVKHGGDVKHGOLQH$VVKRZQLQ)LJXUHEWKHORZHUOLPLWHTXDOVWKHEXLOGLQJ
VRXUFH YHQWLODWLRQ UDWH FIP /V 7KH UHDVRQ RI WKH 6$'&9 VXSSO\LQJ ORZHU YHQWLODWLRQ UDWH WKDQ ZKDW LW LV
UHFRPPHQGHGE\WKH$6+5$(VWDQGDUGLVEHFDXVHRIWKHSRWHQWLDORIUHFLUFXODWLQJWKHEXLOGLQJVRXUFHYHQWLODWLRQUDWH
EDFNWRWKHVXSSO\DLU7KHUHDUHPDQ\ZD\VWRRYHUFRPHRUDOOHYLDWHWKLVSUREOHP7KHVLPSOHZD\LVWRXVHDUHDVRQDEOH
PLQLPXPYDOXHRIRFFXSDQF\RWKHUWKDQ]HURVD\RIGHVLJQ$VVKRZQLQ)LJXUHDDORZHUOLPLWLVIRXQGEDVHGRQ
VWDQGDUG SURFHGXUH XVLQJ RI GHVLJQ RFFXSDQF\ 7KH ORZHU OLPLW FDQ EH GRQH LQ SUDFWLFH E\ FUHDWLQJ DQ RXWGRRU DLU
GDPSHUOLPLWSRVLWLRQ$OWHUQDWLYHO\LWPD\EHHVWDEOLVKHGE\DKLJKVXSSO\&2VHWSRLQWOLPLWGHWHUPLQHGE\(TXDWLRQ
ZLWK WKH 9RW GHWHUPLQHG E\ WKH VWDQGDUG SURFHGXUH XVLQJ RI GHVLJQ RFFXSDQF\ 2WKHU ZD\ FDQ EH GRQH E\ ORZHULQJ
LQWHQWLRQDOO\WKHVXSSO\&2VHWSRLQWE\DFRUUHFWLRQIDFWRUEDVHGRQUHWXUQ&2UHDGLQJDVVKRZQLQ)LJXUHE7KLVIDFWRU
YDULHVOLQHDUO\EHWZHHQWZROLPLWV]HURWRIRULQVWDQW330FDOFXODWHGEDVHGRQ&2VHQVRUUHDGLQJDWUHWXUQDLUVLPLODU
WRWKH$6+5$(XVHU¶VPDQXDOIRULPSOHPHQWLQJ&2EDVHG'&9LQDVLQJOH]RQHV\VWHPDQGGLVFXVVHGLQWKHUHIHUHQFH
0XUSK\2WKHUZD\LVE\VLPSO\XVLQJDIL[HGYDOXHVXFKDV3301DVVLI7KH6$'&9FRQVLGHULQJWKH
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
GHVLJQRFFXSDQF\DQGFRUUHFWLRQIDFWRUEHFRPHVUHDVRQDEO\VLPLODUWRWKHDFWXDORFFXSDQF\VWDQGDUGSURFHGXUH

9HQWLODWLRQUDWH&)0
9HQWLODWLRQUDWH&)0

9HQWLODWLRQUDWH/V
9HQWLODWLRQUDWH/V

3RSXODWLRQ 3RSXODWLRQ
DE
)LJXUH 9HQWLODWLRQUDWHVGHWHUPLQHGE\6$'&9DQG6WDQGDUGSURFHGXUHIRURQH]RQHV\VWHPDVSDFHDUHD
LVQRWFRQVLGHUHG5D EVSDFHDUHDLVVHWWREHIW5D

9HQWLODWLRQUDWH&)0
9HQWLODWLRQUDWH&)0

9HQWLODWLRQUDWH/V
9HQWLODWLRQUDWH/V

6WDQGDUG$FWXDO3RSXODWLRQ
6$'&9
6WDQGDUG'HVLJQ3RSXODWLRQ

DE
)LJXUH 9HQWLODWLRQ UDWHV GHWHUPLQHG E\ 6$'&9 DQG 6WDQGDUG SURFHGXUH IRU RQH ]RQH V\VWHP D 6$'&9
DSSOLHGZLWKDORZHUOLPLWEDVHGRQRIGHVLJQRFFXSDQF\E6$'&9DSSOLHGZLWKDORZHUOLPLWDQGFRUUHFWLRQIDFWRU

$6+5$(
,Q PXOWL]RQH V\VWHPV WKH YHQWLODWLRQ HIILFLHQF\ YDULHV ZLWK ]RQH DLUIORZ UDWHV DQG WKH RSHUDWLQJ SRLQWV ZLOO EH ZLGHO\
VFDWWHUHGDVVKRZQLQ)LJXUH$VDUHVXOWRIWKHYHQWLODWLRQHIILFLHQF\YDULDWLRQWKHYHQWLODWLRQUDWHVE\6$'&9EHFRPHVD
OLWWOHFORVHUWRWKHDFWXDORFFXSDQF\VWDQGDUGSURFHGXUHDVVKRZQLQ)LJXUHD+RZHYHUDVGLVFXVVHGEHIRUHDORZHUOLPLW
DQGDFRUUHFWLRQIDFWRUFRXOGEHDOVRDSSOLHGDVVKRZQLQ)LJXUHE7KH6$'&9SHUIRUPVWKHQDTXLWHVLPLODUWRWKHDFWXDO
RFFXSDQF\VWDQGDUGSURFHGXUHHYHQWKH6$'&9XVHVWKHGHVLJQRFFXSDQF\LQWKHFDOFXODWLRQVRIWKHVXSSO\&2VHWSRLQW

9HQWLODWLRQUDWH&)0
9HQWLODWLRQUDWH&)0

9HQWLODWLRQUDWH/V
9HQWLODWLRQUDWH/V

6WDQGDUG$FWXDO3RSXODWLRQ 6WDQGDUG$FWXDO3RSXODWLRQ
6$'&9 6$'&9
6WDQGDUG'HVLJQ3RSXODWLRQ 6WDQGDUG'HVLJQ3RSXODWLRQ

DE
)LJXUH 9HQWLODWLRQ UDWHV GHWHUPLQHG E\ 6$'&9 DQG WKH $6+5$( VWDQGDUG SURFHGXUH IRU PXOWL]RQH
V\VWHPVD6$'&9DSSOLHGZLWKDORZHUOLPLWEDVHGRQ]HURRFFXSDQF\E6$'&9DSSOLHGZLWKDORZHUOLPLWEDVHGRQ
RIGHVLJQRFFXSDQF\DQGFRUUHFWLRQIDFWRU

,QWKHQH[WVHFWLRQIXOO\HDURSHUDWLQJFRQGLWLRQVFRYHULQJVHYHUDO86$FOLPDWH]RQHVDUHFRQVLGHUHG7KHHQHUJ\VLPXODWLRQ
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
VRIWZDUHH4XHVWLVXVHGWRJHQHUDWHWKHKRXUO\VSDFHORDGVIRUWKRVHILYH]RQHVDQGDVHSDUDWH9$9PRGHOEDVHG$6+5$(
+9$&7RRONLW%UDQGHPXHOHWDOLVXVHGWRGHWHUPLQHWKHORDGRQFRLOV7KHFRROLQJHQHUJ\XVHLVIRXQGEDVHGRQ
WKH'2(FKLOOHUPRGHO+\GHPDQHWDO$OLQHDULQWHUSRODWLRQLVXVHGWRILQGWKHVSDFHORDGRQDVPDOOHULQWHUYDO7KH
VWXGLHV ZLOO EH OLPLWHG WR HVWLPDWH FRROLQJ HQHUJ\ VDYLQJ REWDLQHG E\ DSSO\LQJ WKH 6$'&9 DJDLQVW WKH GHVLJQRFFXSDQF\
SURFHGXUHRI$6+5$(6WDQGDUG7KHVWXG\FRQVLGHUVIRXUGLIIHUHQWRFFXSDQF\SDWWHUQVGXULQJWKHRFFXSLHGSHULRGIURP
$0WR30DVLOOXVWUDWHGLQ)LJXUH$WKHRFFXSDQWVLQVSDFHVDUHDOZD\VIL[HGDWGHVLJQYDOXHVQRVFKHGXOHDSSOLHG
DQGWKHZKROHEXLOGLQJSRSXODWLRQRI%DW\SLFDORFFXSDQF\SURILOHLVXVHGLQHDFKVSDFH&RFFXSDQF\SURILOHLV
RIWKHW\SLFDOSURILOH%DQG'WKHRFFXSDQF\SURILOHLVRIWKHSURILOH%7KHSURILOHUDWLRLVEDVHGRQWKHGHVLJQ
VSDFHRFFXSDQF\LHDQG)LJXUHVKRZVWKHUHVXOWVIRUWKHGHVLJQGD\XVLQJW\SLFDOZHDWKHUFRQGLWLRQV
RI*UHHQVERUR1&7KHVXSSO\DLU&2FRQFHQWUDWLRQVHWSRLQWLVG\QDPLFDOO\UHVHWHYHU\PLQEDVHGRQHTXDWLRQDQG
XVLQJWKHYHQWLODWLRQUDWH 9RW GHWHUPLQHGE\GHVLJQRFFXSDQF\ VWDQGDUGSURFHGXUH7KXVWKHVXSSO\ &2FRQFHQWUDWLRQVHW
SRLQWGRHVQRWYDU\ZLWKWKHRFFXSDQF\FKDQJHVQRFRUUHFWLRQIDFWRULVDSSOLHG7KHG\QDPLFYDULDWLRQRI&2FRQFHQWUDWLRQ
LQWKHUHWXUQDLUGXFWLVVLPXODWHGDQGDOVRLOOXVWUDWHGLQ)LJXUHD$G\QDPLF PRGHOEDVHGRQ&2FRQFHQWUDWLRQEDODQFH
HTXDWLRQILUVWRUGHUGLIIHUHQWLDOHTXDWLRQZLWKWLPHGHOD\LVXVHGWRILQGWKH&2FRQFHQWUDWLRQLQHDFKVSDFHDQGWKHZKROH
EXLOGLQJ $OWKRXJK WKH VXSSO\ DLU &2 FRQFHQWUDWLRQ GRHV QRW YDU\ ZLWK WKH DFWXDO QXPEHU RI SHRSOH WKH UHWXUQ &2
FRQFHQWUDWLRQ DQG FRQVHTXHQWO\ WKH RXWGRRU DLUIORZ EHFRPH OHVV WKDQ WKH GHVLJQ YDOXHV ZKHQ WKH DFWXDO RFFXSDQF\ LV OHVV
WKDQGHVLJQDVLQ%&RU'7KH6$'&9LVDEOHWRUHGXFHWKHDPRXQWRIRXWGRRUDLUDVDIXQFWLRQRIWKHDFWXDORFFXSDQF\
304
ASHRAE Transactions
)LJXUH E DQG WKHUHE\ UHGXFLQJ KHDWLQJ DQG FRROLQJ HQHUJ\ ORDGV ,I WKHUH LV QR ORZHU OLPLW DSSOLHG WKH YHQWLODWLRQ UDWH
FRXOG EHFRPH ORZHU WKDQ WKH UHTXLUHG EXLOGLQJ VRXUFH YHQWLODWLRQ )LJXUH E VKRZV DOVR WKH FDVH ZKHQ WKH ORZHU OLPLW LV
HVWDEOLVKHG7KHVLPXODWLRQUXQVIRUWKHFRROLQJSHULRGZLWKDSSO\LQJWKHFRUUHFWLRQIDFWRUDQGORZHUOLPLWDQGWKHUHVXOWVDUH
VKRZQ LQ )LJXUH 7KH VWUDLJKW OLQH LV D RQH WR RQH OLQH LQGLFDWLQJ DJUHHPHQW EHWZHHQ WKH 6$'&9 DQG WKH VWDQGDUG
SURFHGXUHHLWKHUZLWKGHVLJQRUDFWXDORFFXSDQF\7KH6$'&9VXSSOLHVPXFKOHVVYHQWLODWLRQUDWHWKDQWKHGHVLJQRFFXSDQF\
SURFHGXUHDQGWKHUHE\SURGXFLQJUHODWLYHO\KLJKHQHUJ\VDYLQJ ZKLOHDWWKHVDPHWLPHEHFDXVHRIWKHFRUUHFWLRQIDFWRUWKH
6$'&9VOLJKWO\RYHUHVWLPDWHVWKHYHQWLODWLRQUDWHDVFRPSDUHGWRDFWXDORFFXSDQF\SURFHGXUH

$
%

&
'

5DWLR

+RXUV
)LJXUH 9DULRXVRFFXSDQF\SURILOHV

&2&RQFHQWUDWLRQ330
9HQWLODWLRQUDWH&)0

9HQWLODWLRQUDWH/V

$ 6WDQGDUG'HVLJQ3RSXODWLRQ

%
&
'QROLPLW
6XSSO\DLU&RVHWSRLQW 'PLQLPXPOLPLW

+RXUV +RXUV
DE
)LJXUH 5HVXOWVIRUWKHGHVLJQGD\LQFOXGLQJDWKHVXSSO\DQGUHWXUQDLU&2FRQFHQWUDWLRQIRUGLIIHUHQWRFFXSDQF\
VFKHPHVDQGEWKHDVVRFLDWHGYHQWLODWLRQUDWHV
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(


$6+5$(6WDQGDUG$
$FWXDO3RSXODWLRQ
$6+5$(6WDQGDUG'
'HVLJQ3RSXODWLRQ

FIP6$'&9
/V6$'&9

FIP$6+5$(
)LJXUH &RPSDULVRQEHWZHHQWKH6$'&9DQG$6+5$(VWDQGDUGSURFHGXUHGHVLJQDQQGDFWXDOSRSXODWLRQ

7KH ZKROH \HDU FRROLQJ HQHUJ\ FRQVXPSWLRRQV DUH DOVR GHWHUPLQHG ZLWK WKRVH WKUHH RSWLRQV RI RFFFXSDQF\ SURILOHV
DQG IRU YDULRXV ORFDWLRQV DV VKKRZQ LQ 7DEOH 7KH FRROLQJ HQHUJ\ VDYLQJ LV EDVHG RQ
R WKH GHVLJQRFFXSDQF\
VWDQGDUGSURFHGXUH3URILOH$QRVFKHGXOH%RWKWKH6$'&9DQGVWDQGDUGSURFHGXUHSURGXFHWKHVDDPHUHVXOWVDWIXOOGHVLJQ
RFFXSDQF\ 3URILOH $ DQG WKH VDYLQJ LV ]HHUR +RZHYHU ZKHQ WKH SRSXODWLRQ GURSV DV LQ WKH SURIILOHV RI % & DQG ' WKH
HQHUJ\VDYLQJE\6$'&9LVUHVXOWHG)RUH[[DPSOHZKHQWKHDFWXDORFFXSDQF\IROORZVWKHSURILOH'LQVWHDGRI$WKHFRROLQJ
HQHUJ\VDYLQJE\LPSOHPHQWLQJ6$'&9FRRXOGEHLQUDQJHRI7KHVDYLQJZLOOEHREYLRRXVO\ORZHUDVWKHDFWXDO
RFFXSDQF\LVJHWWLQJFORVHUWRWKHGHVLJQYDDOXH7KHPD[LPXPVDYLQJLVREWDLQHGLQORFDWLRQVZLWK KRWFOLPDWHDQGWKHDUHD
ZKHUHWKHIUHHFRROLQJLVQRWZLGHO\XVHGVXXFKDVLQ2UODQGR/HVVVDYLQJLVREWDLQHGLQ6DQ)UDQFLVFRDQG1HZ<RUNGXHWR
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
HOHYDWHGQXPEHURIKRXUVZKHQWKHHFRQRPLL]HULVDFWLYDWHGDQGIUHVKDLULVPRVWO\EURXJKWIRUWKHIUHHHFRROLQJSXUSRVH

7DEOH$QQXDOFRROLQJHQHUUJ\FRQVXPSWLRQIRURFFXSDQF\SURILOHVDVVKRZ ZQLQ)LJXUH
Orlando Phoenix Greensboro San Franciscco New York
Energy Saving Energy Saving Energy Saving Energy ving
Sav Energy Saving
kWh % kWh % kWh % kWh % kWh %
No Schedule (A) 88130 0 79317 0.0 41490 0 31118 0..0 25790 0
Occupancy profile B 84160 4.5 76023 4.2 39800 4.1 30124 3..2 25000 3.1
Occupancy profile C 80110 9.1 72847 8.2 38790 6.5 28992 6..8 24000 6.9
Occupancy profile D 71600 18.8 65267 17.7 35150 15.3 27352 122.1 22800 11.6

&21&/86,21
7KH PXOWL]RQH YHQWLODWLRQUDWH $6+5 5$(6WDQGDUGSURFHGXUHLQFOXGHVWKH]RQHRFFXSDQQF\QXPEHUDQGDLUIORZ
UDWHLQWKHFDOFXODWLRQV,QSUDFWLFHLWLVKDUGGWRHVWLPDWHDFFXUDWHO\WKHDFWXDOQXPEHURIRFFXSDQWVLLQHDFK]RQHXQOHVVWKHUH
LVD&2VHQVRUORFDWHGLQHDFK]RQHZLWKWKKHKLJKFRVWDVVRFLDWHG7KXVWKHVWDQGDUGSURFHGXUHLVP PRVWO\DSSOLHGXVLQJWKH
GHVLJQRFFXSDQF\SURILOHEXWZLWKWKHDFWXDDODLUIORZUDWHV7KHGHVLJQRFFXSDQF\SURFHGXUHUHVXOWVLLQRYHUYHQWLODWLRQRIWKH
VSDFHV KDYLQJ OHVV WKDQ GHVLJQ RFFXSDQWV DQG LW FRQVHTXHQWO\ OHDGV WR HQHUJ\ ZDVWH $OWHUQDWLYHOO\ WKH 6$'&9 VWUDWHJ\
SURSRVHG LQ WKLV SDSHU FRXOG EH DSSOLHG 7
7KH VWUDWHJ\ LV EDVHG RQ VXSSO\ DLU &2 FRQFHQWUDWLRQ DQG
D LW LQFOXGHV D FRQWURO
306
ASHRAE Transactions
DOJRULWKP WR G\QDPLFDOO\ UHVHW WKH &2 FRQFHQWUDWLRQ VHW SRLQW RQ D SURSHU WLPH LQWHUYDO EDVLV DQG XVLQJ WKH DFWXDO ]RQH
DLUIORZUDWHV$OWKRXJKWKHDOJRULWKPLVEDVHGRQWKHGHVLJQRFFXSDQF\ $6+$5(6WDQGDUGSURFHGXUHWKHVWUDWHJ\LV
DEOHWRWUDFNVRPHZKDWWKHDFWXDORFFXSDQF\SDWWHUQ7KHVWUDWHJ\WHQGVWRXQGHUYHQWLODWHWKHVSDFHEHFDXVHRIUHFLUFXODWLQJ
WKHUHWXUQDLU PL[WXUH WKDWLQFOXGHVDSRUWLRQ IRUEXLOGLQJVRXUFHYHQWLODWLRQUDWH 0DQ\ ZD\VFDQEHGRQHWRDOOHYLDWHWKLV
SUREOHP VXFK DV ORZHULQJ LQWHQWLRQDOO\ WKH VXSSO\ DLU &2 FRQFHQWUDWLRQ VHW SRLQW DQGRU HVWDEOLVKLQJ D ORZHU YHQWLODWLRQ
OLPLW7KHVWUDWHJ\UHTXLUHVD3,FRQWUROOHUDORQJZLWKD&2VHQVRUORFDWHGLQWKHPDLQVXSSO\DLUGXFWWRPRQLWRUDQGFRQWURO
WKH VXSSO\ DLU FRQFHQWUDWLRQ E\ PHDQV RI DGMXVWLQJ WKH RXWGRRU DLU GDPSHU SRVLWLRQ 7KXV WKH VWUDWHJ\ GRHV QRW LQYROYH
VLJQLILFDQWLQLWLDORURSHUDWLQJFRVWDQGLVUHODWLYHO\HDV\WRDSSO\LQ PRVW PXOWL]RQH+9$&V\VWHPVHTXLSSHG ZLWK GLUHFW
GLJLWDOFRQWUROV\VWHP
7KH HQHUJ\ VLPXODWLRQV ZHUH SHUIRUPHG RQ YDULRXV 86$ ORFDWLRQV DQG IRU D W\SLFDO ILYH]RQH RIILFH EXLOGLQJ
FRQGLWLRQHGE\D9$9V\VWHP7KHFRROLQJ VDYLQJFRXOG EHXSWRDVFRPSDUHG WRWKHGHVLJQRFFXSDQF\ $6+5$(
VWDQGDUGSURFHGXUH7KLVVDYLQJYDULHVZLWKWKHORFDWLRQVDQGRFFXSDQF\FKDQJHV7KHVDYLQJZLOOEHORZHUDVWKHDFWXDO
RFFXSDQF\LVJHWWLQJFORVHUWRWKHGHVLJQYDOXH7KHPD[LPXPVDYLQJLVREWDLQHGLQORFDWLRQVZLWKKRWFOLPDWHDQGWKHDUHD
ZKHUHWKHIUHHFRROLQJLVQRWZLGHO\XVHGVXFKDVLQ2UODQGR
5()(5(1&(6
$ODODZL 0 DQG 0 .UDUWL ([SHULPHQWDO (YDOXDWLRQ RI &2±%DVHG 'HPDQG&RQWUROOHG 9HQWLODWLRQ 6WUDWHJLHV
$6+5$(7UDQVDFWLRQV
$6+5$($6+5$(6WDQGDUG9HQWLODWLRQIRU$FFHSWDEOH,QGRRU$LU4XDOLW\$WODQWD$PHULFDQ6RFLHW\RI
+HDWLQJ5HIULJHUDWLQJDQG$LU&RQGLWLRQLQJ(QJLQHHUV,QF
%UDQGHPXHKO0-6*DEHODQG,$QGHUVHQ$7RRONLWIRU6HFRQGDU\+9$&6\VWHP(QHUJ\&DOFXODWLRQ$PHULFDQ
&DUSHQWHU6&(QHUJ\DQG,PSDFWVRI&2±%DVHG'HPDQG&RQWUROOHG9HQWLODWLRQ$6+5$(7UDQVDFWLRQV

&RQJUDGDF9DQG).XOLF+9$&6\VWHP2SWLPL]DWLRQZLWK&2&RQFHQWUDWLRQ&RQWURO8VLQJ*HQHWLF$OJRULWKPV
(QHUJ\DQG%XLOGLQJV±
+\GHPDQ01:HEE36UHHGKDUDQDQG6%ODQF'HYHORSPHQWDQG7HVWLQJRID5HIRUPXODWHG5HJUHVVLRQ%DVHG
(OHFWULF&KLOOHU0RGHO$6+5$(7UDQVDFWLRQV
/X7;/DQG09LOMDQHQ$1RYHODQG'\QDPLF'HPDQG&RQWUROOHG9HQWLODWLRQ6WUDWHJ\IRU&2&RQWURODQG
(QHUJ\6DYLQJLQ%XLOGLQJV(QHUJ\DQG%XLOGLQJV
0XUSK\ - &2%DVHG 'HPDQG&RQWUROOHG 9HQWLODWLRQ ZLWK $6+5$( 6WDQGDUG 7UDQH (QJLQHHUV
1HZVOHWWHU9ROXPH±
1DVVLI 1 6 .DMO DQG 5 6DERXULQ 9HQWLODWLRQ &RQWURO 6WUDWHJ\ 8VLQJ WKH 6XSSO\ &2 &RQFHQWUDWLRQ 6HW SRLQW
,QWHUQDWLRQDO-RXUQDORI+9$& 55HVHDUFK
1DVVLI 1 $ 5REXVW &2%DVHG 'HPDQG&RQWUROOHG 9HQWLODWLRQ &RQWURO 6WUDWHJ\ IRU 0XOWL=RQH +9$& 6\VWHPV
(QHUJ\DQG%XLOGLQJVLQSUHVV9&SSGRLMHQEXLOG
6WDQNH'6\VWHP2SHUDWLRQ'\QDPLF5HVHW2SWLRQV$6+5$(-RXUQDO
7D\ORU67&2EDVHG'&9XVLQJ$6+5$(-RXUQDO
:DQJ 6 DQG ;X ; $ 5REXVW &RQWURO 6WUDWHJ\ RI &RPELQHG '&9 DQG (FRQRPL]HU &RQWURO IRU $LUFRQGLWLRQLQJ
6\VWHPV(QHUJ\&RQYHUVLRQDQGPDQDJHPHQW

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

SA-12-C039
Modeling and Analysis for Mixed-Mode

Cooling of Buildings
Jianjun Hu Panagiota Karava, PhD

ABSTRACT
7KHVWXG\SUHVHQWVDWUDQVLHQWWKHUPDOPRGHOIRUWKHDVVHVVPHQWRI PL[HGPRGHFRROLQJLQEXLOGLQJVZLWKK\EULGYHQWLODWLRQDQGKLJKOHYHOV
RI H[SRVHGWKHUPDOPDVV7KHPRGHOLVEDVHGRQWKHWKHUPDOQHWZRUNUHSUHVHQWDWLRQRI DEXLOGLQJWKDWFDQFDSWXUHWKHQRQXQLIRUPVXUIDFHDQG
DLUWHPSHUDWXUHGLVWULEXWLRQZLWKLQDEXLOGLQJVSDFHFRQVLGHULQJVPDOOHUVLPXODWLRQVXE]RQHV$GYHFWLYHKHDWWUDQVIHUDVVRFLDWHGZLWKDLUIORZLV
XVHGIRUFRPSXWLQJWKHWKHUPDOH[FKDQJHEHWZHHQDLUQRGHVLQDGMDFHQWVXE]RQHV7ZRGLPHQVLRQDOKHDWFRQGXFWLRQLVFRQVLGHUHGWRGHVFULEHWKH
KRUL]RQWDODQGYHUWLFDOKHDWWUDQVIHUZLWKLQWKHPDVVLYHIORRUVODE7KHPRGHOLVYDOLGDWHGZLWKPRQLWRUHGGDWDIURPDQRFFXSLHGLQVWLWXWLRQDO
EXLOGLQJXQGHUUHDOFRQGLWLRQV7KHKHDWUHPRYDOIURPWKHEXLOGLQJWKHUPDOPDVVE\WKHFRRODLUVWUHDPLVFORVHWRDQGWLPHVKLJKHUZLWK
WKHDLUVWUHDPDWDQDYHUDJHWHPSHUDWXUHRI &°)FRPSDUHGWRWKDWZLWKDQDLUVWUHDPDW&°)DQG&°)
UHVSHFWLYHO\7KHQLJKWFRROLQJRSHUDWLRQKRXUVDQGWKHVHOHFWLRQRI ORZWHPSHUDWXUHVHWSRLQWKDYHDVLJQLILFDQWLPSDFWRQWKHIROORZLQJGD\·VFRROLQJ
GHPDQGIRUWKHFDVHVFRQVLGHUHG7KHVLPXODWLRQUHVXOWVLQGLFDWHWKHQHFHVVLW\IRUGHYHORSLQJRSWLPDOPRGHOSUHGLFWLYHFRQWUROVWUDWHJLHVIRUEXLOGLQJV
ZLWKPL[HGPRGHFRROLQJ
INTRODUCTION
1DWXUDOYHQWLODWLRQV\VWHPVVXFKDVZLQGFDWFKHUVDQGFKLPQH\VKDYHEHHQXVHGLQEXLOGLQJVIRUPDQ\FHQWXULHV
:KHQQDWXUDOYHQWLODWLRQLVHPSOR\HGFRRORXWVLGHDLULVGUDZQLQWRWKHEXLOGLQJWRSURYLGHIUHHFRROLQJWKXVUHGXFLQJ
FRROLQJHQHUJ\FRQVXPSWLRQ,WDOVRUHGXFHVWKHHQHUJ\FRQVXPHGE\IDQVWRFLUFXODWHDLULQWKHEXLOGLQJ7KHVDPHEDVLF
SULQFLSOHVDUHXVHGLQPRGHUQPL[HGPRGHFRROLQJRI EXLOGLQJVEDVHGRQDGDSWLYHWKHUPDOFRPIRUWPRGHOV%UDJHUDQG
GH'HDU(PPHULFKHWDOLQWHJUDWLQJQDWXUDOYHQWLODWLRQGULYHQE\ZLQGRUWKHUPDOEXR\DQF\IRUFHVDQG
PHFKDQLFDO V\VWHPV DORQJ ZLWK LQWHOOLJHQW FRQWUROV WR PLQLPL]H HQHUJ\ XVH 0L[HGPRGH FRROLQJ WHFKQLTXHV RIWHQ
LQFRUSRUDWHDVRODUFKLPQH\RUDQDWULXPWKDWLVXVHGWRSURPRWHEXR\DQF\IRUFHVDQGRSWLPL]HDLUPRYHPHQW$QRWKHU
FRPPRQIHDWXUHLVQLJKWFRROLQJZLWKFRRORXWGRRUDLUGULYHQLQWRWKHEXLOGLQJDWQLJKWWKHWKHUPDOPDVVFDQEHFRROHG
WR UHGXFH WKH FRROLQJ ORDG IRU WKH IROORZLQJ GD\ 6SLQGOHU DQG 1RUIRUG D E 0D\2VWHQGRUS HW DO
([LVWLQJWHFKQLTXHVIRUQLJKWFRROLQJRI QDWXUDOO\YHQWLODWHGDQGPL[HGPRGHEXLOGLQJVDUHKHXULVWLFLHEDVHGRQIL[HG
VFKHGXOHV WKDW DUH QRW RSWLPL]HG IRU WKH SDUWLFXODU EXLOGLQJ RU FOLPDWH DQG PD\ OHDG WR LQFUHDVHG RSHUDWLQJ FRVWV RU
RFFXSDQWGLVFRPIRUWGH'HDUDQG%UDJHU7KHVHSUREOHPVFDQEHDYRLGHGE\HPSOR\LQJPRGHOSUHGLFWLYHFRQWURO
VWUDWHJLHVZKLFK DUH EDVHGRQZHDWKHUIRUHFDVW DQGRQHGD\DKHDGFRROLQJORDG SUHGLFWLRQIRUWKHRSHUDWLRQ RI QLJKW
Jianjun Hu is a Ph.D. student and Panagiota Karava is an Assistant Professor in the School of Civil Engineering, Purdue University, West
Lafayette, IN.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
308
ASHRAE Transactions
FRROLQJ 6SLQGOHU DQG1RUIRUG D E 0D\2VWHQGRUS HW DO ,Q WKLV SURFHVVGHYHOR
RSPHQW RI YDOLGDWHG
EXLOGLQJHQHUJ\PRGHOVWKDWFDQEHXVHGIRRUV\VWHPLGHQWLILFDWLRQLVFULWLFDO7KLVSDSHUSUHVHQWVDQ QHQHUJ\PRGHORI D
EXLOGLQJZLWKPL[HGPRGHFRROLQJDQGKLJK KOHYHOVRI H[SRVHGWKHUPDOPDVVEDVHGRQWKHWKHUPDOQHWWZRUNUHSUHVHQWDWLRQ
0HDVXUHPHQWGDWDIURPDQH[WHQVLYHH[SHULLPHQWDOVWXG\FRQVLGHULQJDIXOOVFDOHVHWXSLQDKLJKSHUUIRUPDQFHEXLOGLQJLV
XVHGIRUWKHPRGHOGHYHORSPHQWDQGYDOLGDWWLRQ'LIIHUHQWQLJKWFRROLQJVWUDWHJLHVZLWKYDULDEOHORZWHHPSHUDWXUHVHWSRLQWV
FRQVLGHULQJWKHWKHUPDOUHVSRQVHRI WKHEXLLOGLQJDUHLQYHVWLJDWHGZLWKWKHYDOLGDWHGHQHUJ\PRGHO7 7KHSUHVHQWVWXG\LVD
ILUVWVWHSLQWKHGLUHFWLRQDLPLQJWRGHYHORS
SHIILFLHQWDQWLFLSDWRU\FRQWUROVWUDWHJLHVIRUPL[HGPRGHEXLOGLQJV
BUILDING DESCRIPTION
$VWRU\KLJKLQVWLWXWLRQDOEXLOGLQJOORFDWHGLQGRZQWRZQ0RQWUHDO1:LVFRQVLGHUUHG7KHEXLOGLQJKDV

WZRPDLQODUJHIDoDGHVIDFLQJVRXWKZHVWDQGGVRXWKHDVW7KHWRWDOIORRUDUHDLVDERXWPIWZKLOHWKHIORRU
DUHDFRYHUHGE\SHULPHWHUVSDFHVLVP PIW7KHSHULPHWHU]RQHVLQFOXGHRIILFHVPï ïPïPKLJK
>IWïIWïIW@ODERUDWRULHVPïPïPKLJK>IWïIWïIW@DDQGDQDWULXPRQWKH
VRXWKZHVWIDoDGHRI WKHEXLOGLQJH[WHQGLQJJIURPWKHVHFRQGWRWKHVL[WHHQWKIORRU7KHDWULXPLV VXEGLYLGHGLQWRILYH
VWRU\DWULDVHSDUDWHGZLWKDIORRUVODEZLWKKPRWRUL]HGJULOOHVVRDVWRUHGXFHWKHVL]HRI ILUHVPRNHHFRQWURO]RQHVIORRU
JULOOHVDUHILUHSURRI DQGFORVHLQWKHHYHQWRI DILUH(DFKRI WKHILYHVWRUH\DWULDKDVGLPHQVLRQVRI PïPï
PKLJKIWïIWïIWDQGDZHVWRI 6RXWKRULHQWDWLRQDQGLWLVHTXLSSHGZLWKPRWRUUL]HGUROOHUVKDGHVRQ
DOOJODVVVXUIDFHV
7KHQDWXUDOYHQWLODWLRQGHVLJQIHDWXUHVRI WKHEXLOGLQJLQFOXGHDLQOHWJULOOHVZLWKPRWRUL]HGGDDPSHUVRSHQLQJDUHD
DERXWP>IW@ORFDWHGDWWKHHQGRI WKHFRUULGRUVLQWKHVRXWKHDVWDQGQRUWKZHVWIDoDGHRI HHDFKIORRU)LJXUH
OHIWDQGEILYHVWRU\DWULDWKDWDUHVHSDDUDWHGZLWKDIORRUVODEDQGFRQQHFWHGZLWKJULOOHVZLWKDUHDRI PIW
)LJXUHPLGGOHJULOOHVKDYHPRWRUL]HGGGDPSHUVWRHQKDQFHEXR\DQF\GULYHQIORZ7KHLQOHWJULOOOHVRQWKHIDoDGHDQG
WKHJULOOHVFRQQHFWLQJWKHDWULDDUHFRQWUROOOHGE\EXLOGLQJDXWRPDWLRQV\VWHPDQGEDVHGRQZHDWKHHUGDWDWKH\RSHQRU
FORVHVLPXOWDQHRXVO\7KHK\EULGYHQWLODWLRQFRQFHSWLVVKRZQLQ)LJXUH :KHQWKHEXLOGLQJLVLLQQDWXUDOYHQWLODWLRQ
PRGHDWKHFRUULGRULQOHWJULOOHVDQGWKHJJULOOHVFRQQHFWLQJWKHDWULDRSHQEWKHPHFKDQLFDOVXSS SO\DLUIORZUDWHLQWKH

)LJXUH 0RWRUL]HGFRUULGRULQOHWJULOOHVVOHIWIORRUJULOOHVLQWKHDWULXPPLGGOHFRQFUHWHIORRUVODELQWKHHFRUULGRUULJKW

)LJXUH %XLOGLQJOD\RXWDQGK\EULGYHQWLODWLRQFRQFHSW
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

DWULDLVUHGXFHGWRDPLQLPXPYDOXHFWKHH[KDXVWYHQWLQWKHDWULXPRSHQVDQGGWKHDLUVXSSO\XQLWVORFDWHGDWWKH
FRUULGRUVFORVH,QWKHDWULDVSDFHVDLULVVXSSOLHGXVLQJPHFKDQLFDORUQDWXUDOPHDQVDQGLWLVGLUHFWHGWRRWKHU]RQHV
WKURXJKWKHUHWXUQRULWLVH[KDXVWHGWKURXJKWKHRXWOHWYHQWLQWKHWRSDWULXP7KHQDWXUDOYHQWLODWLRQV\VWHPW\SLFDOO\
RSHUDWHVZKHQWKHRXWGRRUWHPSHUDWXUHLVEHWZHHQ&)DQG&)DQGUHODWLYHKXPLGLW\LVOHVVWKDQ
7KHEXLOGLQJKDVKLJKOHYHOVRI WKHUPDOPDVVLQWKHIRUPRI H[SRVHGFRQFUHWHIORRUVODEVDWWKHFRUULGRUV)LJXUHULJKW
ZKLFKDUHORFDWHGDGMDFHQWWRWKHLQOHWJULOOHVRQWKHVRXWKHDVWDQGQRUWKZHVWHQGVDQGDUHH[WHQGHGDOOWKHZD\WRWKH
DWULXP,WLVH[SHFWHGWKDWWKHORZOLPLWRI &)FDQEHIXUWKHUUHGXFHGDWQLJKWW\SLFDOO\DWWHPSHUDWXUHVUDQJLQJ
IURPWR&>WR)@HQKDQFLQJWKHFRROLQJHIIHFWRI RXWGRRUDLUWRFRROWKHEXLOGLQJWKHUPDOPDVVWKXV
UHGXFLQJWKHFRROLQJORDGRQWKHIROORZLQJGD\
EXPERIMENTAL SET-UP
0HDVXUHPHQWV ZHUH SHUIRUPHG WR H[DPLQH WKH HIIHFWLYHQHVV RI WKH FRRO DLU VWUHDP LQ UHGXFLQJ WKH IORRU VODE
WHPSHUDWXUHLQWKHDEVHQFHRI RFFXSDQWVDQGVRODUKHDWJDLQV7KHFRUULGRUDQGDWULDJULOOHVDUHW\SLFDOO\FRQWUROOHGE\WKH
%XLOGLQJ(QHUJ\0DQDJHPHQW6\VWHP%(06DQGRSHQRUFORVHGHSHQGLQJRQRXWGRRUFRQGLWLRQVWHPSHUDWXUHDQG
UHODWLYHKXPLGLW\7KH%(06ZDVRYHUULGGHQVRDVWRDOORZWKHJULOOHVWRUHPDLQRSHQZKHQWKHRXWGRRUWHPSHUDWXUHZDV
H[SHFWHGWRH[FHHGWKHORZWKUHVKROG7ZRFDVHVZHUHH[DPLQHG
D 7KHFRUULGRULQOHWJULOOHVZHUHNHSWRSHQRQWKH6(DQG1:VLGHRI IORRUDQGZHUHFORVHGRQHYHU\RWKHUIORRU
WRDOORZIRUFRPSDUDWLYHPHDVXUHPHQWVRI WKHFRQFUHWHVODEVXUIDFHWHPSHUDWXUH7KHIORRUJULOOHVFRQQHFWLQJ
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
WKHDWULDZHUHRSHQIURPWKHWKIORRUXSZDUGLQIRXUDWULDWRDOORZWKHDLUWREHH[KDXVWHGWKURXJKWKHWRS

E 7KHFRUULGRULQOHWJULOOHVDQGDWULDIORRUJULOOHVZHUHNHSWRSHQRQDOOIORRUV
0HDVXUHPHQWV ZHUH FRQGXFWHG IRU ZHDWKHU FRQGLWLRQV ZLWK VXQQ\ VNLHV GXULQJ WKH GD\ SUHFHGLQJ WKH QLJKW RI
PHDVXUHPHQWVVRWKDWVRODUKHDWJDLQVDEVRUEHGE\WKHEXLOGLQJWKHUPDOPDVVLQFUHDVHWKHWHPSHUDWXUHRI WKHIORRUVODE
2XWGRRUDLUWHPSHUDWXUHZDVH[SHFWHGWRGURSEHORZ&)DWQLJKWLQRUGHUWRDHQVXUHWKDWWKHFRROLQJHIIHFWRI
WKHLQFRPLQJDLUVWUHDPLVVLJQLILFDQWHQRXJKWREHGHWHFWHGLQVXUIDFHWHPSHUDWXUHPHDVXUHPHQWVEH[DPLQHWKHQLJKW
FRROLQJ HIIHFW IRU FRQGLWLRQV ZLWK ORZHU RXWGRRU WHPSHUDWXUH DQG WKH ZLQG VSHHG ZDV H[SHFWHG WR EH ORZ VR WKDW
WHPSHUDWXUHGLIIHUHQFHLVWKHRQO\GULYLQJIRUFHIRUWKHLQFRPLQJDLUWKURXJKWKHJULOOHV
7KHVXUIDFHWHPSHUDWXUHRI WKHVODEZDVPHDVXUHGDWGLIIHUHQWGLVWDQFHVIURPWKHLQOHWJULOOHVXVLQJDQLQIUDUHGFDPHUD
7KHPHDVXUHPHQWVZHUHFRQGXFWHGRQIORRUVDQGRI WKHEXLOGLQJZKHUHDLUIORZLVPDLQO\GULYHQGXHWRVWDFNHIIHFW
VWDUWLQJODWHDIWHUQRRQRI RQHGD\DQGODVWLQJXQWLOHDUO\PRUQLQJRQWKHIROORZLQJGD\6XUIDFHWHPSHUDWXUHPHDVXUHPHQWV
ZLWKWKHUPRFRXSOHZHUHXVHGWRFRUUHFWWKHLQIUDUHGFDPHUDUHVXOWVZKLFKKDGEHHQRYHUHVWLPDWHGGXHWRUHIOHFWLRQV7KH
DLUYHORFLW\ZDVPHDVXUHGXVLQJDYHORFLW\PHWHU´.DQRPD[µDWGLIIHUHQWKHLJKWVDORQJWKHLQOHWJULOOHVDQGDQDYHUDJH
DLUIORZ UDWH YDOXH ZDV FDOFXODWHG WR EH XVHG LQ WKH DQDO\VLV 7KH WKHUPDO FRQGLWLRQV LQ WKH EXLOGLQJ ]RQHV ZKLFK DUH
VXEMHFWHGWRQDWXUDOYHQWLODWLRQDWULXPDQGFRUULGRUVZHUHDOVRHYDOXDWHGGXULQJWKHGD\ZKLFKDOORZVFDOFXODWLRQRI WKH
FRROLQJORDG0RUHLQIRUPDWLRQRQWKHH[SHULPHQWDOVHWXSFDQEHIRXQGLQ0RXULNL
MODEL DESCRIPTION
7KLVVHFWLRQSUHVHQWVDWUDQVLHQWWKHUPDOPRGHOIRUDORQJFRUULGRUZLWKDKLJKO\PDVVLYHIORRUVODEWKDWLVVXEMHFWHG
WRDFRRODLUVWUHDPRI RXWGRRUDLUIORZLQJLQWRWKHEXLOGLQJWKURXJKWKHPRWRUL]HGLQOHWJULOOHV7KHPRGHOLVXVHG
WRJHWKHUZLWKH[SHULPHQWDOGDWDWRSUHGLFWWKHVODEVXUIDFHWHPSHUDWXUHDWGLIIHUHQWGLVWDQFHVIURPWKHLQOHWJULOOHVDQGWKH
FDOFXODWLRQRI KHDWUHPRYDOIURPIORRUVODEE\WKHFRRODLUVWUHDP7KHUHIRUHLWFDQEHSRWHQWLDOO\XWLOL]HGDVDWRROWRVHOHFW
ORZWHPSHUDWXUHVHWSRLQWVRDVWKHVSDFHLVQRWRYHUFRROHG
7KHFRUULGRUZLWKGLPHQVLRQVRI PïPïPIWïIWïIWLVSDUWRI WKHEXLOGLQJGHVFULEHGDERYH
310
ASHRAE Transactions
DQGKDVRQHH[WHULRUIDoDGHZKHUHWKHLQOHWJJULOOHVDUHLQVWDOOHG7KHFRUULGRUDFWVDV´GXFWµWRDOORZRX XWVLGHDLUFRPLQJLQWR
WKHDWULXP'XHWRLWVODUJHGLPHQVLRQP P>IW@ORQJDVLJQLILFDQWWHPSHUDWXUHGLIIHUHQFHLVDQWLLFLSDWHGLQWKHDLUDQG
VODEVXUIDFHWHPSHUDWXUHVRWKHZHOOPL[HGDVVXPSWLRQFDQQRWEHXVHGLQWKHVLPXODWHGVSDFHDQGWWKHFRUULGRULVHYHQO\
GLYLGHGLQWRVHFWLRQV)LJXUH$QDORJJRXVWRWKHHOHFWULFDOFLUFXLWWKHFRUULGRUFDQEHUHSUHVHQWWHGDVDFROOHFWLRQRI
QRGHVZKLFKDUHFRQQHFWHGZLWKUHVLVWDQFHHV$VFKHPDWLFRI WKHWKHUPDOQHWZRUNIRUVHFWLRQDQ QGRQO\LVVKRZQLQ
)LJXUHDQGWKHQRGHVDUHGHVFULEHGLQ7DDEOH7KHUHDUHIRXUPRGHVRI KHDWWUDQVIHUEHWZHHQWKHHVHQRGHVDGYHFWLRQ
FRQGXFWLRQFRQYHFWLRQDQGUDGLDWLRQ
y $GYHFWLYHKHDWWUDQVIHUEHWZHHQDLUQR RGHV
y 7ZRGLPHQVLRQDOFRQGXFWLYHKHDWWUDQ QVIHUEHWZHHQPDVVQRGHVLQFOXGLQJYHUWLFDOKHDWFRQGXFFWLRQZLWKLQWKHVDPH
VHFWLRQHJEHWZHHQQRGHDQGDQGKRUL]RQWDOKHDWFRQGXFWLRQEHWZHHQDGMDFHQWVHFWLRQVHJEHWZHHQ
QRGHDQG
y &RQYHFWLYHKHDWWUDQVIHUEHWZHHQDLUQ QRGHDQGVXUIDFHV
y 5DGLDWLYHKHDWWUDQVIHUEHWZHHQVXUIDFFHVLQWKHVDPHVHFWLRQHJEHWZHHQQRGHDQGRUEHWZHHQDGMDFHQW
VHFWLRQVHJEHWZHHQQRGHDQG6LQFHWKHYLHZIDFWRUVEHWZHHQVXUIDFHVZKLFKDUHORFDDWHGIXUWKHUDSDUWDUH
VPDOOOHVVWKDQWKHUDGLDWLYHKHDDWH[FKDQJHEHWZHHQWKHVHVXUIDFHVLVFRQVLGHUHGQHJOLJLE EOH
7KHKHDWWUDQVIHUE\FRQGXFWLRQDQGUUDGLDWLRQLVFDOFXODWHGXVLQJWKHUPDOUHVLVWDQFHPRGHOV$ $WKLHQLWLV7KH
DGYHFWLYH KHDW WUDQVIHU DVVRFLDWHG ZLWK DLU IORZ VXFK DV WKH LQILOWUDWLRQ WKURXJK FUDFN DQG DLU IOR RZ EHWZHHQ FRUULGRU
VHFWLRQVFDQEHFRPSXWHGIURP
ܳሶ௜௝
௔ௗ௩
ൌ ‫݉ܨ‬ሶ௜௝ ܿ௣ ൫ܶ௝ െ ܶ௜ ൯
XUH &RUULGRUFRQILJXUDWLRQZLWKFRQWUROYROXPHV
)LJX
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
)LJXUH 6FKHPDWLFYLHZRI WKHUPDOQHWZRUNVHFWLRQDQG
Tab
ble 1 Nodes in thermal network
Nodes type Label in
n network Nodes type Label in network
k
$LUQRGH « )ORRUVXUIDFH «
&HLOLQJVXUIDFH « :DOOVXUIDFH «

)ORRUPDVV
&HLOLQJPDVV «
QRGHVLQVODE
«
:DOOPDVV «
« :LQGRZVXUIDFH 6(6(
ZKHUH)LVWKHXSZLQGRSHUDWRUZKLFKLVXVVHGIRUWKHFRQVLGHUDWLRQRI LQIOXHQFHIURPXSZLQGFRQWUUROYROXPHWRFXUUHQW
$6+5$(

YROXPHL) LI WKHIORZLVIURPYROXPHMWRLDQG) RWKHUZLVH ݉ሶ௜௝ LVWKHPDVVIORZUDWHRI DLUFSLVVSHFLILFKHDWRI
DLU7LDQG7MLVWKHDLUWHPSHUDWXUHLQVHFWLRQLDQGMUHVSHFWLYHO\)RUWKHUDQJHRI UHODWLYHPDJQLWXGHVRI WKH*UDVKRI
QXPEHUEDVHGRQWKHYROXPHRI DLUDQGVODEWHPSHUDWXUHDQGVTXDUHG5H\QROGVQXPEHUEDVHGRQDSSURDFKLQJYHORFLW\
DQG FRUULGRU OHQJWK LQ WKH VWUHDPZLVH GLUHFWLRQ PL[HG FRQYHFWLYH KHDW WUDQVIHU LV DVVXPHG EHWZHHQ WKH IORRU VODE
VXUIDFHV DQG WKH DLU QRGHV &RQVLGHULQJ WKH UDQJH RI 5H\QROGV QXPEHU WKH ERXQGDU\ OD\HU RQ VODE VXUIDFHV FDQ EH
DVVXPHGWREHODPLQDU7KHUHIRUHWKHFRUUHODWLRQIRUWKHIRUFHGFRQYHFWLRQZLWKODPLQDUERXQGDU\OD\HURYHUIODWSODWHLV
XVHG7KHQDWXUDOFRQYHFWLYHKHDWWUDQVIHUFRHIILFLHQWV&+7&DUHFDOFXODWHGDFFRUGLQJWRWKHFRUUHODWLRQVLQ7DEOH
Table 2. Natural convection correlation
Walls “Hot floor” or “Cold ceiling” “Cold floor” or “Hot ceiling”
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
1/3 1/3
1/3 9.482 ⋅ ΔT 1.810 ⋅ ΔT
hn = 1.31 ΔT hn = hn =
7.283 − cos Σ 1.382 + cos Σ
:LWK´+RWµDQG´&ROGµLQGLFDWHWKHFRPSDULVRQZLWKDLUWHPSHUDWXUH
7KHEOHQGLQJIXQFWLRQSURSRVHGE\1HLVZDQJHUHWDOLVXVHGIRUPL[HGFRQYHFWLRQGHVFULSWLRQ
hmx = ( hεf + hnε )

1/ ε

ZKHUHKILVIRUFHG&+7&DQGKQLVQDWXUDO&+7&7KLVIRUPXODKDVEHHQXVHGLQURRPKHDWWUDQVIHUVWXGLHVHJ&DUULOKR
GD*UDFDZLWKƥHTXDOWR7KHHQHUJ\EDODQFHRI DLUQRGHLOHDGVWR
Awi hwi (Twi − Ti ) + Aci hci (Tci − Ti ) + Afi h fi (T fi − Ti ) + FCi (T j − Ti ) = 0

ZKHUH&LLVSURGXFWRI PDVVIORZUDWHDQGVSHFLILFKHDWK7DQG$LVWKH&+7&WHPSHUDWXUHDQGDUHDRI WKHZDOOFHLOLQJ
DQGIORRUVXUIDFH
Table 3 Heat gains on building surfaces
Heat gain label in network Heat gain types
6 6RODUUDGLDWLRQDEVRUEHGE\ZLQGRZH[WHULRUOD\HU
6 6RODUUDGLDWLRQDEVRUEHGE\ZLQGRZLQWHULRUOD\HU
666«6 6RODUUDGLDWLRQJDLQRQIORRU
666«6 6RODUUDGLDWLRQJDLQRQFHLOLQJ
666«6 6RODUUDGLDWLRQJDLQRQZDOO
7KHSUHVHQWHGZRUNLVDIUHHUXQQLQJFDVHZLWK]HURKHDWJHQHUDWLRQLQVLPXODWLRQ]RQHV 7KHPRGHOFDQWUDFNVRODU
UDGLDWLRQRQLQWHUQDOVXUIDFHVWRDFFXUDWHO\FDSWXUHWKHVRODUKHDWJDLQ7DEOH0HDVXUHGKRXUO\VRODUUDGLDWLRQGDWDDWD
QHDUE\ZHDWKHUVWDWLRQZHUHXVHGDVLQSXWVWRWKH3HUH]PRGHO3HUH]HWDO7KHZLQGRZJODVVIDoDGHVKRZQLQ
)LJXUHLVGRXEOHJOD]HGDQG$UJRQILOOHGZLWKDORZHFRDWLQJ7KHZLQGRZSURSHUWLHVEHDPDQGGLIIXVHWUDQVPLWWDQFH
DQGDEVRUSWDQFHZHUHREWDLQHGIURP:,1'2:/%1/7KHKHDWJDLQRQZLQGRZJOD]LQJOD\HUVLV
ͳ = (α ⋅ + α ⋅ ) ʹ = (α ⋅ + α ⋅ )

ZKHUH$ZLQGRZLVZLQGRZDUHDPơHLVH[WHULRUOD\HUDEVRUSWDQFHEHDPUDGLDWLRQơLLVLQWHULRUOD\HUDEVRUSWDQFHEHDP
UDGDLWRQơGLVGLIIXVHDEVRUSWDQFH,GLVWRWDOGLIIXVHLUUDGLDQFHZP,ELVWRWDOEHDPLUUDGLDQFHZP7KHWRWDO
ZLQGRZWUDQVPLWWHGEHDPUDGLDWLRQ7UEDQGGLIIXVLYHUDGLDWLRQ7UGLVFRPSXWHGIURP
Trb = Awindowτ e I b Trd = Awindowτ d I d

312
ASHRAE Transactions
ZKHUHƴHLVHIIHFWLYHEHDPWUDQVPLWWDQFHƴGLVGLIIXVHWUDQVPLWWDQFH7KHWUDQVPLWWHGEHDPUDGLDWLRQDQGGLIIXVHUDGLDWLRQ
WUDQVIHUWRLQWHUQDOVXUIDFHVLVFDOFXODWHGXVLQJWKHYLHZIDFWRUEHWZHHQZLQGRZDQGWKHVXUIDFHZKLOHEHDPUDGLDWLRQLV
XQLIRUPO\GLVWULEXWHGRQWKHVXQOLJKWLQJDUHD7KHVXUIDFHVRODUKHDWJDLQLVWKHVXPRI ZLQGRZWUDQVPLWWHGEHDPDQG
GLIIXVHUDGLDWLRQDQGGLIIXVHUDGLDWLRQIURPRWKHUVXUIDFHV
Alighting _ i
Si = Fw _ iTrd + α ⋅ ⋅ Trb + ¦ ( Fother _ i ⋅ Diffuseother )
¦A lighting

ZKHUH )ZBL LV YLHZ IDFWRU IURP ZLQGRZ WR VXUIDFH L )RWKHUBL LV YLHZ IDFWRU IURP RWKHU VXUIDFH WR VXUIDFH L ơ LV WKH
DEVRUSWDQFH $OLJKWLQJBL LV WKH OLJKWLQJ DUHD RQ VXUIDFH L $OLJKWLQJ LV WKH WRWDO OLJKWLQJ DUHD RQ LQWHUQDO VXUIDFHV 7KH
VXUIDFHUHIOHFWHGEHDPUDGLDWLRQLVFDOFXODWHGIURP
Alighting _ j
Diffuse j = (1 − α ) ⋅ ⋅ Trb
¦A lighting
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
7KLVVHFWLRQSUHVHQWVUHVXOWVIRUWKHPRGHOYDOLGDWLRQDQGWKHDVVHVVPHQWRI GLIIHUHQWQLJKWFRROLQJVWUDWHJLHV
Model validation
7KHHQHUJ\PRGHOZDVXVHGWRSUHGLFWWKHIORRUVODEVXUIDFHWHPSHUDWXUHDWPIWVODEPIWVODE

DQGPIWVODEGLVWDQFHIURPWKHLQOHWJULOOHVIRUWKHIROORZLQJFRQGLWLRQVD6(FRUULGRURQWKHWKIORRUZLWK
VFKHGXOHGQDWXUDOYHQWLODWLRQWKHV\VWHPRSHUDWHVIURPWRDQGVKXWVGRZQIURPWR)LJXUHDE
1:FRUULGRUZLWKWKHVDPHVFKHGXOHDVLQFDVHD)LJXUHEF6(FRUULGRURQWKHWKIORRUZLWKWKHLQOHWJULOOHVRSHQ
IURPWR)LJXUHVFDQGG7ZRZHHNVSUHVLPXODWLRQWLPHZDVXVHGDQGWKHHYDOXDWLRQSHULRGZDVRQ6HS
DQG6HS'XULQJWKLVWLPHPHDVXUHGGDWDIRUWKHDLUIORZUDWHWKURXJKFRUULGRULQOHWJULOOHVZDVXVHGDQLQSXW
7DEOHSUHVHQWVWKHDLUIORZUDWHVDQGRXWGRRUWHPSHUDWXUHVDWGLIIHUHQWWLPHV$QLQFUHDVHLQDLUIORZRFFXUVDVRXWGRRU
WHPSHUDWXUHGHFUHDVHVGXULQJWKHQLJKWGXHWRWKHVWDFNHIIHFW'XULQJWKHSUHVLPXODWLRQSHULRGWKHDLUIORZUDWHZDV
FDOFXODWHGE\DQHPSLULFDOHTXDWLRQZKLFKZDVGHULYHGIURPPHDVXUHGGDWD)RUWKHFDVHZLWKWKHFRUULGRUJULOOHVFORVHG
)LJXUHVDDQGEDQDOPRVWFRQVWDQWVXUIDFHIORRUWHPSHUDWXUHLVREVHUYHGWKURXJKRXWWKHQLJKWDQGWKHVLPXODWLRQ
DQGH[SHULPHQWDOUHVXOWVDUHLQJRRGDJUHHPHQW7KHSUHGLFWHGWHPSHUDWXUHVLQ )LJXUHEDUHVOLJKWO\VPDOOHUWKDQWKH
PHDVXUHGYDOXHVGXHWRWKHLPSDFWRI VRODUKHDWJDLQVIURPWKHDWULXPZKLFKDUHQRWFRQVLGHUHGLQWKHSUHVHQWPRGHOEXW
WKH\ZLOOEHLQFRUSRUDWHGLQWKHIXWXUH'LIIHUHQFHVLQWKHDFWXDOWHPSHUDWXUHYDOXHVVKRZQLQ )LJXUH D DQG EDUH
DWWULEXWHGWRWKHRULHQWDWLRQVRODUKHDWJDLQV6LPXODWLRQUHVXOWVIRUFDVHFZLWKD&+7&FRUUHODWLRQEHWZHHQWKHIORRU
Table 4. Measured temperatures and airflow rates at corridor inlet grilles (floor 6).
Outdoor air temperature Airflow rate SE façade
3
Time °C °F m /s ft3/s

$6+5$(

VXUIDFHVDQGWKHDLUQRGHVGHVFULEHGE\(TXDWLRQDUHVKRZQLQ)LJXUHFZKLOH)LJXUHGSUHVHQWVUHVXOWVIRUWKHVDPH
FRQGLWLRQVDQGDPRGLILHGPL[HGFRQYHFWLRQFRUUHODWLRQ7KLVLVVLPLODUWR(TXDWLRQEXWKDVDIDFWRUWKDWGHSHQGVRQWKH
ORFDODLUIORZSDWWHUQ DQGLWLVHTXDOWRIRUVHFWLRQDQGIRUDOO RWKHUVHFWLRQV7KHPRGLILHG&+7&FRUUHODWLRQ
LPSURYHVWKHDFFXUDF\RI WKHPRGHODQGLWLVDGRSWHGLQWKHIROORZLQJVLPXODWLRQV+RZHYHUDGGLWLRQDOZRUNLVUHTXLUHG
DQGLWLVLQSURJUHVVIRUWKHGHYHORSPHQWRI &+7&LQEXLOGLQJVZLWKQDWXUDOYHQWLODWLRQ2YHUDOOWKHSHUIRUPDQFHRI WKLV
UHODWLYHO\VLPSOHPRGHOZDVFRQVLGHUHGWREHVDWLVIDFWRU\JLYHQWKHXQFHUWDLQW\DVVRFLDWHGZLWKDUHDOEXLOGLQJ
Night cooling
7KH SHUIRUPDQFH RI QLJKW FRROLQJ GHSHQGV RQ WKH ORZ WHPSHUDWXUH SRLQW VHOHFWLRQ DQG LWV GXUDWLRQ RSHUDWLRQ
KRXUV7KHHQHUJ\PRGHOGHVFULEHGDERYHZDVXVHGWRHYDOXDWHQLJKWFRROLQJVWUDWHJLHVZLWKDYDULDEOHWHPSHUDWXUHUDQJH
²&)&DVH²&)²)&DVHDQG²&)²)&DVHZKLFK
HVVHQWLDOO\UHSUHVHQWV´V\QWKHWLFGD\Vµ7KHUHVXOWVVKRZQLQ )LJXUH DLQGLFDWHVLJQLILFDQWSRWHQWLDOZLWKDFRQYHFWLYH
KHDWUHPRYDOIURPWKHFRQFUHWHIORRUVODERI N:KN%WXGXULQJDSHULRGRI KRXUV²IRU&DVH
ZKLOH WKH KHDW UHPRYDO IRU &DVH DQG LV N:K N%WX DQG N:K N%WX UHVSHFWLYHO\ )LJXUH E
SUHVHQWVWKHDLUWHPSHUDWXUHLQDW\SLFDOFRUULGRUVHFWLRQGXULQJWKHIROORZLQJGD\²IRUGLIIHUHQWQLJKW
FRROLQJVFKHGXOHV6FKWRDQG&DVH7KHGDWDVHWVIRU6FKWR6FKUHSUHVHQWDWRWDOQXPEHURI FRROLQJKRXUVIURP
WRZLWKDQLQWHUYDORI KRXUV7KHDLUWHPSHUDWXUHFDOFXODWHGEDVHGRQWKHDGDSWLYHWKHUPDOFRPIRUWUDQJH%UDJHU
DQGGH'HDULVDOVRVKRZQLQWKHILJXUH7KHUHVXOWVVKRZWKDWWKHSUHGLFWHGFRUULGRUWHPSHUDWXUHLVZLWKLQWKH
FRPIRUWUDQJHDOPRVWIRUDOOVFKHGXOHVFRQVLGHUHGEXWRYHUFRROLQJFDQEHH[SHULHQFHGGXULQJPRUQLQJKRXUVZKHQWKH
QLJKWFRROLQJV\VWHPRSHUDWHVIRUKRXUV

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
D6(FRUULGRURI IORRUE1:FRUULGRURI IORRU

F6(FRUULGRURI IORRUZLWKRULJLQDO&+7&G6(FRUULGRURI IORRUZLWKPRGLILHG&+7&
)LJXUH 6XUIDFHWHPSHUDWXUHDWGLIIHUHQWGLVWDQFHVIURPWKHFRUULGRULQOHWJULOOHV
SUMMARY AND CONCLUSIONS
7KHSDSHUSUHVHQWVDWUDQVLHQWWKHUPDOPRGHOIRUDEXLOGLQJZLWKPL[HGPRGHFRROLQJDQGKLJKOHYHOVRI WKHUPDO
PDVV7KHPRGHOLVYDOLGDWHGZLWKPHDVXUHPHQWGDWDIURPDQRFFXSLHGLQVWLWXWLRQDOEXLOGLQJDQGLWLVWKHQXVHGIRUWKH
314
ASHRAE Transactions

)LJXUH 1LJKWFRROLQJDVVHVVPHQW&RQYHFWLYHKHDWUHPRYDOOHIWDQG DLUWHPSHUDWXUHRQWKHIROORZLQJGD\ULJKW
SHUIRUPDQFHHYDOXDWLRQRI YDULRXVQLJKWFRROLQJVWUDWHJLHV7KHPRVWLPSRUWDQWILQGLQJVFDQEHVXPPDUL]HGDVIROORZV
7KHPRGHOLQJDFFXUDF\GHSHQGVRQWKHVHOHFWLRQRI WKHFRQYHFWLYHKHDWWUDQVIHUFRHIILFLHQWWKXVSUHYLRXVNQRZOHGJH
RI WKHDLUIORZSDWWHUQLVHVVHQWLDO
5HGXFLQJWKHORZWHPSHUDWXUHVHWSRLQWRI WKHQLJKWYHQWLODWLRQVFKHGXOHFDQVLJQLILFDQWO\HQKDQFHWKHKHDWUHPRYDO
IURPWKHEXLOGLQJWKHUPDOPDVV7KHFRQYHFWLYHKHDWUHPRYDOLVFORVHWRDQGWLPHVKLJKHUZLWKWKHDLUVWUHDP
DWDQDYHUDJHWHPSHUDWXUHRI &)FRPSDUHGWRWKDWZLWKDQDLUVWUHDPDW&)DQG&)
UHVSHFWLYHO\
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
(3) 7KHQLJKWFRROLQJRSHUDWLRQKRXUVVKRZDQLPSRUWDQWHIIHFWRQWKHIROORZLQJGD\·VFRROLQJGHPDQG$QWLFLSDWRU\
FRQWUROVWUDWHJLHVVKRXOGEHGHYHORSHGWRRSWLPL]HEXLOGLQJSHUIRUPDQFH
REFERENCES
$WKLHQLWLV$.%XLOGLQJWKHUPDODQDO\VLV²0DWKFDGHOHFWURQLFERRNQGHG%RVWRQ0DWKVRIW,QF
%UDJHU *6 GH 'HDU 5- 7KHUPDO DGDSWDWLRQ LQ WKH EXLOW HQYLURQPHQW D OLWHUDWXUH UHYLHZ (QHUJ\ DQG
EXLOGLQJV
&DUULOKRGD*UDFD*6LPSOLILHGPRGHOVIRUKHDWWUDQVIHULQURRPV3K'GLVVHUWDWLRQ8QLYHUVLW\RI&DOLIRUQLD
6DQ'LHJR
GH 'HDU 5- %UDJHU *6 'HYHORSLQJ DQ DGDSWLYH PRGHO RI WKHUPDO FRPIRUW DQG SUHIHUHQFH $6+5$(
7UDQVDFWLRQV$
(PPHULFK6-3ROLGRUR%$[OH\ -:,PSDFWRIDGDSWLYHWKHUPDO FRPIRUWRQFOLPDWLF VXLWDELOLW\RIQDWXUDO
YHQWLODWLRQLQRIILFHEXLOGLQJV(QHUJ\DQG%XLOGLQJV
/%1/:,1'2:9HUVLRQ/DZUHQFH%HUNHOH\1DWLRQDO/DERUDWRU\%HUNHO\&$
0D\2VWHQGRUS 3 +HQ]H *3 &RUELQ &' 5DMDJRSDODQ % )HOVPDQQ & 0RGHOSUHGLFWLYH FRQWURO RI
PL[HGPRGHEXLOGLQJVZLWKUXOHH[WUDFWLRQ%XLOGLQJDQG(QYLURQPHQW
0RXULNL(6RODU$VVLVWHG+\EULG9HQWLODWLRQLQDQ,QVWLWXWLRQDO%XLOGLQJ0DVWHUWKHVLV'HSDUWPHQWRI%XLOGLQJ
&LYLODQG(QYLURQPHQWDO(QJLQHHULQJ&RQFRUGLD8QLYHUVLW\0RQWUHDO
1HLVZDQJHU/-RKQVRQ*$&DUH\93$QH[SHULPHQWDOVWXG\RIKLJK5DOHLJKQXPEHUPL[HGFRQYHFWLRQLQD
UHFWDQJXODUHQFORVXUHZLWKUHVWULFWHGLQOHWDQGRXWOHWRSHQLQJV7UDQVDFWLRQVRI$60(
3HUH] 5 ,QHLFKHQ 3 6HDOV 5 0LFKDOVN\ - 6WHZDUW 5 0RGHOLQJ 'D\OLJKW $YDLODELOLW\ DQG ,UUDGLDQFH
&RPSRQHQWVIURP'LUHFWDQG*OREDO,UUDGLDQFH6RODU(QHUJ\
6SLQGOHU +& 1RUIRUG /. D 1DWXUDOO\ YHQWLODWHG DQG PL[HGPRGH EXLOGLQJV ² 3DUW , 7KHUPDO PRGHOLQJ
6SLQGOHU +& 1RUIRUG /. E 1DWXUDOO\ YHQWLODWHG DQG PL[HGPRGH EXLOGLQJV 3DUW ,, 2SWLPDO FRQWURO
$6+5$(

SA-12-C040
Contaminants in Hotel Room Exhaust Air
W. Brad M. Stanley Bryan K. Ligman

ABSTRACT
Hotels currently exhaust conditioned air to comply with standards such as ASHRAE 62.1-2010 (Std. 62.1). Std. 62.1 requires each
hotel guest room to exhaust the toilet area, which contains conditioned “hotel room air” with minimal contaminant sources present when the room
is vacant. Certain hotel sites have been able to obtain significant energy savings by “recycling” the room exhaust air through a filtration system
(both gas-phase and particulate). This approach may be one way to significantly impact the total energy used by large hotel buildings. However,
many engineers and property owners may be skeptical of this approach, not knowing the types and concentrations of contaminants that are in the
room exhaust air stream.
In order to provide data to the IAQ industry on this concept, the authors performed direct air sampling at a large atrium hotel currently
“recycling” guest room toilet exhaust through an air filtration system for energy savings. This air sampling included 75 common indoor volatile
organic compounds targeted in the EPA TO-15 method, eight aldehydes sampled with the EPA TO-11A method, ammonia sampled using
OSHA ID-188/ID-164, hydrogen sulfide sampled using CAS AQL 110, non-viable and viable fungi sampled using an active impaction
spore trap sampler and an active culture plate impaction sampler, respectively. Bacteria were sampled using an active culture plate impaction
sampler. The contaminant make-up and concentrations will be presented from the sampling upstream of the air filtration system.
INTRODUCTION
Large hotels are a specifically intriguing case because the air volumes exhausted can be large. The guest rooms remain
unoccupied for large portions of time, causing the air exhausted from toilet areas to be essentially free of people generated
contamination. In many cases, the hotel design requires this tempered and dehumidified air to be continuously thrown
away to the outdoors in order to keep building pressure relationships and airflows under control. “Recycling” a portion of
the exhaust air through an air filtration system (both gas-phase and particulate) allows hotels to “recycle” the energy used to
originally condition that exhaust air.
Although the concept is not new, the idea of “recycling” exhaust air is understandably viewed skeptically. There is an
unknown, i.e. what contaminants are in the exhaust air. On the other hand, hotels exist which have used an “air recycling”
strategy through air cleaning for several decades and enjoyed both acceptable air quality and comparably low HVAC energy
costs. Understanding the contaminants in the hotel guest room exhaust air will allow engineers and indoor air quality
professionals to make better informed decisions related to energy savings through “air recycling”.
AIR SAMPLING METHODOLOGY
Contaminants
The authors desired to sample for airborne contaminants that would represent the areas of concern for indoor air
quality. The areas of concern included building product and furnishing emissions, people generated contaminants
(bioeffluents and contaminants generated from various activities), and contaminants perceived to be specifically associated
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
with toilet room exhaust. There was no characterization for environmental tobacco smoke as this was a nonsmoking
Brad Stanley is the Gas-Phase Global Technical Support Leader AAF and currently serves as Vice Chair for ASHRAE TC2.3. Bryan Ligman
is the IAQ Technical Market Manager for AAF Americas and a Council-certified Indoor Environmental Consultant (CIEC).
316
ASHRAE Transactions
facility. The list of contaminants and analytical methods is shown below. The volatile organic compound sampling and
analysis method targeted 75 chemical compounds commonly found in indoor air representing both building material and
furnishing emissions and people generated contaminants. The aldehyde sampling and analysis method targeted 8 aldehydes
representing building material and furnishing emissions and, to a lower degree, people generated contaminants. Ammonia
and hydrogen sulfide represented bioeffluents and sewer system contaminants. The bioaerosols were of particular interest
because of the perception that their concentrations may be elevated in potentially elevated moisture locations such as toilet
(or shower) exhaust air.
1. Airborne Chemicals
a. Volatile Organic Compounds (speciated EPA method TO-15)
b. Aldehydes (EPA method TO-11A)
i. acetaldehyde, benzaldehyde, butyraldehyde, formaldehyde, isovaleraldehyde, n-hexaldehyde,
propionaldehyde, valeraldehyde
c. Ammonia (OSHA ID-188/ID-164)
d. Hydrogen Sulfide (CAS AQL 110 using spectrophotometer)
2. Bioaerosols
a. Non-Viable and Viable Fungi (Spore Trap Active Impaction and SAS Sampler with Anderson Style Active
Plate Impaction)
b. Bacteria (SAS Sampler with Anderson Style Active Plate Impaction). A two plate protocol was used to select
out gram-negative rods, such as E. coli, coliforms or non-fermenters (Pseudomonas species).
Sampling Time Period and Method

For all contaminants, the authors desired to use extended sampling times to allow for detection of lower
concentrations as well as to capture events that may be missed in short grab sample type sampling. As much as possible,
the sampling period began on the afternoon of one day and lasted through the night until the afternoon of the following
day. The purpose was to catch nighttime activities of hotel guests who are not present during the day. Samples were taken
two times in the spring about one month apart. This paper contains the results from the second sample date. Table 1
shows the approximate sample periods and medium used for each contaminant type.
Table 1. Sampling Time Periods and Medium for Each Contaminant Type
Contaminant Example Sampling Period Sampling Medium
volatile organic compounds 24 hrs Summa Canister w/ flow controller
aldehydes 24 hrs Passive Radiello DNHP tubes (code 165)
ammonia 4 hrs Sulfuric Acid-Coated Anasorb-747 (carbon bead) tubes
hydrogen sulfide 24 hrs Passive Radiello tubes (code 170)
Location and Sampling Setup

Figure 1 shows the sampling location within the plenum of the exhaust air handling unit. The hotel has all toilet
exhaust risers ducted to a common air handler location that feeds the air filtration system and discharges to the hotel
atrium. The samples were taken upstream of the air filtration system taking care to place the sampling medium as close to
the center of the airstream as possible.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 1 – Schematic of Air Sampling Location (Plan View)
RESULTS
Tables 2 through 7 show the airborne contaminant concentrations. Table 2 displays the volatile organic compound
results; only 27 of the 75 compounds were above the method detection limits. Table 3 contains the aldehyde results,
indicating the presence of three aldehydes: acetaldehyde, formaldehyde, and n-hexaldehyde. Table 4 puts forth the
ammonia and hydrogen sulfide results, both were below detection limits. Table 5 presents the non-viable fungi (spores)
results indicating the presence of basidiospores, botrytis, Cladosporium, hyphal elements, and Penicillium/Aspergillus type
spores. Table 6 shows the viable fungi results indicating the presences of Acremonium sp., Cladosporium sp., non-sporulating
colonies, Penicillium sp., and Rhodotorula sp. Table 7 displays the bacteria results indicating the presence of coag-negative
Staphylococcus sp. and Micrococcus sp. No samples contained escherichia coli bacterium.
Table 2. Volatile Organic Compound Air Sampling Results

Volatile Organic Compound * CAS # Plenum Concentration
Result MRL Result MRL
(μg/m³) (μg/m³) (ppbv) (ppbv)
1,4-dichlorobenzene 106-46-7 0.2 0.15 0.034 0.026
2-propanol (isopropyl alcohol) 67-63-0 25 1.5 10 0.63
318
ASHRAE Transactions
Volatile Organic Compound * CAS # Plenum Concentration
acetone 67-64-1 49 7.7 20 3.2
acetonitrile 75-05-8 2 0.77 1.2 0.46
alpha-pinene 80-56-8 2.6 0.77 0.46 0.14
benzene 71-43-2 0.85 0.15 0.27 0.048
bromodichloromethane 75-27-4 0.95 0.15 0.14 0.023
carbon tetrachloride 56-23-5 0.48 0.15 0.077 0.024
chloroform 67-66-3 6.6 0.15 1.4 0.032
chloromethane 74-87-3 0.44 0.31 0.21 0.15
dichlorodifluoromethane (cfc 12) 75-71-8 2.6 0.77 0.53 0.16
d-limonene 5989-27-5 11 0.77 2 0.14
ethanol 64-17-5 980 39 520 20
ethyl acetate 141-78-6 5.6 0.77 1.5 0.21
ethylbenzene 100-41-4 2.3 0.77 0.52 0.18
m,p-xylenes 179601-23-1 6.1 0.77 1.4 0.18
n-butyl acetate 123-86-4 2.4 0.77 0.51 0.16
n-heptane 142-82-5 1.1 0.77 0.27 0.19
n-hexane 110-54-3 1.2 0.77 0.33 0.22
n-octane 111-65-9 0.92 0.77 0.2 0.16
o-xylene 95-47-6 1.2 0.77 0.28 0.18
propene 115-07-1 8.3 0.77 4.8 0.45
tetrachloroethene 127-18-4 5.4 0.15 0.8 0.023
toluene 108-88-3 6.2 0.77 1.6 0.2
trichloroethene 79-01-6 0.49 0.15 0.092 0.029
trichlorofluoromethane 75-69-4 1.3 0.15 0.23 0.027
trichlorotrifluoroethane 76-13-1 0.48 0.15 0.063 0.02
* Method evaluated the presence of 75 total volatile organic compounds. There are only 27 compounds indicated
above. All others compounds were below the method reporting limit.
Table 3. Aldehyde Air Sampling Results

Aldehyde CAS# Plenum Concentration
acetaldehyde 75-07-0 18.0 1.7 9.8 0.9
formaldehyde 50-00-0 28.0 1.4 23.0 1.1
n-hexaldehyde 66-25-1 15.0 7.7 3.6 1.9
benzaldehyde + 100-52-7 - - - -
butyraldehyde + 123-72-8 - - - -
isovaleraldehyde + 590-86-3 - - - -
propionaldehyde + 123-38-6 - - - -
valeraldehyde + 110-62-3 - - - -
+ These aldehydes were monitored approximately one month before at the same sampling location and found to be
below the method reporting limit. Therefore, the analytical method was revised to exclude these from this air
sampling visit.
$6+5$(
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---

Table 4. Ammonia and Hydrogen Sulfide Air Sampling Results
Compound CAS# Plenum Concentration
Result ¥ MRL Result ¥ MRL
3
(mg/m , μg/m³) (mg/m³, μg/m³) (ppmv, ppbv) (ppmv, ppbv)
ammonia 7664-41-7 ND 0.087 ND 0.12
hydrogen sulfide 7783-06-4 ND 2.6 ND 3.6
¥ Concentrations for ammonia are in mg/m3 and ppmv. Concentrations for hydrogen sulfide are in g/m3 and ppbv.
ND = nondetect.
Table 5. Non-Viable Fungi (Spore) Air Sampling Results

Spore Identifier Plenum Concentration
(Spores/m3)
Sample 1 Duplicate Sample
algae - -
Alternaria - -
ascospores 75 225
basidiospores 633 633
Botrytis - -
Cercospora - -
Cladosporium 83 142
clear brown - -
Curvularia - -
Curvularia-like - -
Dechslera/Bipolaris group - -
Epicoccum - -
Helicosporium/Helicomyces - -
hyphal elements 8 -
Monochaetia - -
Nigrospora - -
Oidium - -
Penicillium/Aspergillus group 150 83
Peronospora - -
smuts, Periconia, myxomycetes - -
Torula - -
unknown - -
Total 949 1083
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Table 6. Viable Fungi Air Sampling Results

Viable Fungi Identifier Plenum Concentration
(CFU/m3)
Acremonium species - 6
Alternaria species - -
arthrospore former - -
Aspergillus niger - -
Aspergillus species - -
Aureobasidium species - -
Cladosporium species 22 39
Epicoccum species - -
320
ASHRAE Transactions
Viable Fungi Identifier Plenum Concentration
(CFU/m3)
Fusarium species - -
non-sporulating colonies 6 -
Olpitrichum species - -
Paecilomyces species - -
Penicillium species (mixed) 11 -
Rhodotorula species - 6
yeast - -
Total 39 51
Table 7. Bacteria Air Sampling Results

3
Bacteria Identifier Plenum Concentration (CFU/m )
Bacillus species - -
coag-negative Staphylococcus species 22 -
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Corynebacterium species - -
Micrococcus species 22 22
non-fermentative gram negative rod - -
Streptomyces species - -
Total 44 22
Note: The presence of Escherichia coli bacteria was not present in any of the samples collected
CONCLUSION
There currently exists little data on the environmental characterization of hotel room toilet exhaust. Tables 2 through
7 characterize a collective hotel room toilet exhaust airstream over a 24 hour period in terms of airborne chemical presence,
non-viable and viable fungi presence, and bacteria presence. This data can help engineers, building owners, and indoor air
quality professionals become familiar with the types of contaminants and the range of concentrations that may be present in
such airstreams to assist with designs for recirculation of similar exhaust air types.
ACKNOWLEDGMENTS
The authors would like to thank Aflal Rahmathullah and Rob Schick for their assistance in gathering this data.
REFERENCES
ASHRAE. (2010). ANSI/ASHRAE Standard 62.1-2010 Ventilation for Acceptable Indoor Air Quality. Atlanta, GA: American
Columbia Analytical Services. Air Quality Lab Method 110 (CAS AQL 110). Simi Valley, CA.
U.S. Environmental Protection Agency. (1999). Method TO-15: Determination of Volatile Organic Compounds (VOCs) In Air
Collected In Specially-Prepared Canisters And Analyzed By Gas Chromatography/Mass Spectrometry (GC/MS) (EPA/625/R-
96/010b). Cincinnati, OH: Center for Environmental Research Information, Office of Research and Development.
U.S. Environmental Protection Agency. (1999). Method TO-11A: Determination of Formaldehyde in Ambient Air Using Adsorbent
Cartridge Followed by High Performance Liquid Chromatography (HPLC) [Active Sampling Methodology]. Cincinnati, OH: Center
for Environmental Research Information, Office of Research and Development.
U.S. Occupational Safety & Health Administration. (2002). Method ID-188: Ammonia in Workplace Atmospheres - Solid Sorbent.
Sandy, UT: Methods Development Team, Industrial Hygiene Chemistry Division.
$6+5$(

SA-12-C041
(YDOXDWLRQRI7KHUPDO&RPIRUWLQ/RZ/RDG
+RPHV6XSSOLHGE\+LJK6LGHZDOO$LU-HWV

(O+DVVDQ5LGRXDQH3K'
Member ASHRAE
$%675$&7
Mixing of conditioned air with room air is an essential factor for uniform comfort in homes. This can be achieved by
creating the correct combination of temperature and air motion in the occupied zone. Discomfort can be caused by excessive
air motion, excessive room air temperature variations, or failure to deliver or distribute air according to load requirements.
Effective mixing is even more challenging in low load homes with downsized HVAC systems because the design flow rates
are lower.
We used computational fluid dynamics modeling to evaluate the performance of high sidewall air supply for residential
applications in heating mode, and varied the supply velocity and supply temperature. We considered two configurations,
corresponding to an unfurnished and a furnished room and compared the results, then evaluated thermal comfort by
determining the air diffusion performance index (ADPI). The higher the ADPI value, the more comfortable the space is for
most occupants. High sidewall supply inlets can achieve good mixing and provide acceptable comfort levels.
Recommendations are given for the system operating conditions to increase comfort.
,1752'8&7,21
Thermal comfort in homes is directly affected by indoor air velocity, which is specified in thermal comfort indices such
as the predicted mean vote (PMV), the predicted percentage of dissatisfied (PPD), and the air diffusion performance index
(ADPI) (Chow and Wong, 1994). The location where the jet enters the room must be evaluated to minimize the draft risk
(Corgnati et al., 2002), and the occupied and unoccupied zones must be defined. The unoccupied zone should be used to
deliver the high-velocity supply air required for mixing (John, 2011).
Numerous studies during the past two decades primarily predicted indoor airflow patterns with experimental
investigation and numerical calculation. Experimental models have been widely used to predict air distribution performance
in buildings. Chow and Wong (1994) studied the air speed induced by a diffuser in an environmental chamber under different
flow conditions. The measurements showed no obvious correlation between ADPI and the supply flow rate. The diffuser
throw was linearly related to the Archimedes number, which represents the ratio of buoyancy to inertial force exerted on the
fluid. Yu et al. (2007) studied airflow in a ceiling slot-ventilated enclosure in a 1:3 scale model. They measured the velocity
and temperature profiles, temperature decays, and thermal boundary layers. The data was used to develop empirical models.
Zang et al. (2009) used an environmental chamber to mimic a section of an airliner cabin. The experimental model used
tracer gas and particles for contaminant transport simulation. The experiment approximated thermo-fluid boundary conditions
and used heated boxes to simulate passengers. Ultrasonic anemometers were used to measure the air velocity, temperature
distribution, and contaminant concentrations.
Full-scale experimental models limit the simplifying assumptions associated with small-scale and analytical models.
However, they are generally expensive and time consuming and cannot easily cover all possible operating conditions. A
growing trend is to use laboratory experiments to obtain data to validate computer models, and then use the validated models
to predict ventilation performance of HVAC system designs (Chen, 2009). Corgnati et al. (2009) used computational fluid
dynamics (CFD) to perform a numerical study in cooling mode, then validated it against experimental data from a full-scale
test room that simulated a typical two-desk office. They then extended the numerical study to couple mixing ventilation and
cold radiant ceiling panels. The results showed that coupled systems increased the jet throw and significantly decreased the
(O+DVVDQ5LGRXDQH is a Research Associate of the Residential Building Research Group at the National Renewable Energy Laboratory,
Golden, CO.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
322
ASHRAE Transactions
draft risk.
Pure numerical studies dominated the research literature of the past two decades. CFD models in particular have been
widely used to study indoor air distribution, air quality, thermal comfort, and HVAC system performance. Norton et al.
(2006, 2007) conducted reviews of CFD applications for ventilation studies in greenhouses and animal production facilities.
Chung and Lee (1996) investigated three air diffusion models in cooling mode. These include a room with one inlet – one
outlet, two inlets – one outlet, one inlet – two outlets. They evaluated thermal comfort in terms of PPD and the ADPI.
Variations of PPD and ADPI with the inlet velocity and temperature showed that a higher PPD or lower ADPI is associated
with a higher velocity and lower inlet temperature. This study also underscores the importance of inlet/outlet locations for
evaluating indoor thermal comfort.
Lam and Chan (2001) used CFD to analyze the temperature distribution and air movement in an air-conditioned
gymnasium. Different exhaust positions were considered, sharing one inlet located high on a sidewall. They investigated the
effects of thermal stratification on energy performance with respect to the HVAC plant oversizing issue. The smallest cooling
load requirement was in the configuration with ceiling exhaust. This configuration was identified, in terms of energy
efficiency, as the best design arrangement. CFD models have also been used to improve other building simulation tools so
ventilation performance can be more accurately predicted (Chen, 2009). Wang and Chen (2007) coupled CFD with a
multizone airflow program to improve the prediction of airflow and contaminant in an entire building.
Despite the wealth of literature about this problem, information about air distribution in high-performance homes with
low thermal loads is lacking. The objective was to narrow this gap by identifying operating parameters that ensure robust
mixing and design space conditioning systems for these homes to avoid operating parameters that will result in poor air
mixing. We used CFD to evaluate the thermal performance of high sidewall supply jets in heating mode. The three-
dimensional model considers a single room in a high-performance home and evaluates thermal comfort as a function of
operating conditions.
180(5,&$/02'(/
Modeled room dimensions reflected typical residential designs and had a 9-ft (2.74-m) ceiling, 11-ft (3.35-m) width,
transverse to the inlet jet throw, and 12-ft (3.66-m) length. Room length is in the direction of the jet throw. The room is
supplied by an 8-in. (0.2-m) × 1-in. (0.025-m) inlet modeled as an open jet with air discharged in a uniform profile normal to
the inlet face. The inlet was mounted with the upper edge of the opening 9 in. (0.23 m) from the ceiling and was modeled as
uniform velocity boundaries. Actual velocity profiles at the inlet are likely to be non-uniform because of upstream ductwork.
However, the simplified approach is expected to accurately predict the airflow characteristics used to evaluate thermal
comfort. The return outlet was located directly below the supply inlet, with the lower edge of the opening 4 in. (0.1 m) from
the floor. The outlet area was much larger than the inlet, to have a low air speed approaching the opening. The outlet was
modeled as a completely open pressure boundary, set to 0 psig (0 Pa).
Supply velocities varied from 214 fpm (1.1 m/s) to 1430 fpm (7.3 m/s) and supply temperature ranged from 95º to
140ºF (308 to 333 K); the initial room temperature was fixed at 65ºF (291 K). The supply was controlled via a thermostat set
at 71ºF (294.81 K) and mounted on the wall parallel to the inlet air jet at 4 ft (1.22 m) from the floor (see Figure 1).
Simulations ended when 1 air change was supplied to the room. Thermal comfort was then evaluated by determining the
ADPI, which is defined in a zone between 4 in. (0.1 m) and 72 in. (1.83 m) from the floor, 2 ft (0.61 m) from exterior walls,
and 1 ft (0.3 m) from interior walls (ASHRAE 2010). ADPI is a simple comfort criterion based on local temperature, set-
point temperature, and local air speed, and uses effective draft temperature
θ ≡ (T − Tset ) − 0.07(Vx − 30) (1)
where T is local air temperature, Tset is set-point temperature, and Vx is local air speed. Effective draft temperature, ș, must be
–3º (–1.7 K) to 2ºF (1.1 K) and local air speed below 70 fpm (0.35 m/s) for maximum comfort (ASHRAE, 2009). In the zone
of interest, effective draft temperature and air speed were determined for each cell. The number of cells meeting these
requirements was compared to the total number of cells to determine the ADPI for the simulation. The room has an
acceptable ADPI level when its value is greater than 0.7. System operation with ADPI below 0.7 is not recommended.

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

0(7+2'2/2*<
The finite volume method (software ANSYS FLUENT 2010) was used to model airflow and heat transfer in the room.
A fully implicit approach was used for discretizing the transient conservation equations on the grounds of its superior
stability. The Re-Normalization Group (RNG) k-İ model was used for turbulence. The QUICK scheme was used for the
discretization of momentum, k-İ, and energy conservation equations. A second order body-force-weighted scheme was
utilized in the pressure discretization and the SIMPLE scheme was employed in the pressure-velocity coupling. A time step
of 0.5 s was used in the simulations.
A load density of 14.6 Btu/h·ft2 (46.02 W/m2) was applied, which is representative of perimeter zones of high-
performance homes in heating mode. This density was multiplied by the floor area and the outcome was distributed between
the ceiling and at the exterior wall opposite to the air jet (Figure 1). The proportional thermal load distribution between the
ceiling and the wall was based on their respective R-values (R-20 for the wall and R-30 for ceiling). Wall characteristics,
gridding, and other solution parameters were chosen to ensure convergence. Incompressible ideal gas law was used for
density calculations. Figure 1 shows the geometry outline of the computational domain. The room was furnished with three
items: a 33-in. (0.84-m) × 61-in. (1.55-m) × 23-in. (0.58-m) couch, a 24-in. (0.61-m) × 36-in. (0.91-m) × 32-in. (0.81-m) TV
stand, and a 3-in. (0.08-m) × 26-in. (0.66-m) × 21-in. (0.53-m) TV with a heat generation of 239 Btu/h (70 W). The couch is
12 in. (0.3 m) from the exterior wall and the TV stand is 24 in. (0.61 m) from the supply/return wall. These items are centered
on the floor in the direction of the throw. Room walls were considered to be massless and thermal loads were applied at the
inner surfaces of ceiling and exterior wall. The floor and the remaining walls of the room were defined to be adiabatic.
Thermal load
6XSSO\
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
7KHUPRVWDW
7KHUPDOORDG
5HWXUQ
Figure 1 Sketch of the computational domain showing the locations of supply, return, thermostat, and furniture in
the room. Thermal loads are applied only to the ceiling and to the right wall.
5(68/76$1'',6&866,21
This section presents the simulation results in heating mode using an 8-in. (0.2-m) × 1-in. (0.025-m) supply inlet. The
effects of furniture and indoor heat generation sources on mixing are presented and discussed. The choice of this inlet size
was based on a previous analysis (Ridouane, 2011) conducted in an unfurnished room in which we considered three supply
inlet sizes: an 8-in. (0.2-m) × 1-in. (0.025-m), an 8-in. (0.2-m) × 4-in. (0.1-m), and a 10-in. (0.25-m) × 6-in. (0.15-m).
Results, in heating mode, showed that the 8 × 1 inlet performed better than the 8 × 4 and 10 × 6 inlets, and provided
acceptable ADPI at low system capacities (low fan powers). The intermediate (8 × 4) and large (10 × 6) inlets required higher
fan powers to reach the acceptable ADPIs.
All simulations shared uniform room temperature of 65ºF (291.5 K) as an initial condition. The thermostat set-point
was set at 71º (294.82) ± 1ºF (0.56 K). Each combination of supply velocity and temperature defined HVAC system capacity
and was supplied to meet the peak load (heat loss through the building envelope and heat generated by the TV). We explored
a wide range of system capacities, including undersized and oversized systems, to investigate the effects of operating
conditions on ADPI. Results are first presented for the unfurnished room, and constitute a base case for comparison. The

324
ASHRAE Transactions
performance evaluation of the furnished room includes both cases corresponding to with and without thermal radiation
exchange between room surfaces.
8QIXUQLVKHG5RRP
Isotherm plots in the symmetry plane of the unfurnished room are depicted in Figure 2. Two combinations of supply
temperature and supply velocity are presented: 95ºF (308 K) air supplied at 1174 fpm (5.96 m/s) and 120ºF (322 K) air
supplied at 558 fpm (2.83 m/s). These conditions meet the peak load entering the room through ceiling and exterior wall.
Good mixing is seen at low supply temperature of 95ºF (308 K) because of high velocities. The room was highly stratified for
120ºF (322 K) supply air and the jet had no momentum to penetrate into the lower half of the room. Even when the supply
velocity is fixed, lower supply temperatures provide better mixing because of buoyancy.
˚ ˚

Figure 2 Temperature distribution in the symmetry plane of the unfurnished room when system capacity is equal to
peak load: (a) 95ºF (308 K) air supplied at 1174 fpm (5.96 m/s) and (b) 120ºF (322 K) air supplied at 558
fpm (2.83 m/s).
Figure 3 illustrates the time history of air temperature near the thermostat and the corresponding supply velocity for two
supply velocities of 558 fpm (2.83 m/s) and 714 fpm (3.62 m/s) sharing a supply temperature of 120ºF (322 K). The air
temperature increased quickly at high supply velocity of 714 fpm (3.62 m/s) and entered a quasi-steady state regime
characterized by cycling around the thermostat set-point. At low velocity of 558 fpm (2.83 m/s), the air was warming slowly
and its temperature was below the set-point at the time of 1 air change. Here we see two system capacities corresponding to a
15% undersized system running continuously (558 fpm) and a 9% oversized system characterized by cycling (714 fpm). Low
ADPI values were found for these conditions: 0.16 for VS = 558 fpm (2.83 m/s) and 0.24 for VS = 714 fpm (3.62 m/s).

(a) (b) --`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---

Figure 3 (a) Time history of air temperature near the thermostat of the unfurnished room and (b) corresponding
velocity at the inlet when supplied by 120ºF (322 K) air.
)XUQLVKHG5RRP
To visualize the impact of furniture and indoor heat generation on mixing, we compared the temperature distribution at
the symmetry plane of unfurnished (a) and furnished (b) rooms. The plots of Figure 4 are for 105ºF (313.7 K) supply

$6+5$(

temperature and 829 fpm (4.2 m/s) supply velocity. Comparable mixing quality was observed; the unfurnished room had
slight advantage, manifested by a relatively reduced stratification region near the floor.

Figure 4 Comparison of temperature distribution at the symmetry plane of the room supplied by 105ºF (313.7 K) air
at 829 fpm (4.2 m/s): (a) unfurnished room and (b) furnished room.
Figure 5 illustrates the impact of furniture on thermostat readings for two operating conditions: a low supply
temperature of 105ºF (313.7 K) with a high supply velocity of 829 fpm (4.2 m/s) on the left graph, and a high supply
temperature of 140ºF (333 K) with a low supply velocity of 400 fpm (2 m/s) on the right graph. Similar trends are observed
for both conditions and characterized by a faster warm-up of air near the thermostat in the furnished room. This is due to the
change in air jet direction as it impinges the couch surface on its way down and directs warm air toward the thermostat. This,
however, does not necessarily mean better thermal comfort in the furnished room, as other regions of the room may still be
cold.

Figure 5 Comparison of temporal evolution of the thermostat readings in unfurnished and furnished rooms: (a) TS =
105ºF (313.7 K) and VS = 829 fpm (4.2 m/s); (b) TS =140ºF (333 K) and VS = 400 fpm (2 m/s).

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 6 shows the capacity of different systems considered relative to the peak load for unfurnished (a) and furnished
(b) rooms. The peak load was fixed and system capacity was varied with supply velocity and supply temperature. The
difference in the value of the peak load between unfurnished and furnished rooms is due to the internal heat generated by the
TV. The modeled TV generates 239 Btu/h (70 W) and slightly reduces the peak load in the furnished room. System capacities
were distributed around the peak load and include undersized and oversized systems. Undersized and oversized systems were
included to show full range of ADPI trends.

326
ASHRAE Transactions

Figure 6 Capacity of different systems considered relative to peak load: (a) unfurnished room and (b) furnished
room.
The effects of thermostatic control and peak load on the mixing behavior are summarized in Figure 7 by the variations
of ADPI as a function of supply velocity. The figure shows that ADPI started with low values at low velocities. As the
mixing improved because of higher supply velocities, ADPI increased and reached a maximum in the unfurnished room
supplied by 95ºF (308 K) and decreased at high supply velocities. The reduction in ADPI is due to drafts at high supply
velocities. For all other cases ADPI is characterized by a monotonic increase over the range of conditions considered. ADPI
results for 95ºF (308 K) and 105ºF (314 K) supply temperatures (Figure 7(a)) show a fairly broad range of acceptable ADPI,
greater than 0.7, than can be achieved at 120ºF (322 K) and 140ºF (333 K) supply temperatures. For the high supply
temperatures of 120ºF (322 K) and 140ºF (333 K), higher supply velocities (oversized systems) were required to overcome
buoyancy effects and reach acceptable ADPIs in the room. At fixed supply velocity, comparable ADPIs were obtained in
furnished and unfurnished rooms, except for 105ºF (314 K) supply temperatures where the furnished room presented lower
ADPIs.

Figure 7 Comparison of ADPI trends in furnished and unfurnished rooms: (a) low supply temperatures of 95ºF (308
K) and 105ºF (313.7 K) and (b) high supply temperatures of 120ºF (322 K) and 140ºF (333 K).
(IIHFWRI6XUIDFH5DGLDWLRQ
Figure 8 shows the temperature distribution in the furnished room when supplied by 105ºF (313.7 K) air at 829 fpm (4.2
m/s). The figure shows the effect of radiation exchange between room surfaces by comparing the case when surface radiation
is considered (b) to the case when this is ignored (a). A surface emissivity of 0.9 was assigned to all room surfaces. A wider
temperature range occurs when radiation is ignored and may lead to lower comfort levels. Radiation exchange reduced the

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

temperature range and eliminated extreme undesirable temperatures in the room. This resulted in a more uniform temperature
distribution and consequently better mixing.

˚

Figure 8 Temperature distribution in the furnished room without (a) and with (b) thermal radiation exchange
between room surfaces when supplied by 105ºF (313.7 K) air at 829 fpm (4.2 m/s).
To quantify the effect of radiation on thermal comfort, we evaluated ADPI for a selected number of system operating
conditions. For each supply temperature we selected a supply velocity that provided a system capacity equal to peak load.
The outcome is summarized in Table 1 in terms of ADPI for both cases corresponding to “With Radiation” and “Without
Radiation.” Thermal comfort level, measured by ADPI, improved with thermal radiation exchange. This was the case for all
conditions considered. The table also shows that ADPI decreases when supply velocity decreases and when supply
temperature increases. For these conditions, only 95ºF (308 K) supply temperature provided acceptable (greater than 0.7)
ADPI for the room.
7DEOH(IIHFWRIVXUIDFHUDGLDWLRQRQ$'3,LQWKHIXUQLVKHGURRPZKHQV\VWHPFDSDFLW\LV
HTXDOWRSHDNORDG
Supply Supply Velocity ADPI ADPI
Temperature (ºF) (fpm) Without Radiation With Radiation
95 1174 0.92 0.96
105 829 0.32 0.58
120 558 0.16 0.44
140 400 0.11 0.4
&21&/86,216
We used numerical simulations in CFD modeling to evaluate the performance of high sidewall air supply systems in
perimeter zones of residential applications. We simulated a typical residential room, supplied by an 8-in. (0.2-m) × 1-in.
(0.025-m) air inlet, over a wide range of operating conditions. The room was furnished with a couch, a TV stand, and a TV
generating 239 Btu/h (70 W), which was distributed uniformly on its surface. Results are presented in heating mode over a
wide range of supply temperatures and supply velocity. The effect of surface radiation on mixing is also presented. Model
output was used to determine how well the supply air mixes with the room air and quantifies the comfort level using ADPI.
The findings of this study follow:
• For a fixed system capacity, lower supply temperatures of 95ºF (308 K) and 105ºF (313.7 K) provided better mixing and
higher ADPIs for furnished and unfurnished rooms. The rooms were highly stratified for high supply temperatures of
120ºF (322 K) and 140ºF (333 K).
328
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
ASHRAE Transactions
• Comparable mixing quality was observed in furnished and unfurnished rooms; the unfurnished room showed a slight
advantage with relatively reduced stratification region near the floor.
• Thermal comfort level started with low values at low velocities and was characterized by a monotonic increase with
supply velocity. ADPI results for 95ºF (308 K) and 105ºF (314 K) supply temperatures showed a fairly broad range of
acceptable ADPI, greater than 0.7, than can be achieved at 120ºF (322 K) and 140ºF (333 K) supply temperatures. These
trends are directly related to buoyancy.
• Comparable ADPIs were obtained in furnished and unfurnished rooms, except for 105ºF (314 K) supply temperatures
where the furnished room presented lower ADPIs.
• Radiation exchange between room surfaces reduced the temperature range and eliminated extreme undesirable
temperatures in the room. As a result, thermal comfort level measured by ADPI improved for all conditions considered.
In summary, high sidewall air jets were demonstrated to achieve uniform mixing in the room and provide acceptable levels of
ADPI for specific operating conditions.
5()(5(1&(6
ANSYS. 2010. ANSYS FLUENT Manual. Canonsbury, PA : ANSYS, Inc.
ASHRAE. 2009. ASHRAE Fundamentals Handbook, 2009 IP Edition, Atlanta, GA, p. 20.13.
ASHRAE. 2010. ASHRAE Standard 55 – 2010. Thermal Environmental Conditions for Human Occupancy, Atlanta, GA.
Chen, Q. (2009). Ventilation Performance Prediction for Buildings: A Method Overview and Recent Applications. Building
and Environment 44, pp. 848-858.
Chow, W.K. and L.T. Wong. 1994. Experimental Studies on Air Diffusion of a Linear Diffuser and Associated Thermal
Comfort Indices in an Air-conditioned Space. Building and Environment 29(4):523-530.
Chung, K.C. and C.Y. Lee. 1996. Predicting Air Flow and Thermal Comfort in an Indoor Environment under Different Air
Diffusion Models. Building and Environment 31(1):21-26.
Corgnati, S.P., G.V. Fracastro, P.V. Nielsen, and M. Perino. 2002. Jet Drop and Draft Risk in a Mixing Ventilation Office
Experimental and Numerical Investigation. In: Proceedings of ROOMVENT 8th International Conference, Copenhagen,
Denmark.
Corgnati, S.P., M. Perino, G.V. Fracastro, and P.V. Nielsen. 2009. Experimental and Numerical Analysis of Air and Radiant
Cooling Systems in offices. Building and Environment 44, pp. 801-806.
John, D.A. 2011. Designing for Comfort: Selecting Air-Distribution Outlets. ASHRAE Journal 52(9):38-46.
Lam, J.C. and A.L.S. Chan. 2001. CFD Analysis and Energy Simulation of a Gymnasium. Building and Environment 36, pp.
351-358.
Norton, T. and D-W Sun. 2006. Computational Fluid Dynamics (CFD) – An Effective and Efficient Design and Analysis
Tool for the Food Industry: A Review. Trends in Food Science and Technology 17, pp. 600–620.
Norton, T., D-W. Sun, J. Grant, R. Fallon, and V. Dodd. 2007. Applications of Computational Fluid Dynamics (CFD) in the
Modeling and Design of Ventilation Systems in the Agricultural Industry: A Review. Bioresource Technology 98(12):2386–
2414.
Ridouane, E.H. 2011. Evaluation of Air Mixing and Thermal Comfort from High Sidewall Supply Air Jets. NREL/TP-550-
48664. Golden, CO: National Renewable Energy Laboratory.
Wang, L. and Q. Chen. 2007. Theoretical and Numerical Studies of Coupling Multizone and CFD Models for Building Air
Distribution Simulations. Indoor Air 17(5):348–361.
Yu, H., C.-M Liao, H.-M. Liang, and K.-C. Chiang. 2007. Scale Model Study of Airflow Performance in a Ceiling Slot-
ventilated Enclosure: Non-isothermal Condition. Building and Environment. 42(3):1142-1150.
Zhang, Z., X. Chen, S. Mazumdar, T. Zhang, and Q. Chen. 2009. Experimental and Numerical Investigation of Airflow and
Contaminant Transport in an Airliner Cabin Mockup. Building and Environment 44(1):85-94.

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

SA-12-C042
Experimental Study on Ceiling Radiant

Cooling Panel System with Personal Floor
Diffuser
Sei Ito Minoru Kawashima Yoshito Arai
Member ASHRAE
Michiya Suzuki Koji Murakami Tatsuo Nobe

Member ASHRAE
ABSTRACT
A ceiling radiant cooling panel (CRCP) system provides high comfort to occupants since it does not produce draught.
However, some occupants might feel uncomfortable in a room equipped with CRCP system, precisely due to the lack of
draught. To solve this problem, an HVAC system using CRCPs and personal floor diffusers is proposed. An experimental
facility for this HVAC system is constructed, and thermal performance tests as well as thermal environment surveys are
performed. This paper presents an outline of the experimental study on the abovementioned HVAC system and shows that the
system can efficiently remove the cooling load while maintaining comfortable indoor conditions for the occupants.
INTRODUCTION
Energy conservation technology has been in high demand in relation to efforts for realizing (net) zero energy buildings
(ZEBs). Ceiling radiant cooling panel (CRCP) systems represent a key energy conservation technology which can yield high
efficiency, since it uses moderately cold water (16í18°C (60.8í64.4°F)) and does not require energy to propel a fan for
removing the cooling load. Generally, a CRCP system provides high comfort to occupants because it does not produce
draught. However, some occupants might feel uncomfortable in a room equipped with a CRCP system, precisely due to the
lack of draught and the occupants have no chance to select the air current. To solve this problem, an HVAC system using
CRCPs and personal floor diffusers is proposed. It is assumed that CRCPs can remove the cooling load while personal floor
diffusers can provide both conditioned outlet air and air current if desired by the occupants. To demonstrate the satisfactory
performance of the proposed HVAC system, an experimental facility is constructed and used for conducting thermal
performance tests and thermal environment surveys. This paper presents an outline of the proposed system and reports results
of the thermal performance tests and thermal environment survey, which were performed with the aid of participants who
tested the system.
EXPERIMENTAL FACILITY AND THERMAL PERFORMANCE TEST

Experimental facility
The laboratory was composed of an outdoor room and a test room divided by a glass window, as shown in Fig. 1. The
test room occupied an area of about 66 m2 (710 ft2) and simulated the interior of an office. The HVAC system of the test
room was evaluated by applying dummy cooling loads and the air temperature of the outdoor room was controlled to match
the temperature in the test room in order to eliminate the influence of the perimeter area. All surfaces in the test room were
Sei Ito, Minoru Kawashima and Michiya Suzuki are research engineers in Shimizu Corporation, Tokyo, Japan. Yoshito Arai and Koji
Murakami are mechanical design engineers in Shmizu Corporation, Tokyo, Japan. Tatsuo Nobe is a professor in the Department of
Architecture, School of Architecture, Kogakuin University, Tokyo, Japan
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
330
ASHRAE Transactions
insulated to minimize the influence of ambient heat. The ceiling radiant cooling panels, CRCPs (arranged as shown in Fig. 1),
are consisted of general panels and panels with light fixtures. All CRCPs were composed of steel panels, copper pipes and
heat sinks, as shown in Figs. 2 and 3. The CRCPs specifications are listed in Table 1. In CPCPs, The heat sinks promote heat
transmission between the cold water supplied to the CRCP and the steel panel. The active area of the panels with light
fixtures in this case effectively half that of general panels; panels with light fixtures contained three heat sinks while general
panels contained six heat sinks. Accordingly, the active area of the CRCPs was 30 m2 (323 ft2) when cold water was supplied
to only the general panels and 37.5 m2 (404 ft2) when cold water was supplied to both general panels and panels with light
fixtures.
The arrows in Fig. 1 show the flow direction of the cold water, where 15 rows of panels were installed on the ceiling,
with three sequentially connected panels in each row. The flow direction of the supplied cold water was alternated in adjacent
row of panels such that the thermal radiation environment in the test room was uniform. The floor of the test room was raised,
and personal floor diffusers were installed on the floor under each desk in the assumed layout. The personal floor diffusers
supplied both conditioned outlet air and an air current by opening the slider covering the diffuser (Fig. 4). Exhaust air was
extracted from the test room through an opening installed in the space between the ceiling and upper floor slab via fine pores
in the CRCPs. The temperature and the flow rate of the cold water supplied to the CRCPs, as well as the temperature and
volume of the supplied air, were adjusted by the equipment used in the experiment (controllers, valves and air dampers) as
shown in Fig. 6.
mm (34.1 ft) Steel panel
painted white
1695 mm Aluminum
Test Room : 66 m2 (710 ft2) heat sink
(66.7 in)

mm (20.8 ft)
Outdoor
room Copper pipe
600 (23.6 in)
)LJXUH*HQHUDO3DQHO
Opening for
light fixture
: Flow direction of cold water

:General : Panel with : Net pyrradiometer measurement )LJXUH3DQHOZLWKOLJKWIL[WXUH
panel light fixture points
)LJXUH3ODQRIWKHH[SHULPHQWDOIDFLOLW\

7DEOH&5&3VSHFLILFDWLRQ
Closed Open
HOHPHQW ,WHP 6SHFLILFDWLRQ
0DWHULDO 6WHHO )LJXUH3HUVRQDOIORRUGLIIXVHU
6L]H PPPPLQLQ
3DQHO
+ROHVGLDPHWHUPPLQGULOOHGHYHU\
6XUIDFHGHWDLO
PPLQ
0DWHULDO $OXPLQLXP
6L]H PPPPLQLQ^`
+HDWVLQN
3HFHQWDJHRI

^`
KHDWVLQNDUHD
0DWHULDO &RSSHU
7KLFNQHVV PPLQ
3LSH
'LDPHWHU PPLQ )LJXUH3KRWRJUDSKVKRZLQJWKH
/HQJWK PIW^PIW`
LQWHULRURIWKHWHVWURRPLQWKHWKHUPDO
9DOXHIRUSDQHOVZLWKOLJKWLQJIL[WXUHDUHVKRZQLQEUDFHV
SHUIRUPDQFH WHVW
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
$6+5$(

Thermal performance test and experimental conditions
Figure 6 shows the measurement points for the temperature of the supply and exhaust air, the air volume of the air
handling system, the temperature of the supply and return cold water and the flow rate of the water in the CRCP system. The
respective cooling capacities of the air handling system and the CRCP system were determined by taking measurements at
these points. Figure 7 shows the location of the dummy cooling loads, the personal floor diffusers and the measurement
points for the vertical distribution of air temperature. The parameters of the cooling loads are given in Table 2, where the
cooling load corresponding to occupants and equipment was provided by the dummy cooling loads (48 units) in the form of
light bulbs with aluminum cover (Fig. 5), and the load corresponding to lighting was provided by actual light fixtures
attached to the ceiling. To calculate the convective and radiant heat transfer of the CRCPs, the surface temperature of each
CRCP, the walls and the floor were measured together with the air temperature near the surface of each CRCP and the
radiation budget of the CRCPs. The radiation budget was measured with a net pyrradiometer at the measurement points
shown in Fig. 1.
The preset temperature in the test room was 28 °C (82.4 °F) in the thermal performance test. Assuming an office
environment, the cooling load density was taken as 72 W/m2 (6.7 W/ft2) (Table 2). While obtaining readings for the air
temperature in the test room, the flow rate of the CRCPs was adjusted by controlling the opening of valves in order to bring
the room temperature to the preset level.
Ceiling radiant cooling
Pump Valve Exhaust opening
panel system
T
7DEOH,WHPL]DWLRQRIFRROLQJORDGV
Heat RA Air handling
T F T T T T T
T
system ,QWHUQDO /RDGIDFWRU
exchanger 'HVFULSWLRQ
/RDGV >:P @>:IW@
Ceiling Radiant Cooling Panel 3HRSOH SHUVRQP :SHUVRQ

Three
-way Heat
T :Temperature
valve source
F
V
: Flow rate of cold water
: Air volume
(TXLSPHQW &RPSXWHUVPRQLWRUVSULQWHUV
&DOFXODWHGIURPVSHFLILFDWLRQRI
Heat source Air /LJKWLQJ
SA handling OLJKWIL[WXUHLQWHVWURRP
T V
Personal floor diffusers unit
7RWDO
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
)LJXUH'LDJUDPRIWKHH[SHULPHQWDOIDFLOLW\
mm (34.1 f t)
Exhaust opening(900 mm 㽢 900 mm)

Edge of room The center of the room Outdoor
Slab Slab room
mm (20.8 ft)
2750 mm (9.0 ft)

2700 mm (8.9 ft)
2650 mm (8.7 ft) CRCP CRCP
2100 mm (6.9 ft) Room side Room side
Personal floor diffuser

1600 mm (5.3 ft)
: Air temperature
1100 mm (3.6 ft)
: Surface temperature
600 mm (2.0 ft)
: Humidity Dummy
100 mm (0.3 ft)
: Current velocity cooling load
0 mm (0 ft)
)LJXUH/RFDWLRQVRISHUVRQDOIORRUGLIIXVHUVGXPP\FRROLQJORDGVDQGWHPSHUDWXUHPHDVXUHPHQWSRLQWV
Experiment patterns
The experiment was conducted by changing the following parameters (Table 3): (1) supply air volume: 0, 5 and 10
CMH/m2 (0, 16.4 and 32.8 CFM/ft2), (2) temperature of the water supplied to the CRCPs: 16 and 18 °C (60.8 and 64.4 °F),
(3) supply air temperature: 23 and 28 °C (73.4 and 82.4 °F), (4) active area of the CRCPs: 30 and 37.5 m2 (323 and 404 ft2).
The active area of the CPCPs was changed by selecting whether to supply cold water to panels with light fixtures.
332
ASHRAE Transactions
RESULTS OF THERMAL PERFORMANCE TEST
Relation between water temperature, room air temperature and cooling capacity
Figure 8 shows the relation between the supply water temperature, room air temperature and cooling capacity of the
CRCPs. The total cooling capacity was calculated by measuring the flow rate and the difference in temperature between
supply and return water. The values along the Y axis (the cooling capacity in Fig. 8) were calculated by dividing the total
cooling capacity by the active area of the CRCPs. The values of X axis were obtained by subtracting the room air temperature
from the mean supply water temperature and the mean supply water temperature was in turn calculated by taking the average
of the supply water temperature and the return water temperature. Here, the Y axis represents the cooling capacity of the
CRCPs; the most value for the cooling capacity lie on the same line.
Relation between room air temperature, surface temperature and cooling capacity
Figure 9 is analogous to Fig. 8, with the X axis represents values obtained by subtracting the room air temperature from
the mean CRCP surface temperature. The range of the X axis is about 3 °C (5.4 °F) less than that in Fig. 8. The temperature
difference in this case indicates the contribution of supply water to the cooling of the panels. It is seen that the cooling
capacity is lower in the case that there is no supply air (indicated with an open square) as compared with the same value
along the X axis when there is supply air (indicated with a closed diamond). This is attributable to convective heat transfer to
the CRCPs promoted by the supply air.
7DEOH([SHULPHQWDOSDWWHUQV
2SHUDWLQJFRQGLWLRQVRI&5&3 2SHUDWLQJFRQGLWLRQVRI$+8
&RUUHVSRQGLQJV\PEROLQ
3DWWHUQ 6XSSO\ZDWHUWHPSHUDWXUH 3DQHOVZLWKOLJKWIL[WXUHV &5&3DFWLYHDUHD 6XSSO\DLUYROXPH 6XSSO\DLUWHPSHUDWXUH
)LJVDQG

>r&@>r)@ &RROLQJZDWHU >P @>IW @ >&0+P @>&)0IW @ >r&@>r)@
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
6XSSOLHG 2SHQVTXDUH
6XSSOLHG 2SHQGLDPRQG
1RWVXSSOLHG &ORVHGGLDPRQG
6XSSOLHG 2SHQFLUFOH
1RWVXSSOLHG &URVV
1RWVXSSOLHG 'DVK
6XSSOLHG 3OXV
6XSSOLHG 2SHQWULDQJOH
1RWVXSSOLHG &ORVHGWULDQJOH
>%WXKxIW@ >:P@ >%WXKxIW@ >:P@

&RROLQJ FDSDFLW\
&RROLQJFDSDFLW\

>r&@ >r&@
>r)@ >r)@
'LIIHUHQFHEHWZHHQUHIHUHQFHURRPWHPSHUWXUHDQGPHDQSDQHOVXUIDFH 'LIIHUHQFHEHWZHHQUHIHUHQFHURRPWHPSHUWXUHDQGPHDQFRROLQJZDWHU
WHPSHUDWXUH WHPSHUDWXUH
)LJXUH5HODWLRQEHWZHHQZDWHUWHPSHUDWXUHURRP )LJXUH5HODWLRQEHWZHHQURRPDLUWHPSHUDWXUH
DLUWHPSHUDWXUHDQGFRROLQJFDSDFLW\ VXUIDFHWHPSHUDWXUHDQGFRROLQJFDSDFLW\
Heat transfer characteristics of CRCPs

Figure 10 shows the estimated results for the heat transfer characteristics of the CRCPs. The heat transfer was obtained
by separately calculating the downward and upward radiant heat transfer as well as the downward and upward convective
$6+5$(

heat transfer. The downward radiant heat transfer was obtained through measurements with a net pyrradiometer, while the
upward radiant heat transfer was calculated by using the Stefan-Boltzmann equation. The values of the parameters in this
equation were obtained as follows. The difference in temperature between the CRCP surface and other surfaces was
calculated from measurement values, the shape factor was calculated from the room dimensions, and emissivity of each
surface was decided referring the 2009 ASHRAE HandbookíFundamentals. Furthermore the convective heat transfer was
calculated by using the convective heat transfer coefficient obtained by Wilkes experiment (G. B. Wilkes, 1950), where the
difference between the CRCP surface temperature and the air temperature was obtained from measurement values.
The sum of the four types of heat transfer was almost equal to the cooling capacity, which was calculated by measuring
the flow rate and the difference in temperature between supply and return water (indicated with the open diamonds). The
ratio of downward and upward heat transfer was 4:1, and the ratio of radiant and convective heat transfer was 3:2.
6XSSO\DLU 6XSSO\ZDWHU
6XSSO\DLUYROXPH &5&3DFWLYHDUHD 'RZQZDUGUDGLDQWKHDW 'RZQZDUGFRQYHFWLYH 8SZDUGUDGLDQWKHDW 8SZDUGFRQYHFWLYH
3DWWHUQ WHPSHUDWXUH WHPSHUDWXUH
WUDQVIHU KHDWWUDQVIHU WUDQVIHU KHDWWUDQVIHU
>&0+P @ >r&@>r)@ >P @>IW @ >r&@>r)@

>:@

>%WXK@
&RROLQJFDSDFLW\
)LJXUH+HDWWUDQVIHUFKDUDFWHULVWLFVRI&5&3
OUTLINE OF THERMAL ENVIRONMENT SURVEY

Experimental conditions
A thermal environment survey was conducted in the same space as the thermal performance test. Figure 11 shows the
locations of the desks used by the participants, the location of the personal floor diffusers, as well as the measurement points
for air temperature, humidity and air velocity. Table 4 shows the environmental conditions, as well as the number of men and
women who participated. The temperature of the water supplied to the CRCPs was set to a constant 16 °C (60.8 °F), and the
flow rate of the water in the CRCPs was adjusted in such a way that the air temperature at the measurement points (1100 mm
(3.6 ft) above the floor in the center of the space between four desks shown in Fig. 11) was 28 °C (82.4 °F). The supply air
volume was set to a constant 40 CMH (131 CFM) for each diffuser, and the air velocity at 600 mm (2.0 ft) above the diffuser
was 0.35 m/s (1.1 ft/min). Two of the experiment patterns were set by changing the supply air temperature for each personal
floor diffuser (Table 4). The clothing of the participants was set to 0.6 clo, and the cooling load was set to 52 W/m2 (4.8
W/ft2). To reproduce an actual office environment, lights and PCs were arranged appropriately to match the cooling load
conditions, and the participants were asked to assemble jigsaw puzzles to simulate clerical work (Fig. 12).
Experimental procedure
Table 5 shows the experiment schedule. The survey was conducted from 13:30 to 16:30 on each of the experiment days,
and it was divided into three phases. In the adaptation period, the participants were seated for 60 min in order to provide time
for their bodies to become accustomed to the preset conditions. In Phase 1, the participants were exposed to an environment
where only the CRCPs were operated for 45 min, during which they were asked to feedback regarding thermal sensation and
comfort sensation by rating their experience according to the two scales shown in Figs. 13 and 14 at intervals of 5 min.
Similarly in Phase 2 the participants were asked to fill in a questionnaire about their thermal sensation and comfort sensation;
however, in this case they were permitted to use the personal floor diffusers while the CRCP system was operating, and select
opening and shutting freely personal floor diffuser in order to adjust their thermal and comfort sensation. The air velocity
under the desk of each participant was measured during this phase in order to monitor whether the personal floor diffusers
were used.
334 --`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---

ASHRAE Transactions
mm (34.1 ft)
Center of four desk Outdoor

room
mm (20.8 f t)
Slab
2700 mm (8.9 ft)
2650 mm (8.7 ft) CRCP Desk
2100 mm (6.9 ft)
Personal floor diffuse

1600 mm (5.3 ft)
: Air temperature
1100 mm (3.6 ft)
: Surface temperature
Dummy
: Humidity
600 mm (2.0 ft) cooling load
100 mm (0.3 ft) : Current velocity
0 mm (0 ft)
)LJXUH3KRWRJUDSKRIWKHWHVW
URRPWDNHQGXULQJWKHVXUYH\
)LJXUH/RFDWLRQRIWKHGHVNVSHUVRQDOIORRUGLIIXVHUVDQG
7DEOH([SHULPHQWDOSURFHGXUHRIWKH
PHDVXUHPHQWSRLQWV
HQYLURQPHQWDOVXUYH\ZLWKVXEMHFWV
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
7DEOH([SHULPHQWDOFRQGLWLRQIRUWKHWKHUPDOHQYLURQPHQW $FWLRQ 'HVFULSWLRQ
(QWHUURRP
$GDSWDWLRQ $GDSWDWLRQSHULRGWRDOORZSDUWLFLSDQWVWRDGDSW
VXUYH\ SHULRG
PLQ )LOOLQTXHVWLRQQDLUH
WKHWKHWKHUPDOHQYLURQPHQWLQWKHURRP
2SHUDWLQJFRQGLWLRQV 2SHUDWLQJFRQGLWLRQVIRU 6WDUW3KDVH

$PELHQWFRQGLWLRQ 3DUWLFLSDQWV 7KHXVHRISHUVRQDOIORRUGLIIXVHULVQRWSHUPLWWHG
IRU&5&3V SHUVRQDOIORRUGLIIXVHUV
PLQ 3DUWLFLSDQWVSURYLGHIHHGEDFNDERXWWKHUPDO
([SHULPHQW $LU :DWHU 6XSSO\DLU 6XSSO\DLU 3KDVH VHQVDWLRQHYHU\PLQ
3DWWHUQ +XPLGLW\ )ORZUDWH 0HQ :RPHQ
GD\ WHPSHUDWXUH WHPSHUDWXUH WHPSHUDWXUH YROXPH )LQLVK3KDVH
>&0+GLIIXVHU@
>r&@>r)@ >@ >r&@>r)@ >r&@>r)@ PLQ )LOOLQTXHVWLRQQDLUH
>&)0GLIIXVHU@
6WDUWLQJ3KDVH
$
7KHXVHRISHUVRQDOIORRUGLIIXVHULVSHUPLWWHG
% PLQ 3DUWLFLSDQWVSURYLGHIHHGEDFNDERXWWKHUPDO
&$6( &RQWUROHG 3KDVH VHQVDWLRQHYHU\PLQ
& )LQLVK3KDVH
' PLQ )LOOLQTXHVWLRQQDLUH
( PLQ )LOOLQTXHVWLRQQDLUH
&$6( &RQWUROHG
) 7KHUPDOHQYLURQPHQW

VXUYH\LVHQGHG
㻙㻞㻙㻝㻜㻗㻝㻗㻞
&ROG &RRO 6OLJKWO\ 1HXWUDO 6OLJKWO\ :DUP +RW 8QFRPIRUWDEOH 6OLJKWO\ 1HXWUDO 6OLJKWO\ &RPIRUWDEOH
FRRO ZDUP XQFRPIRUWDEOH FRPIRUWDEOH
)LJXUH7KHUPDOVHQVDWLRQVFDOH)LJXUH&RPIRUWVHQVDWLRQVFDOH
7DEOH7KHUPDOHQYLURQPHQWFRQGLWLRQVLQWKHVXUYH\
3DWWHUQ &$6( &$6(
([SHULPHQWGD\ $ % & ' ( )
3KDVH 3KDVH 3KDVH 3KDVH 3KDVH 3KDVH 3KDVH 3KDVH 3KDVH 3KDVH 3KDVH 3KDVH 3KDVH
1RW 1RW 1RW 1RW 1RW 1RW
8VDJHRISHUVRQDOIORRUGLIIXVHUV 3HUPLWWHG 3HUPLWWHG 3HUPLWWHG 3HUPLWWHG 3HUPLWWHG 3HUPLWWHG
SHUPLWWHG SHUPLWWHG SHUPLWWHG SHUPLWWHG SHUPLWWHG SHUPLWWHG
6XSSO\DLUWHPSHUDWXUH>r&@>r)@
$LUWHPSHUDWXUH>r&@>r)@
+XPLGLW\>@
0HDQ5DGLDWLRQ7HPSHUDWXUH>r&@>r)@
&XUUHQWYHORFLW\>PV@>IWPLQ@
3UHGLFWHG0HDQ9RWH
RESULTS OF THERMAL ENVIRONMENT SURVEY

Thermal environment in each pattern
Table 6 shows the mean value for the thermal environment conditions for each phase, test pattern and experiment day.
The MRT (Mean Radiant Temperature) values were calculated from measurements of the surface temperature of the CRCPs
and the walls, and the result was about 26 °C (78.8 °F) in all cases. The predicted mean vote (PMV) was 0.80í0.88 in Phase
1, where the participants were not permitted to use the personal floor diffusers, and was 0.5í0.6 in Phase 2, where the
participants were permitted to use the diffusers.
$6+5$(

Change in thermal and comfort sensation between Phases 1 and 2
Figure 15 and 16 show the change in the thermal sensation and comfort sensation of each participant between Phases 1
and 2. Each plot on the figures is mean value of thermal and comfort sensation of each participant during each phase. The
open symbols show the mean value of thermal and comfort sensation when participants were not permitted to use the
personal floor diffusers, and the closed symbols show the mean value when they were permitted to use the personal floor
diffusers. The dotted arrows show that thermal sensations in Phase 2 are lower (cooler) than that in Phase 1 and comfort
sensations in Phase 2 are higher (more comfortable) than that in Phase 1. The solid arrows show that thermal sensations in
Phase 2 are lower (cooler) than that in Phase 1 and comfort sensations in Phase 2 are lower (less comfortable) than that in
Phase 1. The figures show that thermal sensations of almost all participants in Phase 2 are cooler than that in Phase 1 and
comfort sensations of few participants in Phase 2 are less comfortable than that in Phase 1. (Due to limitations of space, it is
shown only the pattern of CASE 1, but the same tendency is shown in the results of the pattern of CASE 2.)
&RPIRUWDEOH &RPIRUWDEOH
6OLJKWO\ 6OLJKWO\

FRPIRUWDEOH FRPIRUWDEOH
1HXWUDO 1HXWUDO
6OLJKWO\

6OLJKWO\

XQFRPIRWDEOH 2SHQ V\PERO 3KDVH XQFRPIRWDEOH 2SHQ V\PERO 3KDVH
&ORVHG V\PERO3KDVH &ORVHG V\PERO3KDVH
8QFRPIRUWDEOH 8QFRPIRUWDEOH

&RRO 6OLJKWO\ 1HXWUDO 6OLJKWO\ :DUP +RW &RRO 6OLJKWO\ 1HXWUDO 6OLJKWO\ :DUP +RW
FRRO ZDUP FRRO ZDUP
)LJXUH&KDQJHLQWKHUPDODQGFRPIRUWVHQVDWLRQRI )LJXUH&KDQJHLQWKHUPDODQGFRPIRUWVHQVDWLRQRI
HDFKSDUWLFLSDQWEHWZHHQ3KDVHVDQGRQH[SHULPHQW HDFKSDUWLFLSDQWEHWZHHQ3KDVHVDQGRQH[SHULPHQW
GD\VRI$DQG%VXSSO\DLUWHPSHUDWXUH°&°) GD\VRI&DQG'VXSSO\DLUWHPSHUDWXUH°&°)
Comparison of results between men and women

Figure 17 shows the results for the thermal sensation and comfort sensation for men and women when they were not
permitted to use the personal floor diffusers (Phase 1) and the supply air temperature was set to 28 °C (82.4 °F) (Pattern 1).
Figure 18 shows the results for the thermal sensation and comfort sensation for men and women when they were permitted to
use the personal floor diffusers (Phase 2) and the supply air temperature is set 28 °C (82.4 °F) (Pattern 1). These figures
reveal the difference between men and women in terms of thermal sensation and comfort sensation, and also the results of a
variance analysis presented in Figs. 17 and 18 show that the difference between two groups is significant. Similar results are
shown in Figs. 19í21. Generally, in comparison with men and women rated the environment as being cooler (Sami
Karjalainen, 2007), and the same tendency can be inferred from these results.
+RW &RPIRUWDEOH +RW &RPIRUWDEOH
:DUP :DUP
6OLJKWO\FRPIRUWDEOH 6OLJKWO\FRPIRUWDEOH
6OLJKWO\ZDUP 6OLJKWO\ZDUP
1HXWUDO 1HXWUDO 1HXWUDO 1HXWUDO
6OLJKWO\FRRO 6OLJKWO\FRRO
6OLJKWO\XQFRPIRWDEOH 6OLJKWO\XQFRPIRWDEOH
&RRO &RRO

&ROG 8QFRPIRUWDEOH &ROG 8QFRPIRUWDEOH
0HQ :RPHQ 0HQ :RPHQ 0HQ :RPHQ 0HQ :RPHQ
)LJXUH'LIIHUHQFHVLQWKHUPDOVHQVDWLRQDQGFRPIRUW )LJXUH'LIIHUHQFHVLQWKHUPDOVHQVDWLRQDQGFRPIRUW
VHQVDWLRQEHWZHHQPHQDQGZRPHQLQ3KDVH VHQVDWLRQEHWZHHQPHQDQGZRPHQLQ3KDVH
S S
Differences between Phase 1 and Phase 2 in the case of supply air temperature of 28 °C (82.4 °F) (Pattern 1)
Figures 19 and 20 show the results for men and women in terms of indicated thermal sensation and comfort sensation in
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
336
ASHRAE Transactions
Phase 1 when the participants were not permitted to use the personal floor diffusers, and Phase 2, when the participants were
permitted to use the personal floor diffusers for a supply air temperature of 28 °C (82.4 °F). The figures also show that the
thermal sensation in Phase 2 is closer to “Neutral” as compared with that in Phase 1, and comfort sensation in Phase 2 is
higher than that in Phase 1. From these figures, the overall comfort was improved by the air current provided by the personal
floor diffusers. The difference between Phases 1 and 2 in terms of indicated thermal sensation is larger for men than for
women, and the thermal sensation in the case of women is almost “Neutral” in both Phase 1 and Phase 2. This shows that the
effect of the air current is more pronounced for men than for women.
+RW &RPIRUWDEOH +RW &RPIRUWDEOH

:DUP :DUP
6OLJKWO\FRPIRUWDEOH 6OLJKWO\FRPIRUWDEOH
6OLJKWO\ZDUP 6OLJKWO\ZDUP
1HXWUDO 1HXWUDO 1HXWUDO 1HXWUDO
6OLJKWO\FRRO 6OLJKWO\FRRO
6OLJKWO\XQFRPIRWDEOH 6OLJKWO\XQFRPIRWDEOH
&RRO &RRO
&ROG 8QFRPIRUWDEOH &ROG 8QFRPIRUWDEOH
1RWSHUPLWWHG 3HUPLWWHG 1RWSHUPLWWHG 3HUPLWWHG 1RWSHUPLWWHG 3HUPLWWHG 1RWSHUPLWWHG 3HUPLWWHG
)LJXUH'LIIHUHQFHVLQWKHUPDOVHQVDWLRQDQG )LJXUH'LIIHUHQFHVLQWKHUPDOVHQVDWLRQDQG
FRPIRUWVHQVDWLRQDVLQGLFDWHGE\PHQLQ3KDVHV FRPIRUWVHQVDWLRQDVLQGLFDWHGE\ZRPHQLQ3KDVHV
DQGLQ3DWWHUQ S DQGLQ3DWWHUQ S
Difference between results obtained for supply air temperature of 28 °C (82.4 °F) and 25 °C (77 °F)
Figure 21 shows the results for the thermal sensation and the comfort sensation for a supply air temperature of 28 °C
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
(82.4 °F) and 25 °C (77 °F) when the participants were permitted to use the personal floor diffusers. Both the Thermal
sensation and the comfort sensation are closer to “Neutral” for a supply air temperature of 25 °C (77 °F), which suggests that
the thermal environment can be improved by lowering the supply air temperature by several degrees.
+RW &RPIRUWDEOH
:DUP
6OLJKWO\FRPIRUWDEOH
6OLJKWO\ZDUP
1HXWUDO 1HXWUDO
6OLJKWO\FRRO 6OLJKWO\XQFRPIRWDEOH

&RRO
&ROG 8QFRPIRUWDEOH
r& r& r& r&
r) r) r) r)
)LJXUH'LIIHUHQFHVLQWKHUPDOVHQVDWLRQDQGFRPIRUWVHQVDWLRQEHWZHHQ3DWWHUQDQG3DWWHUQ S
CONCLUSION
(1) By performing a thermal performance test, it was shown that the proposed HVAC system with CRCPs and personal floor
diffusers is capable of removing an interior cooling load of 72 W/m2 (6.7 W/ft2) at a room air temperature of 28 °C (82.4 °F).
(2) By conducting a thermal environment survey, it was shown that the proposed HVAC system with CRCPs and personal
floor diffusers is capable of providing a comfortable environment at a room air temperature of 28 °C (82.4 °F) since personal
floor diffusers enable occupants to control the air current in their vicinity independently of other occupants.
REFERENCES
2009 ASHRAE HandbookíFundamentals

G. B. Wilkes. 1950. “Heat Insulation”, John Wiley & Sons
Sami Karjalainen. 2007. Gender differences in thermal comfort and use of thermostats in everyday thermal
environments, Building and Environment, 42, pp. 1594-1603
$6+5$(

SA-12-C043
Determining the Cost-Effectiveness of

Various Sub-Metering Strategies
Christine C. Maurer, PE Peter Rojeski, Jr., PhD, PE Matt Davey
ASHRAE Member ASHRAE Life Member Non Member
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
ABSTRACT
This paper presents the details of a recent study conducted to determine the relative cost-effectiveness of portable energy meters
used to sub-meter Heating, Ventilating and Air Conditioning (HVAC) equipment for energy audits and/or for measurement and
verification of retrofit projects. The results provide an insight as to how fluctuations in voltage and harmonic distortion (caused by
variable speed drives) affect the accuracy of power measurements taken by current meters and relatively inexpensive (<$1,000) watt
meters. The evaluation criteria for this study include data logger and sensor cost, power measurement accuracy, ease of installation within
an electrical breaker panel, and ease of use of the software. The accuracy of the portable meters was tested for a motor controlled by a
variable frequency drive (VFD). The operating range for each test included loads of 25 percent, 50 percent, 75 percent, and 100
percent of full load at 60 Hz and then at 50 Hz and 40 Hz with the corresponding load. Once the motor reached thermal stability at
its rated output, power measurements were taken at the electrical panel with a laboratory-grade power analyzer and with each of the
portable test meters.
INTRODUCTION
Portable energy meters and data loggers are commonly-used tools for short term analysis of energy-use
patterns in buildings and to document the performance characteristics of individual systems and equipment.
Monitoring 25 or more single-phase and three-phase circuits is typical for 100,000 square foot buildings. Data
collection typically occurs over a three- to five-week period. A company specializing in energy-use analyses may be
conducting studies in a dozen or more buildings at the same time. Choosing the best sub-meters for a particular
project is an important task because it not only impacts the quality of the results, but the profitability of the project
as well. For this study, less expensive, non-utility-grade meters/data loggers were investigated. These results were
compared to those from a high accuracy, laboratory grade power meter.
SAFETY ISSUES IN METERING

Protection against electrical shock and arc flash are two important considerations for individuals who
install energy meters in electrified panels. In the experience of the authors, 90 percent of energy meter installations
are accomplished in panels that are electrified; therefore, individuals who install meters should be NFPA 70-E
certified. A minimum hazard rating of Level 2 should be assumed and protective gear should be worn during both
the installation and removal of energy meters from electrified panels.
Care should be taken when installing and removing Current Transducers (CTs) from electrified
conductors. Many times conductors have to be physically separated to allow room for the CTs to be installed. CTs
consist of two pieces that snap together around a conductor. Excessive force should not be used to separate the
conductors because such force may cause the conductors to become disconnected from their terminals. This
disconnection could result in an arc-flash. Care should also be taken to not drop either half of a CT inside an
electrical panel because it may come into contact with exposed, electrified conductors and cause an arc flash.
Similarly, any tools (e.g., pliers or screw drivers) used to install or remove CTs should be insulated. Because of
these safety concerns, the authors considered ease of installation when evaluating meter types.
338
ASHRAE Transactions
METERING OPTIONS AND EVALUATION CRITERIA
The widespread use of variable speed drives in HVAC systems presents a special sub-metering challenge.
The harmonic distortion and the non-sinusoidal current and voltage wave forms created by VFDs can result in
significant metering inaccuracies, especially when low cost meters are used. This study tested four different
metering strategies which will be referenced as follows:
1. Current meters (iron core current transducers)
2. True RMS current and voltage meters
3. Watt meters measuring three-phase power
4. Watt meters measuring only one phase of a three-phase circuit
The evaluation criteria adopted for evaluating various meters and data loggers are listed below.
a) Data-logger and sensor cost. The goal was to evaluate meters that cost less than $1,000 for a three-phase
circuit or three single-phase circuits. For this study, three phase meters with data logging capability that
cost less than $500 are characterized as “low cost.” Units that cost between $500 to $750 are characterized
as “medium cost” and units that exceed $750 are characterized as “high cost.”
b) Power measurement accuracy. All of the tested meters are less accurate than utility grade metering. Test
results are presented later in this paper that show the accuracy of the various metering strategies for motor
loads that are not controlled by VFDs and for motor loads that are controlled by VFDs. For this study,
accuracy results range from 95 percent (high) to 70 percent (unacceptable). As shown in the Results
section, current meters have unacceptable accuracies when used to monitor VFD-controlled motors;
however, these devices produce repeatable results and can be field-calibrated to improve their overall
accuracy.
c) Ease of installation in an electrical panel that is energized. Two factors determine the ease of installation
of a metering system. The first is the ease of installing current transformers. Current transformers that
have an alligator clip design are easily installed with one hand. Hinged current transformers can also be
installed with one hand but installation is not as easy as the alligator clip units. Non-hinged CTs require
two hands and these are harder to install or remove. In some cases the non-hinged CTs are very difficult
to pull apart and often require the use of a screw drive to pry them apart. These units are not
recommended for installation in electrified panels. The second ease of installation factor involves the size
and amount of equipment that must placed inside the electrical breaker panel or the electrical disconnect.
Photographs of the three metering systems are provided to give the reader a sense of size and the amount
of equipment involved with each metering strategy. One main advantage of the current meters is that the
data logger can be installed outside the electrical enclosure because the CTs leads carry low voltage. The
other meters require system voltage inputs (e.g., 277 volt) requiring the meter to be installed within the
electrical enclosure.
d) Ease of use of the software. Some metering systems require data to be exported and manipulated via
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
spreadsheets in order to produce hourly, daily and/or weekly results. This is a time-consuming process
when monitoring occurs over several months. For example, one-minute readings taken for four weeks
results in 120,960 data values for a three-phase circuit. The advantage of watt- meters is that the power
calculations are done internally to the meter and data manipulation via external spreadsheets is not
required.
METERING STRATEGIES
The following four different metering strategies were evaluated for this study. Each metering strategy includes a
metering device(s) and a data-logger.
1) Current Meters (iron core current transducers)

Measure and record circuit amperage at user-selected time intervals. Circuit voltage and power factor are
measured manually but are not monitored.
$6+5$(

True RMS Current and Voltage Meters
Measure and simultaneously record circuit amperage and
voltage at user-selected time intervals. Power factor is
measured manually but is not monitored.
2) Watt Meters
These meters measure voltage, amperage, and power
factor and calculate circuit power several thousands times
per second. Energy consumption is calculated and
recorded at a user-selected interval.
3) Watt Meter (Measuring only one phase of a three-phase

circuit)
Power meters are typically configured to measure three- Figure 1: Current transducers and 4 channel
phase power for a three-phase circuit. However, power data logger. Note: spool of electrical tape used
meters that monitor three single-phase circuits are available to demonstrate the overall size of the metering
and are included in this study. Since the energy of three- package, which typically needs to fit into
phase motors is typically balanced between the phases, three compact electrical panels.
separate motors can be monitored (by monitoring one
phase per motor) using one meter. A typical application
might be an HVAC system that has three chilled water
pumps and the analyst chooses to monitor all three pumps
with one power meter.
Variable Frequency Drives are produced with built-in metering. We

did try to use the output in percent load or percent power from a
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
VFD in the test laboratory. However, the resulting values recorded

did not correlate well and thus the test results are not included. We
do think that this approach is worth further evaluation as it could
prove to be the most economical way to meter a VFD.
A summary of the metering strategies based on the evaluation

criteria is presented in Table 1.
Figure 2: True RMS current and voltage meter
Figure 3: Watt meter and datalogger
340
ASHRAE Transactions
Table 1. Comparison of the Metering Stategies
Metering Strategy
Evaluation Current Meter True RMS Current Watt Mater Watt Meter
Criteria and Voltage Meter (Measuring only
One Phase of a
Three Phase
Circuit)
Cost $380 (Three CTs and $299 Data-logger $635 Data-logger $675 – Data-logger
data-logger) monitoring 1 RMS monitoring a three- monitoring power to
current and 1 voltage phase circuit 3, separate, single-
phase circuits.
Data Storage 43,000 values 250,000 values 500,000 values 500,000 values
Best Lighting and other Short-term (5 days) VFD-controlled Monitoring three,
Application loads where a 15- monitoring of VFD- motors and other VFD-controlled,
minute sampling controlled motors applications where single phase motors
interval is acceptable. where true RMS higher accuracy is or three motors fed
current and voltage required. Refer to from the same
measurements are Results section for electrical panel.
preferred. details.
Advantages Low cost, compact, Compact and has High accuracy at a Relatively low
and the CTs are easy CTs that are easily relatively low price. monitoring cost per
to install and remove. installed and removed Even though data is motor.
from energized collected several
panels. Unit that was thousand times each
tested has a display second, the reporting
that provides real- rate can be set hourly,
time data. daily, etc., without
having to perform
spreadsheet
manipulations.
Disadvantages Unacceptable The meter can The CTs historically The CTs typically
accuracy when used monitor both the sold for use with the used with this
on VFD-controlled circuit voltage and tested meter are product are difficult
equipment. Refer to current, but only for difficult to remove to remove and
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Results section for one phase. Also, the and disconnecting disconnecting power
details. unit that was tested power to the panel is to the panel is
only had a battery life recommended prior recommended.
of five days, but to removal. It is
operating time can be difficult to install
extended using an multiple meters in a
external battery pack single breaker panel
or 120 volt power because of the size of
source. the equipment.
Different meters are
required for 120/208
volt circuits and
277/480 volt circuits.
$6+5$(

DESCRIPTION OF THE TEST PROCEDURE
The accuracy of the portable electrical meters was tested for a
constant speed motor and for a motor controlled by a VFD. The operating
range for each test included six values, 25 percent, 50 percent, 75 percent,
and 100 percent of full load at 60 Hz. The motor was then fully loaded at
60 Hz and the output frequency of the drive was then lowered to 50 Hz
and then 40 Hz without adjusting load on the motor. The reference
electrical meter was a laboratory-grade power analyzer that provides power
readings with an accuracy of 0.02 percent and also provides a power quality
analysis of the line current, voltage, power factor and the total harmonic
distortion of the circuit. One of the main functions this analyzer can
perform is a power quality analysis up to 100 harmonic orders. For this
testing it was useful in determining how voltage and current harmonic
distortions affect metering accuracy and obtain quantifiable information on
those effects. The power readings of the portable meters were compared to
those of the laboratory meter for both the constant speed motor and the
VFD-controlled motor at each of the different loads. Both the laboratory
meter and the test meters were located upstream of the VFD during
testing.
In addition to logging numeric values during the testing, the
laboratory-grade meter provides a graphical output of the voltage and
current waveforms. Figure 4 shows the voltage and current waveforms for
one phase of the test motor without the VFD (top) and with a VFD
(bottom). Without a VFD, the current waveform is sinusoidal. Once the
VFD is connected the current waveform (measured upstream of the drive)
no longer follows the sinusoidal pattern. In the test results, the Current
Total Harmonic Distortion (%I-THD) and the Voltage Total Harmonic Figure 4: Voltage and current waveforms on the
Distortion (%V-THD) are presented for each of the tests. test motor without VFD (top) and with VFD
(bottom).
The same motor was used for each of the metering methods. The
test motor was run at full load until it reached thermal stability. This is defined to be when the peak-to-peak
temperature rise for a 30 minute period is no greater than 1°C. Once the motor reached thermal stability at its
rated output, power measurements were taken at the electrical panel with a laboratory-grade power analyzer and
with each of the portable test meters. We determine percent load to be measured output power over rated output
power.
For these tests we used an Eddy Current (EC) Dynamometer with a built in speed sensor and
appropriately sized load cell to determine output power. Before each test, load cell torque offset and gain values
are calibrated to the expected full load torque of the test motor. This helps to ensure that the most accurate and
repeatable torque readings are obtained. The EC-Dynamometer has an overall torque measurement uncertainty of
0.078 percent calibrated at full scale. Each load point data from the reference meter and test meter was recorded
for a 15 minute period. During this period motor output power was kept as constant as possible.
The motor was then loaded to 100 percent of rated output power and the frequency of the drive was
reduced to 50Hz. The output power of the motor was kept constant at the 50Hz operating speed and data
recorded for a 15 minute period. The output frequency of the drive was then reduced to 40Hz and the process
repeated.

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---

RESULTS
A baseline test was run on the 20 HP motor without a VFD drive. The baseline test was done only for
the current meter and the watt meter. The results are shown in Table 1 and Table 2.
Table 2. Baseline Test on Motor Without a VFD with the current logger
Reference Meter Current Logger
% % %
Current Current Error Error Error
Test Load W-h (Phase 1) PF (Phase 1) W-h W-h* (I) (W-h) (W-h*)
100% 20 Hp 3,991 22.7 0.87 23.2 4,008 4,008 2.2% 0.4% 0.4%
75% 15 Hp 2,959 17.4 0.84 17.7 2,972 3,071 1.6% 0.4% 3.8%
50% 10 Hp 1,964 12.8 0.76 13.1 1,988 2,252 2.1% 1.2% 14.7%
25% 5 Hp 1,018 9.0 0.56 9.3 1,038 1,609 2.8% 2.0% 58.1%
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
*The first W-h calculation uses the power factor measured by the reference meter,
the second W-h calculation uses only the 100% load PF (0.87)
Table 3. Baseline Test on Motor Without a VFD with the Watt Meter and Data-logger
Reference Meter Watt Meter
Test Load W-h PF % I-THD % V-THD W-h % Error (W-h)

100% 20 Hp 3,991 0.87 3.16 2.1 4,077 2.2%
75% 15 Hp 2,959 0.84 3.7% 2.1% 3,029 2.4%
50% 10 Hp 1,964 0.76 4.4% 2.1% 2,022 3.0%
25% 5 Hp 1,018 0.56 5.1% 2.1% 1,073 5.4%
One key point to note from Table 2 is that the current readings on the test meter were fairly accurate
with a maximum of 2.8 percent error. To calculate power, a constant voltage and power factor was used as
measured by the reference meters at full load (100 percent). The assumption that power factor is constant leads to
inaccurate values for energy consumption. Power factor for this test motor varied from 0.87 at 100 percent load to
0.56 at 25 percent load. Because power factor is greatly affected by load, power should be calculated using a
variable power factor. In the field, this task may be done easily only in select situations when a power factor
reading can be taken at different loads. However, in most applications it is difficult to force a motor to these
different loads.
Figure 5 shows the effect of the VFD on the power factor of the motor at different loads. The results
show that the addition of the drive at full load reduces the power factor by 15 percent for the test motor. At low
loads (25 percent) the power factor is improved by the drive system by 14 percent for the test motor.
The results from each meter tested with the VFD are presented in Tables 4–7. The results show how the
accuracy of the portable current meter and the watt meters was affected by the harmonic distortion created by the
VFD drive. The tables present slightly different information as each meter has varied output capabilities relative to
measuring current, voltage, power factor, etc. For the meters which are not power meters, assumptions had to be
made to calculate W-h (i.e., power factor and/or voltage). We used the readings from the reference meter for this
calculation. In the field, it is assumed the person taking measurements would use an instantaneous electrical power
meter to get these measurements when needed. Readers should not take the voltage, amps, PF and calculate their
own W-h readings because actual test time differs.
$6+5$(

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 5. Power factor for test motor with and without the VFD drive.
Table 4. Metering Configuration One. Current Logger Only

Reference Meter Current Logger
% I- % V- % Error (W- % Error
Test Load W-h Current PF THD THD W-h* Current h) (I)
75% 15 Hp 3,103 21.7 0.72 68.3% 1.5% 2198 15.04 -29.2% -30.7%
% Load
50% 10 Hp 2,112 15.5 0.69 71.6% 1.5% 1,512 10.29 -28.4% -33.5%
25% 5 Hp 1,143 8.6 0.64 76.3% 1.9% 818 5.30 -28.4% -38.4%
66.0% 1.7%
Frequency
60Hz 20 Hp 4,096 26.9 0.74 2,884 19.20 -29.6% -28.7%

50Hz 14 Hp 1,936 20.3 0.71 69.3% 1.8% 1,345 13.89 -30.5% -31.7%
10.2
72.0% 1.8%
40Hz Hp 1,440 15.4 0.69 1,005 10.11 -30.2% -34.3%
* Assumes constant Power Factor of 0.74 and Voltage of 462.6 to calculate W-h
Table 5 shows the results from metering with the True RMS Meter. This meter provides a VA reading
directly and watts is calculated by applying the power factor. For the calculations, a constant power factor was
used. Even knowing that this assumption introduces error, the True RMS meter provided W-h values that were
fairly close. The highest error of seven percent was at 75 percent load and there was not a pattern or correlation
with percent error vs. load or frequency.
Table 6 shows the results of the three phase watt meter. The test results show that this meter measured
within one percent of the reference meter at all loads and frequencies. It is the most expensive of meter types
tested but was also the most accurate.
Table 7 shows the results of monitoring only one phase on a three-phase load. For the reference meter,
the three phases are measured separately and then summed to obtain W-h. The watt meter used for testing had
the capability to measure three phases separately. In this case we are treating each phase like it would be a separate
motor in the field. Therefore, for the watt meter the W-h calculated for one phase was multiplied by three to
obtain the power for a three-phase load. The test results show that this approach measured within 7.6 percent at
all loads and frequencies. In this case measuring phase one provided power measurements that were lower than the
344
ASHRAE Transactions
reference while both phase two and three were higher than reference. The phase three measurements were the
most accurate overall.
Table 5. Metering Configuration on Two: One Phase True RMS VA Logger

True
Reference Meter RMS
% I- % V- % Error (W-
Test Load W-h PF THD THD W-h* h)
75% 20 Hp 3,673 0.72 69.9% 1.7% 3,417 -7.0%
Frequency % Load
50% 10 Hp 2,318 0.69 73.9% 2.2% 2,199 -5.1%

25% 5 Hp 1,234 0.64 77.2% 2.4% 1,251 1.4%
66.7% 1.7%
60Hz 20 Hp 4,619 0.74 4,473 -3.2%
50Hz 14 Hp 3,441 0.71 70.4% 1.9% 3,355 -2.5%
40Hz 10.2 Hp 2,481 0.69 73.0% 2.1% 2,423 -2.3%
* Assumes constant Power Factor of 0.73 to calculate W-h
Table 6. Metering Configuration on Three: Three Phase Watt Meter

Reference Meter Watt Meter
% Error (W-
% Load Load W-h PF % I-THD % V-THD W-h PF Voltage h)
75% 15 Hp 3,103 0.72 68.3% 1.5% 3,128 0.72 462.6 0.8%
% Load
50% 10 Hp 2,112 0.69 71.6% 1.5% 2,132 0.69 462.8 0.9%

25% 5 Hp 1,143 0.64 76.3% 1.9% 1,148 0.64 461.7 0.4%
66.0% 1.7%
Frequency
60Hz 20 Hp 4,096 0.74 4,137 0.74 463.2 1.0%

50Hz 14 Hp 1,920 0.71 69.3% 1.8% 1,936 0.71 462.0 0.8%
10.2
72.0% 1.8%
40Hz Hp 1,428 0.69 1,440 0.69 462.5 0.8%
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Table 7. Metering Configuration on Four - Watt Measuring Only One Phase of Three-Phase Load
Reference Meter Watt meter - one phase only
W-h Measured W-h Calculated* % Error - W-h Calculated*
% Load Load All Phases Phase 1 Phase 2 Phase 3 Phase 1 Phase 2 Phase 3
75% 20 Hp 3,232 3110 3419 3307 -3.8% 5.8% 2.3%
Frequency % Load
50% 10 Hp 2,030 1979 2115 2095 -2.5% 4.2% 3.2%

25% 5 Hp 1,128 1042 1206 1192 -7.6% 7.0% 5.7%
60 Hz 20 Hp 4,074 4,025 4,293 4,117 -1.2% 5.4% 1.1%

50Hz 14 Hp 2,911 2,828 3,061 2,972 -2.9% 5.2% 2.1%
40Hz 10.2 Hp 2,177 2,077 2,262 2,262 -4.6% 3.9% 3.9%
$6+5$(

DISCUSSION
The test results show that the three-phase watt meters measured within one percent of the true value for
all motor loads and for all VFD frequencies. When only a single phase of a three-phase motor load was monitored,
the watt meters measured within 5.7 percent of the true value. The accuracy of the true RMS current and voltage
meter is limited by the fact that only a single phase is being monitored, and the meter does not monitor power
factor. The accuracy of this strategy depends on the accuracy of the power factor measurement, which is typically
an independent reading(s) taken with a separate power factor meter. For motors that operate at constant load or
speed, a single power factor measurement should result in an overall accuracy of equal to that of a watt meter that
is monitoring only a single phase. However, when the power factor of the load is not constant, the error associated
with this strategy could approach 10 percent. The error associated with current meters was shown to approach 30
percent, primarily due to inaccurate current measurements. However, the results also show that the current
readings are consistently low and that the measurements are repeatable. Therefore, the accuracy of the current
meter strategy could be increased by field-calibrating the meter using a true RMS current, voltage, and phase angle
meter. Under these circumstances the error would probably be in the range of 10 to 15 percent.
The testing team noticed that the graphical waveform of the sinusoidal current and voltage provided by
the True RMS meter was not well represented by the meter. This feature did not seem to affect the data provided
by the meter. If one wanted to use the approach of measuring one phase, always ensure that the motor load is
balanced by checking the voltage on all three phases of the motor to be monitored. If the monitored motor is not
well balanced then this approach will introduce more error. The three-phase watt meter could be used with high
confidence in most applications. It does not have the accuracy of a utility grade meter but is a cost effective option
for metering when needing high confidence in the readings.
CONCLUSIONS
The cost of sub-metering includes the equipment cost and the labor cost. Current meters are by far the
least expensive, the easiest to install, but they are the least accurate (error approaches 30%). Current meters are
recommended for loads such as lighting loads, and for motors that switch “on” and “off” but that operate at a
constant load when “on.” It is recommended that these meters be field-calibrated in order to improve the accuracy
of the calculated power results. The true RMS current and voltage meter provide accurate volt-amp readings but
separate power factor measurements are required for accurate watt readings (i.e. less than 10% error). These
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
devices work well it situations where the power factor does not vary significantly. The results for the watt meter
tests indicate an error of approximately 1 percent when three phases are monitored and an error of 7 percent when
only one phase is monitored. In conclusion, selecting the “best” metering strategy for the job requires that the
engineer understand the characteristics of the circuit being monitored and the accuracy required for the project.
ACKNOWLEDGEMENTS
The authors would like to acknowledge Advanced Energy for use of the motor testing laboratory
equipment to complete this study.
NOMENCLATURE
W-h – Watt Hour
% I-THD – Percent of Current Total Harmonic Distortion
% V-THD – Percent of Voltage Total Harmonic Distortion
346
ASHRAE Transactions
TECHNICAL PAPERS
Presented at the
June 23–27, 2012
San Antonio, Texas
Technical papers most often report on society-sponsored research and unsolicited research in
HVAC&R technical areas.
All Technical papers undergo critical review by three or more competent specialists to ensure
high quality of content. The ASHRAE Conferences and Expositions Committee decides, based on
these reviews, whether papers are acceptable, need revision, or should be rejected. The name of the
author is withheld from the reviewers and the names of the reviewers are withheld from the author
and from each other to ensure objective comments, which are then transmitted to the author.
The procedure for review of Technical papers for ASHRAE Transactions is as follows:
1. Author submits an electronic version of his/her manuscript through the ASHRAE Web site
using ASHRAE Manuscript Central at http://mc.manuscriptcentral.com/ashrae.
2. Three qualified peer reviewers review the manuscript, deciding on acceptance, minor
changes, mandatory changes, or rejection.
3. Unless a manuscript is immediately accepted or rejected, the author responds to reviewers’
comments and submits a revised manuscript.
4. The review process continues a second and sometimes a third time to confirm responses to
comments. The paper must be accepted by all three reviewers to be approved for the confer-
ence. The ASHRAE Conferences and Expositions Committee makes final decisions on
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
acceptance or rejection of manuscripts.

5. Authors whose papers are approved are invited to present their papers at an ASHRAE
conference, and papers are published in the Transactions of that conference.

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---

SA-12-001
Simulations of Chilled Water Cooling Coil

Delta-T Characteristics
Zhiqin Zhang, PhD, PE Hui Li, PhD, PE Jingjing Liu, PE

Student Member ASHRAE
ABSTRACT and supply temperature difference or delta-T. The benefits of

a high delta-T include reduced pump energy usage, increased
Increasing loop chilled water return and supply temper-
system cooling capacity, and improved chiller performance
ature difference or delta-T is an effective way to improve the
during low loads. However, almost every real chilled water
chilled water system efficiency and save chiller plant operating
costs. In some chiller plants, the delta-T can be expected to fall system is plagued by low delta-T syndrome, particularly at
to about one-half to two-thirds of design at low loads due to low cooling loads. Delta-T degrading is almost inevitable and
various causes, such as coil load, air entering and leaving it can be expected to fall to about one-half to two-thirds of
temperatures, and chilled water supply temperature. Compu- design at low loads (Taylor, 2002). Many factors contribute to
tational simulation is a useful and flexible tool to study cooling the loop chilled water delta-T, such as chilled water supply
coil delta-T characteristics under various conditions. This temperature, cooling coil air entering and leaving tempera-
paper simulated the cooling coil delta-T profiles with different tures, type of flow control valves, tertiary connection types,
geometric configurations and various waterside and airside coil cooling loads, air economizers, etc. However, in most
conditions using an effectiveness-NTU model. The waterside systems, the root cause of low delta-T is at coil side. The first
heat convection coefficient was modeled with a combined factor is the geometric configuration of the coil. A high delta-
correlation. The simulation results show that the existing T coil comes with high air and water hydraulic resistance,
waterside correlations generated significantly different delta- while a low delta-T coil limits the capacity of chillers, wastes
T profiles. With variant geometric configurations, the coil pumping energy, and may be unable to meet the cooling
delta-T at full load may be higher, equal, or lower than that at demand. Life cycle cost studies have shown that life-cycle
part load. The complexity of cooling coil delta-T characteris- costs tend to be optimized at the highest practical delta-T. A
tics indicates that it is hard to make conclusions about the coil cooling coil with a higher design delta-T consumes more
delta-T change with cooling loads. pump and fan energy because of higher airside and water side
pressure drops. However, the energy consumption increase is
INTRODUCTION much lower than the energy reduction due to high delta-T on
the chilled water system (Taylor et al. 1999). The first and
Central chilled water plants are widely employed to
crucial step to improve the system delta-T performance is to
provide cooling to various kinds of facilities like commercial
understand the thermal characteristics of an individual chilled
building districts, university campuses, hospital complexes,
water cooling coil under various operating conditions. It
and industrial plants. In these facilities, chiller plant is a major
includes studying the coil delta-T with different geometric
contributor of the total energy consumption. Improving
parameters and various water- and airside conditions.
chilled water system operating efficiency to minimize energy
consumption is a big challenge for plant operators and Various studies have discussed how to keep a higher delta-
researchers. One of the most effective energy efficiency T for chilled water systems. Kirsner (1995) trouble-shot the
measures is to maximize the system loop chilled water return low delta-T syndrome in a large chiller plant. Some possible
Zhiqin Zhang is a consultant at KEMA, Inc., Oakland, CA. Hui Li is a post doc in the Energy Systems Laboratory at Texas A&M University
in College Station, TX. Jingjing Liu is a project engineer at Nexant, Inc., San Francisco, CA.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
©2012 ASHRAE
349
reasons were analyzed and excluded, such as higher supply ent conditions by using a decoupled model combining the
water temperature, dirty coils, or three-way valves. Finally, it effectiveness-NTU method and the finite element method. It
was found that the main reason was the cooling coil control set was concluded that the main reason for the low chilled water
points of air handling units were ratcheted downward and return temperature was the use of 3-way cooling coil control
blending station were inoperative. Fiorino (1996) recom- valves rather than the partial loads of the coil if the coil was
mended 25 “best practices” to achieve high chilled water designed, operated, and maintained properly. The waterside
delta-T ranged from component selection criteria to distribu- heat convection coefficient was estimated using Dittus-Boilter
tion system configuration guidelines. An example was equation from laminar to turbulent flow. Actually, this correc-
provided that low delta-T will prevent the buildings’ cooling tion should only be applicable for Reynolds Number
load from being satisfied at peak cooling load conditions. It (Re)>10,000.
was pointed out that the chilled water delta-T in a variable flow Various cooling coil models have been proposed to study
hydronic cooling system should be equal to design at full load the coil performance. The most popular one used for plant
and greater than design at part load. Taylor (2002) performed optimization is the effectiveness-NTU model developed by
a comprehensive discussion on the causes and mitigation of Braun (1989). The basic theory of a counter flow cooling coil
low delta-T. He addressed why delta-T degradation would was used, leading to the development of an effectiveness
usually occur and how to design around that eventuality to model used for analyzing the performance of cooling coils.
maintain chiller plant efficiency, despite a degrading delta-T. This model gives satisfactory results for temperature differ-
The focus was to improve chiller low load performance and try ences up to 50°F (27.8°C) between the water inlet temperature
to fully load the chiller. Moe (2005) proposed to apply pres- and the ambient wet-bulb temperature. The advantages of this
sure independent control valves to achieve high delta-T across gray-box approach are its simplicity, accuracy, and consis-
coils. Conventional 2-way control valves were replaced with tency and it has been used by Jiang (2005), Flake (1998), and
pressure independent control valves at coils and used to Yu (2008) to study the chilled water system optimization.
control the process. This valve could eliminate the effect of The objective of this paper is to study the delta-T charac-
sudden pressure difference variations on the coil flow rate teristics of a cooling coil using computational simulation
control or authority distortion. These studies proposed various method. The modified effectiveness-NTU coil model is used
qualitative analyses to delta-T degradation based on practical to simulate the coil performance under various airside and
observation or simplified engineering calculations. However, waterside conditions. The waterside heat convection coeffi-
detailed quantitative analyses on coil performance change at cient is simulated using a combined correlation and the airside
various operating conditions are lack. heat convection coefficient is simulated with the correlations
Some studies claimed that the coil part load performance developed by Elmahdy and Biggs (1979). The simulation
was one of the main contributors to the quick drop of the delta- results provide a straightforward insight into the cooling coil
T at the low cooling load end. During the swing season and delta-T characteristics under various operating conditions.
winter season, the load of the coils in the exterior zone of a
building may drop to an extremely low level, leading to a lami- COOLING COIL MODELING
nar flow in the water tubes and a low delta-T. If an air handling Figure 1 shows the schematic of a single duct VAV air
unit (AHU) is running under an economizer mode, the coil handling unit with a air-cooling and dehumidifying coil consid-
entering air temperature is low causing correspondingly low ered to study the performance of cooling coils. Part of room
return water temperatures. The water film heat transfer resis- return air is exhausted to the outside and the rest is mixed with
tance inside the tube is only a small portion of the overall air- the outside fresh air before entering the cooling coil. The chilled
to-water heat transfer resistance at the design flow rate, but as water flow rate is controlled by a 2-way control valve to main-
the water velocity falls, this resistance rises until, at laminar tain the cooling coil leaving air dry-bulb temperature setpoint.
flow conditions, it accounts for almost 90% of the overall
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
A reheating coil is located after the cooling coil to reheat or heat

resistance. However, this conclusion was not consistent with the cooling coil discharging air when necessary. The fan speed
other researches. Landman (1991) showed that the tempera- is modulated to maintain the air duct static pressure setpoint for
ture of returning chilled water rises at part-load conditions by
simulating the cooling coil chilled water leaving temperature
at 50% design load for variable air volume and constant air
volume systems, respectively, using the ARI-certified Trane
cooling coil performance program. Taylor (2002) showed that
the point at which delta-T began to fall was a function of initial
coil tube velocity, but it typically was at around 40% of load
for variable air volume (VAV) systems and 50% of load for
constant volume systems. During most low-load flow condi-
tions, delta-T will be above design delta-T. Wang, Zheng et al.
(2006) investigated the cooling coil performance under differ- Figure 1 Air handling unit configuration schematic.
350
ASHRAE Transactions
a VAV system. The load reduction is first achieved by reducing conditions, the rated capacity of the previous cooling coil is
airflow volume to 40%, then increasing room supply air temper- 19.7 tons of refrigeration (69.3 kW). The rated chilled water
ature by reheating. flow is 41.7 gpm (9.47 m3/hr) and the corresponding delta-T
A forward cooling coil model calculates the coil cooling is 11.3 °F (6.3°C) when the discharging air is 55°F DB/54°F
capacity from the entering air and water flow rates and temper- WB (12.8°C DB/12.2°C WB).
atures. However, the real control logic is to determine the sole
chilled water flow rate at the given air discharging dry-bulb AHU Operation Schedules
temperature setpoint. The corresponding chilled water leaving The airside and waterside conditions are the extrinsic
temperature will be calculated from the energy conservation factors determining the coil delta-T. In this study, the dry-bulb
principle. As an effectiveness-NTU model is a forward model, temperature of Houston Typical Meteorological Year 3
iterations are applied to find the unique water flow rate at the (TMY3) weather data are divided into 46 bins and the average
given discharging dry-bulb temperature setpoint as well as dry-bulb and wet-bulb temperatures in each dry-bulb bin are
airflow rate, entering air conditions, and entering chilled water calculated and used as the outside air temperature profiles. The
temperature.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
space conditioned by the AHU is maintained at 74°F DB

(23.3°C)/ 55% relative humidity (RH) all year round. The
Cooling Coil Geometry space total cooling load is assumed a second-order function of
The cooling coil geometric configuration is the inherent the outdoor DB between 4.0 ton (14.1 kW) and 13.0 ton (45.7
factor determining the coil delta-T characteristics. The kW). The maximum airflow rate is 6000 cfm (2.832 m3/s) and
selected cooling coil is a full circuit, six-row serpentine coil the minimum flow is 2400 cfm (1.133 m3/s). Reheating will be
with copper tubes and flat aluminum fins. The coil is 50 in. applied when the minimum airflow reaches. The airside econ-
(1.27 m) in length and 30 in. (0.76 m) in height. There are 24 omizer is applied based on the outside air dry-bulb tempera-
feed per row and the 1/2 in. (12.7 mm) outside diameter (OD) ture. The cooling coil discharging air temperature setpoint is
copper tubes are in a 1.25 in. (31.8 mm) equilateral staggered linearly reset between 55°F (12.8°C) and 60°F(15.6°C) and the
pattern. The tube inside diameter is 0.45 in. (11.4 mm) and the chilled water supply temperature is linearly reset between
tube wall thickness is 0.025 in. (0.635 mm). The fin thickness 42°F (5.6°C) and 46°F (7.8°C) both based on the outdoor dry-
is 0.0075 in. (0.19 mm) and the fin spacing is 12 fins per inch bulb temperature. In AHU simulations, the total supply
(fpi) (472 per meter). The maximum airflow rate is 6000 cubic airflow rate is calculated from the enthalpy difference of
feet per minute (cfm) (2.832 m3/s) and the maximum face supply and return air. The calculated total airflow rate is
velocity is 576 feet per minute (fpm) (2.93 m/s). The baromet- capped by the maximum and minimum airflow rates. When
ric pressure is 29.92 inch Hg (101.3 kPa). The thermal conduc- the calculated air flow is less than 2400 cfm (1.133 m3/s),
tivity is 223 Btu/hr·ft2·°F (385.9 W/m·K) for copper at 68°F reheating will be used to increase the supply air temperature.
(20°C) and is 144 Btu/hr·ft2·°F (249.2 W/m·K) for aluminum. The dry-bulb and wet-bulb bin temperatures, total supply
Coil cooling capacity is rated under certain conditions. airflow rate, chilled water supply temperature, coil discharg-
The most popular standard is ARI standard 410-2001 (ARI ing air temperature setpoint, and outside airflow ratio are
2001), which defines the ranges of standard rating conditions inputs of the coil models. Figure 2 shows the profiles of each
for cold water cooling coils: Standard air face velocity is 200 input parameter.
to 800 fpm (1.0 to 4.1 m/s); Entering air dry-bulb (DB) temper-
ature is 65°F to 100°F (18.3°C to 37.8°C); Entering air wet- Convection Coefficient Correlations
bulb (WB) temperature is 65°F to 85°F (18.3°C to 29.4°C); The effectiveness-NTU model is used in simulating the
Tube-side fluid velocity is 1.0 to 8.0 feet per second (fps) (0.3 cooling coil performance. This model simulates the perfor-
to 2.4 m/s); Entering fluid temperature is 35°F to 65°F (1.7°C mance of cooling coils utilizing the effectiveness model for
to 18.3°C). When the Reynolds Number is less than 3100 (1.0 counter-flow geometries. The performance of multi-pass
fps [0.3 m/s] in a 5/8” OD tube), the performance of a cooling crossflow heat exchangers approaches that of counter-flow
coil is somewhat unpredictable. On larger coils, water pressure devices when the number of rows is greater than four. The
drop limits of 15 to 20 ft (44.8 to 59.8 kPa) usually keep such airside convection coefficient is calculated using the correla-
velocities within acceptable limits of 2 to 4 fps (0.3 to 0.6 m/ tions developed by Elmahdy and Biggs (1979). The average
s), depending on circuit design. Dehumidifying coils for heat transfer Colburn J-factor is:
comfort applications are frequently selected in the range of
400 to 500 fpm (2.0 to 2.5 m/s) air face velocity (ASHRAE J = C 1 Re aC 2 (1)
2008). For water coils, circuitry would operate at 4 fps (1.2 m/
s), 42°F (5.6°C) inlet water, 12°F (6.7°C) rise (or 2 gallon per The quantities C1 and C2 are constant for a particular coil
minute (gpm) per ton of refrigeration effect). The standard over the airside Reynolds number (Rea) range of 200 to 2000,
ratings at 80°F dry-bulb (26.7°C) and 67°F (19.4) wet-bulb are corresponding to the air velocity of 1.7 fps (0.5 m/s) to 16.9 fps
representative of the entering air conditions encountered in (5.2 m/s) for the selected coil. Morisot, Marchio et al. (2002)
many comfort operations (ASHRAE 2008). Under these compared this correlation with other correlations of literature.
©2012 ASHRAE
351
Figure 2 Proposed reset schedules and calculated total airflow ratio profile (IP and SI).
It was shown that all results were close except one. Fin effi-
ciencies are calculated using the approximated Bessel func-
tions to calculate heat convection coefficients between the air
stream and the coil.
There are many correlations available in the literature for
the waterside convection coefficient. The most commonly
used one for fully developed turbulent flow inside smooth,
round tubes is the Dittus-Boelter correlation (Dittus and Boel-
ter 1930). Its accuracy is experimentally confirmed to about
±25% for the following conditions: 0.7  Pr  160  10,000
and L  di  10 . Pr is Prandtl number, Re is Reynolds number,
L is the straight pipe length, and di is the pipe internal diameter.
When Re is less than 2300, buoyancy forces become notable Figure 3 Heat transfer coefficients with various
and the Nusselt number (Nu) will be dependent upon water- correlations.
side Reynolds number, the ratio of finned length to tube diam-
eter, and the temperature difference between the bulk fluid and
the tube wall. The Sieder-Tate correlation (Incropera and These methods captured about 70% of their data within an
DeWitt 1990) is considered a good laminar, developing flow error band of ±10% and 97% of their data within ±20%. ARI
model designed to predict the coil performance for (2001) proposed a regression correlation to cover ranges from
Re w  2300 . Mirth, Ramadhyani et al. (1993) found that laminar flow to turbulent flow. When Re  2100 , this equation
Sieder-Tate correlation under-predicted the coil performance evolves into the Sieder-Tate correlation.
by as much as 15% even the length parameter was defined as Figure 3 shows the waterside convection coefficients as
the length of one tube pass. The possible reason is the second- flow varies with various correlations. When Re is higher than
ary flows generated by buoyancy forces and the tube bends are 8000, there is no obvious difference among all correlations.
not taken into account. The waterside heat transfer resistance When Re is between 2000 and 8000, the difference becomes
is usually a small fraction of the total coil heat transfer resis- significant. When Re is less than 2000 where laminar flow
tance for turbulent flow, but it contributes more than 60% occurs, all correlations are close to each other except for
when Re  2500 . The Gnielikski correlation (Gnielinski 1976) Dituus-Boelter correlation where a constant Nu of 4.364 is
was explicitly developed to fit experimental data well for used when Re <2000. In this picture, linear interpolation was
Reynolds numbers down to 2300 and it was considered to be used when Re is between 2000 and 3000 for Dituus-Boelter
more accurate than the Dittus-Boelter correlation over the correlation. In this study, a combined correlation was used for
entire range of turbulent Reynolds numbers (Incropera and the waterside heat transfer coefficient calculations. The ARI
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
DeWitt 1990). It is accurate to about 10% for the following regression correlation is employed for laminar flow
conditions: 0.5  Pr  2000 and 3000  Re  5  10 6 . Ghajar (Re<2100) (ARI 2001):
and Tam (1994) have investigated the transition region
between laminar and turbulent flow in plain tubes experimen- ln j w = A 1 + A 2 ln  Re w  + A 3 ln  Re w  2
tally and proposed some prediction correlations for this (2)
+  A 4 + A 5 ln  Re w  + A 6 ln  Re w  2   ln  ----
L
regime to bridge between laminar methods and turbulent  d i
methods, applicable to forced and mixed convection in the
entrance and fully developed regions for three types of inlet Nu
St = ------------ , j w =  St w  Pr w2  3     tw   w  0.14
configurations: re-entrant, square-edged, and bell-mouth. RePr
352
ASHRAE Transactions
A 1 = 0.62058, A 2 = – 0.66667, A 4 = – 0.33333, indicates that the waterside convection coefficient correlations
A 3 = A 5 = A 6 = 0.0
play an important role in predicting the chilled water delta-T
and various correlations may produce obvious different
The Dittus-Boelter correlation is used for fully turbulent profile. It also illustrates that, when the coil cooling load
flow (Rew > 10,000) (Dittus and Boelter 1930): decreases, the coil chilled water delta-T rises first, then drops,
and finally rises again. However, it should be noted that, such
Nu = 0.023Re 0.8 Pr 0.4 (3) a trending can only represent the delta-T characteristics of this
particular coil under these exactly operating conditions and it
The Gnielinski correlation is applied for transition flow
cannot be generalized to all cooling coils and all conditions.
( 2100  Re w  10,000 ) (Gnielinski 1976):
 f  8   Re – 1000 Pr Delta-T with Various Coil Geometrics
Nu = ------------------------------------------------------------------
- (4)
1 + 12.7  f  8  0.5  Pr 2  3 – 1  The cooling coil geometric configuration determines the
coil thermal characteristics. Figure 5 shows the coil chilled
where f is the Darcy friction factor that can either be obtained
water delta-T profiles with different variants of coil configu-
from the Moody chart or for smooth tubes from correlations:
rations. The baseline coil has 1/2 in. (12.7 mm) OD water
f =  0.79ln  Re  – 1.64  – 2 (5) tubes, 6 rows, 24 feeds per row, 12fpi (472 fin per meter), and
0.0075 in. (0.19 mm) fin thickness. To exclude the effect of
Linear interpolation is used in the transition regions near fluid and airside, the chilled water supply temperature is 42°F
the boundaries of the correlation ranges. (5.6°C) constant, the coil air discharging temperature is 55°F
(12.8°C) constant, and the outside air ratio is 0.2 constant. With
SIMULATIONS AND RESULTS different geometric configuration, the coil delta-T varies in a
wide range. Adding two extra rows can increase the delta-T by
Delta-T with Various Waterside Correlations 2.0°F (1.1°C) averagely. In addition, the cooling load when
Figure 4 shows the simulated cooling coil chilled water laminar flow occurs jumps from 5.5 ton (19.3 kW) to 7.0 ton
delta-T as a function of coil total load with three different (24.6 kW). The delta-T profile of the 8-row coil in the transi-
waterside convection coefficient correlations. The rated cool- tion region is much flatter than that of the baseline coil. The
ing coil performance point is 19.7 ton (69.3 kW) at 11.3°F coil dehumidification capability degrades and the cooling
(6.3°C) delta-T located at the right side of all profiles. The capacity at the maximum flow rate is lower than other coils.
chilled water supply temperature is 42°F (5.6°C) constant, the Selecting thicker fins may improve the delta-T slightly, while
coil air discharging temperature is 55°F (12.8°C) DB constant, larger water tube diameter, fewer feeds per row, and wider fin
and the outside airflow ratio is 0.2 constant. For the selected spacing will decrease the delta-T obviously, especially at the
combined correlation, the chilled water delta-T increases to high cooling load end. Because of lower bulk water tempera-
13.2°F (7.3°C) as the coil load decreases from 18 ton (63.3 kW) ture and associated cooler fin surface, the dehumidification
to 7 ton (24.6 kW), then it drops slightly to 11.4°F (6.3°C) until capability enhances and the coil cooling load increases at the
Re reaches 2100. In the laminar flow region (Re < 2100), the maximum airflow rate. It is observed that, for a coil with 30
delta-T rises to 16.4°F (9.1°C). The delta-T profile for ARI feeds per row, the delta-T decreases with load reducing until
correlation is lower than that of the combined correlation in the laminar flow occurs. This trend is different from other profiles
transition region and overlaps with each other in the laminar in the figure. It also indicates that the coil delta-T character-
region. The results of the Dittus-Boelter correlation are high in istic is so complicated that it is hard to use a general curve to
the transition region and low in the low load end. This figure represent its performance at part loads.
Figure 4 Cooling coil delta-T versus cooling load for two correlations (IP and SI).
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
©2012 ASHRAE
353
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 5 Cooling coil delta-T with different coil geometric variants (IP and SI).
Delta-T with Various AHU Operation Strategies Generally, increasing chilled water supply temperature
can improve the chiller efficiency. However, simulation results
In practice, the cooling coil discharging air temperature show that it will reduce delta-T obviously. Since the chilled
may be reset to a higher value at low loads to minimize simul- water flow rate in the chiller evaporator is capped by an upper
taneous cooling and reheating. The airside economizer is often limit, such a low delta-T could make it difficult to load up
adopted to make full use of the free cooling when the outside running chillers. Another chiller has to be staged on and this
air temperature is low. For a central chiller plant, a widely could only deteriorate the plant performance because each
adopted energy efficiency measure is to reset the chiller chiller will be running at an even lower part load ratio. There-
chilled water leaving temperature to a higher value in winter fore, it should be very cautious when resetting the chilled
to improve the chiller performance. A typical reset schedule is water supply temperature during the winter months. Another
shown in Figure 2. These operating changes will definitely possible reason for a high water supply temperature is that the
change the coil delta-T profile. Figure 6 shows the delta-T return water is mixed with the supply water through the bypass
profiles for constant setpoints, with individual reset schedule, line, three-way valves, or blending stations.
and with all reset schedules. The waterside Reynolds number When an airside economizer is applied, the mixing
profile is also plotted against the coil total cooling load. The temperature of outside air and return air or the coil entering
calculated water flow rate is proportional to the Reynold temperature is lower than that of the baseline case, which
number. At low load end, a higher air discharging temperature results in a lower chilled water delta-T. When the outside air
leads to a higher delta-T. Conversely, if the cooling coil air is cool enough, the mixing temperature drops below the coil
discharging setpoint is ratcheted downward, the coil chilled discharging setpoint and there is little cooling load on the coil.
water delta-T will degrade accordingly. Actually, lowering the The theoretical chilled water flow is very low and the simula-
coil discharging setpoint has been employed by many operation results become unreliable. In actual situation when the
tors trying to solve hot complaints quickly. An out-of-calibra- coil control valve cannot precisely modulate the chilled water
tion discharging air sensor may also exhaust the coil to meet flow, the leaking chilled water will over-cool the air and lead
a low setpoint. to a lower-than-simulated delta-T. This is another important
354
ASHRAE Transactions
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 6 Cooling coil delta-T with different AHU operations (IP and SI).
cause for low delta-T at low loads. The water loop pressure thicker fins will generate higher water return and supply
fluctuation may also push more water through the control temperature difference. When airside and waterside reset
valve and decrease the return water temperature further. schedules are applied, the delta-T profile at low load end devi-
ates widely from the base case. It should be noted that the
SUMMARY simulation results above are only applied to a particular coil
operated under specified conditions and simulated with
Increasing the chilled water loop return and supply
selected convection correlations. The findings cannot repre-
temperature difference is an effective measure to improve the
sent a general conclusion for all cooling coils, but it provides
performance of a chilled water system. However, low delta-T
some insight into the complexity of a chilled water cooling coil
syndrome exists in almost all chiller plants. In a chilled water
delta-T performance.
system, vast majority of chilled water is consumed by cooling
coils to cool down and dehumidify the air. Understanding the NOMENCLATURE
coil delta-T characteristics is critical for improving chilled
water loop delta-T. This paper has presented the chilled water AHU = air handling unit
delta-T simulation results using an illustrative cooling coil as C1, C2 = coefficient
an example. The effectiveness-Ntu coil model with a di = pipe internal diameter
combined waterside convection coefficient correlation is
DB = dry bulb
employed to model the coil chilled water leaving temperature
at given airside and waterside conditions. The simulation f = Darcy friction factor
results show that the different waterside correlations generate fpi = fins per inch
widely different delta-T prediction results, especially in the gpm = gallon per minute
laminar region. With variant geometric configurations, the
J = heat transfer Colburn J-factor
coil delta-T at full load may be higher, equal, or lower than that
at part load. It is no surprise that a coil with more rows, smaller L = straight pipe length
tube diameter, more feeds per row, wider fin spacing, and NTU = number of transfer units
©2012 ASHRAE
355
Nu = Nusselt numbe Ghajar, A. J. and L. M. Tam 1994. Heat transfer measure-
OD = outside diameter ments and correlations in the transition region for a cir-
cular tube with three different inlet configurations.
Pr = Prandtl number
Experimental thermal and fluid science 8: 79-90.
Re = Reynolds number Gnielinski, V. 1976. New equations for heat and mass trans-
RH = relative humidity fer in turbulent pipe and channel flow. Int. Chem. Eng.
St = Stanton Number 16: 359-368.
TMY = Typical Meteorological Year Incropera, F. and D. DeWitt 1990. Fundamentals of heat and
mass transfer. New York, John Wiley and Sons.
VAV = variable air volume
Jiang, W. 2005. Framework combining static optimization,
WB = wet Bulb dynamic scheduling and decision analysis applicable to
jw = heat transfer factor complex primary HVAC&R systems. Ph.D. Disserta-
 tw = dynamic viscosity of water at average tube wall tion, Department of Mechanical Engineering, Drexel
temperature University
w = dynamic viscosity of water at average water Kirsner, W. 1995. Troubleshooting chilled water distribution
temperature problems at the NASA Johnson Space Center. Heating
Piping Air Conditioning 67(2): 51-59.
REFERENCES Landman 1991. A new era of free cooling. Trane Engineers
Newsletter20(3):1-7
ARI 2001. ARI 410: 2001 Forced-circulation Air-heating Mirth, D. R., S. Ramadhyani, et al. 1993. Thermal perfor-
and Air-cooling Coils Arlington, VA, Air-Conditioning mance of chilled-water cooling coils operating at low
and Refrigeration Institute. water velocities. ASHRAE Transaction 99(1): 43-53.
ASHRAE 2008. ASHRAE Handbook—HVAC Systems and Moe, E. M. 2005. Applying pressure independent control to
Equipment Chapter 22 air-cooling and dehumidifying achieve high delta-T. International District Energy
coils. Atlanta, GA, American Society of Heating, Association (IDEA) Annual Conference, Nashville, TN.
Refrigerating and Air-Conditioning Engineers, Inc. Morisot, O., D. Marchio, et al. 2002. Simplified model for
Braun, J. E. 1989. Effectiveness models for cooling towers the operation of chilled water cooling coils under non-
and cooling coils. ASHRAE Transactions 96(2): 164- nominal conditions.
174. Taylor, S., P. Dupont, et al. 1999. The CoolTools Chilled
Dittus, E. J. and L. M. K. Boelter 1930. Publications on water plant design and specification guide. San Fran-
Engineering, Univ. California, Berkeley. cisco, PG&E Pacific Energy Center.
Elmahdy, A. H. and R. C. Biggs 1979. Finned tube heat Taylor, S. T. 2002. Degrading chilled water plant delta-T
exchanger correlation of dry surface heat transfer data. causes and mitigation. ASHRAE Transactions 108(1):
ASHRAE Transactions 85(2): 262-273. 641-653.
Fiorino, D. P. 1996. Twenty-five ways to raise your chilled- Wang, G., B. Zheng, et al. 2006. Impacts on building return
water temperature differential. ASHRAE Transactions water temperature in district cooling systems. Proceed-
102(1). ings of 2006 International Solar Energy Conference,
Flake, B. A. 1998. Parameter estimation and optimal super- Denver, CO.
visory control of chilled water plants. Ph.D. Disserta- Yu, F. W. and K. T. Chan 2008. Optimization of water-cooled
tion, Department of Mechanical Engineering, University chiller system with load-based speed control. Applied
of Wisconsin-Madison Energy 85(10): 931-950.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
356
ASHRAE Transactions
SA-12-002
Using Energy Flow Analysis to Assess

Energy Savings from Recirculation Controls
in Multi-Family Central Domestic
Hot Water Systems
Yanda Zhang, PhD Charlotte Bonneville Stephen Wilson
Member ASHRAE
Michael Maroney, PE Jeff Staller Julianna Yun Wei
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
ABSTRACT water from the central mechanical room to different parts of
the building. Hot water is constantly circulated in the RL so
Recirculation loops (RL) in multi-family central domestic
that building occupants can have quick access to hot water
hot water (DHW) systems bring hot water close to dwelling
whenever it is needed. RL pipe heat loss is a major source of
units to provide tenants with quick access to hot water.
However, long RL pipes make them a major source of system system energy loss since a typical RL includes very long pipes.
heat loss. Previous studies assessed energy savings by different There are several control technologies aimed at reducing RL
RL control technologies based on measurement of overall heat losses. They are particularly effective in existing systems,
system energy consumptions, and they did not reach definite since it is very costly to alter distribution networks. Several
conclusions. Overall energy consumption is affected by many studies have been performed to estimate performance of RL
operational parameters. Without evaluating RL pipe heat loss control technologies (Stone 2008, Gutierrez 2008, and.
directly, energy savings cannot be accurately assessed and Lobenstein 1992). These studies monitored and compared
energy savings mechanisms cannot be well understood. This overall DHW system energy consumptions with and without
paper presents an energy flow analysis method to assess RL controls. Since overall system energy consumption is affected
pipe heat loss based on field measurements of pipe tempera- by many other operational conditions, large variations of
tures and water flows. This method further separates the RL energy savings results were observed. Better understanding of
pipe heat loss into two components, one associated with the control technology performance is required to promote market
recirculation flow heat loss and one associated with hot water adoption of effective control technologies. In addition, it is
draw flow heat loss, so that the direct impact of controls to RL important to understand how system performance is affected
pipe heat loss was revealed. The energy flow analysis method by RL designs so that design guidance can be provided to
also provided a way to characterize central DHW system industry practitioners in new DHW system installations.
performance by based on percentage weighting of four major To address these two issues, the California Energy
energy flow components, including water heater loss, RL pipe Commission sponsored a study on multi-family central DHW
heat loss, branch pipe heat loss and hot water draw energy, in distribution systems, under the Public Interest Energy
the overall system energy consumption. Our field study results Research (PIER) program. This PIER study aimed at provid-
indicated that RL pipe heat loss represented about one third of ing in-depth understanding of RL system performance and
system fuel energy consumption and the average overall quantifying energy savings by control technologies. The
system efficiency was only about 34%. research conducted field performance monitoring studies at
more than thirty (30) multi-family buildings throughout Cali-
INTRODUCTION
fornia. Since RL pipe heat loss could not be directly measured,
In multi-family, hotel, and motel buildings, a centralized an energy flow analysis method was developed to evaluate RL
domestic hot water (DHW) system is usually installed, which pipe loss based on measurements of pipe temperatures and
includes one or more recirculation loops (RL), to bring hot water flows. This paper presents the methodology of the
Yanda Zhang is an associate director; Charlotte Bonneville, Stephen Wilson, and Michael Maroney are project managers; and Jeff Staller
and Julianna Yun Wei are project managers at the Heschong Mahone Group, Inc., Gold River, CA.
©2012 ASHRAE
357
energy flow analysis and the associated field study results. the return temperature drops below a minimum thresh-
Overall DHW system performance were summarized for all old setting, the pump is turned on to recharge the RL
multi-family buildings studied and in-depth analysis of energy with hot water. The pump is turned off when return tem-
savings by controls were provided for selected buildings. perature reaches a maximum threshold.
• Temperature Modulation Control: This technology
CENTRAL DHW SYSTEM CONFIGURATION controls hot water supply temperature, but not the recir-
AND CONTROLS culation pump operation. The control interacts with
Figure 1 provides a schematic of a central DHW system, water heater controls to lower the supply temperature at
which is separated into three sub-systems: water heater, recir- times when hot water demand is expected to be low.
culation loop, and distribution branches. The water heater Control schedules (length of control time and tempera-
subsystem typically includes several integrated storage tank ture settings) are determined by hot water demand pat-
water heaters or one or more boilers connected to a storage terns. Control schedules can be fixed or dynamically
tank. Recirculation loops bring hot water to different parts of adapted based on variations in draw patterns.
the building while distribution branches deliver hot water to • Demand Control: The recirculation pump is turned on
individual fixtures. The RL and branch pipes together make up when hot water draw is detected. This is achieved by
the distribution network. Since they have drastically different installing a flow sensor on the water heater cold water
operation patterns, they are separated into two subsystems for supply pipe, since all hot water draws trigger cold water
performance analysis, and this study focuses on RL perfor- makeup. Demand controls are usually coupled with a
mance. Recirculation systems in actual buildings can be much sophisticated return temperature control. In this case,
more complicated than that illustrated in Figure 1. Multiple the recirculation pump is turned on when both a hot
recirculation loops might exist and one main recirculation water draw is detected and the return pipe temperature is
loop may branch into several loops before merging back into below a threshold value. This combination further
one return path. Figure 2 provides illustrations of several reduces pump operation. Demand control results pre-
central DHW distribution designs obtained from the field sented in this study refer to this last type of control.
performance monitoring study.
There are four types of RL control technologies: FIELD PERFORMANCE STUDY
• Timer Control: A timer control switches off the recir- The study conducted two types of field performance
culation pump according to a fixed schedule, which cor- monitoring, full on-site study and partial on-site study. The
responds to times when the hot water demand is former was a comprehensive investigation of DHW system
expected to be zero, usually during night time. performance, configuration loop design, and energy savings
• Return Temperature Control: The recirculation pump from control technologies in a limited number of buildings.
is controlled based on the RL return temperature. Once The latter focused only on DHW system performance aspects,
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 1 Schematic of a central DHW system.
358
ASHRAE Transactions
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 2 Recirculation loop designs for four multi-family building monitoring by the study.
such as overall system efficiency, for a broader range of multi- rates into each branch were not measured because branch
family buildings. The project team visited and inspected more pipes were mostly inaccessible for flow meter installation. An
than fifty (50) multi-family buildings to identify DHW early investigative test indicated that inserting a thermocouple
systems suitable for monitoring purposes. One of the key crite- into the pipe or firmly attaching a thermocouple on the copper
ria during the building selection process was the presence of pipe surface resulted in the same measurement results. There-
long, straight and accessible pipes so that accurate flow fore, all further temperature measurements were performed by
measurements could be achieved, which were essential to attaching and securing a small gauge thermocouple probe (1/
energy flow analysis. In all, more than thirty (30) multi-family 16 inch [4.23 mm] in diameter) or a bare thermistor to corre-
buildings in different areas of California were selected and sponding water pipe surface. Thermal conducting paste was
studied. At some sites, monitoring led to unexpected service applied between pipe surface and thermocouple/thermistor
interruption and had to be terminated. As a result, not all field tips.
studies were successfully completed.
Full on-site studies tested different control technologies in
Field monitoring measured water flow rates and temper- addition to characterizing the general DHW system perfor-
atures at several locations in the DHW system, as illustrated in mance. Natural gas flow meters were installed to measure gas
Figure 1. Hot water draw flow and recirculation flow rate energy consumption. The project team also investigated RL
measurements are essential to energy flow analysis. Full on- designs of full on-site study buildings through plumbing plan
site studies used in-line magnetic flow meters in order to checks and field inspections. This information was used to
achieve the best measurement accuracy and partial on-site support the development and validation of a RL model. Partial
studies used high-accuracy ultra-sonic flow meters for ease of on-site studies only investigated DHW system performance in
installation, as they do not require plumbing work. Water flow as-found operating conditions, without introducing recirculation
©2012 ASHRAE
359
controls. Full on-site study at each building typically spanned a · ·
(  E Fuel +  E Recirc_Return
full year, while partial on-site monitoring usually lasted for
several weeks. E ControlVolume = +  E· Cold ) – (  E· Recirc_Supply (1)
· ·
ENERGY FLOW ANALYSIS METHOD +  E Draw_Supply +  E Heater_Loss
The first phase of the PIER research involved DHW While different time periods can be used, daily analysis
system performance monitoring and analysis in three multi- makes the most sense since hot water draws follow a daily
family buildings. System performance was evaluated based on pattern. The change of control volume energy, Econtrol Volume,
comparing overall system energy consumption, consistent is determined mostly by storage tank water temperature. In a
with other studies (Stone 2008, Gutierrez 2008, and. Loben- full-day cycle, tank temperature variation is negligible
stein 1992). Just as other studies, large variations of energy compared to other energy flow components in Eq. 1. The
savings were obtained and could not be clearly explained since above equation can be re-arranged to be:
the performance analysis method provided no insight into the ·
 E Heater_Loss + (  E Recirc_Supply
physical processes of RL operation. It was necessary to isolate · · ·
the RL pipe heat loss from the total system energy consump-  E Fuel = –  E Recirc_Return ) + (  E Draw_Supply (2)
tion to understand the impact of different control technologies. ·
–  E Cold )
RL pipe heat loss cannot be measured directly, but can be · · ·
inferred from water flow rates and temperature drops along RL Using E Draw = E Draw_Supply - E Draw_Loop_Loss -
· · ·
pipes. Because water flows in RL pipes are not constant and E Draw_Branch_Loss and E Recirc_Flow_Loss = E Recirc_Supply -
·
hot water draw flows only travel through part of the RL, a E Recirc_Return, the above equation can be further refined as
sophisticated method is needed to implement this concept. follows:
After trying different analysis approaches, the project team · ·
developed a control-volume based energy flow analysis  E Heater_Loss + (  E Recirc_Flow_Loss
· · ·
method that is able to assess daily RL pipe heat loss, as well  E Fuel = +  E Draw_Loop_Loss ) +  E Draw_Branch_Loss (3)
as other major DHW system energy flow components includ- · ·
ing water heater loss, branch pipe loss, and hot water draw +   E Draw –  E Cold 
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
energy. By comparing energy flow components under differ- The four components on the right side of the equation
ent system controls, we obtained better understandings of represent water heater loss, RL pipe heat loss, branch pipe heat
energy savings associated with various control mechanisms. loss, and net delivered hot water energy, respectively. With the
·
Figure 1 illustrates the different energy flow components exception of water heater losses,  E Heater_Loss , all other
in a central DHW system. In particular, hot water flow in RL energy flow components in the above equation can be evalu-
includes two flow components, recirculation flow and hot ated using field measurement data, as following.
water draw flow. The thermal energy carried by the recircula-
·
tion flow at the starting and return point of the RL is labeled
· ·  E Fuel = HV Fuel   V Fuel
as E Recirc_Supply and E Recirc_Return, respectively. The differ-
·
ence between the two, E Recirc_Flow_Loss, is due to RL pipe heat ·   C p   f recirc  t
loss. The energy content of the hot water draw flow at the start-  E Recirc_Flow_Loss =
·  T HW_Supply – T HW_Return 
ing point of the RL is labeled E Draw_Supply. Hot water draws
travel through part of the RL before exiting into branch pipes.
·   C p   f Draw  t
Draw flows suffer heat loss due to RL pipe heat loss,
·  E Draw_Loop_Loss =
E Draw_Loop_Loss, and heat loss due to branch pipe heat loss,  T HW_Supply – T HW_Branch 
·
E Draw_Branch_Loss. The final delivered energy to water fixtures
· · ·   C p   f Draw  t
is E Draw, which equals to E Draw_Supply - E Draw_Loop_Loss - ·
·  E Draw_Branch_Loss =
E Draw_Branch_Loss. Total RL pipe heat loss is the sum of  T HW_Branch – T HW_Fixture 
· ·
E Recirc_Loss and E Draw_Loop_Loss.
All these energy flow components can be correlated to the · ·   C p   f Draw  t
·  E Draw –  E Cold =
water heater energy consumption, E Fuel, using a control  T HW_Fixture – T HW_Cold 
volume analysis. The control volume is defined as the water
heater subsystem, as illustrated by the dashed box in Figure 1. where HVFuel is fuel heat value, VFuel is natural gas volumetric
According to the law of conservation of energy, the change of flow rate, fRecirc is recirculation flow rate,  is water density,
energy within the control volume equals to the difference Cp is water heat capacity,  t is measurement time interval,
between input energy and output energy. System energy fDrawc is hot water draw flow rate, TCold is cold water supply
balance within a period of time can be described by equation temperature, THW_Supply is hot water temperature at the begin-
Eq. 1: ning of the RL, THW_Return is hot water temperature at the end
360
ASHRAE Transactions
of the RL, THW_Branch is average hot water temperature at the the biggest impact on hot water return temperatures. Since
entrance of a branch, and THW_Fixture is average delivered hot pipe heat loss is directly proportional to pipe temperature, hot
water temperature at a hot water fixture. water return temperatures provide some indication of energy
By dividing both sides of equation Eq.3 by the daily fuel savings. The lower the return temperatures, the higher the
energy input, it can be rearranged to calculate overall DHW potential energy savings would be. Measurement of hot water
system efficiency, as well as relative contributions of the three draw flows indicated most multi-family buildings have hot
loss components. water draws throughout the day. This made timer control less
E Draw –  E Cold
useful, since it is designed to stop recirculation flow according

---------------------------------------------
-
· to a fixed schedule of times when there is no hot water draw.
 E Fuel
In some cases, measured cold water supply temperatures
·
= 1–
E Heater_Loss were found to be unexpectedly high when there were no hot
-----------------------------------
· -
water draws, as shown in Figure 2 (a) and (b). This was
 E Fuel
 System = (4) because cold water copper pipes could be heated through
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
· ·
 E Recirc_Flow_Loss +  E Draw_Loop_Loss- conduction from connected storage tanks, boilers, hot water
– ------------------------------------------------------------------------------------------------------
· return pipes, and even hot ambient temperatures in the
 E Fuel
mechanical room. This measurement error was corrected by
·
 E Draw_Branch_Loss- using the averaged cold water temperature when there were
– ---------------------------------------------------
· hot water draws. It should also be noted that a special demand
 E Fuel
control pump was used during demand control and timer
or control operations. As a result, the corresponding recircula-
 System =  WH – f Recirc_Loop_Loss – f Branch_Loss tion flows were higher than those observed in the other two
control scenarios.
where  WH is water heater efficiency, f Recirc_Loop_Loss is RL
pipe heat loss fraction, and f Branch_Loss branch pipe heat loss DHW System Performance Characterization
fraction.
RL pipe heat loss contributes to heat loss of the recircu- As indicated above, some assumptions had to be made to
lation flow and hot water draw flows, and it is equal to the sum complete the energy flow analysis because some hot water
of the following two components: draw temperatures could not physically be measured in actual
buildings. THW_Branch represent the average hot water temper-
·
• Recirculation flow heat loss (  E Recirc_Flow_Loss ): heat ature leaving the RL. It cannot be measured because branch
loss suffered by the recirculation flow pipes were not easily accessible in most buildings. We knew
· that this temperature would be between the RL supply temper-
• Hot water draw RL heat loss (  E Draw_Loop_Loss ):
heat loss suffered by hot water draws as they travel ature and the RL return temperature, or (THW_Supply -
through the RL THW_Branch) = f • (THW_Supply - THW_Return). The fraction, f,
depends on detailed RL plumbing configurations, hot water
Our field study was not able to measure all necessary draw locations and flow rates. As a first-order assumption, we
process parameters required in the above formulae. Certain assumed that, on average, hot water was drawn from the
engineering assumptions had to be made to complete the middle branch along the RL. We also assumed uniform heat
energy flow analysis, and these assumptions will be discussed loss per unit pipe surface area because the temperature gradi-
in details in the following section. ents along the RL pipes were relatively small. For full on-site
It should be noted that a rigorous analysis of branch pipe field studies, where detailed RL plumbing configurations were
heat loss should also include energy wasted during pipe cool available, the average fraction f was estimated to be 0.6. This
down when there is no hot water draws and energy waste asso- value was then used for partial on-site study buildings, where
ciated with hot water damped into drains while building occu- detailed RL designs were not available. Although more rigor-
pants are waiting hot water to come. However, our field studies ous heat transfer assumptions may be used, it was really the
were not able to provide necessary measurements to support RL plumbing configurations that had the largest impact to the
this kind of analysis. The treatment of branch pipe heat loss fraction f.
presented above only provides a very rough estimate of branch Branch pipes are usually much shorter and have smaller
pipe heat loss to illustrate its relative importance. diameters than RL pipes, so branch pipe heat loss is propor-
tionally smaller than RL pipe heat loss. According to pipe
RESULTS AND ANALYSIS surface area estimates, branch pipe heat loss was assumed to
As an overview of measurement results of DHW system be 10% of hot water draw RL heat loss. As indicated in the
operations, Figure 3 (a) through (d) present examples of RL previous section, our analysis method for branch pipe heat loss
operation parameters measured under four different control was only intended to provide a very rough estimate of this
scenarios. Of all measured process parameters, controls had energy flow component, because our field studies could not
©2012 ASHRAE
361
(a) Continuous Pumping: No Control (b) Timer Control
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
(c) Temperature Modulation (d) Demand Control
Figure 3 Measured DHW system operation under four control schemes.
measure hot water use at each end fixture to support detailed higher than hot water draw RL heat loss. Branch heat loss was
analysis. a relatively small component. Water heater loss and total
Full on-site studies measured natural gas flow, which distribution losses were equivalent, in terms of percentages of
allowed complete evaluation of all energy flow components. total system fuel energy input.
For all partial on-site studies, overall water heater efficiency Figure 5 shows two sets of histogram to summarize statis-
was assumed to be 70%. tics of DHW system performance of twenty seven (27) multi-
Using field measurement results and the assumptions family buildings monitored by this research, under either full
noted above, energy flow analysis was performed for all build- on-site or partial on-site studies. It can be seen that each DHW
ings where successful monitoring studies were performed. system had drastically different performance. According to
Figure 4 presents the average DHW system energy flows the histograms for RL loss, in six (6) buildings, or 22% of the
based on analysis results of all full on-site and partial on-site total number of building monitored, RL heat loss represented
study buildings. It corresponds to DHW system operation more than 50% of system natural gas consumption. On the
without controls. As shown in the diagram, average DHW other hand, according to the histogram for draw energy, over-
system overall efficiency was only about 34%. all DHW system efficiency in twenty (20) buildings, or 74%
Figure 4 illustrates that RL pipe heat loss is the dominant of the total number of building monitored, is less than 50%.
component of total distribution loss. As discussed before, RL The study calculated the ratios of daily recirculation flow
pipe heat loss causes heat loss of recirculation flow and hot volume to daily draw flow volume at each building and then
water draw flows. Therefore, the RL pipe heat loss was broken correlated the ratios to the corresponding RL heat loss frac-
into two components, recirculation flow heat loss and hot tion. The results, shown in Figure 6, show that daily recircu-
water draw RL heat loss. Recirculation flow heat loss is much lation flow volumes were much higher than daily draw flow
362
ASHRAE Transactions
Figure 4 Average multi-family DHW system energy flow.
Figure 5 Histogram of recirculation loop heat loss and draw energy fractions.
Figure 6 Correlation of recirculation loop pipe heat loss to water flow ratio.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
©2012 ASHRAE
363
Table 1. Measured Reduction of DHW System Energy Consumption
Timer Temperature Demand RL Pipe Heat Loss

Building
Control Modulation Control Fraction
a 7.6% -4.4% 24% 34%

b 2.1% 0.0% na 42%
c -11% 18% 11% 7%
d 14% na 32% 14%
volumes in most buildings, resulting in high RL heat loss frac- result, energy savings from control technologies cannot be
tions. It implies that, in general, RL systems are very ineffi- determined simply based on the overall DHW system natural
cient since hot water demands in most multi-family buildings gas consumption.
are relatively low compared to the amount of recirculation
The energy flow analysis was able to provide a better
flow needed to keep the RL warm. This is exemplified by a
understanding of control technology performance by reveal-
college dormitory building studied during the summer when
ing changes of RL energy flow components under different
occupancy rate was low, where measurements indicated that
controls. Table 2 through Table 4 show reductions in recircu-
RL heat loss represented 60% of the system natural gas energy
consumption. lation flow heat loss, hot water draw RL heat loss and RL heat
loss, respectively. Again, the baselines for comparison were
Energy Savings By Control Technologies DHW system operation without RL controls. The results of
RL pipe heat loss reduction are consistent with observation of
Table 1 presents the measured system energy savings return temperature reduction shown in Figure 3. Demand
from controls in four full on-site study buildings, where the control had the largest impact on recirculation flow heat loss
details RL designs were available. The baselines for energy (Table 2), since it turned off recirculation flows for the major-
savings calculation were DHW system natural gas consump- ity of the operating time. The large reduction in recirculation
tion without RL controls. Table 1 also provides the RL pipe flow heat loss did not lead to equal levels of RL pipe heat loss
heat loss fraction to indicate that the relative RL pipe heat loss reduction (Table 4), because hot water draw RL heat loss
in these four buildings is quite different. increased at the same time, as shown by the negative savings
From pipe heat transfer point of view, pipe heat loss is values shown in Table 3. Without recirculation flow, hot water
determined by two factors, overall heat transfer coefficient, draws would keep RL supply pipes hot but at slightly lower
UA, and the temperature difference between pipe surfaces and temperatures, because water flow rates inside RL pipes were
the ambient. Pipe UA depends on pipe insulation conditions, lower without recirculation flow. As a result, pipe heat loss
pipe diameter and length, which are not affected by controls. from RL supply pipes was only slightly reduced when there
The only way to reduce pipe heat loss is to reduce pipe surface were hot water draws. In this case, all RL supply pipe heat loss
temperatures. Figure 3 indicates that all controls reduced RL was passed to hot water draw flows and the delivered hot water
return pipe temperatures; therefore, RL pipe heat loss should would have slightly lower temperatures than those with recir-
be reduced accordingly. Demand control, coupled with return culation flow. Therefore, supply pipe heat loss cannot be
temperature control, achieved the highest amount of temper- significantly reduced when there were hot water draws. In
ature reduction and, therefore, should have generated the larg- contrast, without recirculation flow, RL return pipes were
est savings among all three controls. However, measured allowed to cool down to much lower temperatures to reduce
energy savings, shown in Table 1, were not always consistent pipe heat loss. Timer controls had the similar heat loss mech-
with observations of RL return temperatures. A very broad anism as demand control, but with a different pump operating
range of energy savings was obtained without a clear correla- scheme.
tion to relative RL pipe heat loss indicated by RL pipe heat loss
fractions. Timer and temperature modulation controls even Temperature modulation controls adjusted hot water
showed negative energy savings, meaning more energy was supply temperatures so that temperatures of all RL pipes were
used with the controls than those without controls. For build- reduced by about the same amount. The recirculation flow and
ing c and d, the measured percentage energy savings from hot water draw flow heat loss were affected in a similar way.
demand control were more than the corresponding RL pipe Total energy savings were moderate since the amount of
heat loss fractions. One major cause of those unrealistic temperature reduction was moderate and only for a portion of
energy savings was the large daily variation of hot water the day. Even though all three controls reduced water temper-
draws. The associated system energy consumption variation atures, none of them increased hot water uses that were beyond
offset control savings or caused false control savings. As a the observed normal hot water draw variations.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
364
ASHRAE Transactions
Table 2. Reduction of Recirculation Flow Heat Loss
Building Timer Control Temperature Modulation Demand Control
a 17% 1.3% 65%
b 28% 1.5% na
c 14% 11% 73%
d 19% na 49%
Table 3. Reduction of Hot Water Draw RL Heat Loss
Building Timer Control Temperature Modulation Demand Control
a -202% 5.5% -533%
b -109% 6.8% na
c -54% 6% -191%
d -84% na -279%
Table 4. Reduction of RL Pipe Heat Loss
Timer
Building Temperature Modulation Demand Control
Control
a 6.1% 1.0% 36%
b 5.0% 0.1% na
c 7% 0.3% 19%
d 7% na 10%
SUMMARY allowed us to quantify RL pipe heat loss, therefore, to accu-

This study developed an energy flow analysis method to rately assess energy savings from controls. It is important to
facilitate field performance evaluation of multi-family build- understand that total RL pipe heat loss is reflected in both
ing central DHW systems. This method establishes a way to recirculation flow heat loss and hot water draw heat loss. The
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
quantify RL pipe heat loss and to characterize DHW system energy flow analysis revealed that pump controls had more
performance according to system energy breakdown into hot impact to return pipe heat loss than to supply pipe heat loss. An
water draws, water heater loss, RL pipe heat loss and branch optimized recirculation pump control should try to maintain
pipe heat loss. Using this method, the field study revealed that RL supply pipes to a reasonable temperature while avoiding
RL pipe heat loss represented about one third of system fuel filling RL return pipes with hot water.
energy input, and the average overall system efficiency was
The PIER multi-family DHW distribution study has
only about 34%. Many building performed much worse than
developed a detailed RL heat transfer model that is capable of
this average level. Large pipe surface area is the direct cause
of large RL pipe heat loss. Continuous recirculation further assessing recirculation system performance and control
increases RL pipe heat loss by increase RL pipe temperatures, energy savings for different RL designs and hot water draw
especially return pipe temperatures. patterns. Development and validation of this model will be
The energy flow analysis method also provided a good presented in a future publication.
understanding of the performance of RL control technologies.
Simple comparisons of system fuel consumption is not effec- REFERENCES
tive in determining energy savings from controls, because
energy consumption associated with large daily hot water Stone, Nehemiah 2008. Central Domestic Hot Water System
draw variations can offset RL pipe heat loss reduction Study. June 1-3, 2008 – ACEEE Water Heating Confer-
achieved by controls. The energy flow analysis method ence.
©2012 ASHRAE
365
Gutierrez, Jorge and Woo, Kevin 2008. EDC, Inc. Hot Water Domestic Hot Water Temperature Control Strategies in
Recirculating System Study. June 1-3, 2008 – ACEEE Multi-family Buildings with Recirculation Loops.
Water Heating Conference. Report to Energy Resource Center, Legislative Commit-
Lobenstein, M.S, Bohac, D.L., Staller T.J., Dunsworth T.S., tee on Minnesota Resources, Contract # 02140-03535.
and Hancock M.W. 1992. Measured Savings from
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
366
ASHRAE Transactions
SA-12-003
Bayesian Analysis of Savings

from Retrofit Projects
John A. Shonder Piljae Im

Member ASHRAE Associate Member ASHRAE
ABSTRACT of which are common in real-world retrofit projects. No analyt-

ical solutions exist to determine savings or savings uncertainty
Estimates of savings from retrofit projects depend on in this situation, but several simplified formulas have been
statistical models, but because of the complicated analysis proposed. We model the data using a 5-parameter model with
required to determine the uncertainty of the estimates, savings first-order autoregressive errors, and use Bayesian inference
uncertainty is not often considered. Numerous simplified meth- to develop distributions for the model parameters and for the
ods have been proposed to determine savings uncertainty, but reporting period savings, which allows us to determine the
in all but the simplest cases, these methods provide approxi- savings uncertainty. We find the energy savings to be about 5%
mate results only. The objective of this paper is to show that lower than the result obtained by ignoring the autocorrelation.
Bayesian inference provides a consistent framework for esti- In addition, the Bayesian analysis finds the savings uncertainty
mating savings and savings uncertainty in retrofit projects. We to be narrower than the approximate uncertainty calculated
review the mathematical background of Bayesian inference using the simplified formula. These results show that Bayesian
and Bayesian regression, and present two examples of esti- inference can be used to determine savings and savings uncer-
mating savings and savings uncertainty in retrofit projects. The tainty for a wide variety of real-world problems.
first is a simple case where both baseline and post-retrofit
monthly natural gas use can be modeled as a linear function
INTRODUCTION
of monthly heating degree days. The Efficiency Valuation
Organization (EVO 2007) defines two methods of determining The usual objective of a retrofit project is to produce
savings in such cases: reporting period savings, which is an savings of one or more forms: water, energy use, demand,
estimate of the savings during the post-retrofit period; and operating cost, etc. Savings can be broadly defined as the base-
normalized savings, which is an estimate of the savings that line requirement minus the post-retrofit requirement, where
would be obtained during a typical year at the project site. For the requirement refers to the particular quantity intended to be
reporting period savings, classical statistical analysis saved. This simple definition raises a number of issues,
provides exact analytic results for both savings and savings however. In the case of retrofits involving heating, ventilating
uncertainty in this case. We use Bayesian analysis to calculate and air conditioning (HVAC) equipment, energy use depends
reporting period savings and savings uncertainty and show on weather conditions, which vary widely over time in most
that the results are identical to the analytical results. For locations. An objective comparison between the energy use of
normalized savings, the literature contains no exact expression the baseline and post-retrofit equipment requires the introduc-
for the uncertainty of normalized savings; we use Bayesian tion of counterfactual quantities: the amount of energy the
inference to calculate this quantity for the first time, and baseline equipment would have used during the post-retrofit
compare it with the result of an approximate formula. period had it not been replaced; or the amount of energy both
The second example concerns a problem where the base- the baseline and post-retrofit equipment would use under typi-
line data exhibit nonlinearity and serial autocorrelation, both cal weather conditions at the site. Statistical models are
John A. Shonder is a senior staff member and Piljae Im is a staff member at Oak Ridge National Laboratory in Oak Ridge, TN.
--`,`,``,,`,`,```,`,,``,`,
©2012 ASHRAE
367
required to estimate these counterfactual quantities, and the the breakpoint temperatures altogether. Ignoring the uncer-
use of models raises questions of accuracy. Some questions of tainty in the breakpoint temperatures linearizes the problem
interest include: How accurately are the project savings and allows classical statistical techniques to be used, but this
known, given the data available? Can limits be specified such too provides only an approximation to the savings uncertainty.
that there is a 95% probability that the true savings lies within One problem with approximate methods of estimating
them? What is the probability that the savings are greater than savings uncertainty is the difficulty of determining the accu-
some specified level? racy of the estimates they provide. Researchers usually assess
In practice, we find that such questions are infrequently the accuracy of their algorithms by comparing their results
considered. One reason is that classical statistical methods are with simulated problems with known properties. This
difficult to apply to most problems of interest. Bayesian infer- provides no information on the accuracy of the algorithm’s
ence provides an alternative method of characterizing savings results for specific, real-world problems, which may include
and savings uncertainty from retrofit projects. Bayesian tech- features that are outside the range of the simulated datasets
niques have been used previously in the analysis of time series tested by the researcher.
data on building energy use (Chonan et. al, 1996), and are Bayesian inference overcomes many of these limitations,
widely used and accepted in other fields of science. In the providing a flexible, coherent method for determining savings
Bayesian approach, the measured data are used to develop and saving uncertainty for a wide range of linear and nonlinear
probability distributions for parameters such as the slope and problems encountered in retrofit projects. As usually prac-
intercept of a regression line. Numerical techniques are most ticed, Bayesian inference is an approximate technique as well
often used to draw samples from the parameters’ probability since it depends on Monte Carlo integration of probability
distributions, and the samples can be used to estimate various distributions. However, the accuracy of the results can be
properties of the parameters such as their means, standard controlled through the number of samples drawn from the
deviations and quantiles. joint parameter distributions.
In the case of retrofit projects there is usually more inter- This paper begins with a brief introduction to Bayesian
est in functions of a model’s parameters than in the parameters inference and Bayesian regression. The Bayesian approach for
themselves. For example, a typical task in a retrofit project is determining savings and savings uncertainty is compared with
to fit a linear relation between baseline billed monthly natural the classical statistical technique for a simple problem, show-
gas use and monthly heating degree days, with the goal of ing that the two methods provide the same results. Bayesian
predicting annual gas use in a post-retrofit year with a speci- inference is then applied to a non-linear problem where no
fied number of heating degree days. In the Bayesian approach classical solution is available.
described in this paper, the billing data are used to develop a
joint probability distribution for the slope, intercept and BAYESIAN INFERENCE
regression standard error. A large number of samples are then Consider a set of n measured values y =  y1 y 2 ... y n  .
drawn from this joint distribution. Each sample provides a Here y is a column vector, and the prime symbol denotes the
value of slope, intercept and standard error, resulting in a transpose. The data are assumed to be drawn from a probabil-
different linear equation and a different estimate of annual gas ity distribution p  y   where  =   1  2 ... n  . The likeli-
use. The set of estimates so generated can be used to determine hood of the sample y is given by1:
useful quantities such as the mean and standard deviation of
the gas use in the year of interest. Given the actual gas use of n
the post-retrofit equipment in that year, the mean and a 95% p  y   =  p  yi   (1)
confidence interval for the savings can also be determined. i=1
Classical statistical methods handle linear regression In the notation used here the vertical bar signifies “given”
problems well, and provide exact methods for determining the or “subject to”, so that p  y   is the likelihood of the data y,
regression parameters and their uncertainties. However, given the parameter vector  . In classical statistical analysis
models of energy use in buildings are more often fit to non- these parameters are estimated by maximizing the likelihood.
linear functions of weather variables. While many algorithms In Bayesian inference, the data are used to develop probability
exist for estimating the parameters of nonlinear models, there distributions for the unknown parameters  , and the goal is to
is no general method to determine the standard error of a sum characterize the probability distributions. By Bayes’ rule, the
of predicted values, as required to determine savings uncer- distribution of  given the sample y is written as:
tainty in a retrofit project. Ruch at al. (1999) proposed an
approximate technique, but it is not stable in all cases (Kissock p   p  y  
p  y   = ---------------------------- (2)
et. al, 2003). Kissock et. al (2003) proposed the simplified py
numerical technique that is embodied in ASHRAE’s Inverse
Model Toolkit, but this technique is also an approximation, and 1.
To maintain consistency with the Bayesian literature, p (x) is used
the Toolkit addresses only a limited number of model types. to denote the probability density function of x. The functional
Reddy et. al (1998) recommended ignoring the uncertainty in form of p in each case depends on the particular variable x.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
368
ASHRAE Transactions
where p  y   is the likelihood as given in Equation (1), and Equation (7) can be integrated in closed form, and the means
p    is the prior distribution of the unknown parameters. The of the posterior distributions for the regression parameters are
prior distribution is based on what is known about the param- identical to the values obtained using classical statistical meth-
eters  before the data y are measured. When no prior infor- ods; see Gelman et. al, 2004) but in practice the posterior
mation is available, so-called diffuse or non-informative distribution is most often integrated numerically.
priors can be used. During the past few decades, increases in computing
Since the distribution p  y   must integrate to 1, p  y  is power and the development of efficient Markov chain Monte
defined as Carlo (MCMC) algorithms have combined to make it possible
to numerically integrate—or equivalently, to sample from—
py =    p  y   d (3) nearly any posterior distribution. One software package for
implementing MCMC simulation is JAGS (Plummer, 2010).
where  the integral is over all values of  . Because the like- The software is non-commercial, and is freely available via the
lihood as defined in Equation (1) is a function of, Equation (3) JAGS project website (http://www-ice.iarc.fr/~martyn/soft-
integrates to a constant. For the purposes of characterizing the ware/jags/).
distribution of the parameters  this constant can be ignored, JAGS implements a particular form of MCMC called
and it is sufficient to consider Gibbs sampling (in fact JAGS is an acronym for Just Another
p  y    p   p  y   (4) Gibbs Sampler), which is described fully in Casella and
George (1992). Statistical models are specified in JAGS using
In Bayesian terms, p   y  is called the posterior distribu- the BUGS language (Lunn et al., 2000), which is similar to the
tion—i.e., the distribution of the parameters  posterior to the popular (and also non-commercial) R statistical software (R
measurement of the data y —and Equation (4) can be Development Core Team, 2005). In fact, JAGS was chosen for
expressed in words as “the posterior distribution is propor- the present work because it is easily interfaced with the R soft-
tional to the prior distribution times the likelihood”. ware through the use of the rjags library (Plummer, 2011).
Bayesian Regression ENERGY SAVINGS FROM A NATURAL GAS

In a regression problem, in addition to the data y, a matrix RETROFIT
of explanatory variables X of dimension n  m is also given. Consider a project that replaces aging natural gas-fired
The probability model assumes that y is normally distributed heating equipment with new, more efficient equipment in a
with mean X and variance  2 , where  is an m  1 column building. Table 1 presents simulated data on the baseline and
vector. This can be written as: post-retrofit monthly billed gas use and the corresponding heat-
ing degree days in each month. Figure 1 is a plot of the baseline
p  y X,, 2   N  X,, 2 I  (5)
and post-retrofit data. Note that we have not specified the units
In words, Equation (5) means that y—given values for for the natural gas use, or the units and temperature base for the
, 2 and X, and —is distributed normally with mean X and heating degree days; the technique demonstrated here does not
variance  2 I , where I is an n  n identity matrix. The likeli- depend on the units employed, so long as they are consistent.
hood function is: The Efficiency Valuation Organization (EVO 2007)
defines savings in two different ways: reporting period
n
1 1 savings, which is an estimate of the savings during a selected
p  y , 2  =  -------------- exp – --------2-  y i – x i   2 (6)
i=1
 2 2 post-retrofit period; and normalized savings, which is an esti-
mate of the savings during a typical year at the project site.
where x i denotes row i of matrix X. To perform Bayesian Both quantities are estimated here.
inference, prior distributions for  and  2 must be specified.
For the moment these distributions are left unspecified and are Classical Statistical Method for Determining
denoted as p    and p   2  . The joint posterior distribution of Reporting Period Savings
 and  2 is then:
To estimate reporting period savings, the classical method
n requires a regression of monthly natural gas use with monthly
1 1 -
p  , 2 y   p   p   2   -------------- exp – -------------  y – x  2 (7) heating degree days. Figure 1 suggests a simple linear model
 2 2  2  i i of the form E pre = E pre0 + m pre  HHD , where E pre0 is the
i=1
baseline monthly natural gas use and HDD is the correspond-
In order to characterize the distributions of the parameters ing monthly heating degree days. The regression equation is
 and  2 —for example, to determine their means and stan-
found to be:
dard deviations—it is necessary to integrate the posterior
distribution. In some cases, a judicious choice of the prior E pre = 582.6575 + 2.7288  HDD (8)
distributions makes it possible to integrate the posterior in
closed form (and in fact, with certain priors for  and  2 with r2 = 0.982 and standard error of 116.7 units.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
©2012 ASHRAE
369
Table 1. Pre- and Post-Retrofit Monthly The energy use of the baseline equipment in the post-
Natural Gas Use (in Unspecified Units) and retrofit period could be estimated by substituting the heating
Heating Degree Days for a Retrofit Project degree days for each post-retrofit month from Table 1 into
Equation (8) and then summing the twelve monthly values.
Pre-Retrofit Post-Retrofit However, since the equation is linear, the annual energy use
can be calculated as 12E pre0 + m pre  HDDTotal , where
Month HDD Gas use HDD Gas use HDD Total is the sum of the degree days over all twelve months
Jan 906 3,099 963 2,680
of the post-retrofit period. Thus since there were 3,779 heating
degree days in the post-retrofit period, the estimated energy
Feb 529 2,054 682 1,994 use of the original equipment in the post-retrofit period is
(12)(582.6575) + (2.7288)(3,779) = 17,304 units. The new
Mar 564 2,169 504 1,568
equipment consumed 14,674 units during the post-retrofit
Apr 162 963 163 713 period, so the reporting period savings is 17,304 – 14,674 =
2,630 units of natural gas. This is about 15% of the expected
May 20 790 11 691 baseline energy use during the reporting period.
Jun 2 724 4 478 Reddy and Claridge (2000) present a compact matrix
equation for the standard uncertainty E saved of reporting
Jul 0 543 0 539 period savings:
Aug 0 346 0 646
E saved = s pre  1'(X post  X' pre X pre  –1 X' post )1  1 / 2 (9)
Sep 51 841 28 429
where s pre denotes the standard error of the baseline regres-
Oct 252 1,214 212 1,040
sion equation; X pre is a matrix consisting of a column of ones
Nov 326 1,437 373 1,416 prepended to the heating degree days per month in the base-
line period2, X post is a matrix consisting of a column of ones
Dec 565 2,027 839 2,480 prepended to the heating degree days per month in the post-
Total 3,377 16,207 3,779 14,674 retrofit period; I is an identity matrix; and 1 is a column vector
consisting of 1 in every position. Equation (9) does not include
Figure 1 Pre- and post-retrofit monthly natural gas use (in unspecified units) vs. monthly heating degree days for a heating
system retrofit project.
--`,`,``,,`,`,```,`,,``,`,``,,
370
ASHRAE Transactions
2  1'  X –1
tmy  X' pre X pre  X' tmy 1 
the effect of measurement error, and throughout this paper it [s pre
is assumed that measurement error is negligible. E saved  (11)
2  1'  X –1 12
+ s post tmy  X' post X post  X' tmy 1 ]
Using the data of Table 1 and the standard error of the
baseline regression, the standard error of the reporting period where s pre is the standard error of the baseline regression;
savings is found to be 573.8 units. To obtain a 100  1 –  % s post is the standard error of the post-retrofit regression; x pre ,
confidence interval the standard uncertainty must be multi- x post , I and 1 are as defined above, and X tmy is defined anal-
plied by t 1 –   2,n – m , where n is the number of data points and ogously to X pre . In this case the 100  1 –  % confidence
m is the number of parameters in the model. With n = 12, m = interval for each individual regression is calculated by multi-
2, and 100  1 –  % = 95%, a classical 95% confidence interval plying the standard uncertainty by t 1 –   2,n – m with n = 12
for the reporting period savings is found to be 2,630 ± 1,278 and m = 2 , so it is assumed that Equation (11) must also be
units of natural gas. Thus in this case the fractional uncertainty multiplied by t 1 –   2,n – m with n = 12 and m = 2 .
of the reporting period savings is about 49%. Using Equation (11), the data of Table 1 and the standard
errors of the respective regressions, an approximate 95%
Classical Statistical Method for Determining confidence interval for the normalized savings is found to be
Normalized Savings 2,584 ± 1,732 units of natural gas. The approximate fractional
savings uncertainty is about 67%, somewhat higher than the
To estimate normalized savings, a regression of post- fractional uncertainty of the reporting period savings. This is
retrofit natural gas use with heating degree days is also as expected, since the normalized savings depends on two
required. Figure 1 suggests another simple linear model, of the regressions, whereas reporting period savings depends on
form E post = E post0 + m post  HDD . The regression equation is only one. However, the lack of an exact expression for the
found to be: uncertainty of normalized savings makes it difficult to assess
the accuracy of this interval. As will be shown, Bayesian anal-
E post = 516.5162 + 2.2429  HDD (10)
ysis allows the interval to be determined to any desired degree
of accuracy.
with r2 = 0.984 and standard error of 107.3 units.
Given that a typical year at the project site has 3,685 heat- Bayesian Analysis of Savings
ing degree days, Equation (8) gives the normalized baseline The basis for the Bayesian analysis of reporting period
annual energy use as (12)(582.6575) + (2.7288)(3,685) = savings is the joint posterior probability distribution for the
17,047 units. By Equation (10), the normalized post-retrofit baseline regression parameter vector  and the variance  pre 2
annual energy use is (12)(516.5162) + (2.2429)(3,685) = as defined in Equation (7), with y equal to the baseline gas use
14,463 units. The normalized savings for the retrofit is then the and X equal to X pre as defined in Equation (9). The monthly
difference between the normalized baseline annual energy use gas use of the baseline equipment is modeled by the equation
and the normalized post-retrofit annual energy use, or 2,584 E pre = E pre0 + m pre  HHD , so  =  E pre0 m pre ' . To deter-
units of natural gas. This is about 15% of the normalized base- mine reporting period savings, a large number of samples are
line gas use. drawn from the posterior distribution. Each draw j results in a
A search of the literature uncovered no method for deter- set of values E pre0,j , m pre,j , and  pre,j for the slope, intercept
mining the exact uncertainty of normalized savings. Effinger and standard error of the baseline regression equation. These
et al. (2009) proposed calculating the uncertainty of the values are used to predict the energy use of the original equip-
normalized savings as the square root of the sum of the squared ment in each month of the post-retrofit period, and the twelve
uncertainties arising from each regression, but this is only an monthly estimates are summed to give the predicted annual
approximation. It would be exact if the pre- and post-retrofit energy use. For example, since the post-retrofit month of Janu-
normalized energy use were uncorrelated, but since Equations ary had 963 heating degree days, the estimated energy use of
(8) and (10) can be solved simultaneously to yield an equation the baseline equipment in that month for draw j would be:
for E post in terms of E pre only, the two variables are clearly
ˆ Jan
correlated. Nevertheless, the approximate uncertainty is Ej = E pre0,j + 963m pre,j + x j (12)
useful for comparison. Ignoring the correlation and applying
Equation (9) to each regression gives the following expression where x j  N  0, pre,j
2  . In other words, x j is a random variable
for the standard uncertainty: drawn from a normal distribution with mean 0 and variance
 pre,j . The estimated total energy use of the original equip-
ment in the post-retrofit period for draw j is then
1 906 ˆ
2. 1 529 E j = 12E pre0,j + m pre,j  HDD Total +  x j (13)
i.e., given the data of Table 1, X pre =
... ...
where  x j is the sum of twelve random variables drawn from
1 565
a N  0, pre,j
2  distribution. Summing twelve random variables
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
©2012 ASHRAE
371
drawn from a normal distribution with variance  2 is equiv- Although JAGS can select initial values for the Markov
alent to drawing a single random variable from a normal distri- chains based on the specified prior distributions, the algorithm
bution with variance 12 2 . Thus, given that HDDTotal for is more likely to converge to the desired distribution if initial
the post-retrofit period is 3,779 heating degree days, and that values close to the mean of each parameter are specified. Here
the new equipment consumed 14,674 units of natural gas the parameter values obtained from the Inverse Modeling
during that period, estimate j for the reporting period savings Toolkit—as given in Equations (8) and (9)—were used as
initial values for the simulation.
is given by:
E save, j = 12E pre0,j + 3779m pre,j + X j – 14674 (14) Number of Samples Required
The accuracy of the Bayesian estimates of the mean and
where X j  N  0,12 pre,
2
j . the 95% confidence interval for the reporting period and
With a large number of such estimates, the mean and 95% normalized savings depend on the number of samples drawn
interval can be determined for the reporting period savings. To from the respective distributions. Ruppert (2011, §4.3.1) pres-
distinguish them from the confidence intervals calculated ents a theorem which states that for a sample of size n drawn
using classical techniques, intervals calculated using Bayesian from a distribution with probability density function f and
techniques are sometimes called Bayesian credible intervals. cumulative distribution function F, the q th sample quantile
Normalized savings is calculated in a similar manner. The F – 1  q  is distributed approximately normally with mean equal
joint posterior probability distribution for the post-retrofit to the q th quantile of f and variance equal to
regression parameter vector  and the variance  2 is defined qq – 1 -
in Equation (7), with y equal to the post-retrofit gas use and X --------------------------------- (18)
n  f  F–1  q   2
equal to X post as defined in Equation (9). With the monthly gas
use of the new equipment modeled as By elementary sample theory, the number of samples n
E = E post0 + m post  HDD , a large number of samples are required to determine the q th quantile to a precision c at the
drawn from the posterior distribution, resulting in a set of 100  1 –  % confidence level is then approximately:
values E post0,j , m post,j , and  post,j for the slope, intercept and
standard error of the post-retrofit regression equation. The z 12 –   2 q  q – 1 
n = --------------------------------------------------- (19)
estimated total energy use of the new equipment in a typical  cf  F – 1  q  F – 1  q   2
year for draw j is then:
where z 1 –   2 is the two-tailed critical value for a normal
ˆ distribution at the  confidence level. Note that for a distri-
E j = 12E post0,j + m post,j  HDD TMY + Y j (15)
bution symmetrical about some value  , a sample provides
where Y j  N  0,12 post,j
2  and HDD TMY is the total number of two estimates of the quantile of interest: F – 1  q  and
heating degree days in the typical year. By Equation (13), the  – F – 1  1 – q  . Thus we suspect that for any value of q other
estimated energy use of the baseline equipment in a typical than 0.5 (i.e., any quantile other than the median), the number
year for draw j is: of samples required may be as few as one-half the number
given by Equation (19). However, we used Equation (19) as an
ˆ upper bound, realizing that the actual precision of a quantile
E j = 12E pre0,j + m pre,j  HDD TMY + X j (16)
estimate based on this number of samples is likely to be more
where X j  N  0,12 pre,j
2  . The normalized savings for draw j, accurate than the parameter c would suggest.
which is the difference between Equation (16) and Equation To determine the constants F – 1  q  and f  F – 1  q   in
(15), is given by: Equation (19), a large sample is drawn from the target distri-
bution. F – 1  q  is set to the q th quantile of the sample. To find
12  E pre0,j – E post0,j  +  m pre,j – m post,j  f  F – 1  q   , the probability density function f is approximated
E save,j = (17)
HDD TMY + Z j from the sample using a kernel technique; f  F – 1  q   is then
determined through interpolation.
where Z j  N  0,12 pre,j
2 + 12 post,j
2  . The posterior distribu- The required sample sizes turn out to be large in this case.
tions for the baseline and post-retrofit regressions (as given in To estimate the 2.5% quantile (q = 0.025) of the reporting
Equation 7) were implemented in the BUGS programming period savings with 1% accuracy at the 95% confidence level,
language. Separate statements were included to calculate the on the order of 89,000 samples must be drawn from the distri-
reporting period and normalized savings per Equations (14) bution; estimating the 2.5% quantile of the normalized savings
and (17) respectively. The prior distribution of all parameters to the same accuracy and confidence level requires on the
was set to a normal distribution with a mean of zero and vari- order of 356,000 samples. But in practice even more samples
ance of 108. This is a non-informative prior, which essentially are required: accurate determination of the quantiles requires
means that no prior information is known about the parame- an uncorrelated sample from the distribution; samples drawn
ters. The BUGS model was then run using JAGS. using the Gibbs sampler often do exhibit serial autocorrela-
372 --`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---

ASHRAE Transactions
tion. One way of avoiding this is to thin the sampler output by JAGS, one must also ensure that the sampling algorithm has
accepting only a certain fraction of values. Thus in this anal- converged and is providing simulated values that are represen-
ysis, five times the number of samples specified by Equation tative of the target distribution. While this topic is beyond the
(19) were drawn. The resulting sequence was then thinned by scope of this paper (and not often a concern in simple problems
accepting only every fifth value. While the number of samples such as this one), a thorough discussion of the issue is provided
may seem large, the code requires about 3 minutes to run on in Brooks and Gelman (1998).
a rather slow desktop computer. Examination of autocorrela-
tion plots shows that the resulting thinned samples are essen- ENERGY SAVINGS FROM A RETROFIT
tially free of serial autocorrelation. OF ALL-ELECTRIC HOMES
We have shown that in the case of a simple linear model,
Results and Comparison with Classical Results the results of Bayesian analysis of savings are identical to the
The result of Bayesian analysis of a regression problem is classical analytical results. However, most retrofit projects
a joint probability distribution for the regression parameters. involve non-linear models, and can include other complica-
Figure 2 presents histograms of these parameters for the pre- tions such as serial autocorrelation. No analytical formulas
and post-retrofit regressions, as well as for the reporting period exist to determine savings uncertainty for such problems, but
and normalized savings, which are based on the regression Bayesian techniques can be applied directly. Consider
parameters. The marginal distributions of all quantities shown Figure 3, which presents metered daily electrical use vs. daily
appear unimodal and symmetrical. Table 2 compares the average temperature – as well as a 5-parameter fit to the data
values for these quantities as derived through classical statis- – for a group of 40 residences in the Southern United States for
tical methods with the means of the distributions derived using 365 consecutive days. Although too large to present in this
Bayesian inference. For both reporting period savings and paper, the dataset is available online at http://tinyurl.com/
normalized savings, the means of the Bayesian estimates of 3ofq3nt. It can also be obtained directly from the authors.
savings and savings uncertainty are within 1% of the classical The 5-parameter function to which the data were fit has
values. In addition, the means of the distributions for the the following form:
regression coefficients are in close agreement with the classi-
cal estimates. E = Eo + mh  T – Th   H  Th – T  + mc  T – Th   H  T – Tc 
Although the Bayesian estimate of the uncertainty of the (20)
normalized savings agrees with the approximation based on
where E o is the baseline daily electrical use for the group of
Equation (11) to within 1%, examination of the results of
residences; m h and m c are the slopes of the heating and cool-
repeated runs suggested that the Bayesian algorithm was
ing lines, respectively; T h and T c are the heating and cooling
converging to a somewhat lower mean value. When the model
breakpoints, respectively; and H is the Heaviside step func-
was run with 100 times the original number of samples (corre-
tion, equal to zero for negative arguments and one for positive
sponding to 0.1% accuracy for the 95% interval), the Bayesian
arguments. Equation (18) is more commonly written as
estimate of normalized savings was found to be 2,584 ± 1,722
follows:
as opposed to the approximate value of 2,584 ±1,732. Thus in
this case, the approximation suggested by Effinger et al. appar- E = Eo + mh  T – Th  – + mc  T – Tc  + (21)
ently leads to a value for the uncertainty of normalized savings
that is too high by about 0.5%. Nonlinear regression was used to determine the regres-
Having the distribution of savings allows other questions sion parameters. The converged values were found to be E o =
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
to be answered. For example one may wish to determine the 2759.3, m h = -111.29, mc = 93.476, T h = 57.186, and T c =
probability that the savings is greater than some limit, say 2,000 68.140. These parameters were derived by correlating daily
units of natural gas. This is calculated as 85.1% for the report- electricity use with daily average temperature in ºF. Were we
ing period savings and 75.8% for the normalized savings. In to convert the daily average temperatures in our dataset to ºC,
other situations it may be important to establish the similarity different parameters would be obtained, but the same tech-
of two datasets: for example, when comparing the output of a nique would apply.
simulation model with site-monitored data from the actual The residuals of  i Equation (21) are then given by:
facility. The same technique could be used to determine the
probability that the “savings” – i.e., the difference between the E meas,i – 2759.3 – 111.29  T i – 57.186  -
i = (22)
annual energy use corresponding to regression fits of the two + 93.476  T i – 68.140  + )
datasets – was within some specified range that included zero.
In a classical regression problem, the analysis is not where E meas,i is the measured daily electrical use for day i and
complete simply because parameter values have been deter- T i is the average ambient temperature for day i. The RMSE of
mined: one must ensure that the model is consistent with the Equation (21) is 264.5 kWh, but the residuals have a problem.
data, that the residuals appear random and uncorrelated, etc. In Figure 4 plots  i vs.  i – 1 for i = 2 through 365. As is often the
the case of a numerical simulation such as that performed by case with daily data, the residuals exhibit serial autocorrela-
©2012 ASHRAE
373
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 2 Histograms of parameters of pre-retrofit regression (monthly gas use = Epre0 + mpre*HDD), post-retrofit regression
(monthly gas use = Epost0 + mpost*HDD), reporting period savings and normalized savings from Bayesian
analysis.
374
ASHRAE Transactions
tion. The lag-1 correlation coefficient r in this case is found to daily values are then summed to determine the annual baseline
be 0.5734. electricity use. Daily average temperatures for the post-retrofit
Now suppose the 40 residences receive energy efficiency year are also included in the on-line datafile. Given these
upgrades, and in the first year after the upgrades total electric- temperatures and Equation (21) with the converged values for
ity use is measured at 1,175,000 kWh. To calculate reporting the regression parameters, the baseline energy use in the post-
period savings, the usual practice is to ignore the autocorrela- retrofit period is found to be 1,305,037 kWh. By this method,
tion. Daily average temperatures from the post-retrofit year the reporting period savings E save is found to be 1,305,037 –
are substituted into Equation (21) to determine the baseline 1,175,000 = 130,037 kWh. The fractional savings is 130,037/
electricity use for each day in the post-retrofit year; the 365 1,305,037 or about 10%.
ASHRAE Guideline 14 includes an approximate method
Table 2. Comparison of Bayesian and Classical for determining savings uncertainty when the residuals are
Results for Gas Savings Analysis serially correlated. The first step is to calculate the effective
number of observations n' given by:
Bayesian Classical
Parameter 1–r
Result Result n' = n  ----------- (23)
1+r
Reporting period savings and
2,630 ±1,277 2,630 ±1,278
95% confidence interval where n is the original number of observations and r is the
Normalized savings and correlation coefficient. Given n = 365 and r = 0.5734, the
2,584 ±1,725 2,584 ±1,732 effective sample size is calculated as 98.97. The approximate
95% confidence interval
fractional savings uncertainty is then given by:
Pre-retrofit regression equation
1- 0.5
1.26  CV ----  1 + ---
Intercept 582.6 582.7 n 2- ---
E save n'  n' w
Slope 2.729 2.729 ---------------- = t --------------------------------------------------------------- (24a)
E save F
Post-retrofit regression equation
Intercept 516.4 516.5 Although Equation (24a) is also given by Reddy and Clar-
idge (2000), that paper states that “The CV value calculated
Slope 2.243 2.243 using n degrees of freedom has been renormalized by the new
Figure 3 Metered daily electrical use vs. daily average temperature for a group of 40 homes. The lines represent a 5-
parameter fit to the data.
©2012 ASHRAE
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
375
Figure 4 Residual of 5-parameter fit (see Figure 3) plotted against previous day's residual.
degrees of freedom n' .” Thus in this case we believe Equation Similar to the analysis presented in Ruch et al. (1999), the
(24a) should be: data of Figure 3 were assumed to follow a first-order autore-
gressive model:
n–m 1- 0.5
1.26  CV --------------  1 + --- 2- ---
E save n' – m  n'  w Et = Eo + mh  Tt – Th  - + mc  Tt – Tc  + + ut (26)
---------------- = t ------------------------------------------------------------------------- (24b)
E save F
with the residual u t given by:
where w is the number of observations in the post-retrofit u t = ru t – 1 +  t (27)
period, F is the fractional savings and
Equations (26) and (27) can be manipulated to eliminate
n 2 0.5 the u variables. The resulting expression is:
1 i = 1  Ei
– E
CV = --- ----------------------------------- (25)
E n – m rE t – 1 + E o + m h  T t – T h  - + m c  T t – T c  +
Et = (28)
– r  E o + mh  Tt – 1 – Th  - + mc  Tt – 1 – Tc  +  + t
In Equation (25), E i is the measured daily electricity use
for day i in the baseline period, E is the average daily elec- Equation (28) was implemented in the BUGS language,
tricity use for the entire baseline period and m is the number and the model was run under JAGS to sample from the joint
of parameters in the fitted equation (five in this case). In Equa- distribution for r,E o ,m h ,m c ,T h ,T c , and  . Each draw j from
tion (24b), t is either t 1 –   2,n' – m or t 1 –   2,n – m ; neither the joint distribution provides values r j ,E oj ,m hj ,m cj ,T hj ,T cj ,
ASHRAE Guideline 14 nor Reddy and Claridge (2000) spec- and  j . The value  j is used to develop the error series u t per
ify which one to use, presumably because the difference Equation (27):
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
between the two is small when n and n' are greater than about
30. Here it was assumed that t 1 –   2,n' – m . u t = ru t – 1 + N  0, j2  (29)
Given the data and Equations (23) – (25), the approximate
for t = 2 through 365, with u 1 = N  0, j2  . Then u t is substi-
fractional savings uncertainty at the 95% confidence level is
tuted into (26) with the values for E oj ,m hj ,m cj ,T hj , and T cj :
calculated as 0.5180. Thus an approximate 95% confidence
interval for the reporting period electricity savings is 130,037 ˆ
E tj = E oj + m hj  T t – T hj  - + m cj  T t – T cj  + + u t (30)
± 67,356 kWh.
ˆ
With no exact value of the uncertainty available for The predicted daily values of electricity use E tj are
comparison, it is difficult to assess the accuracy of this inter- summed for t = 1 through 365 to give an estimate of annual
val. Here again, Bayesian inference can be used to determine electricity use; the savings estimate for draw j is then the
the uncertainty to any desired degree of accuracy. annual estimate minus the measured post-retrofit electricity
376
ASHRAE Transactions
Table 3. Comparison of Bayesian and Approximate
Classical Results for Electricity Savings Analysis
Parameter Bayesian Result Approximate Result
Baseline energy use

1,298,163 1,305,037
(kWh)
Reporting period
savings and 95% 123,163 ± 39,356 130,037 ± 67,356
confidence interval
Eo 2,876.6 2,759.3
mh -87.619 -111.29
Th 56.951 57.186
mc 78.138 93.476
Tc 67.607 68.140
r 0.7086 0.5734
Figure 5 Histogram of reporting period electricity savings

from retrofit project. relation, 130,037 kWh, does lie within the Bayesian 95% cred-
ible interval of 123,163 ± 39,356 kWh.
Table 3 presents the means of the other regression param-
use of 1,175,000 kWh. The collection of savings estimates is eters, as well as the approximate parameter values obtained
used to determine the mean and a 95% credible interval for the above. It is evident that a model accounting for autocorrelation
reporting period savings. results in parameter values that are quite different from the
results obtained using the approximate techniques recom-
As in the previous case, the regression parameters from mended by ASHRAE Guideline 14.
the 5-parameter fit (as well as the correlation coefficient for There are other methods of determining the parameters
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
the lag-1 residuals) were used as initial values for the Markov r,E o ,m h ,m c ,T h , and T c in Equation (28). Minimizing the sum
chains. The prior distribution of all parameters was set to a of the squared errors in this equation is a non-linear regression
normal distribution with a mean of zero and variance of 108, problem. Applying standard optimization software (see
and a large sample was generated from the distribution of Venables and Ripley 1999, §8.2), we found E o = 2,877.0, m h
reporting period savings. The sample was used to estimate the = -87.26, m c = 79.50, T h = 57.00, Tc = 67.73, and r = 0.699.
probability density function, which in turn was used to gener- These values are all in good agreement with the means of the
ate the parameters required by Equation (19) to determine the respective marginal parameter distributions found using
number of samples required to estimate the mean and the 95% Bayesian techniques, which are presented in Table 3. Using
credible interval to 1% accuracy at the 95% confidence level. the results of the nonlinear regression to determine the base-
In this case, the required number of samples turned out to be line electricity use in the post-retrofit period, the savings are
much lower, on the order of 18,000. On the other hand, since calculated as 124,318 kWh. This is within 1% of the Bayesian
the sample generated by JAGS included significantly more result.
autocorrelation, it was decided to draw 10 times this number Of course, it is likely that the post-retrofit data would
of samples and to thin the result by accepting only every tenth include serial autocorrelation as well, and thus a more compli-
value. The calculation of the required number of samples was cated analysis, akin to the calculation of normalized savings,
an iterative process; the figure of 18,000 samples was derived would be required to account for the autocorrelation in both
from an uncorrelated sample. the pre- and post-retrofit data. The same techniques applied
Figure 5 is a histogram of the reporting period savings as here could be used to calculate savings and savings uncertainty
derived from the Bayesian analysis. Again, the distribution for that case.
appears symmetric and unimodal. The mean savings of
123,163 kWh is about 5% lower than the value calculated by CONCLUSIONS
ignoring the autocorrelation. Additionally, the 95% interval In this paper, we have shown that where classical analyt-
for savings uncertainty is about 40% lower than the approxi- ical solutions are available for savings and savings uncertainty
mate value, meaning that the approximate savings uncertainty in retrofit projects, Bayesian inference using non-informative
is quite conservative in this case. We also note that the value prior distributions provides identical results. We have also
for reporting period savings obtained by ignoring the autocor- shown that Bayesian analysis can be applied to complicated
©2012 ASHRAE
377
nonlinear problems where no analytical solution is available. w = number of observations in post-retrofit period
We propose that Bayesian analysis is similarly reliable in the X = a matrix of predictor variables
latter case, and can provide values for the savings and savings
x = a random variable
uncertainty that are unavailable analytically. As opposed to the
mix of approximate formulas presently used to calculate X = a random variable
savings and savings uncertainty, Bayesian inference provides y = a column vector of measured response variables
a coherent, consistent methodology that can be applied across Y = a random variable
a wide range of linear and nonlinear problems. The numerical
technique we have demonstrated can also be used to account
Z = a random variable
for other sources of uncertainty such as measurement error and  = a column vector of regression parameters
weather variability. Thus we believe that Bayesian inference  = regression residual
should be put to more widespread use in estimating savings  = correlation coefficient
and savings uncertainty in retrofit projects.
 = a column vector of statistical parameters
NOMENCLATURE i = the individual elements of 
CV = coefficient of variation  = standard deviation
c = precision 1 = a column of ones
E post = monthly post-retrofit natural gas use
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
E post0 = constant in linear regression of E post with monthly REFERENCES
heating degree days American Society of Heating, Refrigerating and Air-Condi-
E pre = monthly pre-retrofit natural gas use tioning Engineers, Inc. (ASHRAE). 2002. ASHRAE
E pre0 = constant in linear regression of E pre with monthly Guideline 14-2002, Measurement of Energy and
heating degree days Demand Savings.
E saved = standard uncertainty of savings Brooks, S.P. and A. Gelman, 1997. General methods for
monitoring convergence of iterative simulations. Journal
F = cumulative probability density function of Computational and Graphical Statistics, 7, 434-455.
f = a probability density function Chonan, Y., Nishida, K., and T. Matsumoto, 1996. Great
H = Heaviside function Energy Predictor Shootout II. A Bayesian Nonlinear
HDD = heating degree days (base 65 °F) Regression with Multiple Hyperparameters. ASHRAE
I = an identity matrix Transactions 1996, Vol.102, Part 2, Paper number SA-
k = the number of elements in  96-3-2, 405-411.
EVO (Efficiency Valuation Organization), 2007. Interna-
kWh = kilowatt-hours
tional Performance Measurement and Verification Pro-
m = the number of parameters in the model tocol. Available at http://evo-world.org/.
mc = slope of cooling line in 5-parameter fit Gelman, A., Carlin, J., Stern, H., and D.B. Rubin, 2003.
mh = slope of heating line in 5-parameter fit Bayesian Data Analysis. Second Edition. Chapman &
m post = slope of linear regression of E post with monthly Hall/CRC.
heating degree days Kissock, J.K., Haberl, J.S., and D.E. Claridge, 2003. Inverse
m pre = slope of linear regression of E pre with monthly Modeling Toolkit - Numerical Algorithms. ASHRAE
heating degree days Transactions, vol. 109, pt. 2.
N = the normal distribution Lunn, D.J., Thomas, A., Best, N., and D. Spiegelhalter,
2000. WinBUGS -- a Bayesian modeling framework:
n = number of observations, or number of samples
concepts, structure, and extensibility. Statistics and
n = adjusted number of data points Computing, 10:325--337.
px = probability density of the variable x Plummer, M., 2010. JAGS Version 2.2.0 user manual. Down-
q = quantile of a distribution or sample loaded from http://iweb.dl.sourceforge.net/project/
r2 = correlation coefficient mcmc-jags/Manuals/2.x/jags_user_manual.pdf.
s = standard error Plummer, M., 2011. Package ‘rjags’ (documentation).
T = temperature, °F Downloaded from http://cran.r-project.org/web/pack-
ages/rjags/rjags.pdf.
Tc = cooling breakpoint temperature, °F
R Development Core Team 2005. R: A language and envi-
Th = heating breakpoint temperature, °F ronment for statistical computing, reference index ver-
t = critical value of Student’s t-distribution sion 2.9.0. R Foundation for Statistical Computing,
u = autocorrelated residual Vienna, Austria. URL http://www.R-project.org.
378
ASHRAE Transactions
Reddy, T.A. and D.E. Claridge, 2000. Uncertainty of “Mea- ing energy use models with autocorrelated residuals.
sured” Energy Savings from Statistical Baseline Mod- ASME Journal of Solar EnergyEngineering 121(1):63-
els. HVAC&R Research, Vol. 6, No. 1. 68.
Reddy, T.A., Kissock, J.K., D.K. Ruch, 1998. Uncertainty in
Ruppert, D., 2011, Statistics and Data Analysis for Financial
Baseline Regression Modeling and in Determination of
Engineering. Springer-Verlag, New York, NY.
Retrofit Savings. Journal of Solar Energy Engineering,
August 1998, Volume 120, Issue 3, 185. Venables, W. N. and Ripley, B. D., 1999. Modern Applied
Ruch, D.K., Kissock, J.K. and T.A. Reddy, 1999. Model Statistics with S-PLUS. Third Edition. Springer-Verlag,
identification and prediction uncertainty of linear build- New York, NY.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,
©2012 ASHRAE
379
SA-12-004
Energy Savings Potential of Variable

Condenser Water Flow Systems
Zhiqin Zhang, PhD Jingjing Liu

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
ABSTRACT Refrigeration Institute) 550/590 (ARI 1998). The ARI rating
conditions are frequently used as design conditions for chilled
Most chilled water systems are designed and operated
water systems. However, they may not represent the best
with constant condenser water flow rates. Recently, there is a
design conditions to use for every project.
trend to apply VSDs to all the components to achieve higher
operation performance when specially tailored operating In recent years, there has been considerable debate on the
strategies are incorporated. This paper studied the energy advantages of designing lower nominal condenser water flow
savings potential of varying condenser water flow rates by rates in order to optimize system performance. Although
simulating an example condenser water loop. The simulation lowering condenser water flow from 3.0 gpm/ton (0.194 m3/
results show that, if the pump water flow and fan airflow are hr·kWt) to 2.0 gpm/ton (0.129 m3/hr·kWt) will penalize the
throttled by valves or dampers, the system could consume more energy performance of the chiller, it is claimed that the energy
energy when reducing the condenser water flow. If the pump savings on the cooling towers (CT) and condenser water
and cooling tower are equipped with VSDs, the optimal pumps will outweigh the chiller losses. It is also suggested that
condenser water flow is highly dependent on the design condi- the capital cost could be reduced due to reduced equipment
tions and the sensitivity of chiller performance to the size. Waller (1988) recommended a condenser water flow rate
condenser water flow. The local climate has a minor effect on of 1.5 gpm/ton (0.097 m3/hr·kWt) for centrifugal chillers to
the optimal condenser water flow rate. At around the optimal achieve “the lowest life cycle cost with no energy penalty
CW flow rate, the system annual total energy usage is not sensi- when compared to 3 gpm/ton (0.194 m3/hr·kWt).” The initial
tive to the CW flow rate. Compared to operation with a performance of the chiller, tower fan, and CW pump is 0.631
constant optimal CW flow, the savings of applying variable kW/ton, 0.103 kW/ton, and 0.0623 kW/ton, respectively.
optimal CW flows are negligible. This method can be applied Shelton and Weber (1991) found, compared to 3 gpm/ton
to actual systems to find the optimal condenser water operating (0.194 m3/hr·kWt), 2 gpm/ton (0.129 m3/hr·kWt) could reduce
strategies. system peak demand by 3.5% and annual energy usage by
10.5%. The chiller performance degrades by 4% when the
INTRODUCTION condenser water flow decreases. Shelton and Joyce (1991)
further indicated that the chiller manufacturer’s minimum of
The condenser water (CW) loop of a chilled water (ChW)
1.5 gpm/ton (0.097 m3/hr·kWt) was optimum under a range of
system is normally designed at 3.0 US gallons per minute
conditions in terms of lowest capital cost and annual energy
(gpm) (0.68 m3/hr) per ton of refrigeration (ton). A nominal
cost. Kinter-Myer and Emery (1995) also recommended using
cooling tower ton is defined as cooling 3.0 gpm (0.68 m3/hr)
the manufacturer’s minimum to save pumping costs.
of condenser water from 95°F (35°C) to 85°F (29°C) at a 78°F
(26°C) entering air wet-bulb (WB) temperature (ASHRAE Some people have proposed variable condenser water
2008). These have long been the standard rating conditions for flow by employing a variable-speed drive (VSD) on condenser
water-cooled chillers as found in ARI (Air-Conditioning and water pumps not only for balancing purposes but also for
Zhiqin Zhang is a senior project engineer, and Jingjing Liu is a project engineer at Nexant Inc., San Francisco, CA.
380
ASHRAE Transactions
whole system energy savings. There is a trend to apply the condenser water flow was artificially underestimated in
variable-speed drivers to all components to achieve higher Kirsner’s example. The pump and tower power totally saves
operation performance when specially tailored operating 0.03 kW/ton with 2 gpm/ton (0.129 m3/hr·kWt). It is interest-
strategies are incorporated (Hartman 2001). ASHRAE recom- ing to note that all the theoretical chillers have penalties near
mends a minimum condenser water flow velocity of 3.3 feet 0.045 kW/ton, or 8.0%, for going to 2.0 gpm/ton (0.129 m3/
per second (fps) (1.0 m/s) to maintain turbulent velocity and hr·kWt). In fact, compared with the theoretical efficiencies,
prevent formation of deposits in the condenser tubes. This actual performance averages 8.9% worse at 2.0 gpm/ton
minimum flow rate is well below the velocity of 6 to 8 fps (1.8 (0.129 m3/hr·kWt). However, the selected chiller in Kirsner’s
to 2.4 m/s) encountered in today’s chiller designs. Heat example degrades only 6.3% because the impeller was not
exchange calculations show us that the main factor in condens- machined to the correct diameter for the duty. At off-design
ing the hot refrigerant gases is a large surface area requirement conditions prevailing for nearly 99% of operating hours, the
for the copper tubes found in most condensers. The water net effect of the varying energy penalties and savings depends
velocity at the typical design flow rates is a small and nearly on the chiller part load ratio. For a system with constant speed
negligible factor in the heat transfer equation. chiller and 55°F (13°C) minimum entering condenser water
It is generally accepted that reducing the condenser water temperature, the break-even point occurs at approximately
flow rate can reduce the condenser water pump power and 60% load. With a variable-speed chiller utilizing 55°F (13°C)
cooling tower fan power while the chiller compressor power minimum entering CW temperature in a 2 gpm/ton (0.129 m3/
increases. However, some researchers found that it was not hr·kWt) system, energy savings start below 83% load. As a
necessarily true that total operating costs and first cost can be result, it is concluded that there are no miracle savings avail-
saved by reducing condenser water flow. The legitimacy of the able and that the old standard of 3 gpm/ton (0.194 m3/hr·kWt)
above claims was first scrutinized by Kirsner (1996). He will result in a more economic design most of the time.
conducted an impartial analysis of the impact of 3.0 gpm/ton Other researchers also conducted some studies on the
(0.194 m3/hr·kWt) versus 2.0 gpm/ton (0.129 m3/hr·kWt) optimization of a CW loop, which deserve to be mentioned.
condenser water flow rate and drew some different conclu- Soylemez (2004) presented a thermo-hydraulic performance
sions. It is shown that a 3.0 gpm/ton (0.194 m3/hr·kWt) optimization analysis, yielding a simple algebraic formula for
condenser water flow rate will most often lead to a smaller estimating the ratio of mass flow rate of water to dry air of
full-load power draw for system components compared to 2.0 counter flow mechanical draft wet cooling towers. At the opti-
or 1.5 gpm/ton (0.097 or 0.129 m3/hr·kWt). Although a lower mum point, the maximum cooling tower effectiveness arises.
condenser water flow rate can often result in lower annual The fixed and variable parameters are listed as the thermo-
energy cost, this result is highly dependent on the specific site physical properties of water and air, ambient pressure and
utility rates, pumping head, and the chiller load profile and so average water temperature throughout the cooling tower.
cannot be generalized without specifying these factors. The There exist two important parameters, the ratio of heat capac-
power per full load of the selected chiller is approximately ity rates of the flow and the number of transfer units, in opti-
0.60 kW/ton at ARI conditions. Kirsner showed that the kW mizing the operating conditions of a counter current cooling
per ton penalty for chillers due to the reduction in condenser tower. The ambient pressure, or local elevation, has an impor-
water was from 8.0% to 9.6% at full load. The pump power per tant effect on the optimum ratio of mass flow rate of water to
full load ton is 0.0375 kW/ton, and it drops to 0.0143 kW/ton dry air. The average temperature of the tower water has a much
at the lower condenser water flow. For the selected cross-flow more significant effect on the ratio of mass flow rate of water
cooling tower with draw through fan, the air flow rate reduces to dry air. The results are checked for a typical cooling spray
by 15% to achieve the same leaving water approach when pond problem and the best water leaving temperature for
reducing condenser water flow. The fan power per full load ton which minimum fan and pump costs occur are presented
drops from 0.0373 kW/ton to 0.0224 kW/ton. Finally, at full (Stoecker 1989). Lu and Cai et al. (2004) presented a model-
load, the chiller kW penalty will generally exceed savings based optimization strategy for the CW loop of centralized air-
from the pump and cooling tower. At part load, the savings conditioning systems. A modified generic algorithm for this
from the pump and cooling tower will catch up with and particular problem was proposed to obtain the optimal
outstrip the chiller penalty. Therefore, the total energy cost setpoints of the process. Simulations and experimental results
savings depend on the electric rate structure and chiller load on a centralized pilot plant showed that the operating cost of
profile. It is recognized that more research is needed to know the CW loop could be substantially reduced compared with
if increased fouling will result from reduced condenser water conventional operation strategies. As mentioned above, the
flow rates. cooling tower load equates condenser water flow multiplying
Some further discussions were made based upon condenser water return and supply temperature difference and
Kirsner’s example (York Inc.). It is confirmed that the first a coefficient of around 500. A lower condenser water flow rate
costs and energy costs cannot be reduced simultaneously and results in an increased condenser water supply temperature,
the goal should be maximizing energy-cost savings. This which, in turn, increases the temperature difference between
paper pointed out that the chiller kW penalty at the lower the condenser water inlet and ambient environment to meet the
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
©2012 ASHRAE
381
required heat rejection rate. Although it makes the chiller effect of some factors on the optimum condenser water flow
consume slightly more energy, the advantages of increased profile, such as chiller minimum condenser water entering
heat exchange efficiency of the cooling tower and decreased temperature, chiller load profile, weather, etc.
pump and fan power make the whole system operate more effi-
ciently. SYSTEM MODELING
The review above shows that the energy savings of vari-
A simplified chilled-water system example is employed
able condenser water flow depend on a lot of factors, such as
in this study. It consists of an induced-draft counter-flow evap-
equipment size, sensitivity of equipment performance to
orative cooling tower equipped with a VSD, a variable-speed
condenser water flow rate change, and chiller part-load ratios.
centrifugal CW pump, and a constant-speed centrifugal
Most previous studies are focused on comparing two different
chiller. Table 1 shows the parameters of each component.
CW flow rate scenarios. The equipment performance changes
Details can be found in the following sections.
due to CW flow reduction are analyzed using some general
rules or study results, and some general conclusions are made.
Chiller Modeling
For a specific chilled water system, some questions may be
raised, such as what the optimum variable-condenser water Because of its high weight factor and significant
flow rate is for this system, what the savings potential of vari- complexity, more effort should be made on chiller perfor-
able condenser water flows is compared with constant flow, mance modeling. In this study, a Gordon-Ng model for vapor
and what the optimum constant condenser water flow rate is. compression chillers with variable condenser flow is selected.
To answer these questions, a systematic simulation and anal- It can apply to unitary and large chillers operating under
ysis is required. In this paper, a simplified one-chiller system steady-state variable condenser flow conditions. This model is
was employed to search for the optimum condenser water flow strictly applicable to inlet guide vane capacity control (as
rates at various given conditions. This paper also studied the against cylinder unloading for reciprocating chillers, or VSD
Table 1. Parameters of the Example ChW System
Rated capacity Qrate 2250 ton 7913 kW
Rated power input Prate 1433 kW 1433 kW
Rated performance 0.637 kW/ton 5.5 COP
Chiller model coefficient 0 c0 –2.81E–01 –2.81E–01

Chiller model coefficient 1 c1 1.02E+01 1.02E+01
Chiller Chiller model coefficient 2 c2 1.74E+03 1.74E+03
Chiller model coefficient 3 c3 2.71E–03 2.71E–03

Rated CW flow Vcw 6,000 gpm 1362.7 m3/hr
Rated ChW temperature TChW,S 42 ºF 5.6 ºC
Chiller max CW enter temp. TCW,max 85 ºF 29.4 ºC
Chiller min CW enter temp. TCW,min 60 ºF 15.6 ºC
CT coefficient 1 c 3 3
CT coefficient 2 n –0.6 –0.6
Cooling Rated air flow at full speed Vair 550,000 cfm 260 m3/s
Tower Fan power at full speed Pfan 135 hp 101 kW
Air flow when fan is off Vair,off 60,000 cfm 28.3 m3/s
Approach setpoint T app sp 7.0 Delta ºF 3.9 Delta ºC
Rated pump flow Vpump 6700 gpm 1521.7 m3/hr
CW Pump Rated pump head Hcwp 53.7 ft 23.3 m
Tower height Hct 10.0 ft 3.0 m
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
382
ASHRAE Transactions
for centrifugal chillers) (Jiang and Reddy 2003). It is in the ularly, the overall number of transfer units (NTU) can be
following form: correlated with the following form:
y = c0 + c1 x1 + c2 x2 + c3 x3 1+n
 m· w
NTU = c  ------
· -  (4)
 -----------
1   ma 
T cho T cdi – T cho  COP- + 1 Q ChW
x 1 = -------------- x 2 = --------------------------- x 3 = ------------------------------------------
Q ChW Q ChW T cdi T cdi (1) The value of coefficient c is between 1.0 and 3.0 for
 -----------
1 - + 1 T  -----------
1  towers, and coefficient n ranges between –0.4 and –0.8
 COP  cho 1  COP- + 1 Q ChW (Kreider, Curtiss et al. 2002). If a typical value of n is assumed,
y = --------------------------------------- – 1 – ----------------------------------------- ------------------------------------------
T cdi  V CW per  w c pw  T cdi the value of c can be determined from air mass flow rate (ma)
and water mass flow rate (mw) at nominal design conditions.
where Tcho is chiller chilled water leaving temperature in Once c and n are known for a particular cooling tower, the
K, Tcdi is chiller condenser water entering temperature in K, cooling tower performance can be predicted at any operating
QChW is chiller cooling load in kWt, Vcw is condenser water condition given the water inlet temperature, the ambient air
flow rate in m3/s, ρw is water density in kg/m3, and cpw is water wet-bulb temperature, and the air and water flow rates.
thermal capacity in kJ/kg·K. Typical values of airflow from manufacturer literature
It is noted that the actual chiller condenser water flow is range from about 200 standard cubic feet per minute per ton
limited by the upper and lower limits of condenser flow rate. (scfm/ton) (97 m3/hr·kWt) to about 300 scfm/ton (145 m3/
The upper limit is intended to prevent erosion, and the lower hr·kWt). Typical values of fan horsepower for induced-draft
limit is to prevent fouling in the tubes. In standard condenser towers range from about 0.04 hp/ton to about 0.08 hp/ton
tubes of some new chiller models, the minimum water velocity (Stout 2003). For this particularly system, the air flow of the
can be as low as 1.5 ft/s (0.46 m/s) and as high as 12 ft/s (3.66 selected fan at full speed is 550,000 cfm (934,956 m3/hr). The
m/s). rated fan power is assumed at 135hp or 0.06 hp/ton. The power
of the fan at lower speed is a function of the fan speed ratio
Pump Modeling (xfan). Shelton and Weber (1991) report a correlation of fan
break horsepower to the air flow rate raised to the 3.2 power.
The general calculation equation of the CW pump power
In principle, one should calculate the fan power rigorously
is:
from the fan flow and pressure drop. In practice, the power of
 H- the VFD-equipped cooling tower fan can be calculated with
P pump = 0.746V
--------------------------- (2) the following regression curve:
3960 all
where  all is the overall efficiency including the efficiencies P fan 2 3

- = 0.1643x fan – 0.05735x fan + 0.9657x fan
-------------------- (5)
of pumps, motors, and VSDs. The pump flow is given, and the P fan rate
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
pump head can be calculated from the intersection point of the

pump head curve and the system hydraulic curve. At design Weather and Chiller Load
CW flow rate, this CW pumps are sized for approximately
53.7 feet (23.3 m) of head, or 0.064 kW/ton, with 10 ft (3.0 m) Weather conditions play a significant effect on the opti-
of static head from tower sprayers to the water surface in the mum operation of a water-cool chiller plant. It is assumed that
sump. The efficiencies of the CW pump and variable-speed this system is located in Houston, Texas, with a typical hot and
drive can be simulated as a function of pump flow rate or humid climate. The Typical Meteorological Year 3 (TMY3)
pump speed. A typical motor efficiency as a function of name- hourly weather data (NREL 2008) for Houston are used to
plate loading percentage for large size motors (>25 hp) is generate wet-bulb based bin data. In each bin, the hour number
given by ASHRAE (1996). Finally, the pump power ratio can is counted, and the average dry-bulb and wet-bulb tempera-
be regressed as a function of pump flow rate ratio (xcw) as tures in each bin are calculated. The chiller cooling load is a
follows: second-order polynomial function of the ambient wet-bulb
temperature. The chilled water flow is varied, while the return
P cwp 3 2 and supply chilled water temperatures are maintained at 42°F
- = 0.7264x cw + 0.0951x cw + 0.1094x cw + 0.0705 (3)
---------------------- (5.6°C) and 54°F (12.2°F), respectively. The condenser water
P cwp rate
flow rate can be varied from 1.0 gpm/ton (0.065 m3/hr·kWt) to
3.0 gpm/ton (0.194 m3/hr·kWt). The tower leaving water
Cooling Tower Modeling
temperature is maintained between 60°F (15.6°C) and 85°F
The mass and heat transfer process in a cooling tower is (29.4°C) with a constant approach setpoint of 7 delta°F (3.4
fairly complicated. The effectiveness-NTU model is the most delta°C). These data are used as inputs for the condenser water
popular model in cooling tower simulations, but iterations are loop simulation program. Figure 1 shows the profiles of dry-
required to obtain a converged solution (Braun 1989). Partic- bulb, wet-bulb, chiller load, and hour number in each bin.
©2012 ASHRAE
383
Figure 1 Profiles of DB, WB, chiller load, and hour number in each bin (IP/SI).
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 2 kW per ton versus chiller PLR for each component Figure 3 System total kW per ton versus chiller PLR at two
at 3.0 gpm/ton (IP and SI). different CW flow rates (IP and SI).
SIMULATION AND RESULTS Figure 3 compares the kW per ton profiles of the system
for two constant CW flow rate scenarios: 2.0 gpm/ton (0.129
Constant CW Flow m3/hr·kWt) and 3.0 gpm/ton (0.194 m3/hr·kWt). No signifi-
cant difference on the kW per ton can be observed when the
First of all, the scenarios with constant condenser water chiller is almost fully loaded. When the chiller is 74% loaded,
flow all year round are simulated and analyzed. Figure 2 shows the difference reaches its peak of 0.05 kW/ton. This indicates
the kW per ton profile for each component and system total as that, if the chiller is highly loaded most of time, the energy
a function of chiller part load ratio (PLR) when the CW flow savings potential will shrink dramatically.
rate is 3.0 gpm/ton (0.194 m3/hr·kWt). The chiller perfor- For this particular example system, Table 2 shows that the
mance reaches its lowest value at around 0.74 PLR. For the annual total power usage can decrease by 480,719 kWh or
CW pump, the kW per ton decreases approximately linearly 4.7% of the baseline power usage when the CW flow rate
with the chiller PLR because the absolute value of pump drops to 2.0 gpm/ton (0.129 m3/hr·kWt). Particularly, the
power is constant when the chiller load drops. As a reset pump power usage drops by 60.1%, chiller power usage
schedule is applied to the CT condenser water leaving temper- increases by 5.0%, and tower power usage drops by 37.6%. It
ature, the kW per ton profile for the tower increases from 0.027 is recognized that the majority of kWh savings comes from the
kW/ton to 0.068 kW/ton when the chiller PLR declines from CW water pump. The CW pump and the fan motors are all
1.0 to 0.74. After that point, the CT condenser water leaving equipped with variable-speed drives for this example system.
temperature is maintained at 60°F, and the fan speed reduces If the capacities of CW pump and cooling tower fan are
when the chiller PLR falls further. The chiller dominates the controlled by throttling valves, a simple analysis shows that
total power consumption of the system, the kW per ton profile the pump power savings dive to 138,968 kWh or 12.5% of the
of the system total is similar to that of the chiller. As chiller pump power baseline, the fan power savings shrink to 101,692
load declines, the system total kW per ton drops quickly from kWh or 9.4% of the fan power baseline. The total power
0.71 to 0.54 and then rises slightly to 0.55. increases by 1.7% of the total power baseline. As a result, if the
384
ASHRAE Transactions
Table 2. Energy Comparison for Two CW Flow Scenarios
Pump, kWh Chiller, kWh Cooling tower, kWh Total, kWh
3.0 gpm/ton 1,113,806 8,495,092 633,095 10,241,993
2.0 gpm/ton 443,899 8,922,593 394,782 9,761,275
Difference 669,907 –427,501 238,313 480,719
Savings percent 60.1% –5.0% 37.6% 4.7%
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 4 Annual average kW per ton for each component versus CW flow rate (IP/SI).
system is designed at 3.0 gpm/ton (0.194 m3/hr·kWt), it may The same analysis can be applied to other places with
not make economic sense to reduce the CW flow rate by throt- different climates. It is found that the optimum constant
tling valves. condenser water flow rate is 1.8 gpm/ton (0.116 m3/hr·kWt)
for Phoenix, 2.1 gpm/ton (0.135 m3/hr·kWt) for Miami, 1.8
Variable CW Flow gpm/ton (0.116 m3/hr·kWt) for Denver, 1.9 (0.123 m3/hr·kWt)
gpm/ton for Chicago, and 1.8 (0.116 m3/hr·kWt) gpm/ton for
For a system with VSD-equipped CW pumps and CT Los Angeles. This may indicate that the local climate has a
fans, it is possible to modulate the condenser water flow rate minor effect on the optimal condenser water flow rate.
to improve the system performance. Simulations and analysis
are performed on a variable condenser water system. The In this example, the chiller annual average performance
annual average kW per ton profiles for each component as well degrades only by 4.7% when the condenser water flow drops
as system total versus CW flow rate is shown in Figure 4. The from 3.0 gpm/ton (0.194 m3/hr·kWt) to 2.0 gpm/ton (0.129
curve labeled “Total new” corresponds to the total kW per ton m3/hr·kWt). This penalty is much lower than 8.0% to 9.6%
profile of a scenario with a better condenser water loop design. used by other researchers. The possible reasons are various,
The rated CW pump head can drop by 25% to 40.3 ft (17.5 m), such as the Gordon-Ng model might not accurately reflect this
and the rated tower motor power reduces by one third to 90 hp penalty, the model training data are flawed, or the chiller is a
or 0.04 hp/ton. The regression curves for the chiller and cool- special one. If the selected chiller is more sensitive to the
ing tower kW per ton as a function of CW flow rate are close condenser water flow rate as shown in some previous studies,
to a secondary-order polynomial. The cooling tower perfor- the chiller performance penalty could increase by 8.0% to
mance curve shows a linear correlation with the CW flow rate. 9.0% when the CW flow rate reduces. For a better condenser
The system total performance reaches its lowest value at water loop design, Figure 4 shows that such a new system
around 2.0 gpm/ton (0.129 m3/hr·kWt), which is lower than performs best at around 2.6 gpm/ton. If the CW flow rate drops
the rated value of 2.7 gpm/ton (0.174 m3/hr·kWt). If the CW from 3.0 gpm/ton (0.194 m3/hr·kWt) to 2.0 gpm/ton (0.129
flow rate decreases from 3.0 gpm/ton (0.194 m3/hr·kWt) to 2.0 m3/hr·kWt), the energy penalty is around 1.7% of the annual
gpm/ton (0.129 m3/hr·kWt), simulation results indicate that total baseline usage. This indicates that the optimum
the system total kWh will decline by 4.9%, and the chiller condenser water flow rate highly depends on the system
performance degrades by 4.7% when it is fully loaded. If the design as well as the sensitivity of chiller performance to the
cooling tower minimum leaving temperature is raised from condenser water flow rate. At around the optimal CW flow
60°F (15.6°C) to 70°F (21.1°C), the optimum CW water flow rate, the system annual total energy usage is not sensitive to the
rate is still around 2.0 gpm/ton (0.129 m3/hr·kWt). CW flow rate.
©2012 ASHRAE
385
Figure 5 Profile of optimal CW flow rate versus chiller part load ratio (IP/SI).
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
For a specific chiller PLR, an optimal CW flow rate can Arlington, VA: Air-Conditioning and Refrigeration
be found to minimize the system total power. Figure 5 shows Institute.
the optimal condenser water flow rate as a function a chiller ASHRAE. 1996. ASHRAE Handbook—HVAC Systems and
part load ratio. The optimal value varies between 1.3 gpm/ton Equipment Chapter 38 Centrifugal Pumps. Atlanta, GA:
(0.084 m3/hr·kWt) and 2.5 gpm/ton (0.161 m3/hr·kWt). Gener- American Society of Heating, Refrigerating and Air-
ally, a higher CW flow rate is preferred at a higher chiller PLR. Conditioning Engineers, Inc.
Compared to the operations with constant optimal 2.0 gpm/ton ASHRAE. 2008. ASHRAE Handbook—HVAC Systems and
(0.129 m3/hr·kWt) all year round, the operations with variable Equipment Chapter 39 Cooling Tower. Atlanta, GA:
optimal CW flow rates can only save the annual energy use by American Society of Heating, Refrigerating and Air-
0.4%. Consequently, it makes more sense to fix the CW flow Conditioning Engineers, Inc.
rate at an annually optimal value rather than to vary the Braun, J.E. 1989. Effectiveness models for cooling towers
condenser water flow from time to time. and cooling coils. ASHRAE Transactions 96(2): 164–
174.
SUMMARY AND CONCLUSION Hartman, T.B. 2001. All-variable speed centrifugal chiller
With broad application of VSDs, some people have plants. ASHRAE Journal 43(9): 43–52.
considered applying variable condenser water flow strategy in Jiang, W., and T.A. Reddy. 2003. Re-evaluation of the Gor-
a chilled water system. This paper studies the effect of variable don-Ng performance models for water-cooled chillers.
condenser water flow on the performance of a sample chilled ASHRAE Transactions 109(2): 272–287.
water system. Appropriate equipment models are selected for Kinter-Myer, M. and Emery, A. 1995. Cost optimal design
chillers, CW pumps, and cooling towers. The bin method is for cooling towers ASHRAE Journal 37(4): 46–55.
adopted to simulate the annual energy usage within moderate Kirsner, W. 1996. 3GPM/ton condenser water flow rate:
efforts. The simulation results show that, if pump water flow Does it waste energy? ASHRAE Journal 38:63–69.
and fan airflow are throttled by valves or dampers, the system Kreider, J.F., P.S. Curtiss, et al. 2002. Heating and cooling of
could consume more energy when reducing the condenser buildings: Design for efficiency 2nd edition. New York,
water flow. If the pump and cooling tower are equipped with McGraw-Hill.
VSDs, the optimum condenser water flow is highly dependent Lu, L., W. Cai, Y.C. Soh, L. Xie, and S. Li. 2004. HVAC sys-
on the design conditions and sensitivity of chiller performance tem optimization––Condenser water loop. Energy Con-
to the condenser water flow. The local climate plays has a version and Management 45(4): 613–630.
minor effect on the optimal condenser water flow rate. At NREL. 2008. National Solar Radiation Data Base 1991–
around the optimal CW flow rate, the system annual total 2005 Update: Typical Meteorological Year 3. From
energy usage is not sensitive to the CW flow rate. Compared http://rredc.nrel.gov/solar/old_data/nsrdb/1991-2005/
to operating with a constant optimal CW flow, the savings of tmy3/.
applying variable optimal CW flow are negligible. This Shelton, S., and Weber, E.D. 1991. Modeling and optimiza-
method can be applied to actual systems to find the optimal tion of commercial building chiller/cooling tower sys-
condenser water operating strategies. tems. ASHRAE Transactions 97(2):1209–1216.
Shelton, S., and Joyce, C. 1991. Cooling Tower Optimization
REFERENCES
for centrifugal chillers. ASHRAE Journal 33(6):28–36.
Soylemez, M.S. 2004. On the optimum performance of
ARI. 1998. ARI Standard 550/590, 1998 Standard for water forced draft counter flow cooling towers. Energy Con-
chilling packages using the vapor compression cycle. version and Management 45: 2335–2341.
386
ASHRAE Transactions
Stoecker, W.F. Design of thermal systems. 3rd ed. New York: York Inc. Reduced condenser-water flow rate: energy-saving
McGraw-Hill; 1989. miracle or mirage? From: http://www.york.com/prod-
Stout, M.R. 2003. Cooling tower fan control for energy effi- ucts/esg/updates/eng-Updates/22.pdf.
ciency. M.S. thesis, Department of Mechanical Engi- Waller, B. 1988. Various flow rates in condenser water cycle.
neering, North Carolina State University. ASHRAE Journal 30:30–34.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
©2012 ASHRAE
387
SA-12-005
Method for Estimating Energy Savings

Potential of Chilled-Water Plant
Retro-Commissioning
Zhiqin Zhang, PhD William D. Turner, PhD, PE
ABSTRACT crucial for the success of a retro-commissioning project.

Generally, these measures interact with each other and the
Chilled-water plants are major contributors of total
performance of chiller (CHLR) plant equipment is nonlin-
energy consumption for facilities. Various energy-efficiency
ear, making it difficult to conduct an accurate savings poten-
measures have been proposed to improve plant performance
in retro-commissioning projects. This paper introduces a tial estimation using simplified engineering calculations. At
method to estimate the savings potential of some popular the same time, building a detailed plant model for each proj-
measures by building a forward simulating model for a ect is time consuming. Sometimes it is even impractical due
chilled-water plant without storage. This model is based on to the lack of enough information or divergence of itera-
a wire-to-water efficiency concept to simulate the plant tions. In response to this challenge, this paper introduces a
power for producing cooling. The wire-to-water efficiency of method to estimate the savings potential of some popular
each type of equipment is calculated with selected models or measures by building a forward simulating model for a
equations. The fluctuations of the chilled-water loop supply chiller plant without storage.
and return temperature difference are also considered to A chiller plant produces ChW and transports it to end
reflect their impacts on the chiller loading and pumping users (such as air-handling units [AHUs]) through piping.
power. By finding the optimal reset schedule of some Central ChW plants are widely employed because they have
controlled variables, the utility billing cost of a chiller plant long been promoted as an energy-efficient and low-operating-
can be minimized. The variables to be optimized are cooling cost means of rejecting heat from building air-conditioning
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
tower approach temperature, chiller chilled-water leaving systems to the atmosphere. A large portion of the power
temperature, and chiller condenser water flow rate. This is a required to run a facility is consumed in the central chiller
nonlinear programming problem and can be solved with the plant. However, the energy performance of most existing
generalized reduced gradient nonlinear solver. The applica- chiller plants is not very efficient. It was estimated that about
tion of this method is illustrated with an example chilled- 90% of water-cooled, centrifugal central plants operated in the
water system. 1.0 to 1.2 kW per ton (2.9 to 3.5 coefficient of performance
[COP]) “needs improvement” range, while a highly efficient
INTRODUCTION plant can reach 0.75 kW per ton (4.7 COP) (Erpelding 2006).
Retro-commissioning has been widely used for central All kinds of problems are to blame, such as the low delta-T
chilled-water (ChW) plants to improve their performance. syndrome (Kirsner 1995), low chiller part-load ratio (PLR,
Some measures are very popular and effective, such as chW which equals chiller actual cooling load over design cooling
supply temperature reset, cooling tower (CT) approach capacity), significant mixing, valve and pump hunting, higher
temperature reset, secondary loop differential pressure (DP) than needed pump pressure, etc. In addition, there are other
reset, and condenser water (CW) flow rate reset. Accurately reasons for a lack of plant optimization, such as equipment
estimating the energy savings potential of each measure is performance degrading with age, load changes (Taylor 2006),
Zhiqin Zhang is a senior project engineer at Nexant Inc., San Francisco, CA, and William D. Turner is a professor in the Department of
Mechanical Engineering, Texas A&M University, College Station, TX.
388
ASHRAE Transactions
plant expansion in an unorganized manner, and energy cost The system-based method has been adopted by Ahn and
fluctuations. Today, more and more people are aware of Mitchell (2001) to find the influence of the controlled variables
sustainability and the importance of saving energy. The fluc- on the total system and component power consumption. A
tuating energy prices also make plant owners and operators try quadratic linear regression equation for predicting the total
various ways to enhance the performance of chiller plants so cooling system power, in terms of the controlled and uncon-
as to reduce the energy costs. trolled variables, was developed using simulated data collected
Automatic control systems have been widely applied in under different values of controlled and uncontrolled variables.
chiller plants to achieve robust, effective, and efficient operation The trade-off among the components of power consumption
of the system on the basis of ensuring thermal comfort of occu- resulted in the total system power use in that both simulated and
pants and satisfying indoor air quality. Generally, all the control predicted systems were minimized at lower supply air, higher
methods used in heating, ventilating, and air-conditioning ChW, and lower condenser-water temperature conditions.
(HVAC) systems can be divided into supervisory control and Bradford (1998) developed linear, neural network, and
relational control. Supervisory control, often named optimal quadratic system-based models and a component-based model
control, seeks stable and efficient operation by systematically to predict the system energy consumption including demand
choosing properly controlled variable setpoints, such as flow, side. It has been shown that the use of component-based models
pressure, or temperature. These setpoints can be reset when for either on-line or off-line optimal control is viable and robust.
uncontrolled variables (such as ambient air wet bulb [WB] Although the system-based plant model is much simpler
temperature, dry bulb [DB] temperature, and building cooling than the component-based model, the objective function under
load) are changed, and they are maintained by modulating each feasible combination of discrete control modes has to be
control variables (speed of components) through proportional generated, and considerable regression error as well as solu-
integral derivative (PID) controllers or sequencing. This method tion difficulty may exist. The component-based model is more
is easy to understand and is widely accepted in practice due to accurate, but it takes a longer time to build the model for each
the simplicity and effectiveness. The fundamentals of supervi- project. Iterations are inevitable and convergence could be a
sory control strategies have been comprehensively introduced problem. Some sophisticated algorithms are also required to
in ASHRAE Handbook—HVAC Applications (2003) and are optimize such a system.
widely applied in practice. Most of these controls originated
The objective of this paper is to develop a method to esti-
from the supervisory control methodology developed by Braun
mate the savings potential of some popular measures by build-
(1988). Relational control is determining continuous and
ing a forward simulating model for a ChW plant without
discrete control variables directly according to uncontrollable
storage. Its application is illustrated with an example system
variables or equipment power input, such as demand-based
to find the optimal reset schedule of the chiller ChW leaving
control (Hartman 2001) and load-based control (Yu and Chan
temperature, CW flow rate, and CT approach temperature.
2008). It was claimed by the authors that these controls could
The energy and billing cost savings potential of several energy
realize tremendous energy savings.
conservation measures can also be estimated.
The chiller plant models can be divided into component-
based and system-based. The component-based model simu- METHOD
lates plant performance by modeling each plant component
individually. Sometimes, it takes too much effort to build a A central ChW plant consists of cooling towers, CW
component-based model or the necessary data may not be pumps, chillers, and primary and secondary ChW pumps.
available. An alternative way to simulate plant performance is Figure 1 shows the simplified general physical configuration
to simulate the plant power with one function. This methodol- of a primary-secondary loop ChW system. Most valves and
ogy was first advanced by Braun et al. (1989) when they devel- fittings are omitted for simplicity purposes. All the variables
oped a system-based optimization based on results from shown in Figure 1 are setpoints that could be optimized by
component-based optimization. The method involves correlat- implementing reset schedules. In practice, these setpoints are
ing overall cooling plant power consumption using a quadratic maintained by adjusting the equipment speed or control valve
function form. The inputs are uncontrolled variables and position with a PID controller. Except for continuous
controlled continuous variables while the outputs are total cost. controlled variables, discrete control variables will also need
Separate cost functions are necessary for each operating mode. to be optimized, such as the sequencing of chillers, cooling
Minimizing this function leads to linear control laws for towers, and pumps. The constraints on the equipment opera-
controlled continuous variables in terms of uncontrolled vari- tions, such as maximum and minimum flow rates, limit the
ables. The empirical coefficients of this function depend on the number of feasible combinations of control variables. For
operating modes so that these constants must be determined for example, chillers have lower and upper limits on the evapora-
each feasible combination of discrete control modes. The tor flow rate. The number of chillers running simultaneously
determined controlled variables, such as ChW flow rate, will be should ensure the actual evaporator flow be within this range.
maintained by modulating continuous control variables, such This limits the possible number of chillers staged on simulta-
as valve open percentage and motor speed. neously.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
©2012 ASHRAE
389
Figure 1 Configuration of a primary-secondary loop ChW
system.
Plant Power Modeling
The cooling system total power can be divided into plant

power and non-plant power. The electricity consumed by
ChW production is considered plant power while all other
electricity consumptions (such as plant AHU fans, lighting,
and plug loads) are non-plant power. Figure 2 Flow chart of ChW plant power simulation.
Figure 2 is a flow chart of the ChW plant power simu-
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
lation model. All the variables on the left are the inputs, while
the output is the plant total power. The plant model deter- Some regression formulas, together with energy conser-
mines the plant total power consumption in response to a set vation laws, are used to simulate the WTW efficiency of
of external parameters and a set of plant parameters. The pumps and CTs. This forward plant model can be easily set up
driving force of chiller plant operations is the ChW flow and used for plant energy simulation. Since it is based on
demanded by users on the secondary loop. This demand basic physical definitions and conservation laws, it has an
cannot be controlled by operators. For each given plant total explicit physical meaning. Its application is not restricted by
ChW flow demand, the ChW plant model will export the total the equipment number and sequencing strategies. All calcu-
plant power under the given conditions. In this study, an lations are explicit expressions and no iterations are required.
equipment performance-oriented plant model is proposed to One prerequisite is that the pumps are well sequenced and
calculate the plant power under predefined conditions. This controlled so that the pump flow rate is within the normal
model is based on a wire-to-water (WTW) plant efficiency operation range.
concept. The WTW efficiency of a pump was first introduced In most cases, the optimization target of a chiller plant
by Bernier and Bourret (1999). It was originally used to operation is to minimize the operating cost within a billing
quantify the whole performance of a ChW plant. In this period, such as a year. This target is achieved by implementing
study, it is used to define the transportation efficiency of the optimal reset schedules of some controlled variables, such
plant equipment except for secondary pumps (SPMPs). The as ChW supply temperature, CW flow rate, and CT approach
power of SPMPs is determined by the secondary loop ChW temperature. This is a nonlinear programming (NLP) problem
flow rate, pumping system total efficiency, and pressure drop and can be solved with the generalized reduced gradient
of the secondary distribution system including pipe friction (GRG) nonlinear solver. This method and specific implemen-
and valve losses. tation have been proven in use over many years as one of the
The system total power can be calculated from the follow- most robust and reliable approaches to solve difficult NLP
ing formula: problems.
Chiller Modeling
P sys =   CT +  CWP +  CHLR +  PPMP Q Plant ChW
(1) This paper selected the Gordon-Ng model (Gordon and Ng
+ P SPMP + P non_plant 2000) to simulate chiller performance. This model is valid for
390
ASHRAE Transactions
vapor-compression chillers with variable condenser flow. It can Pump Modeling
apply to unitary and large chillers operating under steady-state The general calculation formula of the ChW pump
variable condenser flow conditions. This model is strictly appli- power is:
cable to inlet guide vane capacity control (as opposed to cylinder
unloading for reciprocating chillers or variable-speed drive 0.746V  H
P pump = -------------------------- (4)
[VSD] for centrifugal chillers) (Jiang and Reddy 2003). It is in 3960 all
the following form:
where all is the overall efficiency including pumps, motors,
y = c0 + c1 x1 + c2 x2 + c3 x3 (2)
and VSDs.
where The total cooling transported by the water pump is
approximately calculated by:
 -----------
1
- + 1 Q ChW
T cho T cdi – T cho  COP  V ChW T ChW
x 1 = --------------- , x 2 = ---------------------------, x 3 = -------------------------------------------, Q ChW = -------------------------------
- (5)
Q ChW Q ChW T cdi T cdi 24
The pump WTW efficiency is:
and
P pump 0.746V  H 24
- = --------------------------  --------------------------------
 pump = -------------- (6)
Q ChW 3960 all V ChW T ChW
 -----------
1
- + 1 T cho For primary pumps (PPMPs), V = VChW. The WTW effi-
 COP 
y = ---------------------------------------- ciency is:
T cdi
H PPMP
 -----------  PPMP = 0.004521 ------------------------------------ (7)
- + 1 Q ChW
1  PPMP T ChW
1  COP 
where – 1 – -------------------------------------------- -------------------------------------------
 V  per  w c pw  T cdi For condenser water pumps (CWPs), V = Npump · VCW, per .
T cho =  TCW CHLR ChW S – 32   5  9 + 273.15 , Consequently, the CW pump WTW efficiency is:
T cdi =  T CHLR CW S – 32   5  9 + 273.15 ,
H CWP V CW per
 CWP = 0.0001884 --------------- ------------------------ (8)
12,000Q CHLR per  CWP Q ChW per
Q ChW
Q ChW = --------------------------------------------
- , COP = ----------------
-,
3412 P CHLR For SPMPs, the pump power is:
0.746V Lp_ChW  2.31   DP Lp + eV Lp_ChW 2 
and P SPMP = --------------------------------------------------------------------------------------------------------------
3960 SPMP
V CW = 0.00006309V CHLR CW per (9)
The head and efficiency of pumps can be simulated as func-
tions of pump flow rate or be constant. Obviously, the energy
The chiller WTW efficiency (kW per ton) is: consumption of SPMPs is subject to the loop-side operation
and is not determined by plant operations.
P CHLR
 CHLR = -------------------------------
- =
3412 Cooling Tower Modeling
Q ChW ----------------
12,000
The mass and heat transfer process in a cooling tower is
  fairly complicated. The effectiveness-NTU model is the
  (3)
  c 0 + c 1 x 1 + c 2 x 2 + 1 T cdi  most popular model in CT simulations, but iterations are
3.517  ---------------------------------------------------------------------------------------- – 1 required to obtain a converged solution (Braun 1989). To
 Q ChW 
 – c 3 Q ChW + T cho – --------------------------------  overcome this obstacle, a simple CT fan power model is
  V CW  w c pw   proposed to calculate the tower WTW performance:
It is recognized that the actual chiller ChW flow is also

limited by the upper and lower limits of evaporator ChW flow P CT  d2 
 CT = -------------- -  1 + 0.2843 CHLR  (10)
- =  d 1 + ---------------
rate. The upper limit is intended to prevent erosion and the Q ChW  T App
lower limit is to prevent freezing in the tubes.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
©2012 ASHRAE
391
The actual CT approach temperature is obtained from the APPLICATION
following formula:
Example System Description
T App = T CT CW R – T wb (11) The example system investigated in this study is a
central utility plant called Energy Plaza (EP) located in the
Dallas-Fort Worth metropolitan area. The EP consists of six
The coefficients d1 and d2 are regressed from the trended
5500 ton (19,343 kWt) constant-speed centrifugal chillers,
data, and TApp = TCT,CW,R – Twb is maintained by sequencing
called on-site manufactured (OM) chillers; one 90,000
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
the cooling towers or modulating fan speed.
ton·h (316,517 kW t·h) naturally stratified ChW storage
tank; six 150 hp constant-speed primary pumps; four
Loop Delta-T
450 hp variable-speed secondary pumps; eight 400 hp CW
The loop ChW delta-T is subject to many factors, such as pumps; and eight 150 hp two-speed cooling towers.
chiller ChW leaving temperature, cooling coil air leaving Figure 3 shows a schematic diagram of this ChW system.
temperature, type of flow control valves, coil design parame- This study only deals with the ChW system operated with-
ters and degrading due to fouling, tertiary connection types, out ChW thermal storage.
coil cooling load, air economizers, etc. Considering the diffi- The monthly electricity billing cost consists of a meter
culties in developing a physical model to simulate the loop charge, current month non-coincident peak (NCP) demand
delta-T, a linear model regressed from the trended data is used charge, four coincident peak (4CP) demand charge, and
in this study. energy consumption charge. The total monthly electricity
billing charge (CTotal) is:
n
T Lp = C Total = C delivery_point + R 4CP D 4CP
 hi xi + h0 (12)
(13)
i=1 + R NCP D NCP + R energy E consumption
where xi are the variables that could be the dominant factors of

The rates R4CP , RNCP , and Renergy for each month are subject
the loop delta-T model, such as ChW supply temperature,
to minor adjustments, and the rates from a selected one-year
loop total cooling load, ambient DB and WB temperatures,
period are used in the simulation. The meter charge
hour of the day, weekday or weekend, and month. The air
Cdelivery_point is constant for each month. All demand kilo-
system side parameters, such as coil air leaving temperature,
watts used have been adjusted to 95% power factor. The
total airflow rate, coil design delta-T, and sensible load ratio,
monthly average power factors during this period will be
are not included due to diversity or unpredictability. used in the power factor correction.
The exact form of the regression model may vary for The 4CP demand kilowatt is the average of the plant’s
different projects. It could be necessary to build different integrated 15 minute demands at the time of the monthly Elec-
models to accommodate air-conditioning system operation tric Reliability Council of Texas (ERCOT) system 15 minute
changes at different seasons. A constant delta-T can be used in peak demand for the months of June, July, August, and
a rough, first-order simulation. September (called summer months) of the previous calendar
Figure 3 Schematic diagram of the EP ChW system.
392
ASHRAE Transactions
year. The exact time will be announced by ERCOT. The plant’s To check the accuracy of the system model, the EP
average 4CP demand will be updated January 1 of each calen- monthly utility bills are compared with the simulation
dar year and remain fixed throughout the calendar year. The results. A good match is found, although minor differences
NCP kilowatt applicable shall be the kilowatt supplied during exist in several months. This could be attributed to imperfec-
the 15 minute period of maximum use during the billing tion of models, inaccurate parameter inputs, or operations
month. The current month NCP demand kilowatt shall be the different from actual situations. The present system power
higher of the NCP kilowatt for the current billing month or model can reasonably predict the monthly electricity
80% of the highest monthly NCP kilowatt established in the consumption.
eleven months preceding the current billing month. Loop-Side Modeling. The parameters and inputs for the
The ChW flow through the chiller evaporator is controlled thermal energy storage (TES) system loop side are shown in
by modulating flow control valves on the leaving side of the Table 1. The upper and lower limits of the loop end DP, as well
evaporator. The sequencing of the constant-speed PPMPs is as the loop flow rate change points, are subject to hydraulic
dedicated to the corresponding chillers. The variable- requirements and operating experiences. If the loop total ChW
frequency device speed of the SPMPs is modulated to main- flow rate is equal to or lower than 10,000 gallons per minute
tain the average of differential pressures at the loop ends at a (gpm) (0.631 m3/s), the DP setpoint is 22.0 psid (151.7 kPa).
given setpoint. This setpoint is manually adjusted to be If the rate is equal to or higher than 16,000 gpm (1.009 m3/s),
between 25 and 48 psid (172.4 and 330.9 kPa) all year round the DP setpoint is 28.0 psid (193.1 kPa). The ChW secondary
to ensure there are no hot complaints. The cooling tower stag- DP setpoint is reset linearly from 22 to 28 psi (151.7 to
ing control in place is to stage the number of fans and select 193.1 kPa), when the secondary ChW flow is between 10,000
high and low speeds of fans to minimize the chiller compressor and 16,000 gpm (0.631 and 1.009 m3/s). A loop load factor is
plus CT fan electricity consumption. defined to test the savings when the actual load profile is
different from the one used in the simulation.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
System Modeling A temperature rise exists between the loop supply

temperature and the chiller ChW leaving temperature, which
System Power. The trended historical data are used for is due to pumping heat gain and piping heat losses. The
system modeling. The electricity consumed by OM chillers, trended data show that the temperature rise fluctuates between
CT fans, CW pumps, PPMPs, and SPMPs is considered plant 0.0°F and 2.0°F (0°C and 1.1°C) most of time and the annual
ChW electricity load while all other electricity consump- average temperature rise is 1.0°F (0.6°C).
tions, such as EP air-conditioning, air compressors, lighting, When the loop end DP setpoints are determined, a loop
and plug loads, are non-plant electricity loads. hydraulic coefficient is required to calculate the differential
Table 1. Parameters of TES System Loop Side
LP end DP upper setpoint DPh 28.0 psid 1.009 kPa
LP end DP lower setpoint DPl 22.0 psid 151.7 kPa

LP end DP upper shift flow Vupper 16,000 gpm 3634 m3/s
Loop (LP) Hydraulic LP end DP lower shift flow Vlower 10,000 gpm 0.631 m3/s
LP hydraulic coefficient 1 e1 1.00E-07 1.0E-07
LP supply temperature rise Ts 1.0 Delta°F 0.6 Delta°C
LP DT coefficient 0 h0 32.1898 32.1898
LP DT coefficient 1 (TLP,ChW,S) h1 –0.5439 –0.5439

Loop Delta-T
LP DT coefficient 2 (QLP,ChW) h2 6.86E-05 6.86E-05
(LP DT)
LP DT coefficient 3 (Twb) h3 6.34E-02 6.34E-02
LP max DT TLp,max 22.0 Delta°F 12.2 Delta°C
LP min DT TLp,min 12.0 Delta°F 6.7 Delta°C
©2012 ASHRAE
393
pressure before and after the SPMPs. Three hydraulic coeffi- assumed that all pumps are sequenced reasonably to ensure
cients are regressed from trended data corresponding to one, that the running pumps are operated around the design points.
two, or three SPMPs running scenarios. The coefficients can The CT coefficients are obtained from the regression
be regressed from a plot of piping DP losses versus loop total results of the historical data. The CT model fitting curve is
flow rate. shown in Figure 5. It should be noted that the coefficients
Equation 14 is a linear regression model developed to obtained from the trended historical data are only applicable
simulate the loop delta-T as a function of ChW loop supply to the current cooling tower operation strategy. If a new CT
temperature (x1), loop total cooling load (x2), and ambient WB operation strategy is used, the coefficients are subject to
temperature (x3). A higher loop supply temperature, a lower adjustment.
WB temperature, and a lower loop total ChW load lead to a The coefficients of the Gordon-Ng chiller model are
lower loop delta-T, which is consistent with the observations. obtained by regressing with the trended historical data of the
An upper and a lower limit are defined to avoid unreasonable OM chillers. The rated CW flow rate is equal to the average
regression results when an extrapolation is applied. The of the trended data. In this study, the total available chiller
system error of the loop delta-T can be used to check the effect number is limited to six. The chiller ChW leaving temper-
of loop delta-T prediction deviations on the system total ature default setpoint is 36°F (2.2°C). The ChW flow rate
energy and costs. limits and CW entering temperature limits are based on the
chiller design specifications. Figure 6 is a comparison
T Lp = h 0 + h 1 x 1 + h 2 x 2 + h 3 x 3 (14) between measured and predicted motor power using the
Gordon-Ng model.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 4 is a comparison of the measured and predicted Non-Plant Power Modeling. The non-plant power is
ChW supply and return temperatures. If the model accurately composed of two segments. When the ambient DB temper-
fits the data on which it was trained, this type of evaluation is ature is lower than 60ºF (15.6ºC), the non-plant power is
referred to as internal predictive ability. The external predic- 750 kW constant. Otherwise, a second-order polynomial is
tive ability of a model is to use a portion of the available data used to calculate the total non-plant power contributed by
set for model calibration, while the remaining data are used to plant HVAC, glycol cooling systems, air compressors, etc.
evaluate the predictive accuracy. The root mean square errors The coefficients are shown in Table 3, and Equation 15
(RMSEs) of the internal and external predictions are both shows the mathematical form of the regression model.
1.1°F (0.63°C). The coefficient of variations (CVs) of the
internal and external predictions are 6.86% and 6.93%, P non – plant =
respectively.
 750 whenDB  60F  15.6C   (15)
Plant-Side Modeling. Table 2 shows the main parame-
 g + g T + g T2 
ters and inputs for the plant side. The efficiencies of all pumps  1 2 DB 3 DB whenDB  60F  15.6C  
are assumed constant or determined from pump efficiency
curves. The overall efficiency is a product of motor efficiency,
Simulation and Results
shaft efficiency, and pump efficiency (and variable-frequency
drive efficiency for SPMPs). The pump heads are determined The selected variables to be optimized are chiller ChW
from pump head curves. The primary side flow rate is leaving temperature, CT approach temperature, and chiller
controlled to be equal to the secondary side flow rate. It is CW flow rate. The default and upper and lower limits of each
(a) (b)
Figure 4 Comparison of measured and predicted loop delta-Ts: (a) I-P and (b) SI.
394
ASHRAE Transactions
Table 2. Parameters of ChW System Plant Side
SPMP overall efficiency spmp 75% 75%

SPMP
SPMP design flow rate Vspmp 8,000 gpm 0.505 m3/s
PPMP overall efficiency ppmp 80% 80%

PPMP
PPMP head Hppmp 80 ft 239.1 kPa
CHLR coefficient 0 c0 –2.81E-01 –2.81E-01
CHLR coefficient 1 c1 1.02E+01 1.02E+01
CHLR coefficient 2 c2 1.74E+03 1.74E+03
CHLR coefficient 3 c3 2.71E-03 2.71E-03
CHLR condenser water flow Vcw 10,300 gpm 0.650 m3/s
CHLR ChW leaving temperature TChW,S 36°F 2.2°C
CHLR ChW low limit Vchw,min 4,000 gpm 0.252 m3/s
CHLR ChW high limit Vchw,max 7,400 gpm 0.467 m3/s
Motor max power input Pmtr,max 3,933 kW 3,933 kW
Max CW entering temperature TCW,max 83.0°F 28.3°C
Min CW entering temperature TCW,min 60.0°F 15.6°C
CT coefficient 1 d1 0.01 0.01
CT CT coefficient 2 d2 0.16 0.16
Approach default setpoint Tapp,sp 6.0°F 3.3 Delta°C
Pump head Hcwp 92 ft 275.0 kPa

CWP
Pump overall efficiency cwp 82% 82%
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
(a) (b)
Figure 5 Cooling tower regression model: (a) I-P and (b) SI.
©2012 ASHRAE
395
variable are shown in Table 4. Different default values are used (0.6°C to 2.8°C), the monthly total electricity consumption
to test their effects on the optimal solutions found by the of the plant reduces by 82,242 kWh. Particularly, the chiller
solver. The monthly optimal setpoints, as well as the energy electricity consumption reduces by 1.1% of the chiller
and utility billing cost savings, are shown in Table 5.
monthly consumption due to a higher chiller efficiency. This
In the winter months, the loop cooling load and loop is consistent with the rule of thumb that for each 1°F (0.6°C)
delta-T are both pretty low. Running chillers are lightly increase in ChW temperature, constant-speed chiller effi-
loaded most of time and the evaporator flow rate has reached
ciency increases 1.0% to 2.0%. The SPMPs consume 4.3%
the upper limit. A lower ChW leaving temperature is preferred
to increase the loop delta-T, which benefits the part-load ratio more of the monthly electricity usage due to a higher ChW
(PLR) of the chiller. This may sound a little counterintuitive, flow rate as a result of a lower loop delta-T. The minor
as most plants will increase supply ChW temperature to changes to cooling towers, condenser water pumps, and
improve chiller performance. When the efficiency improve- primary pumps are due to chiller staging change.
ment due to higher PLR dominates the degrading due to lower
supply temperature, we will see improved overall chiller effi- If the cooling tower approach decreases by 1.0°F
ciency. This phenomenon has been observed for this project. (0.6°C), the chiller will consume 0.9% less electricity due to
However, it should be recognized that this is a case-by-case a lower CW entering temperature. However, the cooling
result and cannot be generalized to other systems. In the tower fan will consume 12.2% more electricity as it needs to
summer months when the loop cooling load is high, the chiller run harder to approach closer to the ambient WB tempera-
can still be loaded up to the optimal PLR; even the ChW
ture. As a result, chillers consume 78,075 kWh less and tower
supply temperature rises to 44°F. A higher ChW leaving
temperature can improve the chiller efficiency. The optimal fans consume 63,705 kWh more. Overall, this measure can
CT approach temperature is around 4.6°F (2.6°C) and the still save 23,414 kWh in August. This result is also consistent
optimal chiller condenser water flow rate is 11,000 gpm with the annual optimizations that the optimal approach is
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
(0.694 m3/s) all year round. Compared to the scenario with around 4.6°F (2.6°C).
default setpoints, the optimal reset schedule can reduce elec-
If the chiller CW flow rate increases to 11,000 gpm
tricity energy consumption by 2,559,426 kWh per year and
reduce electricity billing costs by $261,387.00 (U.S.) per (0.694 m3/s), the chiller efficiency is improved due to
year or 3.3% of the baseline annual electricity costs. enhanced heat transfer in the condenser. The kilowatt–hour
As the change of each variable may place different savings from chillers are 70,625 kWh. For this particular proj-
impacts on the performance of each component, a single- ect, the CW pumps consume even less energy when the flow
variable sensitivity study is performed using August as an
example. Table 6 shows the electricity consumption change Table 3. Parameters of Non-Plant Power
if the setpoints are different from the default values. The
percentage is the change of monthly kilowatt–hour Coefficient 1 g1 1266.3 1266.3
consumption for each corresponding component. If the
chiller ChW leaving temperature increases by 1.0°F to 36°F Coefficient 2 g2 –4.4327 –4.4327
Non-
Plant Coefficient 3 g3 0.1983 0.1983
Power
Winter shift DB Twb,shift 60°F 15.6°C
Winter base power Pw,base 750 kW 750 kW
Table 4. Plant Optimization Controlled Variables
Lower Upper
Variable Unit Default Value
Limit Limit
°F 36 36 44
TChW,S
°C 2.2 2.2 6.7
°F 6 3 10
TApp
°C 3.3 1.7 5.6
gpm 10,300 11,000

Figure 6 Comparison of OM chiller measured and Vcw
predicted motor power. m3/s 0.650 0.568 0.694
396
ASHRAE Transactions
Table 5. Monthly Results of Plant Optimization
Energy Billing Cost Cost Savings

Month TChW,S TApp Vcw
Savings Savings Percentage
°F °C °F °C gpm m3/s kWh $
1 36.0 2.2 4.6 2.6 11,000 0.694 27,093 $6912 1.9%
2 36.0 2.2 4.6 2.6 11,000 0.694 30,229 $7159 1.9%
3 39.8 4.3 4.6 2.6 11,000 0.694 138,451 $16,132 2.8%
4 40.2 4.6 4.6 2.6 11,000 0.694 136,701 $15,563 2.8%
5 39.8 4.3 4.6 2.6 11,000 0.694 233,601 $23,172 3.1%
6 42.3 5.7 4.6 2.6 11,000 0.694 377,957 $34,289 3.9%
7 42.9 6.1 4.6 2.6 11,000 0.694 399,130 $36,903 3.9%
8 44.0 6.7 4.7 2.6 11,000 0.694 567,537 $51,490 4.9%
9 42.0 5.6 4.6 2.6 11,000 0.694 354,537 $32,368 3.8%
10 41.0 5.0 4.6 2.6 11,000 0.694 214,461 $21,568 3.2%
11 37.0 2.8 4.6 2.6 11,000 0.694 47,094 $8483 1.7%
12 36.0 2.2 4.6 2.6 11,000 0.694 32,634 $7347 1.9%
Total 2,559,426 $261,387 3.3%
Table 6. Results of Single-Variable Sensitivity Study
kWh Usage Change Tchw = 37.0°F (2.8°C) TApp = 5.0°F (2.8°C) Vcw = 11,000 gpm (0.694 m3/s)
CHLR 1.1% –88,228 –0.9% –78,075 –0.9% –70,625
CT –0.1% –284 12.2% 63,705 –0.1% –755
CWP –1.0% –6596 –1.0% –6596 –2.8% –19,564
PPMP –0.3% –1028 –0.8% –2449 –0.8% –2449

SPMP 4.3% 13,894 0% 0 0% 0
Total (kWh) –82,242 –23,414 –93,393
rate increases. This is associated with the particular perfor- annual utility costs can be reduced by 4.9%. Generally, the
mance curves of the selected CW pumps and may not be valid temperature of the ChW reaching the end users is higher than
for other systems. that produced by the chiller. This could be due to piping heat
This model is also used to test the energy saving sensitiv- losses, mixing, or pumping power. In the baseline case, the
ity of several energy conservation measures, such as reducing loop supply temperature rise is assumed to be 1.0°F (0.6°C).
the loop end DP setpoint upper limit (DPh), the loop cooling If the temperature rise can be reduced by half, 1,765,763 kWh
demand, and the loop ChW supply temperature rise (Ts). can be saved each year and the annual utility costs savings is
Table 7 shows the annual billing cost savings potential of each 2.0%. These results can be used to estimate the payback of
measure. It is estimated that, when the loop DP upper limit each measure.
decreases by 2.0 psid (13.8 kPa), the annual electric usage of
SPMPS can be reduced by 80,534 kWh or 3.7% of the SPMP SUMMARY AND CONCLUSION
power consumption. The utility cost decreases by $7487 per A chilled-water plant is a high-energy-density facility.
year or 0.1% of system annual total utility costs. If the cool- However, the energy performance of most existing chilled-
ing demand by the end users can be cut by 5%, the electric water plants is not very efficient. With volatile energy prices,
consumption can be saved by 3,467,316 kWh per year and the improving the plant efficiency to save utility costs is an urgent
©2012 ASHRAE 397

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---

Table 7. Billing Cost Savings Estimation of Energy Conservation Measures
Variable Unit Default Value Lower Value Savings, kWh Savings, $ Savings, %
psid 28 26
DPh 80,534 $7487 3.7%
kPa 193.1 179.3
fLp,cooling – 1.00 0.95 3,467,316 $302,236 4.9%
°F 1.0 0.5
Ts 1,765,763 $156,382 2.0%
°C 0.6 0.3
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
task for chiller plant owners and operators. Some effective N = number
retro-commissioning measures have been widely adopted to P = power, kW
reduce plant energy usage by 3% to 5%. Estimating the energy Q = cooling load, ton (kW)
savings potential of these measures is crucial for the success
R = electricity energy or demand rate, $/kWh or $/kW
of a retrocommissioning project.
Modeling and simulating is a popular method to optimize T = temperature, °F (°C)
the plant operation and estimate the savings potential of vari- V = flow rate, gpm (m3/s)
ous measures. The system-based model is simple but not accu- x = independent variables
rate, while the component-based model is accurate but  = temperature difference, °F (°C)
complicated. This paper proposes a forward plant model based
on a wire-to-water efficiency concept. The wire-to-water effi- Greek Symbols
ciency of each type of equipment is calculated with selected  = efficiency
models or equations. This is a nonlinear programming prob-
 = wire-to-water efficiency, kW/ton
lem and can be solved with the generalized reduced gradient
nonlinear solver. This forward plant model can be easily set up  = density, kg/m3
and used for plant energy simulation. Its application is not
Subscripts
restricted by the equipment number or sequencing strategies.
All calculations are explicit expressions and no iterations are App = approach
required. It is also easy to implement this model for different cdi = condenser water inlet
systems. The coefficients of each equipment model as well as cho = chilled-water outlet
the upper and lower limits should be updated to reflect the d = demand
performance of targeted systems.
db = dry bulb
The application of this method is illustrated with an exam-
ple system. The chilled-water system is modeled and three e = energy
variables are selected for optimization. Compared to the base- Lp = loop
line, 3.3% of the annual total utility costs are saved by imple- max = maximum
menting the new reset schedules of the controlled variables. A min = minimum
single-variable sensitivity study shows that the reset of each mtr = motor
variable can affect the energy consumption of several types of
R = return
equipment. The saving potentials of reducing loop end dew-
point setpoint upper limit, loop cooling demand, and loop S = supply
chilled-water supply temperature rise are also analyzed and sys = system
presented. sp = setpoint
t = thermal
NOMENCLATURE wb = wet bulb
C = cost, $
Cp = water heat capacity, kJ/kg·K (Btu/lbm · °F) REFERENCES
d = cooling tower model coefficients Ahn, B.C., and J.W. Mitchell. 2001. Optimal control devel-
e = loop hydraulic performance coefficient opment for chilled-water plants using a quadratic rep-
resentation. Energy and Buildings 33(4):371–78.
g = non-plant power model coefficients ASHRAE. 2003. ASHRAE Handbook—HVAC Applica-
h = loop delta-T model coefficient tions, Chapter 41. Atlanta: American Society of Heat-
H = water head, ft (m) ing, Refrigerating and Air-Conditioning Engineers.
398
ASHRAE Transactions
Bernier, M.A., and B.Bourret. 1999. Pumping energy and Erpelding, B. 2006. Ultra efficient all-variable-speed
variable frequency drives. ASHRAE Journal chilled-water plants. Heating Piping Air Conditioning
41(12):37–40. 78(3):35–43.\
Bradford, J.D. 1998. Optimal supervisory control of cool- Gordon, J.M. and K.C. Ng. 2000. Cool Thermodynamics.
ing plants without storage. PhD Dissertation, Depart- Cambridge, UK: Cambridge International Science
ment of Civil, Environmental and Architectural Publishing.
Engineering, University of Colorado, Boulder, CO.
Braun, J.E. 1988. Methodologies for the design and control Hartman, T.B. 2001. Ultra-efficient cooling with demand-
of central cooling plants. PhD Dissertation, Depart- based control. HPAC Engineering 73(12):29–35.
ment of Mechanical Engineering, University of Wis- Jiang, W., and T.A. Reddy. 2003. Re-evaluation of the Gor-
consin–Madison, Madison, WI. don-Ng performance models for water-cooled chillers.
Braun, J.E. 1989. Effectiveness models for cooling towers ASHRAE Transactions 109(2):272–87.
and cooling coils. ASHRAE Transactions 96(2):164–74. Kirsner, W. 1995. Troubleshooting chilled-water distribu-
Braun, J.E., J.W. Mitchell, S.A. Klein, and W.A. Beckman. tion problems at the NASA Johnson Space Center.
1987. Performance and control characteristics of a Heating Piping Air Conditioning 67(2):51–59.
large central cooling system. ASHRAE Transactions
93(1):1830–52. Taylor, S.T. 2006. chilled-water plant retrofit—A case
Braun, J.E., S.A. Klein, J.W. Mitchell, and W.A. Beckman. study. ASHRAE Transactions 112(2):187–97.
1989. Methodologies for optimal control to chilled- Yu, F.W., and K.T. Chan. 2008. Optimization of water-
water systems without storage. ASHRAE Transactions cooled chiller system with load-based speed control.
95(1):652–62. Applied Energy 85(10):931–50.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
©2012 ASHRAE
399
SA-12-006
ASHRAE/IES Standard 90.1 Metal Building

U-factors—Part 5: Mathematical Modeling of
Wall Assemblies and Validation by
Calibrated Hot Box Measurements
Manoj K. Choudhary, PE Chris P. Kasprzak David E. Musick
Mark J. Henry N. Douglass Fast
ABSTRACT INTRODUCTION
A three-dimensional mathematical model developed The work described in this paper is a part of a compre-
earlier by Choudhary et al. (2010) was used to investigate heat hensive project to develop an experimentally validated, three-
transfer in two metal building wall insulation assemblies. The dimensional mathematical model for heat transfer in metal
model predictions for the overall heat transfer coefficients (the building roof and wall assemblies that contain fiberglass as the
U-factors) were tested with measurements from calibrated hot main insulating component. In a previous publication (Choud-
boxes for these assemblies. One of the hot boxes consisted of hary et al. 2010), the mathematical model and its validation
a single layer of R-25 (R-4.40 K.m2/W) fiberglass insulation with measured data from a calibrated hot box were described.
installed over girts 48” (1.2192 m) apart. The other consisted In that paper, model predictions of the overall heat transfer
of R-13 (R-2.29 K.m2/W) fiberglass and R-30 (R-5.28 K.m2/W) coefficient (the U-factor) and temperatures at several locations
fiberglass insulation layers installed over girts 60” (1.524 m) were compared with hot box measurements performed for a
apart. The measurements were conducted with the insulation standing seam roof (SSR) assembly. The model was used to
side facing the metering chamber and the wall side facing the calculate the overall heat transfer coefficients (U-factors) for
climate chamber of the hot box. The metering chamber average several metal building roof insulation assemblies. A selection
air temperature was 102.2 oF (312.15 K) for the single insu- of the computed results on the U-factors and a simplified,
lation layer assembly and 100.04 oF (310.95 K) for the two correlation based approach to predict the U-factors were
insulation layer case. The climate chamber average air presented by Choudhary and Kasprzak (2010). In this paper,
temperatures for the two assemblies were 48.4 oF (282.26 K) modeling results and hot box data are presented for two metal
and 49.97 oF (283.13 K) respectively. building wall insulation assemblies. One assembly contained
The predicted and measured heat transfer coefficients (U- a single layer of R-25 (R-4.40 K.m2/W) fiberglass insulation
factors) were in 93.3% agreement for the single insulation layer and the other contained two layers of fiberglass insula-
layer wall assembly and 95.5% agreement for the two insulation [R13 (R-2.29 K.m2/W) and R30 (R-5.28 K.m2/W)]
tion layer assembly. The effect of air motion in the flutes of the installed between the metal wall panels and the metal struc-
metal wall panel of the R-13 (R-2.29 K.m2/W) /R-30 (R-5.28 tural elements, called ‘girts’.
K.m2/W) assembly on the overall heat flow was found negli- In general, energy savings considerations have not been a
gible. The excellent agreements for the MBI wall assemblies significant factor in the design and construction of metal build-
observed in the present study, in conjunction with the 93%- ings. That situation is changing due to concerns about higher
97% agreement reported previously by Choudhary et al. 2010) energy prices and a better understanding of building science in
for the MBI roof assemblies, provide further confidence in the the industry. Proper use of insulation materials in metal build-
use of the mathematical model to study heat transfer in and ing roofs and walls will likely improve the energy efficiency
calculate U-factors for the MBI assemblies that contain fiber- of these structures. Fiberglass insulation with an appropriate
glass as the key insulation component. facing is predominant in metal building roofs and is also
Manoj K. Choudhary is a member of senior technical staff and David E. Musick is a senior technologist at Owens Corning Science & Tech-
nology Center, Granville, OH. Chris P. Kasprzak is a product manager at Owens Corning World Headquarters, Toledo, OH. Mark J. Henry
is a senior research engineer at Butler Manufacturing, Grandview, MO. N. Douglass Fast is a research engineer at Johns Manville, Littleton,
CO.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
400
ASHRAE Transactions
widely used for the walls. During the late 1990s, the American building insulation assemblies. The measured and calculated
Society of Heating, Refrigeration, and Air Conditioning Engi- U-factors for the assemblies presented in this paper are not
neers (ASHRAE) Standing Standard Project Committee intended to confirm or refute any values in the current
(SSPC) 90.1 Envelope Subcommittee incorporated specific ASHRAE/IES Standard 90.1.
maximum allowable U-factors for metal building roofs and
walls into ANSI/ASHRAE/IESNA Standard 90.1-1999 HOT BOX SET-UP AND RESULTS FOR MBI WALL
(ASHRAE/IESNA, 1999). Recently, ASHRAE and other ASSEMBLIES
industry organizations have recognized that Standard 90.1
The Hot Box measurements were performed by Fast et al.
should be revised and updated to account for a more accurate
(2010) for the R25 (R-4.40 K.m2/W) wall assembly and by
understanding of the installation of insulation in metal build-
Henry and Krueger (2010) for the R13 (R-2.29 K.m2/W) / R30
ing roof and wall assemblies. This recognition for a better
(R-5.28 K.m2/W) wall assembly. Figures 1 and 2 show sche-
understanding of the thermal performance of the metal build-
matics of the two hot box test frames used in this study. In the
ing insulation (MBI) assemblies led to a symposium in 2010
following sections, the test assembly used by Fast et al. (2010)
where the previous two papers (Choudhary et al. 2010, Choud-
will be referred to as the single-layer assembly, and the one
hary and Kasprzak, 2010) were presented.
used by Henry and Krueger (2010) will be called the two-layer
The present study is the continuation of the earlier studies. assembly.
Here, we have applied the mathematical model to two wall The test frame or the single-layer assembly (see Figure 1)
assemblies and tested its predictions with measurements from was a 120 in x 96 in (3.048 m x 2.438 m) panel with the girts
calibrated hot boxes for these assemblies. Since, relevant liter- mounted in the 120 in (3.048 m) direction. Two girts 4 ft (1.219
ature on mathematical modeling of heat transfer in MBI m) apart were used to secure the insulation in the test frame.
assemblies and the formulation of the model used in this study The test assembly consisted of 26 gauge fluted metal wall
have been described previously (Choudhary et al. 2010), in panels attached to 8 in (0.203 m) steel girts faced with 1/8 in
this paper, we shall focus on the model results and hot box (0.003 m) foam tape. Unfaced R-25 (R-4.40 K.m2/W) MBI
measurements for the two wall assemblies. fiberglass filler blanket was installed within the girt cavities
To reiterate, this work is a part of an ongoing effort to and supported with impaling type insulation supports. Two-
develop and apply a comprehensive and validated mathemat- sided tape was installed over the interior girt flange before the
ical model to calculate the thermal performance of metal installation of the PSK facing. Steel bands were then attached
Figure 1 Side view of the R-25 (R-4.40 K·m2/W) assembly.

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
©2012 ASHRAE
401
perpendicular to the girts, 24 in (0.610 m) on center, next to the blanket was installed within the girt cavities and supported
PSK Facing. The hot box testing by Fast et al. (2010) was with impaling type insulation supports. Two sided tape was
sponsored by the North American Insulation Manufacturers installed around the perimeter of the wood test frame to facil-
Association (NAIMA). itate installation of a Polyethylene (PE) vapor barrier. The
vapor barrier was installed over the interior of the girts held in
The test frame used in the two-layer assembly (see
place with bands attached perpendicular to the girts, 48 in
Figures 2a and 2b) was a 96 in x 120 in (2.438 m x 3.048 m)
(1.219 m) on center, over the PE Facing. The entire perimeter
panel with the girts mounted in the 96 in (2.438 m) direction.
was then sealed with duct tape. The testing by Henry and
Two girts 5 ft (1.524 m) apart were used to secure the insula-
Krueger (2010) was performed at the request of Owens Corn-
tion in the test frame. Two insulation layers were used. The test
ing.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
assembly consisted of 26 gauge fluted metal wall panels with
faced NAIMA 202-96 R-13 (R-2.29 K.m2/W) placed perpen- Photographs of the single-layer and two-layer test panels
dicular to the exterior of the 8 in (0.203 m) steel girts prior to are shown in Figures 3 and 4 respectively. Both are views from
wall panel attachment. A permeable facing will be required as the metering side of the assembly. The principal dimensions of
a carrier for the insulation to facilitate field installation of the the two test assemblies and the ambient air temperatures in the
system. Unfaced R-30 (R-5.28 K.m2/W) MBI fiberglass filler metering and the climate chambers are summarized in Table 1.
(a)
(b)
Figure 2 Views of the R-13 (R-2.29 K·m2/W)/R-30 (R-5.28 K·m2/W) assembly: (a) top, (b) side.
402
ASHRAE Transactions
Table 1. Principal Dimensions and Ambient Air
Temperatures in the Hot Box Tests
Parameter Single-layer Two-layer
96 in 120 in
Test Frame Length
(2.438 m) (3.048 m)
120 in 96 in
Test Frame Width
(3.048 m) (2.438 m)
0.018 in 0.018
Wall Panel Thickness
(4.6 x 10-4 m) (4.6 x 10-4 m)
36 in 36 in
Wall Panel Width
(0.914 m) (0.914 m)
8 in 8 in
Girt Depth
(0.203 m) (0.203 m)
Figure 3 Interior (metering chamber) side of the R-25 2.75 in 2.625

Girt Flange Width
(R-4.40 K·m2/W) assembly. Also shown are the (0.070 m) (0.067 m)
thermocouples. 0.0934 in 0.082
Girt Thickness
(2.40 x 10-3 m) (2.1 x 10-3 m)
120 in 120 in
Fastener Spacing
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
(3.048 m) (3.048 m)
Girt Mounting
Outside Flange Up Outside Flange Up
Direction
Girt Orientation Horizontal Horizontal
48.0 in
Girt Spacing 60.0 in (1.524 m)
(1.219 m)
0.125 in R13
Insulation Between
(0.003 m) (R-2.29 K.m2/W)
Girt and Wall Sheet
PE Foam Tape Fiberglass
R25 R30
Insulation within
(R-4.40 K.m2/W) (R-5.28 K.m2/W)
Girt Cavity
Fiberglass Fiberglass
Average Climatic
48.4 oF 49.97 oF
Chamber
(282.26 K) (283.13 K)
Air Temperature
Figure 4 Interior (metering chamber) side of the R-13 Average Metering
102.2 oF 100.04 oF
(R-2.29 K·m2/W)/R-30 (R-5.28 K·m2/W) Chamber
(312.15 K (310.95 K)
assembly. Air Temperature
Mean 75.3 oF 75.01 oF
Air Temperature (297.21 K) (297.04 K)
As shown above in Table 1, the single-layer assembly was
put inside the climate simulator with the climatic chamber on
exterior with the average air temperature of 49.4 o F (282.26 K) MODELING RESULTS
and the metering chamber (insulation and girt side) on the inte-
The formulation of the model as well as the details on the
rior with the average air temperature and at 102.2 o F (312.15
numerical solution have been discussed in detail by Choud-
K). The two-layer assembly was put inside the climate simu-
hary et al. (2010) and will be omitted here.
lator with the climatic chamber on exterior and at 49.97 o F
(283.13 K) and the metering chamber (insulation and girt side)
Modeling Results for the Single-Layer R-25 (R-4.40
on the interior and at 100.04 o F (310.95 K).
K·m2/W) Wall Assembly Hot Box
The heat balance and the measured surface temperature
data for the single and double layer hot-box assemblies are Figure 5 shows the modeling domain used for this case. It
shown in Tables 2 and 3 respectively. is 6 in (0.01524 m) wide and extends 22.75 in (0.5779 m) on
©2012 ASHRAE
403
Table 2. Heat Balance and Surface Temperature Data for the Single-Layer Hot-Box Setup
Item IP Value SI Value
Fan power 25.82 Btu/hr 7.56 W

Heater power 325.39 Btu/hr 95.28 W
Total power to Metering Chamber 350.87 Btu/hr 102.83 W
Heat gain through Metering Chamber wall 4.38 Btu/hr 1.28 W
Net heat entering Metering Chamber 355.25 Btu/hr 104.02 W
Flanking loss 10.93 Btu/hr 3.20 W
Infiltration heat flow 23.61 Btu/hr 6.91 W
Specimen heat flow 320.71 Btu/hr 72.5 W
Average Metering Chamber air temperature 102.2 oF 312.15 K
Average Climate Chamber air temperature 48.4 oF 282.26 K
Temperature difference between Metering

53.8 oF 29.89 K
and Climate Chamber air
Test specimen area 80.38 ft2 7.47 m2
Overall U factor 0.074 Btu/ft2 hr oF 0.425 W/m2 K
Metering Chamber surface temperature* 98.37 oF 310.02 K
o
Climate Chamber surface temperature* 49.62 F 282.94 K
* Area weighted surface temperature
Table 3. Heat Balance and Surface Temperature Data for the Two-Layer Hot-Box Setup
Fan power 21.76 Btu/hr 6.38 W

Heater power 224.73 Btu/hr 65.86 W
Total power to Metering Chamber 250.80 Btu/hr 73.50 W
Heat gain through Metering Chamber wall -0.03 Btu/hr -0.01 W
Net heat entering Metering Chamber 250.77 Btu/hr 73.49 W
Flanking loss 70.15 Btu/hr 20.56 W
Infiltration heat flow Not measured Not measured
Specimen heat flow 176.44 Btu/hr 51.71 W
Average Metering Chamber air temperature 100.04 oF 310.95 K
Average Climate Chamber air temperature 49.97 oF 283.13 K
Temperature difference between Metering and Climate

50.07 oF 283.19 K
Chamber air
Test specimen area 80.00 ft2 7.43 m2
Overall U factor 0.044 Btu/(ft2 hr oF) 0.250 W/(m2 K)
Metering Chamber surface temperature* 98.61 oF 310.16 K
Climate Chamber surface temperature* 50.79 oF 283.59 K
* Area weighted surface temperature.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
404
ASHRAE Transactions
Table 4. Thermal Conductivity and Boundary Conditions for the Single-Layer Assembly Model
Parameter IP Value SI Value
Thermal Conductivity of Fiberglass R-25 insulation 0.3589 Btu·in./(ft2·hr·°F) 0.05175 W/(m·K)

2
Thermal Conductivity of Polyethylene Foam Tape 1.734 Btu·in./(ft ·hr·°F) 0.250 W/(m·K)
Thermal Conductivity of Steel 313.683 Btu·in./(ft2·hr·°F) 45.241 W/(m·K)

2
Thermal Conductivity of Air 0.180 Btu·in./(ft ·hr·°F) 0.026 W/(m·K)
Heat Transfer Coefficient on the Climate Chamber Side 3.24 Btu/(ft2·hr·°F) 18.39 W/(m2·K)
Ambient Air Temperature in the Climate Chamber 48.4°F 282.3 K
Heat Transfer Coefficient on the Metering Chamber Side 1.06 Btu/(ft2·hr·°F) 6.02 W/(m2·K)
Ambient Air Temperature in the Metering Chamber 102.2°F 312.2 K
and the calculated average climate and metering chamber side

temperatures are also in fairly good agreements [49.6oF vs.
49.4 oF (282.93 K vs. 282.8 K); 98.4 oF vs. 99.8 oF (310.04 K
vs. 310.8 K)].
Modeling Results for the Two-layer R-13/R-30 Wall

Assembly Hot Box
Figure 6 shows the modeling domain used for this case. It
is 24” (0.610 m) wide and 60” (1.524 m) long. Both the insu-
lation layers [R-13 (R-2.29 K.m2/W) is attached to the metal
sheet, R-30 (R-5.28 K.m2/W) is next to it] have profiles in the
longitudinal [i.e. in the 60” (1.524 m)] direction. The R-30 (R-
5.28 K.m2/W) layer also has profile in the width [i.e. in the 24”
(0.610 m)] direction. The thickness of the R-13 (R-2.29 K.m2/
W) insulation on the girt was 0.25” (0.00635 m). The maxi-
mum thickness of the R-13 (R-2.29 K.m2/W) layer was 3.196”
(0.081 m) on the left side, and 3.88” (0.099 m) on the right. The
corresponding thicknesses for the R-30 (R-5.28 K.m2/W)
were 9.157” (0.233 m) and 8.473” (0.215 m) respectively. The
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
thickness profiles for both the layers were assumed parabolic,

Figure 5 Mathematical model domain for the R-25 (R-4.40 As mentioned above, the R-13 (R-2.29 K.m2/W) insulation
K·m2/W) assembly. layer thickness varied in the longitudinal direction and the R-
30 (R-5.28 K.m2/W) layer thickness varied in both the longi-
tudinal and the width directions. This is indicated in Figure 6,
either side from the edge of the 2.5 in (0.0635 m) long girt (i.e. which shows the modeling domain for the two-layer assembly.
the center-to-center spacing between the girts is 48 in The properties and boundary conditions used for this case are
(1.2192 m). The properties and boundary conditions used for summarized in Table 6.
this case are summarized in Table 4. As shown in Figure 1, The modeling results on the heat balance and surface
adopted from Fast et al. (2010), the insulation thickness was temperatures are summarized in Table 7. The calculations
assumed constant at 8 in (0.203 m). Figure 5, the sketch of the were carried out for the two cases; with or without the natural
modeling domain, shows the constant thickness of the insula- convection of the air in the flutes. The overall heat balance for
tion used in the model. Natural convection of air in the flutes the two cases differed only by about 0.4%. The results given
was accounted for. in Table 7 are for the case of stagnant air in the flutes.
The modeling results on the heat balance and surface It is seen from Tables 3 and 7 that the measured and the
temperatures are summarized in Table 5. predicted overall U-factors are in 95.5% agreement [0.042
As seen in Tables 2 and 5, the measured and the model Btu/ft2·hr·oF (0.236 W/m2·K) Vs 0.044 Btu/ft2·hr· oF (0.250
predicted overall U-factors are in 93.3% agreement W/m2·K)]. The measured and the calculated average climate
[0.074 Btu/ (ft2 hr oF), 0.069 Btu/ (ft2 hr oF)]. The measured and metering chamber specimen surface temperatures are also
©2012 ASHRAE
405
Table 5. Predicted Heat Balance and Surface Temperature Data for the Single-Layer Assembly Model
Heat in through the insulation surface 4.948 Btu/hr 1.450 W
Heat in through the girt surface 2.484 Btu/hr 0.728 W
Total heat in 7.432 Btu/hr 2.178 W
Heat loss through wall panel 7.398 Btu/hr 2.168 W
Heat loss through fastener 0.034 Btu/hr 0.010 W
Total heat out 7.432 Btu/hr 2.178 W
Surface area 2 ft2 0.186 m2
Overall U-factor 0.069 Btu/(ft2 hr oF) 0.392 W/(m2 K)
Area averaged insulation surface temperature 99.8 oF 310.8 K

o
Area averaged wall surface temperature 49.4 F 282.8 K
present work, we applied the mathematical model to two wall

assemblies and tested its predictions with measurements from
calibrated hot boxes for these assemblies. One of the hot boxes
consisted of a single layer of R-25 (R-4.40 K.m2/W) fiberglass
insulation installed over girts 48 in (1.219 m) apart. The other
consisted of R-13 (R-2.29 K.m2/W) fiberglass and R-30 (R-
5.28 K.m2/W) fiberglass insulation layers installed over girts
60 in (1.524 m) apart.
The test assemblies were put in chambers with the insu-
lation side facing the metering chamber and the wall side
facing the climate chamber. For the single layer wall assembly,
the metering chamber average air was temperature was 102.20
o
F (312.15 K) and the climate chamber average air tempera-
ture was 48.4 oF (282.26 K). The corresponding average air
temperatures for the two insulation layer assembly were
100.04 oF (310.95 K) for the metering chamber and 49.97 oF
(283.13 K) for the climate chamber. Heat flow through the test
frame was calculated from a heat balance analysis involving
measured data on power input to the metering chamber and
estimates of heat gain through the wall and flanking heat loss.
Temperatures at various locations on both the bottom and top
surfaces were measured.
Mathematical modeling involved the numerical solution
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 6 Mathematical model domain for the R-13 (R-2.29

K·m2/W)/R-30 (R-5.28 K·m2/W) assembly. of the steady state-heat conduction equation with appropriate
boundary conditions for the geometries of the two hot box
assemblies. The computational fluid dynamics software,
in fairly good agreements [50.8 oF vs. 50.3 oF (283.6 K vs. Fluent was used for the numerical solution. The model incor-
283.8 K); 98.6 oF vs. 98.7 oF (about 310.2 K for both cases)]. porated the relevant geometry details (e.g., girts, insulation
profiles, foam strip in the single layer assembly, fasteners, wall
SUMMARY AND DISCUSSION panels) and allowed for the fact that thermal conductivity of
the fiberglass insulation depends upon its density (i.e., its
In a previous study by Choudhary et al. (2010), a three- compressed thickness). The wall panel in the two layer assem-
dimensional mathematical model for heat transfer in metal bly had a fluted construction and this case was modeled for
building insulation assemblies was developed and tested with two scenarios: with air in the flute being stagnant or moving.
measurements conducted on the hot box set up of a standing The overall heat flow, and thus the U-factors, for the two
seam roof assembly consisting of fiberglass insulation. In the scenarios differed only by 0.4%.
406
ASHRAE Transactions
Table 6. Thermal Conductivity and Boundary Conditions for the Two-Layer Assembly Model
Parameter IP Value SI Value
Thermal Conductivity of Fiberglass R-13 insulation on the left* 0.2624 Btu-in/(ft2 hr oF) 0.03783 W/m K
2 o
Thermal Conductivity of Fiberglass R-13 insulation on the right* 0.2504 Btu-in/(ft hr F) 0.03611 W/m K
2 o
Thermal conductivity of Fiberglass R-13 insulation on the girt** 0.2364 Btu-in/(ft hr F) 0.03408 W/m K
Thermal Conductivity of Fiberglass R-30 insulation on the left* 0. 2893 Btu-in/(ft2 hr oF) 0.04172 W/m K
2 o
Thermal Conductivity of Fiberglass R-30 insulation on the right* 0. 2877 Btu-in/(ft hr F 0.0418 W/m K
2 o
Thermal Conductivity of Steel 313.683 Btu-in (ft hr F) 45.241 W/m K
Heat Transfer Coefficient on the Climate Chamber Side 1.467 Btu/(ft2 hr oF) 8.328 W/m2 K
Ambient Air Temperature in the Climate Chamber 49.97 oF 283.13 K
Heat Transfer Coefficient on the Metering Chamber Side 1.039 Btu/(ft2 hr oF) 5.898 W/m2 K
Ambient Air Temperature in the Metering Chamber Side 100.04oF 310.95 K
* Average value for the profile. The calculation of the average is explained in Choudhary et al. (2010).
** The insulation is compressed to 0.25” (0.00635 m) above the girt.
Table 7. Predicted Heat Balance and Surface Temperature Data for the Two-Layer Model
Heat in through the left R-30 insulation surface 7.047 Btu/hr 2.065 W
Heat in through the right R-30 insulation surface 7.159 Btu/hr 2.088 W
Total heat in through the R-30 surface 14.206 Btu/hr 4.163 W
Heat in through the girt surface 5.90 Btu/hr 1.729 W
Heat in through the brace surface 0.674 Btu/hr 0.198 W
Heat in through the fastener in the metering chamber 0.033 Btu/hr 0.01 W
Total heat in 20.813 Btu/hr 6.100 W
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Heat loss through the wall panel 15.571 Btu/hr 4.564 W

Heat loss through the flute surfaces 5.194 Btu/hr 1.522 W
Heat loss through fastener in the climate chamber 0.047 Btu/hr 0.014 W
Total heat out 20.813 Btu/hr 6.100 W
2
Surface area for heat transfer 10 ft 0.929 m2
Overall U-factor 0.042 Btu/ft2 hr oF 0.236 W/m2 K
Area averaged R-30 surface temperature 98.6oF 310.2 K
Area averaged wall surface temperature 51.1oF 283.8 K
The predicted and measured heat transfer coefficients (U- for the MBI assemblies involving fiberglass as the key insu-
factors) were in 93.3% agreement for the single insulation lation component.
layer wall assembly and 95.5% agreement for the two insula-
tion layer assembly. These excellent agreements for the MBI ACKNOWLEDGEMENTS
wall assemblies, in conjunction with the 93%-97% agreement The authors appreciate many helpful comments and
reported previously (Choudhary et al. 2010) for the MBI roof support on this project from Merle McBride, Neil Hettler,
assembly, give us further confidence in the use of the mathe- Todd Carlson, Gary Milosovich, Jason Fokens, and Todd
matical model to study heat transfer in and calculate U-factors Romain.
©2012 ASHRAE
407
REFERENCES lation or assembly construction. Would this have a significant
ASHRAE/IESNA Standard 90.1-1999: Energy Standards for impact on U-factor calculation?
Buildings Except Low-Rise Residential Buildings, Manoj K. Choudhary: Yes, it might. We have fully recog-
American Society of Heating, Refrigerating and Air- nized this, and Part I of this work, which was presented and
Conditioning Engineers, Inc., Atlanta. published in 2010, describes the approach to account for the
effect of insulation thickness on thermal performance. The
Choudhary, M.K, C. Kasprzak, R. Larson, R., and R. Venutu-
citation is as follows: M.K. Choudhary, C. Kasprzak,
rumilli. 2010; ASHRAE Standard 90.1 Metal Building U-
R.H. Larson, and R. Venuturumilli. 2010. ASHRAE Standard
Factors- part 1: Mathematical Modeling and Validation
90.1 metal building U-factors—Part 1: Mathematical model-
by Calibrated Hot Box Measurements. ASHRAE Transac-
ing and validation by calibrated hot box measurements.
tions 2010, Vol. 116, Part 1, pp 157-168.
ASHRAE Transactions 116(1):157–68.
Choudhary, M.K. and C. Kasprzak. 2010. ASHRAE Stan-
Andrew Price, Utility Infrastructure Market Leader,
dard 90.1 Metal Building U-Factors- part 2: A Systems
Affiliated Engineers, Madison, WI: One of the issues with
Based Approach for Predicting the Thermal Perfor-
using fiberglass and metal wall buildings in cold climates is
mance of Single Layer Fiberglass Batt insulation
condensation within the wall assembly. Can your modeling
Assemblies. ASHRAE Transactions 2010, Vol. 116, Part
methods be used to calculate surface temperatures throughout
1, pp 169-176.
the wall/ roof assembly?
Fast, D., Larson, R., and J. Owens. 2010: Metal Building
Choudhary: Our modeling approach can, in principle, be
Wall Thermal Values for Cavity Filled
used to account for the effect of condensation within the wall
Henry, M. J. and L. Krueger. 2010: ASTM C 1363 Thermal assembly. However, the model has to be augmented with addi-
Performance Test Report, Test Number: 2010-29, Butler tional details, such as the rate of water vapor diffusion, its
Technical Center, Kansas City, Missouri. condensation, and the increase in the thermal conductivity due
Liner Systems, Johns Manville Research Report-E, Number: to the presence of moisture. We do not have those features
436-04415, Johns Manville Technical Center, Littleton, included in the model. Incidentally, we have published a paper
Colorado. that looked at condensation in insulation used for pipes carry-
ing chilled water. The citation for that paper is as follows:
DISCUSSION M.K. Choudhary, K.C. Karki, and S.V. Patankar. 2004. Math-
Mehmood Ismail, Mechanical Engineer (Project ematical modeling of heat transfer, condensation, and capil-
Manager), Research in Motion, Burlington, Ontario, lary flow in porous insulation on a cold pipe. International
Canada: Insulation material can get squeezed during instal- Journal of Heat and Mass Transfer 47:5629.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
408
ASHRAE Transactions
SA-12-007
Review and Economic Feasibility Study of

the Currently Practiced New Housing
Constructions in Ontario
Aya Dembo Alan S. Fung, Ph.D., P.E.

This paper presents the overview of the currently practiced With an escalated awareness for limited energy resources,
new housing constructions in Ontario based on a survey to local the urgent need to reduce energy consumption in the residen-
homebuilders in the Greater Toronto Area (GTA), which was tial sector in Canada has been brought to the forefront. Dong
conducted as part of the proposed methodology to determine the et al. (2005) stated that the residential sector is the third largest
suitability of the potential optimal upgrades considered for the consumer of end-use energy in Canada (Farahbakhsh et al.,
analysis. Also presented in this paper is the economic feasibility 1998, Natural Resources Canada [NRCan], 2000). One of the
study of the identified optimal upgrades from the survey for the immediate solutions is to develop potential energy efficiency
selected cities in Ontario with varying climatic conditions. upgrades that can be implemented into the currently practiced
new housing constructions to achieve significant energy
Based on the results of the survey, it was concluded that
savings through the optimization in energy related costs.
the participant builders have already implemented or consid-
ered for implementation some of the advanced technologies
Literature Review
into their current building practices to achieve the energy
performance level that exceeded the minimum requirements as Numerous studies have been conducted worldwide to
specified by the existing Ontario Building Code (OBC). determine the optimal upgrade solutions among various
Furthermore, the results of the economic feasibility study improvement measures through the optimization in energy
showed that, based on the life cycle analysis of 30 years, the costs using two distinct methods: a brute force sequential
combination of optimal upgrades comprised of increasing the search (BFSS) method, and a genetic algorithm (GA) method.
thermal resistance of the building envelope components, and For instance, Sambou et al. (2009) applied the latter to deter-
installations of the most energy-efficient heating, ventilating, mine the optimization of a building wall by maximizing ther-
and air-conditioning (HVAC) system could achieve up to 31% mal insulation and thermal inertia, while Verbeeck & Hens
reduction in both the estimated total annual energy consump- (2007) used such a methodology to estimate energy costs over
tion and greenhouse gas (GHG) emissions in comparison to the life cycle of the building. Based on the results from these
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
the same building configuration, but built in accordance with studies, it can be concluded that the genetic algorithm (GA)
the OBC, and without the implementation of the identified opti- method is useful when there are more than one objectives
mal upgrades. Furthermore, with the installation of a residen- required to achieve optimization, in other words, energy, envi-
tial photovoltaic (PV) system combined with these upgrades, ronment, and overall cost.
the estimated total profit of up to $89,035 could be achieved Another method is the brute force sequential search
through Ontario’s micro Feed-in-Tariff (FIT) program, allow- (BFSS) method, and this is the methodology used in this study.
ing a homeowner to pay for the installation of additional In the United Kingdom, the same methodology was used in the
upgrade(s) to further improve the energy efficiency level of a study by Gorgolewski (1995) to examine the relationship that
home, ultimately reaching the net zero energy status. existed among the various energy efficiency upgrade solutions
Aya Dembo recently completed her MASc degree in the building science program at the Department of Architectural Science and Alan S.
Fung is an associate professor in the Department of Mechanical and Industrial Engineering, Ryerson University, Toronto, ON, Canada.
©2012 ASHRAE
409
as more than one upgrade interacted with another upgrade, and mined based on the sum of monthly energy costs (whether it
by doing so, identified a combination of optimal upgrade solu- was natural gas and/or electricity), and the mortgage payments
tions. The methodology, which Gorgolewski (1995) referred at the end of each year. The annual energy cost and mortgage
to as the “optimization process” was used not only to identify payment are shown in Figure 1 by a red line and blue line,
the optimal upgrade (i.e., the one with the highest saving-to- respectively. It should be noted that all of the costs were
investment ratio [SIR]), but also, to evaluate the effects of expressed in terms of net present value (NPV), similar to the
combining the optimal upgrade solution with the remaining current study. Also similar to the current study was the fact that
solutions by making adjustments to their capital costs, and Anderson et al. (2006) determined the minimum total annual
determining a new SIR for each upgrade. Gorgolewski (1995) cost based on the lowest guaranteed results from all possible
continued this process until all the identified upgrade solutions upgrade solutions considered in their study.
were either accepted, resulting in a SIR of greater than or equal Lastly, net zero energy status was achieved when the 100%
to 1, or rejected if any upgrade resulted in a SIR of less than increase in energy efficiency and the resultant zero energy cost
1, after readjusting their capital investment cost. were exclusively a result of installing an onsite renewable
Another example of using the BFSS method would be the energy system, as shown by an arrow in the top-right corner of
studies done in the United States (Anderson et al., 2006; Chris- Figure 1. Anderson et al. (2006) considered the residential PV
tensen et al., 2005), where the optimization process was devel- system as an onsite renewable energy system for their study;
oped as a result of combining advanced building envelope although the authors believed that other renewable energy
designs with the implementation of renewable energy systems systems would soon be available in the U.S. housing market.
into the space heating/cooling systems to achieve net zero The objective of this paper is to identify the optimal
energy status. This methodology, referred to in the study as the upgrade solutions that meets various levels of thermal perfor-
“optimal least-cost path to net zero energy”, as shown in mance while considering the use of local materials and
Figure 1, was developed. Anderson et al. (2006) explained the resources, ease of construction, but most importantly, adop-
conceptual marginal curve of the optimal least cost path to net tion by Ontario housing industry through the optimization in
zero energy as follows: First, the energy performance of a energy related costs using a brute force sequential search
baseline building was determined using a building that was (BFSS) method. All subsequent references to “optimization”
built either 1) in accordance with either the building code; or in this paper should be interpreted as cost optimization.
2) currently practiced new building constructions, or 3) a
representative of some other reference building. The baseline METHODOLOGY
building resulted in 0% increase in the energy efficiency, as This paper is the continuation of the work being carried
shown by a purple line in Figure 1. out by Dembo et al. (2012) and Dembo (2011), which inves-
From the baseline case, energy consumption of the base- tigated the methodology developed to identify the most cost-
line building decreased due to the outcome of the identified effective specifications to achieve improved energy efficiency
energy efficiency upgrade solutions, resulting in a minimum standards in Ontario new housing constructions using the life
total operating energy cost with the estimated increase in cycle cost analysis. The methodology with respect to the
energy efficiency of between 20 to 40%. The total operating details of the study house, determination of the building codes
energy cost, as shown by a green line in Figure 1, was deter- used to develop a baseline case model for the building energy
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,
Figure 1 Conceptual path to net zero energy.
410
ASHRAE Transactions
simulation analysis, and the assumptions made on occupancy, cooling degree days of the selected cities in Ontario that were
appliance and lighting should be referred to their studies. As considered for the study.
such, the methodology for this paper is considered and broken As Table 1 summarizes, the results of the survey can be
down into following two parts: extrapolated to Ottawa, and Windsor based on the similarity in
their climatic conditions to that of Vaughan, and of Toronto
1. Conducting a survey to local homebuilders in Ontario to and/or Oakville, respectively. Thunder Bay, on the other hand,
determine the suitability of the considered optimal (i.e., due to its extreme cold climatic conditions, the results of the
the lowest life cycle cost) upgrades that can be imple- survey could not be applied and, therefore, considered as a
mented or considered for implementation into the special case in this study.
currently practiced new housing constructions to achieve
Although it was important to account for the variance in
the improved energy efficiency standards: and
the climatic conditions of the selected cities in Ontario, it was
2. Applying the same methodology used in the studies by reported in the study by Dembo (2011) that the difference in
Dembo et al. (2012) and Dembo (2011) to determine a the climatic condition for even one city was less effective in
combination of considered optimal upgrades using the terms of identifying the final combination of optimal upgrades
BFSS method for the other cities (Ottawa, Windsor and than that of the economic parameters used for conducting a life
Thunder Bay) in Ontario. cycle cost analysis (LCCA) of those upgrades. The results will
be discussed later in this paper.
In the first part of the paper, the survey was developed to
conduct a detailed review of the suitability of considered Assumptions Made
upgrades by analyzing what types of advanced technologies
the local homebuilders in Ontario have already put into prac- To conduct the LCCA of 30 years, the economic param-
tice. With the introduction of the 2012 and the future 2016 eters comprised of mortgage rate, discount rate, and fuel price
Ontario Building Codes (OBCs), it was critical to understand escalation rates for electricity and natural gas, respectively,
the current state of new housing constructions in Ontario, and was determined based on the historical data obtained from the
based on the analysis, the survey was developed to identify studies by Dembo et al. (2012) and Dembo (2011). Table 2
what would be the major changes in the current practice if the summarizes the identified economic parameters used for the
local homebuilders were to consider implementing any one of analysis.
the identified upgrades to meet the requirements of the new Also part of a sensitivity analysis was the estimated reduc-
codes. The survey was conducted, over the phone and/or in- tion in appliance and lighting, as well domestic hot-water
person, by seven local homebuilders in various locations (from (DHW) consumptions to further investigate the variances in the
Toronto to Vaughan to Oakville) in Ontario, but all within the estimated total annual energy consumption, percent savings
Greater Toronto Area (GTA), the largest metropolitan area in per year, greenhouse gas (GHG) emissions, and the total life
Canada (Organization for Economic Co-operation and Devel- cycle cost (in net present value) of the identified combinations
opment [OECD], 2009). Despite the low sample size, these of upgrades for the selected three cities in Ontario. The
were the large volume builders with each building up to 300 assumptions were based on the study done by Tse et al. (2008).
houses a year, thus making the survey result more representa- Table 3 summarizes the estimated reduction in appliance, light-
tive of the housing market1. ing and DHW consumptions used for the analysis.
For the second part, the combination of optimal upgrades It was reported in the study by Dembo (2011) that the
were determined using the BFSS method for various cities in overall total energy consumption of the Case 3 decreased by
Ontario, namely Ottawa, Windsor and Thunder Bay with
distinct climatic conditions, following the same methodology
Table 1. Climatic Conditions of the Selected Cities in
used in the studies by Dembo et al. (2012) and Dembo (2011).
Ontario (Environment Canada, n.d.)
According to the Environment Canada (n.d.), both Ottawa and
Thunder Bay have colder climatic conditions than Toronto
Location in Heating Degree-Days Cooling Degree-Days
with the estimated annual heating degree days (below 18°C) of Ontario: (below 18°C [64°F]) (above 18°C [64°F])
4602, and 5677, respectively; whereas Windsor, its climatic
condition is warmer than that of Toronto with the estimated Oakville 3897 (7015) 230 (414)
annual heating degree days (HDD) of 3525. Table 1 summa-
Ottawa 4602 (8284) 245 (441)
rizes the climatic conditions as defined by both the heating and
Thunder Bay 5677 (10,219) 70 (126)
1.
It was reported by the CMHC (2011) that the total number of Toronto 3570 (6426) 359 (646)
single-family dwellings built in the Greater Toronto Area (GTA)
in year 2010 was 11,079, while the number of new houses built by Vaughan 4137 (7447) 232 (418)
the participant builders was estimated to be 1,653, thus represent-
ing 15% of the housing market. Windsor 3525 (6345) 422 (760)
--`,`,``,,`,`,```,`,,``,`,``,,,-
©2012 ASHRAE
411
Table 2. Summary of the Economic Parameters Considered for the Study
Economic Parameters Case 1 Case 2a Case 2b
Mortgage rate (%) 7.87 5.05 5.05

Inflation rate (%) 1.99 3.10 3.10
Fuel price escalation rate for:
Electricity (%) 3.50 3.50 3.50[1]
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Natural gas (%) 3.02 5.50 5.50
Electricity price (monthly): $/kWh (before taxes)
6
Less than 800 kWh (2.7x10 Btu) 0.1038 0.1038 0.1038
6
Greater than 800 kWh (2.7x10 Btu) 0.1138 0.1138 0.1138
Natural gas price (monthly): $/m3 (before taxes)
Distributor 1: First 30 m3 (1059 ft3) 0.2892 0.7053 0.5300
3 3
Next 55 m (1942ft ) 0.2843 0.7053 0.5300
Next 85 m3 (3002ft3) 0.2804 0.7053 0.5300
3 3
Over 170 m (6003ft ) 0.2774 0.7053 0.5300
Distributor 2: First 100 m3 (3531 ft3) 0.2573 0.7152 0.5300
3 3
Next 150 m (5297ft ) 0.2550 0.7152 0.5300
3 3
Next 250 m (7063ft ) 0.2496 0.7152 0.5300
Note: The Ontario Ministry of Energy recently announced the projected long-term electricity price increase of 3.50% annually for the next twenty years (Ministry of Energy [MIE],
2010). Therefore, 3.50% will be used as the identified electricity escalation rate for the proposed methodology.
Table 3. Percent Reduction in the Electrical Consumption between the Two Scenarios
Cases 1 and 2[1]: Case 3[2]:

Breakdown of Baseloads Unit Standard Operating Reduced % Reduction
Conditions Consumption
Appliances kWh/yr 5110 3066 40%

Lighting kWh/yr 1095 274 75%
Other Electric Devices kWh/yr 1095 1095 –

Exterior Use kWh/yr 1460 675 54%
Total Baseloads Consumptions kWh/yr 8760 5110 42%

Notes:
1. The standard operating conditions used in HOT2000 were based on the occupancy of two adults and two children, at present 50% of the time and using 24 kWh per day of electricity
consumption (NRCan, 2010)
2. The assumptions were based on the study done by Tse et al. (2008), which reported that 40% reduction in major house appliances was achieved by installing energy-efficient
appliances; while 75% reduction in lighting was achieved by installing Energy Star approved compact fluorescent light (CFL) bulbs.
2%, despite the fact that the estimated space heating consump- PV system in Ottawa, Windsor, and Thunder Bay, respec-
tion increased from the previous cases (1 and 2) by 15%, while tively. For this analysis, NRCan’s RETScreen® version 4 was
the estimated space cooling consumption was reduced by 16% used in addition to the HOT2000 program due to the lack of
as a result of the 42% reduction in the appliance and lighting
possible inputs that the latter requires (Tse et al., 2008). The
consumptions.
Lastly, additional analysis was conducted to determine assumptions made for the simulation of a residential PV using
the overall energy performance of the study house, after the the RETScreen were same as the ones identified in the studies
installation of the identified combination of upgrades with a by Dembo et al. (2012) and Dembo (2011).
412
ASHRAE Transactions
REVIEW OF THE CURRENTLY PRACTICED NEW upgrading it to 0.80 EF, although the remaining 86%
HOUSING CONSTRUCTIONS IN ONTARIO— were not considering upgrading their existing systems
SURVEY RESULTS unless such an upgrade were to be required by the future
Prior to conducting the survey, it was important to under- 2012 or 2016 OBCs or the ESNH Standard.
stand the current state and requirements of the commonly • A total of 43% of the participant builders were installing
adopted energy efficiency standards in Ontario, because these the heat recovery ventilator (HRV) system with 55%
standards, although not mandated by the federal/provincial efficiency, and out of these, 29% were considering
governments in Canada, have been practiced by the local upgrading it to 70% efficiency, although the remaining
homebuilders, thereby forming the basis of today’s commonly 71% were not considering upgrading their existing sys-
practiced new housing constructions in Ontario. Based on the tems to higher efficiency due to the lack of market
results of the survey, it was apparent that all participant build- demand at this point, unless otherwise required by the
ers who have responded to the survey were building their future 2012 or 2016 OBCs or the ESNH Standard.
homes to the existing 2006 OBC and/or the ENERGY STAR
for New Homes Standard (NRCan, 2011). Some of the results ECONOMIC FEASIBILITY STUDY OF THE
to highlight were as follows: IDENTIFIED OPTIMAL UPGRADES
As in the studies by Dembo et al. (2012) and Dembo
• A total 43% of the participant builders responded that (2011), the combination of the optimal upgrades was deter-
their current exterior wall construction practices used mined using the BFSS method for the climatic condition of
timber-framed construction, where the studs were Ottawa, Windsor and Thunder Bay, respectively. The results for
spaced at 400mm (16 in.) on-centre, and with RSI 3.34 the sensitivity analyses are presented and discussed as follows:
(R19) wall insulation placed in between the studs. Out
of these, 29% were considering upgrading the existing Case 1—Using the Average Economic Parameters
wall insulation to RSI 3.87 (R22), and the remaining
14% were considering upgrading it to RSI 4.23 (R24) Figure 2 show the selection of the final upgrade by round
(comprising installing RSI 3.34 (R19) interior wall insu- using the BFSS method for Ottawa, Windsor, and Thunder Bay,
lation, and RSI 0.90 (R5) exterior wall insulation) in the respectively. For a complete listing or glossary of the considered
near future. upgrades, refer to Glossary (Dembo et al., 2012; Dembo 2011).
• For the attic construction, 43% of the participant build- As shown in Figure 2, the identified combination of opti-
ers were considering upgrading their existing attic/ceil- mal upgrades was almost identical in all three cities, except for
ing insulation of RSI 7.00 (R40) to RSI 8.80 (R50) in the HRV with 70% efficiency [V1], which was the most opti-
the near future, while the remaining 57% were reluctant mal upgrade solution at the end of Round 2 in Thunder Bay,
to upgrade their current attic/ceiling construction prac- Round 3 in Ottawa, and Round 4 in Windsor, respectively.
tices by installing an insulation of RSI 8.80 (R50) or This could imply that the colder the weather the better the
higher. This could be attributed to the fact that such an HRV as an upgrade.
upgrade is neither required by the homeowners and, It was also interesting to note that the estimated annual
therefore, no benefit to the builders, nor specified by the total household energy consumption for all three cities did not
existing 2006 OBC or the ESNH Standard at this point. change after the installation of an AC unit at SEER 14 (at the
• A total of 43% of the participant builders responded that end of Round 7). This could imply that the effectiveness of the
they were installing a furnace with 90% annual fuel uti- change in the efficiency of the unit (from SEER 13 to 14) was
lization efficiency (AFUE) as per the existing 2006 very minimal and resulted in a very small reduction in the
OBC and/or ESNH Standard. Out of these, 43% overall energy consumption, while increased the total life
responded that they were considering upgrading their cycle cost by 3% for all three cities. This could be due to the
existing furnace to one with higher AFUE of 92%, 94% increase in the installation cost associated with the replace-
and 95%, respectively in the near future, while the ment of the unit as a result of its life expectancy of 12 years,
remaining 43% responded that they have begun to install or the fact that the energy required for air-conditioning a typi-
the same system with AFUE of 92 or higher. It should cal residential home is a very small fraction of overall energy
be noted that consensus (i.e., 100%) could not be use, considering the total cooling degree-days (above 24°C
achieved with regard to the participants’ views on [75°F]) for all three cities, which are very small in comparison to
upgrading their existing furnace due to the fact that the those of the heating degree-days (refer to Table 1). Therefore, it was
remaining 14% had no response. This could imply that concluded that this upgrade was removed from the rest of the
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
these houses were built with no furnace installed or such analyses.

equipment was on lease. For the case in Thunder Bay (Figure 2, bottom), the
• As for the DHW heater, 57% of the participant builders above-grade wall insulation of RSI 5.10 (R29), placed in
were installing the system with 0.57 Energy Factor (EF) between the studs at 600 mm (24 in.) on-centre [AGW7] was
or higher, and out of these, 14% were considering the most optimal upgrade solution at the end of Round 5. It
©2012 ASHRAE
413
Figure 2 Least cost upgrade by round using the BFSS method for Ottawa (top), Windsor (middle), and Thunder Bay (bottom)
(Case 1).
should be noted that some of the potential upgrades, in partic- Table 4 summarizes the estimated total annual household
ular, those with respect to the above-grade walls were elimi- energy consumption, percent savings per year, reduction of the
nated due to the baseline case having a higher thermal GHG emissions, and the total life cycle cost in net present
resistance (RSI) level (of 4.22) as mandated by the existing value for Ottawa, Windsor, and Thunder Bay, respectively, in
OBC than the rest of the cities with their estimated annual comparison to Toronto, for the Case 1.
HDD of 5000 or lower (i.e., Toronto, Ottawa, and Windsor). It Figure 3 shows the total life cycle cost in net present value
should also be noted that the incremental costs for these with respect to the increase in energy efficiency for the
upgrades were adjusted properly to account for such changes selected cities (Ottawa, Windsor, and Thunder Bay) in
in the baseline case. Ontario, compared against that of Toronto.
414
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
ASHRAE Transactions
Table 4. Simulation Results using the Average Economic Parameters (Case 1)
Location in Ontario
Toronto
Results (Units)
(Reference)
Ottawa Windsor Thunder Bay
Energy Consumption (kWh/year) 24,359 29,302 22,097 31,378
Energy Consumption (MBTU/year) 83 100 75 107
% Savings (% per year) 25 21 24 23
GHG Emissions[1] (Ton/CO2) 5.59 7.45 6.24 7.76
EnerGuide Rating (ERS) 82.6 81.2 82.3 82.7
Total LCC[2] in NPV ($) 103,811 102,795 96,787 104,796

Notes:
1. The GHG emission calculation was done using the carbon dioxide (CO2) factors determined from the study by Fung & Gill (2011). In their study, 1.856 kg/m3 equivalent CO2
(NRCan, 2006) was used for natural gas, and the average emission factor for electricity was 226.35 tons CO2/total gigawatt-hour (GWh) generation (Gordon & Fung, 2009).
2. The lifecycle cost (LCC) was determined based on the identified average mortgage rate of 7.87%, average inflation rate of 1.99%, and the fuel escalation rates for electricity and
natural gas of 3.50%, and 3.02%, respectively.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 3 Total cost in NPV ($) vs. increase in energy efficiency (%) for the selected cities in Ontario (Case 1).
©2012 ASHRAE
415
Case 2a—Using the Extreme Economic Parameters Case 2b—Using the Extreme Economic Parameters
with 2006 Natural Gas Prices with 2008 Natural Gas Prices
In this case, the combination of optimal upgrades was In this case, the combination of optimal upgrades was the
almost identical for Ottawa and Thunder Bay, except for same for Ottawa and Windsor, but not for Thunder Bay, as
Windsor, as shown in Figure 4. shown in Figure 6.
Table 5 summarizes the estimated total annual household Table 6 summarizes the estimated total annual household
energy consumption, percent savings per year, reduction of the energy consumption, percent savings per year, reductions of
GHG emissions, and the total life cycle cost in net present the GHG emissions, and the total life cycle cost in net present
value for Ottawa, Windsor, and Thunder Bay, respectively, in value for Ottawa, Windsor, and Thunder Bay, respectively, in
comparison to Toronto, for the Case 2a. comparison to Toronto, for the Case 2b.
As shown in Table 5, similar to the results of the previous As shown in Table 6, similar to the results of the previous
case was the estimated energy consumption and the GHG cases were the estimated energy consumption of the baseline
emissions of the baseline case in all three cities, which case in all three cities, which decreased due to the outcome of
decreased due to the outcome of the identified upgrades, the identified upgrades, resulting in decrease in the total life
resulting in decrease in the life cycle cost in net present value. cycle cost in net present value. As for the EnerGuide Rating
As for the EnerGuide Rating (ERS), it was not sensitive, and (ERS), it was not sensitive, and remained at EGR of 82 with
remained at EGR of 82 with a difference in percent savings of a difference in percent savings of 5% in Ottawa, 1% in Wind-
5% in Ottawa, and 4% in Thunder Bay, respectively, while sor, and 4% in Thunder Bay, respectively.
remained at EGR of 83 with a difference in percent savings of It was interesting to note that RSI 5.11 (R29) exterior wall
2% in Windsor. Figure 5 shows the life cycle cost in net present insulation was identified as the most optimal upgrades not
value with respect to the increase in energy efficiency for the only in Thunder Bay, but also, Ottawa and Windsor, particu-
selected cities in Ontario, compared against that of Toronto. larly in the case of using the extreme parameters (i.e., maxi-
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 4 Least cost upgrade by round using the BFSS method for Ottawa and Thunder Bay (top), and Windsor (bottom) (Case 2a).
416
ASHRAE Transactions
mum discount rate, and natural gas price escalation rate). Figure 7 shows the total life cycle cost in NPV with
However, this particular upgrade was not considered for respect to the increase in energy efficiency for the selected
implementation by the local homebuilders in Ontario based on cities in Ontario, compared against that of Toronto.
the results of the survey. In fact, none of the participant build-
ers considered the implementation of any insulation material Case 3a—Reduced Appliance and Lighting
with nominal RSI of 4.75 (R27) or higher into their current Consumptions
building practices, unless otherwise required by the building Based on the results of the reduced appliance and lighting
code and/or the existing energy efficiency labeling programs consumptions, and using the economic parameters identified
such as the ESNH. in Case 1, the combination of optimal upgrades in net present
Table 5. Simulation Results using the Extreme Economic Parameters (Case 2a)
Location in Ontario
Toronto
Results (Units)
(Reference) Ottawa Windsor Thunder Bay
(MBTU/year) 83 89 68 107
GHG Emissions (Ton/CO2) 6.58 6.89 5.87 7.76
Total LCC[1] in NPV ($) 143,216 149,056 128,905 159,286

Note:
natural gas of 3.50%, and 5.50%, respectively. Furthermore, the commodity price in 2006 was used for natural gas.
Figure 5 Total cost in NPV ($) vs. increase in energy efficiency (%) for the selected four cities in Ontario (Case 2a).
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
©2012 ASHRAE
417
Figure 6 Least cost upgrade by round using the BFSS method for Ottawa and Winds or (top), and Thunder Bay (bottom) (Case 2b).
Table 6. Simulation Results using the Extreme Economic Parameters (Case 2b)
Location in Ontario
Toronto
Results (Units)
(MBTU/year) 83 89 75 107

[1]
Total LCC in NPV ($) 124,333 128,733 113,943 134,840
Note:
natural gas of 3.50%, and 5.50%, respectively. Furthermore, the commodity price in 2008 was used for natural gas.
value was determined using the BFSS method, which was Despite the fact that the estimated total household energy
almost identical to the one identified in the previous case using consumption decreased at the end of each round, it was inter-
the standard operating conditions in HOT2000 for all three esting to note that the overall household consumption
cities. increased by approximately 10% when compared it against
that of the previous case without the reduction in the appliance
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
418
ASHRAE Transactions
Figure 7 Total cost in NPV ($) vs. increase in energy efficiency (%) for the selected cities in Ontario, after the reduction in
the appliance and lighting consumptions.
Table 7. Simulation Results for the Reduced Appliance and Lighting Consumption
Location in Ontario
Toronto
Results (Units)
(MBTU/year) 94 109 85 119

Total LCC[1] in NPV ($) 83,173 79,908 75,668 85,041

Note:
The lifecycle cost (LCC) was determined based on the identified average mortgage rate of 7.87%, average inflation rate of 1.99%, and the fuel escalation rates for electricity and
and lighting consumptions, while the total life cycle cost in net ances and lighting. Table 7 summarizes the estimated total
present value decreased by up to 22% (20% in Ottawa, 22% in annual household energy consumption, percent savings per
Windsor, and 19% in Thunder Bay). This could imply that the year, reductions of the GHG emissions, and the total life cycle
effectiveness of the reduction in the appliance and lighting cost in net present value for Ottawa, Windsor, and Thunder
consumptions by 42% resulted in decrease in the energy Bay, respectively, in comparison to Toronto, for the Case 3a.
required for space cooling, but not for space heating purposes, Figure 8 shows the total life cycle cost in net present
where the latter accounted for the heat gains from the appli- value, after the reduction in the appliance and lighting
--`,`,``,,`,`,```,`,,``,`,
©2012 ASHRAE
419
Figure 8 Total cost in NPV ($) vs. increase in energy efficiency (%) for the selected four cities in Ontario (Case 2b).
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
consumptions, with respect to the increase in energy efficiency for all three cities. As for the EnerGuide Rating (ERS), it was
for the selected cities in Ontario, compared against that of not sensitive, and remained at EGR of 83 with a difference
Toronto. percent savings of 3% in Ottawa, and EGR of 84 with a differ-
ence in percent savings of 4% in Windsor, and 3% in Thunder
Case 3b—Reduced Appliance, Lighting and DHW Bay, respectively.
Consumptions
For the cases in Ottawa and Windsor, it was interesting to
In addition to reducing the amount of electricity used for note that, with the reduction in the average hot water load, the
appliances and lighting, the average hot water load was combo boiler with 0.85 EF was the most optimal upgrade (at
reduced from 225L/day to 100L/day based on the assumption the end of Round 1) than the same system with higher effi-
used in the study by Tse et al. (2008). ciency (of 0.9 EF), where the latter was identified as the most
Based on the results of the reduced appliance, lighting and optimal upgrade in the previous case without the reduction.
hot water consumptions using the economic parameters iden- This could imply that with 56% reduction in the average hot
tified in Case 1, the combination of optimal upgrades in net water load, the combo boiler with lower efficiency would be
present value was identified using the BFSS method, which sufficient when a significant amount of hot water load was
was almost identical to the one identified in the previous case reduced.
using the standard operating conditions in HOT2000 for all Figure 9 shows the life cycle cost in net present value,
three cities. after the reduction in the appliance, lighting and hot water
Table 8 summarizes the estimated total annual household consumptions, with respect to the increase in energy efficiency
energy consumption, percent savings per year, reductions of for the selected cities in Ontario, compared against that of
the GHG emissions, and the total life cycle cost in net present Toronto.
value for Ottawa, Windsor, and Thunder Bay, respectively, in
comparison to Toronto, for the Case 3b. Case 4—Combining the Installation of a Residential
As shown in Table 8, similar to the results of the previous PV System
case with the reduction in the appliance and lighting consump-
tions were the estimated energy consumption and the GHG Based on the results using the Hot2000 and RETScreen,
emissions of the baseline case in all three cities, which the estimated electricity exported to grid for the cases in
decreased due to the outcome of the identified upgrades, Ottawa, Windsor, and Thunder Bay was 7667 kWh, 8317
resulting in decrease in the life cycle cost in net present value kWh, and 8901 kWh per year, respectively.
420
ASHRAE Transactions
With the introduction of Ontario's microFIT program, value of the total electricity cost of what a homeowner would
the price of $0.802 was applied for every kilowatt-hour of pay if the PV system were not installed (hence the savings)
electricity generated by the PV. After subtracting the cost of was estimated to be $12,760 for Ottawa, $13,883 for Windsor,
installing a PV system (assumed $7/W for a total of $42,560) and $14,756 for Thunder Bay. Therefore, with the installation
from the income, the net present value of the total income for of the PV system, the total profits, over the 30-year mortgage
20 years was estimated to be $58,081 for Ottawa, $66,613 for period, were estimated to be $70,841 for Ottawa, $80,496 for
Windsor, and $74,279 for Thunder Bay. Windsor, and $89,035 for Thunder Bay, which would allow a
Accounting for the electricity production from the PV homeowner to pay for the implementation of additional
system from 21 to 30 years, the difference in the net present upgrade(s).
Table 8. Simulation Results for the Reduced Appliance, Lighting and DHW Consumptions
Location in Ontario
Toronto
Results (Units)
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
(MBTU/year) 85 101 77 109

[1]
Total LCC in NPV ($) 80,553 77,349 73,187 82,324
Note:
The lifecycle cost (LCC) was determined based on the identified average mortgage rate of 7.87%, average inflation rate of 1.99%, and the fuel escalation rates for electricity and
Figure 9 Total cost in NPV ($) vs. increase in energy efficiency (%) for the selected four cities in Ontario, after the reduction
in the appliance, lighting and hot water consumptions.
©2012 ASHRAE
421
Table 9. Estimated Total PV Profit over 30 years for Ottawa, Windsor, and Thunder Bay
Location
Results Unit
Ottawa Windsor Thunder Bay
Estimated electricity exported to grid kWh/yr 7667 8317 8901

Case 1: Using the Average Parameters
Total income for 20 years $ 58,081 66,613 74,279
Total PV profit after 30 years (mortgage period) $ 70,841 80,496 89,035

Case 2: Using the Extreme Parameters
Total savings from 21 to 30 years $ 9,672 10,525 11,187

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Case 3: Using the Reduced Appliance, Lighting and DHW Consumptions
Table 9 summarizes the estimated total profit as a result of COMPARISON OF THE CURRENT STUDY AGAINST
installing a PV system in each of the selected cities (Ottawa, THE 2012 OBC
Windsor, and Thunder Bay) using Ontario’s microFIT
program, assuming such a program will continue onward. This paper used the proposed prescriptive-based specifi-
The overall energy performance of the study house, after cations of the 2012 OBC as the energy performance target
the installation of the identified combination of optimal against the existing 2006 OBC as the baseline case for compar-
upgrades with a PV system, was re-evaluated to account for ison purposes. Despite the fact that the official release of the
the reduction in the appliance, lighting and hot water 2012 OBC was not available at the start of the current study as
consumptions. it was released in the beginning of year 2010 (MMAH, 2010),
Based on the results using RETScreen, the estimated elec- the identified combination of optimal upgrades have met or, in
tricity exported to grid was the same as the previous cases some cases, exceeded the minimum efficiency level as
without the reduction for all three cities. However, the net prescribed by the building code. Table 10 summarizes the effi-
present value of the total income for 20 years, after subtracting ciency level with respect to the thermal resistance (RSI) of the
the cost of installing a PV system from the income, was building envelope components, and of the HVAC systems of
slightly less than that of the previous cases without the reduc- the identified upgrades against the minimum efficiency level
tion, where the former was estimated to be $58,081 for as prescribed by the 2012 OBC.
Ottawa, $66,613 for Windsor, and $74,279 for Thunder Bay, It was concluded in the study by Dembo (2011) that the
respectively. Furthermore, the difference in the net present energy performance level of the Canadian Centre for Housing
value of the total electricity cost of what a homeowner would Technology (CCHT) Reference House (Ouazia et al., 2009;
pay if the PV system were not installed (hence the savings), Swinton et al., 2007), built in accordance with the existing
while accounting for the electricity production from the PV 2006 OBC was equivalent to the EnerGuide Rating of 78 using
system from 21 to 30 years, was slightly less than that of the the HOT2000 simulation program. In order to meet the
previous case without the reduction, where the former was proposed requirements of the 2012 OBC (the EnerGuide 80),
estimated to be $11,656 for Ottawa, $12,075 for Windsor, and the implementation of additional energy efficiency measures
$11,330 for Thunder Bay, respectively. Therefore, with the was mandatory for the same house to achieve, at minimum,
installation of the PV system, the total profits, over the 30-year two additional points on the EnerGuide scale, or 12% reduc-
mortgage period, were estimated to be slightly less than those tion in the total annual energy consumption. As it was the
of the previous case without the reduction, where the former intent of this paper to determine the most optimal solutions
were $69,737 for Ottawa, $878,688 for Windsor, and $85,609 that could be adapted by the large-volume builders in Ontario,
for Thunder Bay, respectively. the outcome of the paper was the indication that the energy
422
ASHRAE Transactions
Table 10. Comparison of the Potential Optimal Upgrades against the 2012 OBC
Zone 1 (Less than 5000 Degree-Days[1]) Zone 2 (5000 or More Degree-Days[2])
2012 OBC[3] Current Study 2012 OBC[4] Current Study

Component
Compliance Compliance Upgrade Upgrade Compliance Upgrade
Package Package Package Package Package Package
(A) (I) (1) (2) (A) (1)
Ceiling Attic Space 8.81 8.81 8.81 8.81 8.81 8.81

Minimum RSI (R)-Value (R50) (R50) (R50) (R50) (R50) (R50)
Exposed Floor 5.46 5.46 5.46 5.46 5.46 5.46

Walls Above Grade 4.23 3.87 4.23 3.87 5.11 5.11
Basement Walls 3.52 3.87 3.52 3.52 3.52 3.52

Below Grade Slab

0.88 2.11 2.11 0.88 2.11
Entire surface > 600 mm below grade –
(R5) (R12) (R12) (R5) (R12)
Minimum RSI (R)-Value
Edge of Below Grade Slab
1.76
 600 mm below grade – – – – –
(R10)
Minimum RSI (R)-Value
Windows and Sliding Glass Doors
1.6 1.8 1.6 1.6 1.6 1.6
Maximum U-Value
Space-Heating Equipment
90% 92% 90% 90% 90% 90%
Minimum AFUE
HRV
– 55% 70% 70% – 70%
Minimum Efficiency
DHW Heater
0.57 0.62 0.85 0.90 0.57 0.90
Minimum EF
Notes:
Weather condition for Toronto, ON, was used.
Weather condition for Thunder Bay, ON, was used.
Refer to the Table 2.1.1.2.A (MMAH, 2010)
Refer to the Table 2.1.1.3.A (MMAH, 2010)
efficiency level of the proposed 2012 OBC could well be new housing constructions to achieve significant energy
achieved by using the existing conventional new housing savings.
construction practices, and thereby expecting a savings in The objective of this paper was to identify the optimal
return with the implementation of the additional upgrades over upgrade solutions that could meet various levels of thermal
the lifetime of the ownership of the home or the mortgage performance while considering the use of local materials and
period. resources, ease of construction, but most importantly, adop-
tion by the Canadian housing industry through the optimiza-
CONCLUSION tion in energy related costs using a brute force sequential
search (BFSS) method.
It can be concluded based on numerous studies that most Potential energy efficiency upgrade solutions consid-
currently practiced new housing constructions in Canada are ered in this paper were identified as either implemented or
not as close to the optimal levels in terms of energy perfor- considered for implementation by the local homebuilders in
mance for the same overall cost to the homeowners over the Ontario into their current practices of building new homes, as
lifetime of the ownership or mortgage period. One of the it was evidenced by the results of the survey. However, as the
immediate solutions is to develop potential energy efficiency requirements of both the national and provincial building
upgrades that can be implemented into the currently practiced codes continue to change, some of the identified optimal
©2012 ASHRAE 423

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---

upgrades considered in this paper may have very minimal financial incentives and rebates, this will not only allow the
effect on the overall reductions in the energy consumption homebuilders and manufacturers adapt to the implementation
and/or the GHG emissions; for instance, RSI 5.11 (R29) of these technologies into their current practices without the
exterior wall insulation was identified as the optimal burden to pay for the additional expenses, but also, potentially
upgrade, not only in Thunder Bay, but also Ottawa and Wind- reduce the overall costs down as the adaption by the local
sor. However, based on the results of the survey, none of the homebuilders and manufacturers becomes the norm in
participant builders considered the implementation of any constructing new houses that are far more energy efficient than
insulation material with nominal RSI of 4.75 (R27) or higher the recently built houses, leading to the successful reduction in
into their current building practices due to the potential the overall energy consumption and the GHG emissions from
increase in the material and installation costs, unless other- the residential sector in Canada.
wise required by the future building codes and/or the existing
energy efficiency labeling programs. Furthermore, an ACKNOWLEDGMENTS
upgrade with the large incremental cost was eliminated by
The authors would like to acknowledge Canada Mortgage
using the BFSS method, despite the fact that such an upgrade
and Housing Corporation (CMHC) External Research
had the most reduction potentials in the overall energy
Program (ERP), Natural Sciences and Engineering Research
consumption and/or the GHG emissions, as shown in
Council (NSERC) of Canada, Ryerson Centre for Urban
Figure 10. Such an upgrade was the improvement in the exte-
Energy (CUE), and Mr. John Overall of Union Gas Ltd. for
rior wall insulation by using the SIPs, the insulated concrete
their financial support.
forms (ICFs) for the basement wall, high-efficiency boiler
and solar-assisted DHW heater for both the space and DHW The authors would also like to acknowledge the following
heating. organizations for their participations in the survey: Aspen
Ridge Homes, Country Wide Homes, Heathwood Homes,
This raises a need for the government to support their
Mason Homes, Rosehaven Homes, and Terra Brook Homes.
local homebuilders and manufacturers in a form of financial
incentives and rebates to offset the high initial (material and
GLOSSARY
installation) costs associated with the implementation of new
or unconventionally used technologies. With the support of the C1 RSI 8.81 (R50) Ceiling Insulation
Figure 10 Total cost in NPV ($) vs. increase in energy efficiency (%) reaching the net zero energy (Case in Windsor, using the
extreme economic parameters with 2006 natural gas prices).
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
424
ASHRAE Transactions
C2 RSI 10.57 (R60) Ceiling Insulation tions to Achieve Improved Energy Efficiency Standards.
AGW1 RSI 3.87 (R22) Above Grade Wall Insulation MASc Thesis, Ryerson University, Toronto, ON.
AGW2 RSI4.22 (R24) Above Grade Wall Insulation Dembo, A., Khaddad, F., & Fung, A. S. (2012). Least Cost
Upgrade Solutions to Achieve Improved Energy Effi-
AGW3 RSI 4.75 (R27) Above Grade Wall Insulation
ciency Standards for Residential New Housing in Can-
AGW4 RSI 5.10 (R29) Above Grade Wall Insulation ada. American Society of Heating, Refrigerating and
AGW5 RSI 4.22 (R24) Above Grade Wall Insulation @ 600 Air-Conditioning Engineers (ASHRAE) Transactions
mm (24 in.) o/c 2012, Vol. 118, Part 1.
AGW6 RSI 4.57 (R26) Above Grade Wall Insulation @ 600 Dong, B., Kennedy, C., & Pressnail, K. (2005). Comparing
mm (24 in.) o/c life cycle implications of building of retrofit and
AGW7 RSI 5.10 (R29) Above Grade Wall Insulation @ 600 replacement options. Can. J. Civ. Eng. (32), 1051-1063.
mm (24 in.) o/c Environment Canada. (n.d.). Canadian Climate Normals or
Average 1971-2000. Retrieved March 1, 2010, from
AGW8 RSI 7.00 (R40) Above Grade Wall Insulation
Environment Canada: http://climate.weatherof-
BGW1 RSI 2.11 (R12) Below Grade Wall Insulation, Exterior fice.gc.ca/climate_normals/index_e.html
BGW2 RSI 3.5 (R20) Below Grade Wall Insulation Farahbakhsh, H., Ugurasal, V. I., & Fung, A. S. (1998). A
BGW3 RSI 3.5 (R20) Below Grade Wall Insulation, Exterior residential end-use energy consumption model for Can-
BGW4 RSI 3.87 (R22) Below Grade Wall Insulation ada. International Journal of Energy Research (22),
1133-1143.
BGW5 RSI 4.22 (R24) Below Grade Wall Insulation
Fung, A. S., & Gill, G. S. (2011). Energy and Environmental
BGW6 RSI 4.22 (R24) Below Grade Wall Insulation (ICFs) Analysis of Hot Water Systems: A Study for Ontario,
EF1 RSI 5.45 (R31) Exposed Floor Insulation Canada. ASHRAE Transactions, In Press.
EF2 RSI 5.10 (R29) Exposed Floor Insulation Gordon, C., & Fung, A. S. (2009). Hourly emission factors
BS1 RSI 2.11 (R12) Basement Slab Insulation from the electricity generation sector - A tool for ana-
BS2 RSI 3.5 (R20) Basement Slab Insulation lyzing the impact of renewable technologies in Ontario.
CSME Transactions, 33 (1), 105-18.
V1 HRV @ 70% Efficiency
Gorgolewski, M. (1995). Optimising Renovation Strategies
AC1 Air-conditioning @ SEER of 14 for Energy Conservation in Housing. Building and Envi-
GH1 Furnace w/ ECM @ 90% AFUE ronment, 30 (4), pp. 583-589.
GH2 DHW Heater @ 85% AFUE Ministry of Energy [MIE]. (2010). Ontario's Long-Term
GH3 DHW Heater @90% AFUE Energy Plan. Queen's Printer for Ontario.
Organization for Economic Co-operation and Development
GH4 Solar-Assisted DHW @ 85% AFUE
[OECD]. (2009). OECD Territorial Reviews: Toronto,
GH5 Solar-Assisted DHW @ 90% AFUE Canada OECD Publishing, ISBN 92-64-07940-8, p37
GH6 Drain Water Heat Recovery @ 55% Efficiency Ouazia, B., Barhoun, H., Haddad, K., Armstrong, M.,
Marchand, R., & Szadkowski, F. (2009). DESICCANT
REFERENCES EVAPORATIVE COOLING SYSTEM FOR RESI-
Anderson, R., Christensen, C., & Horowitz, S. (2006). Anal- DENTIAL BUILDINGS. 12th Canadian Conference on
ysis of Residential System Strategies Targeting Least- Building Science and Technology, (pp. 1-12). Montreal,
Cost Solutions Leading to New Zero Energy Homes. Quebec.
ASHRAE 2006 Annual Meeting Session: How Low Can Sambou, V., Lartigue, B., Monchoux, F., & Adj, M. (2009).
You Go? Low-Energy Buildings through Integrated Thermal optimization of multilayered walls using
Design, (pp. 1-6). Quebec City, Canada. genetic algorithms. Energy and Buildings, 41, 1031-
Christensen, C., Horowitz, S., Givler, T., Courtney, A., & 1036.
Barker, G. (2005). BEopt: Software for Identifying Natural Resources Canada [NRCan]. (2000). Energy in Can-
Optimal Building Designs on the Path to Zero Net ada 2000 [online]. Retrieved June 12, 2002, from http://
Energy. ISES 2005 Solar World Congress, (pp. 1-6). oee.nrcan.gc.ca/infosource/pdfs/SHEU_e.pdf
Orlando, Florida. NRCan. (2011). Improving Energy Performance in Canada.
CMHC. (2011, August). Housing Now - Greater Toronto Ottawa, ON: Her Majesty the Queen in Right of Canada.
Area (previously Toronto). Retrieved September 9, 2011, NRCan. (2010). HOT2000 Procedures Manual.
from Publications_and_Reports:https://www03.cmhc- Swinton, M. C., Manning, M. M., Elmahdy, A. H., Parekh,
schl.gc.ca/catalog/productDe- A., Barry, C., & Szadkowski, F. (2007). Field assess-
tail.cfm?lang=en&cat=70&itm=41&fr=1315582512906 ment of the effect of different spectrally selective low
Dembo, A. (2011). Least Cost Analysis for Canadian New emissivity glass coatings on the energy consumption in
Housing - Identifying the Most Cost-Effective Specifica- residential application in cold climates. 11th Canadian
©2012 ASHRAE 425

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---

Building Science and Technology Conference, (pp. 1- In our analysis, we did not consider any incentives since such
16). Banff, Alberta. incentives are very different from region to region and utility
Tse, H., Siddiqui, O., Fung, A. S., & Masoumi Rad, F. to utility and such programs usually do not last very long.
(2008). Simulation and Analysis of a Net-Zero Energy However, for example, in the province of Ontario we have a
Townhome in Toronto. Proc. 3rd SBRN and SESCI 33rd microFIT that the government pays CAN$0.802/kWh from
Joint Conference. Fredericton, 8 pages. PV-generated electricity for 20 years. We have a separate
financial analysis for the net present worth income just for
Verbeeck, G., & Hens, H. (2007). Life Cycle Optimization of this.
Extremely Low Energy Dwellings. Journal of Building
Physics, 31 (2), 143-177. Manoj Choudhary, Senior Technical Staff, Owens Corn-
ing, Granville, OH: What are the federal/provincial policies
and programs for greenhouse gas reduction and energy effi-
DISCUSSION ciency in Canada?
Hugo Hens, Professor Emeritus, University of Leuven, Fung: In our study we did not consider greenhouse gas reduc-
and BPh Consult bvba, Leuven, Belgium: Did you consider tion policies. However, we did have the estimation of green-
injection, replacement costs, and maintenance costs? Be care- house gas emission reduction potentials due to the improved
ful with energy. Starting from a low or high energy cost is not efficiency. For the energy efficiency policy in Canada, there is
really correct.... Insulation adds secondary construction costs a new building code standard in the province of Ontario where
such as larger roof surface (reason: thicker walls), more the energy efficiency level equivalent to Energuide 80 is
complex window reveals, etc. Finally, take care with incen- required. One of the objectives of our project is to seek the cost
tives. optimal upgrade “package” to meet this new standard and the
proposed standard in 2016.
Alan S. Fung: In our study, all envelope-related upgrades
Mehmood Ismail, Mechanical Engineer (Project
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
(such as insulation) were assumed to last as long as the build-

ings (in our case, a 30-year mortgage period was assumed) Manager), Research in Motion, Burlington, Ontario,
while all mechanical equipment (such as the furnace, central Canada: What is the payback period of upgrading insulation
air conditioner, hot-water heaters, etc.) would have defined over building code for obtaining higher energy savings?
life expectancy and corresponding injection of replacement Fung: We did not use the measure of payback period for our
costs. In our study, the potential insulation upgrades were analysis. Instead we used life-cycle cost analysis to rank the
identified from a builder survey and they did not include potential upgrade or combination of upgrades.
extremely thick insulation types that required major structural Richard Ferguson, Engineer, Toyota, Erlanger, KY: How
changes to conventional construction practice. We have does square footage impact the cost/payback?
confirmed many of these issues and costs with a few major Fung: Since this is only a case study for “typical” new single
builders in the Greater Toronto Area. For potential window detached housing in Canada, a representative CCHT House
upgrades, the local builders could not reach a consensus on the located in Ottawa was used for the analysis. It should be noted
upgrade costs; therefore, we did not consider such potential that the majority of new low-rise housing in Canada is two-
upgrades in our analysis. All the identified potential upgrades story single detached type with approximately 2000 ft2 of
were selected based on the survey consensus from local build- living area, excluding the basement. However, our methodol-
ers, and potential upgrade costs were checked and verified ogy can be extended to any type of low-rise housing, including
with multiple sources. row or town houses.
426
ASHRAE Transactions
SA-12-008
Reliability Modeling for Data Center

Cooling Systems
Bob Schuerger, PE Gardson Githu, PE

Bob Arno Christopher Kurkjian, PE
Member ASHRAE
ABSTRACT Mean time between failures (MTBF) is the mean (aver-

The purpose of this paper is to provide a broader under- age) exposure time between consecutive failures of a compo-
standing of Reliability Engineering and how it can be success- nent. [1]
fully used as a tool for designing better mechanical cooling Electrical components and many other types of equip-
systems. We will also point out a few of the pitfalls when using ment have a relatively constant failure rate over much of its
reliability modeling to be avoided along the way. The use of useful life. Some types of mechanical components exhibit fail-
probabilistic methods of predicting failure developed in Reli- ure distributions that are not constant. Bearings, pump impel-
ability Engineering has greatly assisted the design of reliable lers and other components that exhibit wear typically have an
electrical power distribution systems for data centers. These increasing failure rate with time. However, the parts that wear
same proven techniques have been applied to the mechanical may get periodically replaced, so that the whole assembly
cooling systems. A review of the metrics of Reliability Engi- (pumps, chillers, cooling towers, etc.) has a relatively constant
neering will also be included. failure rate.
The common assumption for reliability analysis is that all
RELIABILITY TERMINOLOGY AND METRICS the equipment in the system to be analyzed has failures that are
random and the failure rate is constant. The calculations
Availability (A): Availability is the long-term average provided in this paper assume a constant failure rate for all the
fraction of time that a component or system is in service and equipment.
satisfactorily performing its intended function. [1] For exam-
For the case of a constant failure rate:
ple, if the electricity is off for one hour in a year, but the rest
of the year the electricity is on, the availability of electrical MTBF = 1  
power for that year is 8759 hours divided by 8760 hours, which
is 0.999886. Mean time to repair (MTTR) is the mean (average) time
An availability of 0.99999 could mean that the system to repair or replace a failed component. Logistics time asso-
was down for 5.3 minutes (or 315 seconds) per year. It would ciated with the repair, such as parts acquisitions, crew mobi-
make no difference in the availability calculation if there was lization, are not included. [1]
one 5.3 minute outage, or 315 one-second outages. It could Availability is mathematically defined as the mean time
also be one outage of 1.77 hours in 20 years. In all three cases, between failures divided by the mean time between failures
the availability is 0.99999. plus the mean time to repair:
The failure rate (  ) is defined as the mean (arithmetic A = MTBF   MTBF + MTTR 
average) number of failures of a component and/or system per
unit exposure time. The most common unit in reliability anal- Reliability (R) is the probability that a component or
ysis is hours or years. Therefore, failure rate is expressed in system will perform required functions under stated condi-
failures per hour or failures per year. [1] tions for a stated period of time without failure. [1]
Bob Schuerger is a principal reliability analysis corporate lead, Gardson Githu is a senior mechanical engineer, Bob Arno is a principal, and
Christopher Kurkjian is a chief technologist at HP Critical Facility Services in Los Angeles, CA, Frankfort, NY, and Albany, NY.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
©2012 ASHRAE
427
Table 1. MTBF of Outages Examples
Number outages Failure rate MTBF MTTR Reliability

Availability
per year (failure/hour) (years) (hours) (1 year)
0.99999 315 3.60E-02 0.0032 0.00028 0%
0.99999 1 1.14E-04 1.0 0.0883 36.78%
0.99999 0.05 5.71E-06 20 1.77 95.12%
For a constant failure rate  , reliability as a function of Reliability Block Diagram: An RBD is a block diagram in
time R(t) is: which the major components are represented by blocks that
are connected together in the same manner as they are in a one-
R(t) = e –t line or piping diagram. Each of the blocks have the failure and
As shown in the equation above, reliability is time depen- repair data for that component included in the block. The junc-
dent. The longer the time, the lower the reliability, regardless tions connecting the block are set according to the system
of what the system design is. The better the system design, the redundancy (e.g. “one out of two” when there are two compo-
higher the probability of successful operation for a longer nents and only one is required to carry the load). Quantitative
period of time. results (reliability, availability, etc.) for the RBD are obtained
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Table 1 shows three examples of Availability = 0.99999. by performing the series and parallel combinations of the
The first is a system that failed 315 times a year for one second blocks. RBD is a very good methodology for analyzing
(top row). The second is one that failed once in a year for 5.3 systems that have an obvious “flow” to them. In this paper,
minutes (middle row). The last is one that failed once in 20 RBD will be use to analyze the reliability of the electrical
years for 1.77 hours (bottom row). power to the mechanical cooling system and to analyze the
“Reliability is the “probability of success” for a given reliability of the equipment and configuration of the mechan-
period of time. Reliability is a metric directly related to how ical cooling system.
often (or how fast) the system fails. The last column in Table 1 Shown in Figure 1 is a RBD for a simple chilled water
shows how very different the Reliability is for these three cooling system consisting of a water cooled chiller, with its
systems, even though all three have the same Availability. associated cooling tower and pumps, and an air cooled chiller
From the equations shown above and Table 1, we see that as back up (see Figure 5 for the associated piping diagram).
there are five important factors to define the “reliability” of a Either the water or air cooled chiller are capable of carrying the
system; MTBF, MTTR, availability, reliability and time. It can load, which in this RBD consists of Computer Room Air
also be seen how these five factors are interrelated. Handling (CRAH) units. Eight of the ten CRAH are capable
of providing the required cooling in the computer room.
RELIABILITY ENGINEERING METHODOLOGIES
The purple junctions with “1:1” are input junctions that
FOR THE ANALYSIS OF MECHANICAL
connect the blocks for the two possible paths, the “1:2” junc-
COOLING SYSTEMS
tion determines that 1 of the 2 blocks is required for “success”
The first thing to understand about Reliability Engineer- and the “1:2Sb” is a standby junction. The standby junction
ing is that it uses “stochastic methodologies,” which means it sets which one of the two possible paths to use as “preferred,”
uses statistical methods to analyze processes that can have and that only one of the two sources is required.
many “answers” and predict the probability of the occurrence
of each of those “answers.” This is significantly different than In the RBD, the Makeup water is used before the storage
most of the methodologies in engineering. The vast majority tank and the water cooled chiller is used before the air cooled
of engineering calculations are “deterministic methodolo- chiller.
gies,” which calculate a single “answer” for a given set of Fault Tree Analysis is a top-down approach, where an
conditions. Determining how much water a certain size pump undesirable event is identified as the “top event” in the “tree”
with a specific head pressure will move is a deterministic and the potential causes that could lead to the undesirable
calculation. Determining the probability that the pump will event are identified as “branches” below it. Boolean Algebra
still be available after it has been running for two years is a is used to connect the potential causes of failure in the
stochastic process. branches to other branches and the top event. Figure 2 shows
There are a number of methods that can be used to the symbols used in the fault tree. An “OR” gate requires only
perform reliability analysis. We are going to discuss three of one of the inputs below it to be true for the output to be true.
them; Reliability Block Diagram (RBD), Fault Tree Analysis An “AND” gate requires all of the inputs to be true for the
(FTA) and Failure Modes, Effects and Criticality Analysis output to be true. Initiating events and undeveloped events are
(FMECA). both the basic cause of the failure; undeveloped events are
428
ASHRAE Transactions
Figure 1 RBD of water and air cooled chilled water plant.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
©2012 ASHRAE
429
ure for the electrical power as loss of power to the critical loads
being modeled is easily doable.
With the loss of mechanical cooling, the effect is seldom
immediate, since the critical load has to overheat sufficiently
that it impacts its operation. In the case of a sparsely populated
computer room with legacy IT equipment, it may take quite
Figure 2 Fault tree symbols: OR, AND, initiating event and some time for the IT equipment to overheat to a point where
undeveloped event. it significantly impacts the operation. However, with blade
servers closely packed into a few racks, the equipment can
overheat in a matter of a few minutes.
areas in which the fault tree could be expanded into more Therefore in using reliability analysis to evaluate designs
detailed analysis. for mechanical cooling systems, the factor of how fast the crit-
If the failure rate and repair data is available for all of the ical equipment overheats should be included in the analysis.
initiating and undeveloped events in the Fault Tree, quantita- In the Reliability Block Diagram method failure is
tive results (reliability and availability) can be calculated for defined as loss of adequate cooling to the critical loads
the “top event” and each of the branches. FTA is an excellent (usually IT equipment). By adequate cooling we mean the
tool to analyze specific failures that have critical importance to number of components required to cool the critical loads. If
the 7X24 Facility. By working backwards from the failure to there are four Computer Room Air Handling (CRAH) Units
required to cool the computer room and one of the four fails,
be prevented, down to all of the items that could cause this
that would constitute a failure of the mechanical cooling
particular failure, interfaces between equipment and systems
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
system.
can be brought to light that may be overlooked using other
The RBD analysis will only show how likely the system
types of analysis. FTA can also be a very effective reliability
is to fail. It does not address how quickly the loss of cooling
tool for qualitative analysis for items such as human errors that
becomes a failure for the critical IT processes. That requires
cannot be easily quantified with failure and repair rates. For another type of analysis (air/heat flow modeling), called
example, a fault tree can be used to show the interrelationship Computational Fluid Dynamics (CFD). CFD modeling is not
between the automatic and manual controls on a mechanical what is normally considered “Reliability Engineering,” since
cooling system, as shown in Figure 3. it is a “deterministic” methodology. CFD is often used in
Failure Modes, Effects and Criticality Analysis conjunction with reliability analysis, to determine the actual
(FMECA) can be done on the components of the equipment to redundancy of the design and to evaluate how critical the fail-
be analyzed or the steps of a procedure to for a process. To ure is in terms of impact to the IT equipment.
analyze the equipment, it would consist of examining individ- Another pitfall in using reliability analysis to evaluate
ually and collectively each of the components in the system to designs is that manual intervention by an operator is very diffi-
determine the effect when one or more components fail. This cult to address in the reliability analysis. The functions that are
is a bottom-up approach: each of the components is examined, automatic are much easier to take into account.
all of the ways it can fail are listed (failure modes) and the The importance of manual intervention by an operator is
effect of each failure to the components itself and on the over- directly related to how fast the equipment overheats. There-
all system is predicted. Then a criticality level is assigned for fore, in selecting the best system for a particular application it
each failure mode, base on the overall effect on the system. An is important to remember these two factors about the analysis.
interdisciplinary team is required and it is time consuming in If the equipment overheats very quickly, the importance of
direct proportion to how thorough and to what level of detail having a very reliable mechanical cooling system will be high.
the analysis is taken. If the system takes a significant amount of time to overheat, a
less reliable system may be acceptable, since manual interven-
Table 2 shows a partial FMECA for a chilled water tion by an operator can be part of the mitigation strategy.
system. The example takes a high level look at the major fail-
ure modes and what the effect they would have on the overall RELIABILITY OF MECHANICAL SYSTEMS
data center. The criticality of the failure is also listed in the
The reliability analysis has been performed for the chilled
“Severity of failure” column.
water plants shown in Figures 4 and 5. Figure 4 is a chilled
water cooling system that consists of an “N+1” design in
IMPORTANT FACTORS AND PITFALLS IN
which there is a redundant component for each of the major
RELIABILITY ANALYSIS FOR MECHANICAL
equipment. In this design there are two cooling towers,
COOLING SYSTEMS
condenser water pumps, chillers, and primary chilled water
The first pitfall is a subtle one: What is the definition of pumps providing chilled water to ten Computer Room Air
failure? Handling (CRAH) units. Eight of the ten CRAH units are
In doing reliability analysis for electrical power to the required to provide adequate cooling to the data center.
critical loads, a loss of power to the critical load causes an Figure 5 is also an “N+1” design, but an air cooled chiller
immediate failure of that critical load. Therefore defining fail- has been used to replace one of the water cooled chillers. In
430
ASHRAE Transactions
Figure 3 Partially developed fault tree of chilled water cooling system.
this design there are a cooling towers, condenser water pumps RELIABILITY OF POWER
and water cooled chiller plus an air cooled chiller providing TO MECHANICAL SYSTEMS
chilled water to two primary chilled water pumps and ten The mechanical design is only one part of the reliability
Computer Room Air Handling (CRAH) units. Eight of the ten equation for the cooling system. Electrical power is always a
CRAH units are required to provide adequate cooling to the SPOF for the cooling system; if power is lost the cooling
data center. system is lost.
Table 4 shows the reliability calculations for three exam-
The reliability of both designs is quite good (Table 3). The ples of how the power to the air cooled and water cooled
significant increase in the reliability with the air cooled chiller chilled water system could be configured.
is due to the fact that loss of water is no longer a single point For the first example, a standby generator provides back
of failure (SPOF) to the cooling system, as it was with the up to utility power. There are two Automatic Transfer
design using two water cooled chillers. Switches (ATS), one to provide power for the air cooled chiller
©2012 ASHRAE 431

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---

432
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Table 2. Example of partial FMECA for Chilled Water Plant

Failure Mode Severity of failure Component(s) Failure Local Effect of Failure Next Effect of Failure End Effect of Failure
Sustained loss of
Provided by IHS Markit under license with ASHRAE

Major chilled water pipe rupture, mul- Affected equipment out of Cascading failures of the All IT equipment overheats, shuts
mechanical cooling sys- I. Catastrophic
tiple equipment failures service chilled water plant down; data center complete failure
tem
No reproduction or networking permitted without license from IHS

Sustained loss of power
Electrical power for Mechanical sys- Affected equipment out of Cascading failures of the All IT equipment overheats, shuts
to the mechanical equip- I. Catastrophic
tems RBD service chilled water plant down; data center complete failure
ment

Part or all of the data center Some or all of IT equipment over-
Loss of air flow from Temperature rises above 95°
II. Critical CRAH, AHU or power to them without cooling other than heats, shuts down; depending on
CRAHs or AHUs F in IT room
fans in IT equipment how long without air flow
BMS failure, fewer CT/chillers/pumps/ Some of IT equipment overheats,
Temperature above 95° Part of the data center with- Temperature continues to
II. Critical CRAH operational than required for shuts down; depending on how
F in IT room out required cooling rises in IT room
load long without required cooling
Some of IT equipment overheats,
Part of the data center with- Temperature continues to
Loss of water flow III. Serious SCHWP or PCHWP or power to them shuts down; depending on how
out required cooling rises in IT room
long without required cooling
Operator intervention
Central plant on manual Equipment operating at the Results depend upon operator
III. Serious BMS failure required for any changes in
control last set point provided actions
plant
Loss of redundancy Results depend upon whether sys-
Affected equipment out of None - provided equipment is
(system is now "N" for a III. Serious Redundant system failed or power to it tem is restored to normal before a
service repaired before next failure
critical path) significant failure
Set point incorrect, BMS failure, con-
Not for Resale, 12/29/2018 20:39:41 MST

Results depend upon whether sys-
IT room air temperature sistent upsets to central plant causing Some IT equipment may be Equipment operating above
IV. Important tem is restored to normal before a
outside of set point fewer CT/chillers/pumps/CRAH oper- operating above 95° F 95° F shuts down
significant failure
ational than required for load
Chilled water tempera- Set point incorrect, BMS failure, con-
Results depend upon whether sys-
ture to the IT room consistent upsets to central plant causing Some IT equipment may be Equipment operating above
IV. Important tem is restored to normal before a
sistently outside of set fewer CT/chillers/pumps/CRAH oper- operating above 95° F 95° F shuts down
significant failure
point range ational than required for load
Licensee=Hong Kong Polytechnic University/9976803100, User=WONG, Kit Chun

None - provided equipment is
Loss of redundant com- Results depend upon whether sys-
Redundant component failure or power Affected equipment out of repaired before another fail-
ponent in multiple IV. Important tem is restored to normal before a
to component service ure occurs to a system operat-
redundant systems significant failure
ing at N
ASHRAE Transactions
Figure 4 Two water cooled chillers.
and associated primary chilled water pump and the other to CONCLUSIONS
proved power to the cooling tower fans, condenser water
From the discussion above, it is obvious that reliability
pump, water cooled chiller and associated primary chilled engineering can be a useful tool in designing or evaluating a
water pump. There are also two power panels, each fed from cooling system for a critical facility.
one of the ATS and supplying power to five CRAH units. The To fully specify “reliability” requires five major metrics:
reliability of this system is quite low, since loss of either ATS MTBF, MTTR, availability, reliability and time. These metrics
or power panel causes five CRAH units to lose power. The are significantly impacted by what the definition of “failure”
system design requires eight CRAH units to provide sufficient is for the system to be modeled.
cooling, so loss of five is a failure for the RBD. The reliability of the system is also significantly impacted
For the second example, the standby generator, ATS, and by the size of the facility. The larger the facility, the more
power to the two chilled water systems is the same as the first equipment there is that can fail. Therefore two identical
designs will have significantly different reliability metrics if
example. For this example each of the CRAH units have two
one facility is twice the size of the other.
contactors installed in such a manner that two sources of
Reliability modeling can be a very effective tool when
power can be provide either one of which can provide power
used for comparison between mechanical systems. Under-
to the CRAH unit. The reliability of this system is much better,
standing of the basic concepts involved is necessary to
since now the loss of an ATS or panel does not cause the loss correctly model the systems and utilize the data provided to
of five CRAH units. reach the proper conclusions.
For the last example, a second standby generator has been
added, making the electrical system providing power to the REFERENCES
mechanical cooling system 2N. With this design, it is very 1. IEEE Gold Book, Standard 493-1997/2007 IEEE Recom-
unlikely for electrical power to cause loss of the cooling mended Practice for the Design of Reliable Industrial
system. and Commercial Power Systems
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
©2012 ASHRAE
433
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 5 Air cooled and water cooled chillers.
Table 3. Reliability Calculations for Chilled Water Systems
Probability of Failure in 5
Description of RBD MTBF (years) MTTR (hours) Availability
years
Two Water Cooled

72.2 7.3 0.9999885 2.39%
Chillers
Air + Water Cooled Chiller 307.2 10.0 0.9999963 1.38%
Table 4. Reliability Calculations for Electrical Power to Chilled Water Plant
Probability of Failure in 5
Description of RBD MTBF (years) MTTR (hours) Availability
years
Electrical Power to Air + Water

5.1 2.4 0.9999463 62.02%
Chiller System
Electrical Power to Air + Water

Chiller System - 2 sources to 24.9 1.8 0.9999918 17.59%
CRAH
Electrical 2N Power Air + Water

236.4 0.7 0.9999997 2.25%
Chiller - 2 sources to CRAH
434
ASHRAE Transactions
2. “What Five 9’s Really Mean and Managing Expectations,” 4. “Reliability of Example Mechanical Systems for Data
by Robert Arno, Peter Gross, PE and Robert Schuerger Center Cooling Selected by Tier Classifications,” by
PE, IEEE Industry Applications Society Conference Robert Arno, Gardson Githu, PE, Peter Gross, PE and
2008 Robert Schuerger PE, and Scott Wilson PE, IEEE Indus-
trial Applications Society Conference 2010
3. “Reliability of Example Data Center Designs Selected by 5. “Reliability Engineering Applied to Critical Operations
Tier Classifications,” by Robert Arno, Addam Friedl, Power Systems (COPS),” by Michael Anthony PE, Rob-
PE, Peter Gross, PE and Robert Schuerger, PE, IEEE ert Arno, Mark Beirne, Patrick Saad Saba and Robert
Industrial and Commercial Power Systems Conference Schuerger, PE, IEEE Industrial and Commercial Power
2010 Systems Conference 2011
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
©2012 ASHRAE
435
SA-12-009 (RP-1487)
The Development of Simplified

Rack Boundary Conditions for
Numerical Data Center Models
John Z. Zhai, PhD Knud A. Hermansen Salah Al-Saadi

Member ASHRAE Associate Member ASHRAE Associate Member ASHRAE
This paper is based on findings resulting from ASHRAE Research Project RP-1487.
ABSTRACT approximately 61 billion kWh)1. As current computing trends

As cloud computing and computational needs grow, data place more emphasis on cloud computing and data storage this
centers will continue to become a larger part of our energy percentage is only going to increase. The layout and design of
load. Proper design and layout are crucial to efficient energy a data center can make a significant difference in its energy use
use in data centers. Modeling the rack is one of the critical and the consequences of improper data center design can be
pieces in the design process. Often this is done as a black box dramatic. Cooling energy in poorly designed data centers can
rather than modeling the rack in detail. Modeling a computer constitute up to 50% of its energy use2. Computational Fluid
rack as a black box has been done in numerous data center Dynamics (CFD) plays an important role in aiding the layout
studies, but rarely has it been validated against experimental design and management of data centers. While the use of CFD
temperature and velocity data. This study looks at two simpli- modeling is common in data center design, there are some
fied rack models and compares them against a rack composed important issues that need to be resolved. Due to the fact that
of four 10U server simulators. The first model is an open box many data centers are considered mission critical facilities, it
model that has a heating and fan plate and allows air to flow is often hard to generate real data sets against which to
through the rack. The second model is a black box model that compare models3. This also makes it extremely hard to test
allows no flow through the rack and imposes a constant pres- hypotheses under varying conditions since few data center
operators are willing to allow their data centers to undergo any
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
sure boundary across the inlet and exhaust. The black box
model adds the rack load from the upwind cells at the rack inlet experimentation which could potentially disrupt operations.
plate to generate the exhaust temperature profile. Both models Most modelers use some form of a black box to simulate
were tested across a range of airflows and rack loads. Agree- the rack. In the black box model, room inputs are put into the
ments were found to be within 5.4°F (3°C) and 40 fpm (0.2 m/ front (inlet) of the rack and the added enthalpy outputs come
s), on average, over all experiments. An interesting finding of from the back of the rack (outlet). As discussed above, one of
this study was the importance of correctly capturing the bound- the central issues with this approach is the question of which
ary conditions at the perforated floor tile. Modeling the perfo- boundary conditions for the rack produce acceptable accuracy.
rated floor tile as a nozzle diffuser using the momentum method The goal of this project was to develop a set of easily repro-
described in ASHRAE RP-1009 was found to produce accept- ducible boundary conditions and validate them against sets of
able results for airflow from the perforated floor tile. experimental data. This project did not address cooling within
INTRODUCTION 1.
(EPA, 2007)
2.
A 2007 study by the EPA found that 1.5% of the nation’s (Uptime Institute, 2011)
3.
energy use was being used by data centers (consuming (Shrivastava, Iyengar, Sammakia, Schmidt, & VanGilder, 2006)
John Z. Zhai is an associate professor, Knud A. Hermansen is a master’s candidate, and Salah Al-Saadi is a PhD candidate in the Department
of Civil, Environmental, and Architectural Engineering at the University of Colorado at Boulder, Boulder, CO.
436
ASHRAE Transactions
the racks or efficiency of cooling systems. These are issues for Among the major internal components of the server simu-
overall data center design and this project is simply trying to lators were the fans and heating element. The fans were
address how an individual component of this overall model, arranged in a two-by-two array with each fan approximately 6
the rack, can be modeled with sufficient accuracy and effi- in (15.2 cm) in diameter (See Figure 2(b)). Combined, the fans
ciency in the larger model. for each server simulator drew approximately 150 W. Each
server simulator set of fans could move between 200-650 cfm
LABORATORY (0.094-3.07 m3/s) of air. The heating element in each server
simulator consisted of seven vertical heating fins
Experiment Setup (Figure 2(b)). Each server simulator heating element could
This project considered a single rack populated with four output up to 5.75 kW in 250 W increments. Each server simu-
server simulators in an isolated test chamber. Figure 1 shows lator heating element had its own dedicated 208 V circuit.
the set-up of the test chamber as viewed from the entrance, or Figure 2(c) shows the controls for each server simulator. For
the “front-right” corner as well as the dimensions. The back the fans in particular it should be noted that the number on the
and right walls were R-50 and the left and front walls were dial corresponded to a flow rate that was both in manufac-
dark plastic sheeting attached to wood framing. The ceiling turer’s data and independently verified, but that this dial did
and the floor were sealed to minimize air infiltration / exfil- not lock in but was set by sight.
tration.
Air was supplied by means of a 24 in (61 cm) square Data Acquisition
perforated floor tile with a 25% open area. The air was deliv- Table 1 summarizes the data acquisition tools used in this
ered to the perforated floor tile via a 15 in (38 cm) raised floor project, their accuracies and purpose.
plenum (RFP), which in turn had its air delivered via the duct
The data acquisition for this project was broken into two
seen in Figure 1. The return air grille was also two foot (61 cm)
categories:
square located approximately 6 inches (15 cm) behind the
back of the rack.
1. Inputs required for the boundary conditions of the CFD
For this article, the rack refers to the entire cabinet that
model.
enclosed the server simulators. The rack itself was 24 in
(61 cm) wide, 3 ft 6 in deep (1.07 m) and 7 ft (2.13 m) high. 2. Parameters to validate the CFD model.
The front and rear grills had approximately a 60% open area.
The server simulators were 10U (17.4 in or 44.2 cm) in height
and the rack was capable of mounting four server simulators.
Each server simulator was approximately 19 in wide (48.3 cm)
and 27.5 in (69.9 cm) deep. When mounted in the rack there
was approximately a 3 in (7.6 cm) gap between the front of the
server simulators and the front grille of the rack and a 1 ft
(30.5 cm) gap between the back of the server simulators and
the back grille of the rack. Figure 2(a) shows the rack.
Figure 2 Various aspects of the rack and server

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,
Figure 1 The Test Chamber Set-up as viewed from the simulators: (a) rack (b) fans and heating coils (c)
“front-right” of the room. controls.
©2012 ASHRAE
437
In order to establish the boundary conditions for the CFD pole measurements were taken around the room (Figure 3(a)).
model the following parameters were measured: Each temperature pole consisted of five to seven type-T ther-
mocouples (Figure 3(b)) spaced evenly along the vertical
1. Supply Air Flow Rate length of the 6.5 ft (2 m) pole.
2. Supply Air Temperature Air velocities in the model were validated by four hot-
3. Flow rate through the rack wire anemometer poles whose locations are shown in
4. Rack power (and distribution) Figure 4(a). Each velocity pole consisted of four anemometer
measurements, spaced at approximately 15 in (38 cm) vertical
The supply air flow rate was measured using a flow hood increments. Figure 4(b) shows one of the anemometer poles.
while supply air temperature was measured using a type-T Since there were only two, each one would take measurements
thermocouple. The flow rate through the rack was determined at 60 Hz for five minutes and then each be moved to the next
by taking hotwire anemometer measurements over an evenly vertical location.
distributed area over of the front of the rack. For this measure-
ment the rack was tested at a series of fan settings that corre- EXPERIMENTS
sponded to the numbers one through ten on the control
(Figure 2(c)). The flow rates from this experiment were used Supply Air Quantification and Other Experiments
to set the flow rates in the models for all subsequent experi- One interesting aspect that came out of the initial exper-
ments. For the rack power, the settings on the front panel were iments and models was the importance of correctly modeling
verified via a power quality analyzer at the panel box. the supply air. This applied not just to the model, but also in
The CFD model was validated against temperature and taking accurate measurements with the flow hood. In order to
velocity pole measurements. For validation, ten temperature accurately model the performance of the perforated floor tile
Table 1. Primary Data Acquisition Equipment – Purpose and Accuracy
Equipment Purpose Accuracy*

Thermocouples, Type T Temperature measurements ± 0.9 °F (0.5 °C) or 4%
Flow Hood Measure supply air volumetric airflow ± 3%
Power Quality Analyzer Measure power drawn by server simulators ± 2.5%
Hot-Wire Anemometer Air velocity measurements ± 3% over ±3 ft/min (±0.015 m/s)

*Based on the equipment reading
Figure 3 (a) Location of temperature (thermocouples) poles (b) picture of one thermocouple pole.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
438
ASHRAE Transactions
Figure 4 (a) Location of velocity poles (b) picture of velocity pole.
it was first necessary to quantify the following variables that For this experiment, X could be assumed to be equal to or
categorize its performance: close to the height of the room and VX can be assumed to be
zero or close to zero. V0 can be measured by getting an average
• throw velocity measurement across the perforated floor tile using
• terminal velocity procedure described in ASHRAE’s Standard 70-2006 Method
• decay constant of testing for rating the performance of air outlets and inlets.
• turbulence intensity The trickiest value to get for this formula is the effective area,
A0. (Chen & Srebric 2001) note that this is not necessarily the
The most important of the variables to classify was the actual area. However, for this type of “diffuser”, the value for
throw. In these experiments terminal velocity was chosen as the perforated floor tile can be assumed to be close to the effec-
minimal detectable air movement in the room (essentially no
tive area or within 90%4.
air movement) – or the point at which there was no noticeable
effect from the jet. From chapter 33 of ASHRAE Fundamen- Using Equation 1 with values from experimentation for
tals, we have the formula for the decay constant in zone 3, the X, V0, VX and A0, gave a decay constant, K, of around 1.8.
main zone, in IP units (Equation 1). If one converts this to the While there was some uncharacterized uncertainty in this
formula for SI units and solves for the decay constant K, then value, this research was focused on developing simplified rack
we get formula (Equation 2). boundary conditions, not full characterization of supply air
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
conditions, so this number was considered acceptable for the

V 1.13  K  A
IP: -----X- = ---------------------------------0 (1) CFD models.
V0 X
V X Experiments
X
SI: K = --------------------
- (2)
V0  A0 Ten experiments were conducted to test the CFD models.
Of these ten, three will be discussed in this paper. Table 2 lists
Where:
these experiments. For a full list see the report for RP-1487.
• VX is the velocity at point X (in this case the terminal For all experiments listed here the rack load was evenly
velocity) [fpm, m/s] distributed across all sever simulators (i.e., in Experiment 1,
• X is the height above the jet [ft, m] each server simulator had a load of 1 kW).
• V0 is the initial velocity in the main zone [fpm, m/s]
• A0 is the effective area of the perforated floor tile [ft2, 4.
(Chen & Srebric, Simplified Diffuser Boundary Conditions for
m2] Numerical Room Airflow Models, ASHRAE RP-1009, 2001)
©2012 ASHRAE
439
Table 2. Select Validation Experiments
Fan Speed, Supply Air

Rack
Exp.
Load, Remarks
No.
kW cfm (m3/min) fpm (m/s) Temp, °F (°C) cfm (m3/min)
1 4 800 (23) 110 (0.56) 61.2 (16.2) 790 (22) medium powered rack with low fan speed
3 8 1920 (54) 209 (1.06) 53.6 (12.0) 804 (23) higher powered rack with medium fan speed
9 4 1120 (32) 138 (0.70) 60.8 (16) 840 (24) top two server simulators removed
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
MODELING was modeled as a multiple nozzle diffuser according to the

procedures outlined in ASHRAE RP-1009.
This team developed two distinct CFD models, an open
box model (OBM) and a black box model (BBM). Both Open Box Model (OBM)
models were designed to be simple and require minimal user
Figure 5 shows the layout of the Open Box Model (OBM).
inputs. All models were developed using commercially avail-
All sides of the rack were modeled as adiabatic plates. The
able CFD software. The OBM was developed first and its
front and back plates were modeled with a percent open area
purpose was to be an interim model to inform the development
which allowed restricted airflow to pass through. Both the fan
of the BBM. While it was very simplified in its detail, the
and heating plates were broken up to allow different heat
OBM was an approximation of the server simulator which
fluxes and flow rates for each server simulator. The fan plates
allowed for air to flow through the rack. The BBM, by
were given defined X velocities and were modeled without any
contrast, was a solid box. It took inputs at the rack inlet and
swirl. The heating plates were given a defined heat flux
outputted modified values at the rack outlet. Its assumptions
(divided among each server simulator) with one-half of the
were tested both against the experimental data and the OBM.
heat flux coming out of each side of the plate.
Both models shared the Boussinesq approximation and the ?-
? turbulence model used in the RANS equations (RANS, Black Box Model (BBM)
Reynolds-averaged Navier–Stokes equations, being the time
averaged equations that were used to model the conservation The rack inlet boundary conditions were defined by a
of energy, momentum and mass in the domain). velocity normal to the front rack plate (i.e. – the rack door) face
and a temperature profile. A uniform velocity was imposed
One parameter that was not considered was radiation. All normal to the plate. Smoke pen tests conducted during the
of the sources reviewed during the background research for experiment as well as validated models confirmed this as a
this paper looked at forced convection as the dominant form of reasonable assumption. At the rack inlet face this velocity was
heat transfer. Only one paper, (Rambo 2005) even mentioned defined by the fan speed divided by the porosity as shown in
it – and only to then to say that it was ignored to simplify the Equation 3. The same velocity condition was imposed at the
model. Parametric models run with radiation confirmed this. rack exhaust boundary plate.
Boundary Conditions V fan

V face = ------------------------------------- (3)
plate porosity
The normalized root mean squared error (NRMSE) was
used to analyze the results of different levels of meshes to find where V face is velocity normal to the rack inlet or exhaust at
the grid independent solution. Resolutions of 72,000, 244,800, the plate, V fan is the velocity imposed by the rack fans, and
576,000 and 1,150,000 cells were tried. Grid independence plate porosity is expressed as a percent open area.
was found at 244,800 cells with the average cell length coming At the rack inlet plate the temperature was taken from the
out to approximately 1.6 in (4 cm). All surfaces were modeled adjacent upwind cell. For the rack exhaust plate this temperate
as adiabatic surfaces and radiation models were not used. (at the same y and z coordinates) was taken with the appropri-
ate amount on enthalpy added, shown in equation 4.
While there is not sufficient space to discuss it in this
paper, the modeling of the perforated floor tile was found to be q server
very important for determining the velocity vector field in T ex = T in + ---------------
·- (4)
cp  m
front of the rack (and to a lesser degree, the temperature distri-
bution). Simply modeling the perforated floor tile as an inlet where Tex is exhaust temperature for cell on rack exhaust plate
was found to be insufficient. Instead the perforated floor tile [°C], Tin is temperature at rack inlet plate for same y,z [°C],
440
ASHRAE Transactions
Figure 5 Open box model.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 6 Black box model.
qserver is heat added by the server [W] (for BBM total heat ments. A full analysis of all results can be found in ASHRAE
generated by each server was assumed to be evenly distributed RP-1487.
over the server cross sectional area), cp – specific heat capacity
There was very little difference between the results for the
of air, and m· is mass flow rate across the cell.
Figure 6(a) shows the set-up for the BBM while open box model and the black box model. This was considered
Figure 6(b) shows the translation of temperatures and velocity to be a good sign since it indicates that mass and momentum
values from the front plate to the rear plate. decoupling required black box programming is not necessar-
ily required to produce good modeling results. The only caveat
RESULTS of using an open box model is that people who read the results
While ten total experiments were conducted and validated need to understand that while it allows airflow through the
against the open-Box and black box models, the results below rack for purposes of the simulation, it is not intended to give
are for three select experiments. These three were chosen to results for rack-internal airflow and heat transfer – only room
give a representative sample of the full range of rack experi- level results.
©2012 ASHRAE
441
Temperature and Velocity Agreement (Figure 8 (c)). These could possibly be resolved by a more
complete modeling of the perforated floor tile.
The average temperature agreement between all experi-
ments and model results were within 5.2 °F (2.9 °C), on aver-
Experiment 3: 8kW Rack, Fan Speed of 1.06 m/s
age. Velocity predictions across all models were found to be
within 40 fpm (0.2 m/s), on average. Experiment 3 is an 8kW rack with a fan speed of 1.06 m/
Temperature results shown in this paper are normalized s or a volumetric flow rate of 1920 cfm (0.906 m3/s). In spite
against the supply air temperature and the exhaust air temper- of a volumetric flow rate through the rack being around twice
ature. While temperature rise across a rack is a more familiar the flow through the perforated floor tile of 802 cfm (0.379 m3/
metric for those in the server industry, the paper chose to use s), temperature agreements are within 8 °F (4.5 °C). Table 4
the supply air temperature instead of the rack inlet temperature shows the average temperature differences between the
as the lower boundary. This is due to the fact that the supply models and the experimental results.
air will always be the lowest temperature in the room. Figure 9 shows that the biggest error came near one of the
Velocity results were examined in an absolute sense. This poles to the right of the rack and that it involved a consistent
was due to the fact that velocity was a much more difficult and under-prediction. This could to be linked to issues that these
uncertain measurement to take and therefore was considered models may have at resolving horizontal re-circulations at
more of a secondary comparison. higher rack velocities (as compared to perforated floor tile
flow). The other notable issue here is the under-prediction of
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Experiment 1: 4kW Rack, Fan Speed of 0.56 m/s inlet and outlet temperatures below 0.75 m (29.5 in). This
could be due to a slight over-prediction of the velocity exiting
The 4kW Rack had an even power distribution and a fan
the perforated floor tiles. This is a common issue with the
speed setting of 110 fpm (0.56 m/s). As Table 3 shows, all
momentum method for grille/nozzle diffusers as noted in
poles were within 5 °F (2.8 °C) on average. The differences
ASHRAE RP-1009.
listed in each column are the average differences, over the
entire pole for each point along the thermocouple pole with
Experiment 9: 4kW Rack, Fan Speed of 0.7 m/s, Top
what each model predicted at those locations. Figure 7 shows
Server Simulators Removed
normalized results for selected temperature poles.
For this experiment one can see that the poles on the left It is a common occurrence in data centers to have racks
and right show very good agreement. On the front and back of that are not fully populated with servers. While many manu-
the rack there is some under-prediction. For the front pole this facturers recommend blanking panels, not all data center oper-
under-prediction may be due to a slight under-prediction of the ators follow this. Experiments were conducted where the top
throw for the perforated floor tile. The effects of the under- two server simulators were removed and no blanking panels
predictions of the models for the rack-inlet pole may have were used as well as an experiment where blanking panels
carried over to some of the under-predictions on the rack outlet were installed. Experiment 9 shows the rack with the top two
pole. server simulators removed (Figure 10 (a)). Interestingly, it was
While there is not space to show velocity agreement observed during the experiment that there did not seem to be
results for all of the select experiments, it is worthwhile to any significant recirculation through the rack (Figure 10 (b)).
consider at least one set of results. Figure 8 shows the velocity With this observation in mind, this experiment was
pole comparisons for this experiment. In Figure 8 (a) one can modeled with the top portion of the rack as completely
see the possible issues with the throw calibration. In Figure 8 blocked. Table 5 shows that agreements were as good with this
(b) these issues start to resolve themselves but there are still experiment as with any of the other models, although the front
some issues at the 1.2 m (47 in) height at the rack inlet pole averages showed the largest gap in agreement.
Table 3. Experiment 1 Temperature Differences Between Models Experimental Results (°F)
Pole OBM, °F (°C) BBM, °F (°C)
F Avg 4.9 (2.7) 4.1 (2.3)
B Avg 4.3 (2.4) 4.5 (2.5)
L1 4.5 (2.5) 4.9 (2.7)
L2 1.8 (1.0) 2.5 (1.4)
R1 3.2 (1.8) 3.4 (1.9)
R2 2.7 (1.5) 4.1 (2.3)
442
ASHRAE Transactions
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 7 Selected temperature poles for Experiment 1.
This can also be seen in Figure 11 (a) where the lower rate. For the rack porosity a reasonable value of between 0.5
portion has an under-prediction while the top has a sharp over- to 0.75 can be assumed. The fan speed can be determined by
prediction of temperatures. The temperature profiles from the looking at the volumetric flow rate (from manufacturer’s liter-
experiment had a much lower jump in temperature increase ature) and the cross sectional area of the server. The velocity
than was predicted in the model. This could indicate the need on the inlet face is set uniformly across the rack inlet face as
for increased grid refinement near the front of the rack when the rack airflow rate divided by the porosity (or percent open
there are sudden jumps in pressures (from a servers with area) of the rack cabinet door as shown in Equation 3.
running fans to no servers at all). However, temperature poles Step 2: Set the velocity boundary condition at the rack
in the rest of the room showed good agreement overall. exhaust face. In order to conserve momentum, this is set as the
same uniform velocity condition which was specified for the
SIMULATION GUIDELINES rack inlet in step 1.
Since the black box rack model is the most common Step 3: Set the temperature boundary condition at the rack
method of simulating a rack in full-room data center CFD exhaust face. Here one takes the temperature from cell adja-
models, steps were taken to codify the procedures that were cent to the rack inlet cell at the same vertical and horizontal
found to be effective in this research. position on the rack inlet face and adds the appropriate amount
Step 1: Set velocity boundary conditions at rack inlet. In of heat based on the thermal load of the server. One must know
order to do this you need a reasonable estimate of the porosity the mass flow rate which can be determined from the volumet-
of rack cabinet door (percent open area) and the rack airflow ric flow rate of the fans in the server and the density of air for
©2012 ASHRAE
443
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 8 Selected velocity poles for Experiment 1.
Table 4. Experiment 3 Temperature Differences Between Models Experimental Results (°F)
F Avg 5.4 (3.0) 5.4 (3.0)
B Avg 5.9 (3.3) 5.6 (3.1)
L1 6.7 (3.7) 6.5 (3.6)
L2 5.6 (3.1) 4.3 (2.4)
R1 8.1 (4.5) 8.1 (4.5)
R2 3.6 (2.0) 6.3 (3.5)
444
ASHRAE Transactions
Figure 10 Experiment 9 set-up.

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
©2012 ASHRAE
445
Table 5. Experiment 9 Temperature Differences Between Models Experimental Results
F Avg 6.7 (3.7) 7.4 (4.1)
B Avg 4.7 (2.6) 3.8 (2.1)
L1 2.7 (1.5) 0.5 (0.3)
L2 5.6 (3.1) 8.8 (4.9)
R1 2.5 (1.4) 3.2 (1.8)
R2 4.0 (2.2) 3.8 (2.1)
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
446
ASHRAE Transactions
the server location. Equation 4 shows the relationship used DISCUSSION
across each server (or across each cell).
Mark Seymour, Director, Future Facilities, London, UK:
1) Did the velocity variation across the perforated tile have a
CONCLUSIONS
significant effect? 2) Would you expect it to be more signifi-
Both the open box model and black box model produce cant in a realistic scenario? Obviously the black box/simpli-
acceptable results as validated against ten different sets of fied models ignore internal recirculation.
experimental data for a rack populated by four 10 U server Knud A. Hermansen: This study focused on understanding
simulators. The steps for setting up the boundary conditions the boundary conditions of the rack, and to the extent that we
for a black box rack model in this paper are easily reproducible examined other boundary conditions it was to make them
and require minimal user inputs of rack load and airflow. It is uniform and simple to model. This was done to minimize
hoped that these steps will give data centers designers a better sources of error from boundary conditions that were not the
ability to develop models with confidence in their accuracy. focus of the experiment. Therefore, our team made every
effort to ensure a uniform velocity distribution across the
ACKNOWLEDGMENTS perforated floor tile and did not study variation there. We
This team would like to thank Haidong Wang, Dr. Ling tested the perforated floor tile to ensure that we were achiev-
and James McNeil of the University of Colorado at Boulder ing a uniform velocity distribution and then focused our para-
for their assistance in conducting background research and in metric studies on variables concerning the rack.
providing technical advice on CFD modeling. This team Section 4.1.2.1 of the final report for ASHRAE RP-1487
would also like to acknowledge Mr. James VanGilder of APC (Hermansen, K., and S. Al-Saadi, 2012, The Development of
for his technical assistance and help in setting up the experi- Simplified Rack Boundary Conditions for Numerical Data
ments and for chairing TC 4.10 which sponsored this project. Center Models, ASHRAE RP-1487 Final Report, Atlanta:
ASHRAE) discusses our work on ensuring uniformity of
REFERENCES velocity distribution across the perforated floor tile (this
section is included here for reference). [Editor’s note: final
ASHRAE. (2009). ASHRAE Handbook-Fundamentals. reports of ASHRAE Research Projects are available for free
Atlanta: American Society of Heating Refrigeration and download to members at www.ashrae.org.]
Air Conditioning Engineers, Inc.
4.1.2.1 Air Supply
ASHRAE. (2006). Method of testing for rating the perfor- The supply air was delivered through a perforated floor
mance of air outlet and inlets, ANSI/ASHRAE Standard tile was a standard two foot by two foot (61 cm x 61 cm) dimen-
70-2006. Atlanta: American Society of Heating, Refrig- sioned tile. It was located one-half tile length (~ 1 ft or 30.5
eration and Air Conditioning Engineers. cm) from the front wall and 1 tile length (2 ft or 61 cm) from
Chen, Q., & Srebric, J. (2001). Simplified Diffuser Boundary the right wall.
Conditions for Numerical Room Airflow Models, The perforated floor tile itself was the AF200 CS-1103
ASHRAE RP-1009. Atlanta: American Society of Heat- from ASM products. It has a twenty-five percent open area.
ing, Refrigeration and Air Conditioning Engineers, Inc. This percentage was chosen to ensure a high velocity and give
EPA. (2007). Report to Congress on Server and Data Center a sufficient throw to get cool air up to the server simulators at
Energy Efficiency: Public Law 109-431. Washington, the top of the rack (Figure 4-17).
D.C.: U.S. Environmental Protection Agency: Energy Both the supply and return air are delivered through ducts
Star Program. that run through the suspended ceiling. Since most data centers
Shrivastava, S., Iyengar, M., Sammakia, B., Schmidt, R., & use under-floor air distribution, the supply air was ducted
VanGilder, J. (2006). Experimental-Numerical Compar- down from the ceiling to the under-floor air distribution system
ison For A High-Density Data Center: Hot Spot Heat (UFADS) via a circular duct which can be seen in Figure 4-14.
Fluxes In Excess Of 500 W/Ft^2. IEEE. San Diego, CA: Since the test chamber was only a partition of the entire East
IEEE. Zone and this team did not want to deal with leakage across the
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Uptime Institute. (2011). Heat Density Trends in Data Pro- floor of the entire zone, part of the raised floor plenum was
cessing, Computer Systems and Telecommunications partitioned in order to make it easier to seal and maintain air
Equipment. Retrieved March 2009, from Uptime Insti- pressure through the perforated floor tile (Figure 4-18).
tute: http://uptimeinstitute.org/ The primary concern in partitioning the raised floor
Zhang, X., VanGilder, J., Iyengar, M., & Schmidt, R. (2008). plenum was that this would create an uneven air distribution
Effect of rack modeling detail on the numerical results over the perforated floor tile. In order to address this concern
of a data center test cell. 2008 11th Intersociety Confer- the team designed the partition so that the edges of the parti-
ence on Thermal and Thermomechanical Phenomena in tion were always at least one tile's length away from the perfo-
Electronic Systems (ITHERM '08) (pp. 1183-90). Piscat- rated floor tile (Figure 4-19). The air distribution was tested
away, NJ: Institution of Engineering and Technology. using hot-wire anemometer measurements taken at multiple
©2012 ASHRAE
447
measuring points across the tile space and found to be uniform
with maximum variations across the tile of around five
percent.
The return air grille was also two foot by two foot (61 cm
by 61 cm) and was located on the ceiling, one foot (30.5 cm)
from the back wall and 31.5 in (80 cm) from the left wall. It had
a porosity of approximately 56%.
Andrew Louie, Mechanical Engineer, Parsons Brincker-
hoff, New York, NY: 1) What grid size did you use? 2) Did
you do grid sensitivity?
Hermansen: Our team used the normalized root mean
squared error (NRMSE) to examine four meshes. Each mesh
contained approximately twice as many cells as the previous
mesh. Our team found very little difference in the NRMSE
between the meshes, so we chose a mesh that optimized
computation time and grid density (for detailed results). This
mesh had an average cell size of 1 in. (2.5 cm). Section 5.2.5
of the final report for ASHRAE RP-1487 (Hermansen, K., and
S. Al-Saadi, 2012, The Development of Simplified Rack
Boundary Conditions for Numerical Data Center Models,
ASHRAE RP-1487 Final Report, Atlanta: ASHRAE) directly
addresses the second question (this section is included here for
reference). [Editor’s note: final reports of ASHRAE Research
Figure 4-17: Perforated Floor Tile Projects are available for free download to members at
www.ashrae.org.]
5.2.5 Grid Independence
The normalized root mean squared error (NRMSE) was
used to analyze the results of different levels of meshes to find
the grid independent solution. Equation (5-6) shows how the
NRMSE is reached.
 in = 1  p ic – p i2c  2
NRMSE  P c ,P 2c  = ------------------------------------------
 in = 1  p if  2
p ic = prediction of coarse grid at point i
p i2c = prediction of doubled grid at point i
p if = prediction of fine grid at point i
Figure 4-18: Raised Floor Plenum Partition Construction For grid independence analysis, the rack inlet anemom-
eter pole was used. This was due to the fact that there is more
variability in the velocity measurements versus the tempera-
ture measurements and if sufficient accuracy and resolution
due to grid refinement is achieved here it can be reasonably
assumed to apply for temperature as well. Grids were chosen
by roughly doubling the number of total cells in the domain
with the exception of the first interval (72,000 to 244,800). The
cells were proportionally increased in each of the cardinal
directions.
Table 5-16 shows the results of the grid independence
analysis. The third column shows the percentage of computa-
tional time that the lower cell count had versus the highest
(1,150,000 cells). Interestingly the NRMSE stayed around the
same range and actually went up slightly. This team ultimately
decided to go with 244,800 cells since this still had a reason-
able computational time but would give greater resolution
Figure 4-19: Raised Floor Plenum Partition Space than 72,000 cells.

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---

Table 5-16: Grid Independence Analysis to overestimate velocities in the immediate vicinity of the
nozzle diffuser. However, once the jets from the individual
Cell Comparison NRMSE Computational Time nozzles have merged, this model gives a much better repre-
sentation of the airflow and decay in the main section of the jet
72,000 vs. 244,800 0.084 6%
flow. We assumed a turbulence intensity of 5%, as this is a
244,800 vs. 576,000 0.109 25% fairly common assumption for indoor airflow. Section 4.3.1 of
the final report for ASHRAE RP-1487 (Hermansen, K., and S.
576,000 vs. 1,150,000 0.112 100%
Al-Saadi, 2012, The Development of Simplified Rack Bound-
ary Conditions for Numerical Data Center Models, ASHRAE
David Schowalter, Principal Engineer and Director of RP-1487 Final Report, Atlanta: ASHRAE) goes into greater
Business Development, Alden Research Laboratory, Inc., detail on the modeling of the perforated floor tile.
Holden, MA: What turbulence boundary conditions were For the rack outlet we used all of the assumptions previ-
used for the perforated tile and rack outlet? ously mentioned (to include a turbulence intensity of 5%). The
most interesting issue to come out of modeling the rack outlet
Hermansen: Our team conducted an extensive literature was the need for increased mesh density near the surface of the
review that found that nearly all data center models used the rack outlet to allow the model to resolve the velocity jump that
standard  turbulence models, with the Boussinesq approx- was simulated as the air passed through the porous surface of
imation for buoyancy in a RANS set of equations. We adopted the rack outlet. Section 5.2.4 of the final report for ASHRAE
these standard assumptions, as we wanted our models to use RP-1487 addresses this issue in greater detail (an excerpt from
common practices and assumptions. this section is included here for reference). [Editor’s note: final
For modeling the perforated floor tile we used the reports of ASHRAE Research Projects are available for free
momentum method described in the final report for ASHRAE download to members at www.ashrae.org.]
RP-1009 (Chen, Q., and J. Srebric, 2001, Simplified Diffuser ....Since the inlet and outlet grills for the rack had a
Boundary Conditions for Numerical Room Airflow Models, certain porosity there is a velocity increase that occurs as the
ASHRAE RP-1009 Final Report, Atlanta: ASHRAE). The air passes through the grill. In reality this occurs within milli-
momentum method specifies mass, momentum, energy, and meters of the grill. In the CFD program this takes four to five
concentration at the supply diffuser surface—in this case, the cells to resolve. The initial unrefined grid had the inlet pole
perforated floor tile. It avoids a detailed representation of the within the first cell of the rack inlet (Figure 5-54) causing it to
perforated floor tile geometry. It is not as accurate for the area over-predict velocities when compared with experimental
immediately outside of the perforated floor tile (as compared results. The grid had to be refined at both the rack inlet and
to the remainder of the space), and the authors of the final outlet so that the CFD pole readings were at least four cells
report for ASHRAE RP-1009 noted that it did have a tendency away from the rack inlet and outlet.
Figure 5-54: Velocity in Front of Rack Inlet Grill with Unrefined Mesh
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
©2012 ASHRAE
449
SA-12-010
Simulating Central Chiller Plant Loop Delta-T

with Quadratic Regression Models
Zhiqin Zhang, PhD, PE Jingjing Liu, PE

ABSTRACT efficiency measures have been proposed and implemented for

central chiller plants, such as optimal dynamic chiller
Loop delta-T is a critical parameter for the simulation and
sequencing (Olson 1993; Chang, Lin et al. 2005; 2009), all-
optimization of a central chilled water system, especially for
variable-speed chiller plant (Hartman 2001a; Erpelding 2006;
those with chilled water storage tanks and those with high
pump power installation. This paper proposes to model the Yu and Chan 2008), optimal condenser water and chilled water
loop delta-T using a quadratic regression model. The modeling temperature reset (Ahn and Mitchell 2001; Deng, Turner et al.
procedure is divided into three steps: variables selection and 2002), pump differential pressure setpoint reset (Moore and
filtering, best regression screening, and model adequacy Fisher 2003); and primary-only system (Taylor 2002). In addi-
check. Eight variables are selected as candidates of significant tion, one of the most effective and proven measures is to maxi-
variables. The variables with less significance are screened mize the system loop chilled water return and supply
out using P-value criteria or statistic criteria R2adj and RMSE. temperature difference or delta-T.
The remaining variables are combined with their correspond- Some researchers (Fiorino 1999; WMGroup Engineers
ing cross-product terms to form quadratic regression models. 2008) have discussed the benefits of a higher delta-T including
The best quadratic model is identified using Mallow's Ck reduced chilled water pump energy usage, increased system
statistic. Finally, the adequacy of the model is checked with cooling capacity, and improved chiller performance at low
residual analysis and influential point examination. This loads. Studies have shown that life-cycle costs tend to be opti-
procedure was illustrated using data from two real chilled mized at the highest practical delta-T (Taylor, Dupont et al.
water systems and reasonably accurate models were obtained. 1999). For a chilled water storage system, chilled water loop
delta-T plays a significant impact on system operations
INTRODUCTION (Zhang 2010). The chilled water tank should store as much
Improving chilled water plant operating efficiency and cooling capacity as possible to meet the cooling demand
minimizing electric energy usage is a big challenge for plant during the high electric charge period. The cooling energy
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
operators and researchers. Promoted as energy efficient and stored in the tank is proportional to the loop delta-T providing
low operating cost means of rejecting heat from cooling the tank volume is constant. Therefore, a degraded delta-T will
systems to the atmosphere, central chilled water systems are result in a lower tank tonnage capacity and cooling supply
widely employed in the United States to provide cooling to deficiency during the high cost period. For the same cooling
various types of facilities, such as university campuses, hospi- load, a higher delta-T also corresponds to a lower chilled water
tal complexes, industrial campuses, and central business flow rate, which could improve chiller efficiency. Braun
districts. The central chiller plants usually contribute a major (1988) showed that, within the normal operating range of a
part of energy consumption in these facilities. As more and chiller, the effect of variations in chilled water flow rate on the
more people are aware of the importance of reducing energy overall chiller performance was relatively small (< 2%). It is
consumption and greenhouse gas emissions, various energy interesting to note that reducing the evaporator flow shows an
Zhiqin Zhang is a consultant for KEMA Inc., Oakland, CA. Jingjing Liu is a project engineer of Nexant Inc., San Francisco, CA.
450
ASHRAE Transactions
improvement in the chiller performance. For a given cooling hundreds of cooling coils or even more connected to a central
load and chilled water leaving temperature setpoint, a lower cooling system loop, and it is impractical to apply this method.
flow gives a higher chilled water return temperature. Consid- Fong, Hanby et al. (2006) assumed that similar air-side equip-
ering that this effect dominates over the reduced heat transfer ment would run in the same operating mode. Therefore,
coefficient effect, the evaporation temperature will rise and the several air-handling units and fans were lumped into one air-
chiller performance will improve. Ho (2006) also pointed out handling unit, one outdoor air fan, one return air fan, and one
that, a properly designed variable-primary-flow system could free cooling fan. The feasibility of such simplification is not
operate as a well-balanced and efficient chilled water system. verified.
A lower water flow rate in chillers generally reduces the heat The objective of this paper is to propose a procedure to
transfer and hence chiller efficiency. This however was construct a quadratic regression model used to simulate
compensated by the larger log-mean temperature difference chilled water loop delta-T. It correlates the loop delta-T with
between the water and refrigerant sides, particularly during several variables, which are selected based on previous expe-
part load conditions. Overall, keeping the delta-T as high as riences and underlying theoretical considerations. The model
possible is beneficial to the central cooling system energy coefficients are identified using the ordinary least squares
performance. linear regression applied to the data sets. A stepwise regres-
Loop delta-T is a critical parameter in the study of simu- sion method is used to screen out the best quadratic regression
lation and optimization of central chilled water systems, espe- model from all possible combinations. The performance of the
cially for those with chilled water storage tanks and those with model built for individual project can be quantified by statistic
high pump power usage. In practice, it has been widely recog- criteria, such as adjusted coefficient of determination (Radj2)
nized that many chilled water systems are plagued by low and root mean square error (RMSE). The adequacy of the
delta-T syndrome, particularly at low cooling loads. Delta-T model can be checked using residual plots or examination of
degrading is almost inevitable over time and it can fall to about influential points. This paper illustrates this whole procedure
one-half to two-thirds of the design value at low loads (Taylor using trend data from two real central chilled water systems
2002). Many factors can contribute to chilled water loop delta- and reasonably accurate delta-T models were developed. It is
T degrading, such as chilled water supply temperature, cool- easy to apply this procedure to other central chilled water
ing coil entering and leaving air temperatures, flow control systems.
valves type and performance, tertiary connection types, coil
MODEL BUILDING PROCEDURE
cooling loads, air economizers, etc. In most systems, the root
causes of low delta-T syndrome are primarily related to build- Low delta-T is a complicated phenomenon and there are
ing operations and, in many cases, are not in the control of so many factors and uncertainties involved that it is practically
chilled water plant operators. Many researchers have tried to very difficult to be simulated with an explicit physical model.
use computer simulation methods to understand and optimize This paper proposes to model the loop delta-T using a
chilled water plant operations. However, to accurately simu- quadratic regression model generally in the following form:
late the plant performance change due to applications of
k k k
energy efficiency measures, a reasonably accurate loop delta-
T =  0 +   i x i +    i,j x i x j +  (1)
T model is indispensable.
i=1 i=1 i=1
A literature review shows that most of previous
researches on chilled water system optimization ignored or The terms in the model include constant β0, variable term
paid little attention to the loop delta-T modeling (Braun 1988; xi, and quadratic terms xixj, which are necessary when the
Hydeman and Zhou 2007; Jiang 2007a; Morris and Blaine correlationship is curvilinear. If the total number of variables
2008; Yu and Chan 2008). Other researchers chose simplified is k, the model has (k2+3k+2)/2 coefficients to be solved. The
methods to model the delta-T. For example, to predict the random error term ? is usually assumed to be an independent
required chilled water flow rates of each individual zone on a normal distribution with a mean of zero. The procedure of
campus under different operating conditions, Ma and Wang finding the best regression model is divided into three steps:
(2009) assumed a fictitious global air-handling unit to repre- variables selection and filtering, best regression screening,
sent all terminal units. An empirical formula was used to and model adequacy check. This paper adopted a statistics
predict the total chilled water flow as a function of the cooling program named SAS to fulfill this procedure.
load, zone airflow rate, and air handling unit inlet air and water
temperatures. The loop delta-T was calculated from the total Variable Selection and Filtering
cooling load and predicted chilled water flow rate. This model This is the first and also the most crucial step in building
was verified in a virtual simulation environment but it is a good model as it involves selecting the set of variables xi that
unknown if this model is tested by field data. Lu and Cai et al. will be used in the model. It is important to find truly indepen-
(2005c) adopted the cooling coil model developed by Wang, dent variables. If one variable is found highly correlated with
Cai et al. (2004) to calculate the coil load at given air and water one or more other variables, it can result in inflated variance in
flow rates and temperatures. However, there are typically the estimated regression coefficients and unstable predictions
©2012 ASHRAE 451

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---

outside the experimental region. Unfortunately, most variables Two quantifiable and easily obtainable variables are the
available to us are correlated to each other. For example, the loop chilled water supply temperature and the loop total cool-
loop total cooling load is highly dependent on the ambient ing load. Figure 1 shows a scatter plot of the loop delta-T as a
weather conditions, the chilled water supply temperature could function of these two variables for a real chilled water system.
be reset based on the ambient dry bulb temperature, the internal A lower coil cooling load induces a lower delta-T. During the
load consisting of occupancy, plug load, lighting, etc is obvi- winter time, raising the chilled water supply temperature will
ously correlated with hour-of-day. The only remedy is to lead to a delta-T drop. At the higher load end, the delta-T
collect as wide as possible range of data to regress the model. approaches a relatively steady value, which, in most cases,
At the same time, keep in mind the limitation of such models equals the coil design delta-T. All these characteristics derive
and cautions should be used when it is necessary to project the from the part-loaded cooling coils, which are majority
delta-T outside the range of regression data. consumers of chilled water in this system.
The loop supply and return pressure differential or delta-
Some information affecting the delta-T is not available to P, ambient dry-bulb and wet-bulb temperatures are also good
researchers. For example, occupancy reflects the internal variable candidates. For a central chilled water system with
cooling load for office buildings but occupant number is not variable-speed secondary pumps, the pump speed is modu-
trended in most cases. However, majority of occupants will lated to maintain a pre-defined loop delta-P setpoint. This
follow a known working schedule and the occupancy factor setpoint may be reset continuously based on the ambient dry-
can be replaced with hour-of-day. Recognized experiences in bulb temperature. A higher loop delta-P can lead to a higher
the industry and theoretical considerations of the underlying pressure across the cooling coil control valves and possibly a
problem can help select the most significant set of variable higher leaking rate through the valves. Leaking through
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
candidates to be included in a particular problem. control valves is one of major causes for delta-T degrading.
Figure 1 Loop DT as a function of the loop cooling load and supply temperature (IP and SI).
452
ASHRAE Transactions
When the ambient dry-bulb temperature is lower than the The second-order terms like xixi are included because
setpoint, economizer mode may be triggered. Under econo- some correlationships may be curvilinear instead of linear.
mizer mode, the cooling coil entering temperature is low, The cross-product terms like xixj are excluded temporarily to
which will bring down the return water temperature. During reduce the number of regression model candidates. The ordi-
the swing season or winter season, the coil loads in the exterior nary least squares linear regression method is adopted to iden-
zones of a building may drop to an extremely low level, lead- tify the model coefficients from the data set. The exact form of
ing to laminar flow in the coil tubes. The much lower heat the regression model may vary in different projects.
transfer coefficient with laminar flow will lead to a lower After the variables are identified, the model with only the
delta-T. xi and xixi terms may be improved by adding the xixj terms in
In addition, the internal load consisting of occupant load, the model. The best quadratic regression model can be iden-
lighting, process load, and plug load, and external load such as tified using all possible regressions or stepwise regressions. If
solar heat gain and infiltration will also affect the loop delta- the number of variables is small, the all-possible regression
T. However these pieces of information may not be available. method is recommended. This method can always find the real
An alternative way is to choose some variables, for example best regression equation because it will not be distorted by the
month, hour-of-day, or day-of-week, to substitute for these dependencies among the variables, which may happen in step-
factors. The internal and external load factors are highly corre- wise-type methods (Montgomery and Runger 2002).
lated with these time-based variables. For example, the However, stepwise regression and its variation forward selec-
academic calendar of a university campus follows several tion and backward elimination are the most widely used vari-
semesters and breaks. The throughput of a manufacturer able selection technique. The best model can be identified
campus may show some seasonal patterns. These patterns may using Mallow's Ck statistic, which is a measure of the total
be related with chilled water system operation change. mean square error for the regression model. The best regres-
However, attention should also be paid for using time based sion model is either a model with minimum Ck or a model with
variables as proxies for important variables that are unavail- Ck approximately equal to k, where k is the variable number
able. of the model (Mallows, 1973).
Some well-recognized factors may be significant to the Model Adequacy Check

loop delta-T but it is difficult to either collect enough data or
Finally, it is necessary to check the adequacy of the
include them into the model. For example, the three-way
selected model. Residual analysis is a powerful and effective
control valves, coil leaving air temperature setpoint, total air
method. Plot the residuals against the predicted delta-T and
flow rate, sensible load ratio, economizer mode, the leaking
against possible candidate variables not included in the model
through the control valves due to various reasons, and building
for a pattern check. Non-neutral patterns in the plot to a candi-
blending valves.
date variable indicate that the model may be improved by
In summary, the practical independent variable candi- adding that candidate variable. A normal probability plot of
dates are the loop chiller water supply temperature, loop total the residuals can be used to check the assumption of normal
cooling load, loop delta-P, ambient dry-bulb and wet-bulb errors. A linear plot indicates that the normal distribution is
temperatures, month, hour-of-day, and day-of-week. It is consistent with the data. If the errors are normally distributed,
possible to include other variables into the model if trended approximately 95% of the standardized residuals should fall in
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
data are available, and they may be found significant. the interval (-2, +2). If the sample size is large enough, a
normalized frequency histogram of the residuals may also be
Best Regression Screening constructed as an approximate check of normality (Montgom-
ery and Runger 2002).
It is expected that not all the proposed variable candidates Examination of influential points is another way to check
are necessary to adequately model the delta-T. We would like the model adequacy. An influential point is one that has great
a model with as few regression variables as possible but it effect on the estimated coefficients or on the estimated mean
performs satisfactorily. This compromise is achieved with response. One technique of identifying influential points is
variable screening. P-value criteria are used to screen the Cook’s distance. A value larger than 1 would indicate that the
candidate variables and obtain the best subset of independent point is influential. If the point turns out to be “bad” data, it
regression variables. The terms with a P-value less than a mini- should be eliminated from the data set. On the other hand,
mum threshold is removed from the model. Statistic criteria there may be nothing wrong with these points, but we would
R2adj and RMSE can also be used to filter alternative subset of like to determine whether or not they produce results consis-
regression variables (Montgomery and Runger, 2002). The tent with the rest of the data (Montgomery and Runger 2002).
linear model examined is:
Model Application
k k
T =  0 +   i x i +   i,i x i x i +  (2) The purpose of constructing a delta-T regression model
i=1 i=1 is to predict the loop delta-T based on knowledge of the
©2012 ASHRAE
453
regression variables values. This application is only justifi- round screening. All terms left show high significance with P-
able for variable values within the range of the data used for values less than 0.007. Further tests show that no other vari-
developing the regression model. It is unknown whether the ables can be removed without obvious degrading on Radj2 and
linear relationship holds true outside this region. This is the RMSE. The basic linear regression model with seven variables
shortage of regression modeling. The best estimation occurs and fifteen terms is obtained as follows.
in the center of the experimental value range of the data set.
0.0.5904 – 0.07576x˜ 1 + 0.2252x˜ 2 – 0.15681x˜ 3
APPLICATIONS – 0.34176 x˜ + 0.15383x˜ – 0.59988x˜ – 0.06679x˜
˜ 4 6 7 8
T = (4)
One year hourly trend data from the energy management – 0.44295x˜ 12 – 0.07709x˜ 22 + 0.03749x˜ 32 – 0.09955x˜ 42
system (EMS) of the chilled water system of an airport in the
United States is used as an example to illustrate the procedure – 0.05621x˜ 2 + 0.57953x˜ 2 + 0.06875x˜ 2
6 7 8
of seeking a good chilled water loop delta-T model. The total
The above model can be notably improved by introducing
cooling capacity of the plant is 33,000 ton (116,056 kW) with
cross-product terms of xixj. The number of variable terms
a design delta-T of 24°F (13.3°C). The chilled water supply
increased from 15 to 35. Forward stepwise selection and Ck
temperature is reset between 36°F (2.2°C) and 46°F (7.8°C)
statistic was adopted to find the best quadratic model. Finally,
following a predefined time schedule. The secondary chilled
a quadratic model with 27 terms was selected with Radj2 equal
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
water pumps are equipped with variable speed drives and the
to 0.8306 and RMSE equal to 0.055.
pump speed is modulated to maintain the pressure differential
Figure 2 shows the scatter plot of the predicted value
between the supply and return piping headers. The pressure
versus the measured value of the loop delta-T. The points on
differential is also reset following predefined time schedules.
the solid line indicate a perfect match, while those far away
Based on the analysis on the chilled water system opera-
from the line reveal the model’s incapability in those zones. A
tions as well as available information, eight variables are
reasonable match can be observed for the majority of the
selected as shown in Table 1. Max-min normalization was
observations. Figure 3 shows the profiles of measured delta-T
performed on all input data to transfer each variable into a
and predicted delta-T with both the models with and without
range between 0 and 1. The transformation formula used here
the cross-product terms. The randomly selected time period
is:
used in the plot spans one week or 168 hours and the measured
x i – x i, min delta-T fluctuates between 15°F (8.3°C) and 19°F (10.6°C).
x˜ i = -----------------------------------
- (3) During most of this period, the profiles predicted by both
x i, max – x i, min
models overlap the measured profile nicely. The model with
Where xi,max and xi,min are the maximum and minimum cross-product terms provides obvious corrections, but it over-
values of variable xi data set. The first round screening shows shoots during the low delta-T periods. Both models missed the
that the ambient wet-bulb temperature terms (x5 and x5x5) are big spikes of the measured delta-T profile occurred after noon.
the least significant and removed because the P-values of these The possible explanation is that the cause of these spikes has
two terms are 0.3726 and 0.03, respectively. The Radj2 and not been addressed using the existing variables.
RMSE of the model is 0.8188 and 0.057, respectively. No To check the adequacy of the model, a normalized
noticeable change occurs on Radj2 and RMSE in the second frequency histogram of the residuals was created with the one-
Table 1. Variable Information and Variable Filtering
Unit Max Min Unit Max Min
Loop delta-T Y °F 22.3 4.9 °C 12.4 2.7
Loop supply T x1 °F 47.2 35.9 °C 8.4 2.2
Loop cooling load x2 Ton 17839 2493 kW 62,737 8,767
Loop delta-P x3 psid 50.6 24.0 kPa 349 165
Ambient dry-bulb x4 °F 104 24 °C 40.0 -4.4

Ambient wet-bulb x5 °F 80 21 °C 26.7 -6.1
Hour of day x6 23 0 23 0
Month x7 12 1 12 1
Day of week x8 7 1 7 1
454
ASHRAE Transactions
Figure 2 Scatter plot of predicted versus measured delta-T (IP and SI).
year data set and given in Figure 4. The standardized residuals As a comparison, the same procedure was applied to a
were calculated as follows: central chilled water system serving a university campus in the
United States. There is no chilled water supply temperature
 T i, measured – T i, predicted  reset and the loop delta-P profile is unknown. The selected
d i = --------------------------------------------------------------------------
- (5)
̂ independent variables are the loop cooling load, ambient dry-
bulb, time-of-day, month, and day-of-week. The best
Where  Ti,measured is measured delta-T,  Ti,predicted is quadratic model selected includes 15 terms with Radj2 equal to
predicted delta-T, and ˆ is the RMSE of the regression. The
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
0.9677 and RMSE equal to 0.38. Figure 5 shows the scatter

histogram is compared with a standard normal distribution, plot of predicted delta-T against measured delta-T and
i.e. the mean is zero and the variance is 1. Since the normalized Figure 6 shows the one-week delta-T profiles. A good match
residual frequency histogram approximately matches the stan- is observed. However, it should be recognized that quadratic
dard normal curve, it is concluded that there is no severe depar- regression method cannot guarantee a satisfactory fit of the
ture from normality and no serious model inadequacies are data set. The performance of a quadratic regression is highly
indicated. The standardized residuals within the interval (-2, dependent on the system characteristics, available variables in
+2) is counted at 96.3%. the trend data, and the quality of the trend data.
At the end, Cook’s distance is calculated for each point in
Figure 2 to filter out any influential observations. Influential SUMMARY
points that were identified as outliers or bad points were
removed and the above procedure was repeated to update the Loop delta-T is a critical parameter in the simulation and
model terms and the corresponding coefficients. optimization of a central chilled water system because delta-T
©2012 ASHRAE
455
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 3 Profiles of measured and predicted delta-T with two models (IP and SI).
Figure 4 Frequency plot of standardized residuals and normal distribution (IP and SI).
456
ASHRAE Transactions
degrading over time is prevailing in many real chiller plants. data from two real chilled water system and reasonably accu-
Existing studies either ignored loop delta-T fluctuations or rate models were obtained. It is recognized that the regression
modeled it using an imaginary air-handling unit. The feasibility model is only justifiable for variable values within the range of
of such simplification is not verified in practical applications. the data used for regression modeling. Therefore, data set
This paper proposes to model the loop delta-T using a covering a wide range of values is important for the success of
quadratic regression model. The whole procedure is divided this modeling approach. It also should be recognized that the
into three steps: variable selection and filtering, best regres- model should be updated periodically or continuously when
sion screening, and model adequacy check. Eight variables are time based proxies are used for field applications.
given as candidates for screening and significance tests. They
are loop chiller water supply temperature, loop total cooling REFERENCES
load, loop delta-P, ambient dry-bulb and wet-bulb tempera-
Ahn, B. C. and J. W. Mitchell 2001. Optimal control devel-
tures, month, hour-of-day, and day-of-week. The variables
opment for chilled water plants using a quadratic repre-
with less significance are screened out using P-value criteria
sentation. Energy and Buildings 33(4): 371-78.
or statistic criteria R2adj and RMSE. The remaining terms are
combined with their corresponding cross-product terms to ARI 2001. ARI 410: 2001 Forced-circulation Air-heating
form quadratic regression models. The best model is selected and Air-cooling Coils Arlington, VA, Air-Conditioning
using Mallow's Ck statistic. Finally, the adequacy of the and Refrigeration Institute.
selected model is checked with residual analysis and influen- Braun, J. E. 1988. Methodologies for the design and control
tial point examination. This method was illustrated using the of central cooling plants. Ph.D. Dissertation, Depart-
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 5 Scatter plot of predicted versus measured delta-T for a campus plant (IP and SI).
©2012 ASHRAE
457
ment of Mechanical Engineering, University of Wiscon- Hartman, T. B. 2001a. All-variable speed centrifugal chiller
sin-Madison plants. ASHRAE Journal 43(9): 43-52.
Chang, Y., F. Lin, et al. 2005. Optimal chiller sequencing by Ho, S. 2006. Maximizing chiller water plant efficiency. Eco-
branch and bound method for saving energy. Energy Librium.
Conversion and Management 46(13-14): 2158-2172. Hydeman, M. and G. Zhou 2007. Optimizing Chilled Water
Deng, S., W. D. Turner, et al. 2002. Retrocommissioning of Plant Control. ASHRAE Journal (June): 44-54.
central chilled hot water systems. ASHRAE Transac- Hyman, L. B. and D. Little 2004. Overcoming low delta T,
tions 108(2): 75-81. negative delta P at large university campus. ASHRAE
Journal 46(2): 28-34.
Erpelding, B. 2006. Ultra efficient all-variable-speed chilled
Jiang, W. 2007. General Methodology Combining Engineer-
water plants. Heating Piping Air Conditioning 78(3):
ing Optimization of Primary HVAC&R Plants with
35-43.
Decision Analysis Methods--Part I Deterministic Analy-
Fiorino, D. P. 1999. Achieving high chilled-water Delta Ts. sis. HVAC&R Research 13(1): 93-117.
Fiorino, D. P. 1999. ASHRAE Technology Awards Boiler Jiang, W. and T. A. Reddy 2003. Re-evaluation of the Gor-
Plant Replacement Reduces Steam Costs 35%. ASHRAE don-Ng performance models for water-cooled chillers.
Journal(9): 57-59. ASHRAE Transactions 109(2): 272-287.
Fong, K. F., V. I. Hanby, et al. 2006. HVAC system optimiza- Kirsner, W. 1995. Troubleshooting chilled water distribution
tion for energy management by evolutionary program- problems at the NASA Johnson Space Center. Heating
ming. Energy and Buildings 38: 220-231. Piping Air Conditioning 67(2): 51-59.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 6 Profiles of measured and predicted delta-T for a campus plant (IP and SI).
458
ASHRAE Transactions
Lu, L., W. Cai, et al. 2005c. HVAC system optimization—in- Taylor, S., P. Dupont, et al. 1999. The CoolTools Chilled
building section. Energy and Buildings 37(1): 11-22. water plant design and specification guide. San Fran-
Ma, Z. and S. Wang 2009. An optimal control strategy for cisco, PG&E Pacific Energy Center.
complex building central chilled water systems for prac- Taylor, S. T. 2002. Degrading chilled water plant delta-T
tical and real-time applications. Building and Environ- causes and mitigation. ASHRAE Transactions 108(1):
ment 44(6): 1188-1198. 641-653.
Mallows, C.L. Some Comments on CP. Technometrics 15 Taylor, S. T. 2006. Chilled water plant retrofit—a case study.
(4) (1973) 661-675 ASHRAE Transactions 112(2): 187-197.
Moe, E. M. 2005. Applying pressure independent control to
Wang, Y.-W., W.-J. Cai, et al. 2004. A simplified modeling
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
achieve high delta-T. International District Energy
Association (IDEA) Annual Conference, Nashville, TN. of cooling coils for control and optimization of HVAC
Montgomery, D. C. and G. C. Runger 2002. Applied Statis- systems. Energy Conversion and Management 45(18-
tics and Probability for Engineers 3rd. New York, John 19): 2915-2930.
Wiley & Sons, Inc. Wang, Z., G. Wang, et al. 2007. Achieving high chilled water
Moore, B. J. and D. S. Fisher 2003. Pump differential pres- delta T without blending station.
sure setpoint reset based on chilled water valve position. Yao, Y. 2004. Optimal operation of a large cooling system
ASHRAE Transactions 109(1): 373-379. based on an empirical model. Applied Thermal Engi-
Morris, T. and S. Blaine 2008. Chiller Plant Optimization neering 24(16): 2303-2321.
Tool ASHRAE Journal: 54-60. Yu, F. W. and K. T. Chan 2008. Optimization of water-cooled
Olson, R. T. 1993. A dynamic procedure for the optimal chiller system with load-based speed control. Applied
sequencing of plant equipment part1 algorithm funda- Energy 85(10): 931-950.
mentals. Engineering Optimization 21(1): 63-78.
Reddy, T. A. and K. K. Andersen 2002a. An evaluation of DISCUSSION
classical steady-state off-line linear parameter estima-
tion methods applied to chiller performance data. Doug McLellan, Mechanical Engineer, Hewlett-Packard,
HVAC&R Research 8(1): 101-124. Herndon, VA: Which variable was the key variable?
Sun, Y., S. Wang, et al. 2009. Chiller Sequencing Control Zhiqin Zhang: The key variables for most plants are loop
with Enhanced Robustness for Energy Efficient Opera- total cooling load and chilled-water supply temperature.
tion. Energy and buildings doi:10.1016/ Other key variables are case by case and it depends on the
j.enbuild.2009.07.023. actual operations of the plant.
©2012 ASHRAE
459
SA-12-011
Calculation of Evaporation
from Indoor Swimming Pools:
Further Development of Formulas
Mirza Mohammed Shah, PhD, PE

Fellow ASHRAE
ABSTRACT parameters to take it into account. As was shown by Shah

(2002, 2008), its agreement with fully occupied pool data is
The author had earlier published formulas for evapora-
reasonable but shows poor agreement with data for unoccu-
tion from occupied and unoccupied indoor swimming pools.
pied and partially occupied pools.
The formula for unoccupied pool was derived from the analogy
between heat and mass transfer during natural convection. It The present author presented a formula for evaporation
was shown to be in good agreement with all published test data. from an undisturbed water pool into quiet air derived by the
However, it was applicable only to positive density difference direct application of the analogy between heat and mass trans-
(i.e. when the density of room air was greater than that of air fer (Shah 2002, 2008). It was shown to be in good agreement
at the pool surface). It has now been extended to apply to nega- with almost all available test data. However, it is applicable
tive density differences. It has also been extended to include the only to positive air density difference (i.e. when the density of
case when air is forced over the surface of the pool to remove room air is larger than the density of air at the surface of water).
off gases. Calculation method is also provided for various
There are situations in which the density difference is nega-
types of occupied pools. Tabulations of calculated evaporation
tive. The calculation methodology has now been modified to
rates from unoccupied pools at typical design conditions.have
include such situations.
been provided.
In recent years there has been a concern that gases given
INTRODUCTION off the water surface have harmful effect on pool occupants.
One of the ways proposed to counter this problem is to blow
Accurate calculation of evaporation from swimming
air along the pool surface to carry away these gases. This will
pools is needed to ensure proper sizing of HVAC equipment as
increase the evaporation rate. A tentative formula has been
well as for the estimation of energy consumption. Unoccupied
indoor swimming pools are one important application. developed for calculation with forced air flow.
Numerous empirical correlations have been presented for In the following, the further development of the author’s
unoccupied swimming pools, the best known among which is formulas to extend them to negative density differences,
the correlation of Carrier (1918). None of them has been found forced air flow, and various types of pools, is presented. Tables
to be accurate beyond the data on which they were based (Shah of calculated evaporation rates at typical design conditions are
2002). ASHRAE Handbook (ASHRAE 2007) has provided provided for the convenience of designers. The ASHRAE
multiplication factors (called activity factors) to be applied to Handbook method is discussed first as it is widely used and is
the Carrier correlation to bring predictions closer to those
involved in the developments presented here.
experienced. However, attempts at such corrections cannot be
fully successful as evaporation occurs mainly by natural The paper is written in SI units. the recommended calcu-
convection while the Carrier formula does not include the lation procedure is given In Inch-Pound units in the Appendix.
Mirza Mohammed Shah is a consultant in Redding, CT.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
460
ASHRAE Transactions
ASHRAE HANDBOOK METHOD OF CALCULATION  w)  0. This shortcoming has been remedied in the further
The ASHRAE method is based on the following formula development discussed in the next section.
given by W. H. Carrier (1918)
FURTHER DEVELOPMENT FOR POOLS
 0.089 + 0.0782u   p w – p r  WITH TYPICAL VENTILATION SYSTEMS
E = ------------------------------------------------------------------
- (1)
i fg
Physical Model
This formula was based on tests done on an unoccupied Evaporation occurs by two mechanisms:
pool across which air was blown. In the past, it was widely
used for both occupied and unoccupied pools with or without • Natural convection
forced air flow. Experience showed that it greatly over- • Convection due to air currents caused by ventilation sys-
predicted evaporation from unoccupied pools as well as from tem
some occupied pools. ASHRAE therefore gave the multipliers
(called activity factors) listed in Table 1 to correct the predic- These two mechanisms work independent of each other.
tions of Equation 1. These activity factors are intended for use It is postulated that all evaporation occurs by the stronger of
in sizing the dehumidification equipment and were based on the two mechanisms.
consensus among ASHRAE Technical Committee members,
Evaporation by Natural Convection
not on any documented tests. Further, Equation 1 is modified
in ASHRAE Handbook to the following form which is recom- Air in contact with the surface of water becomes satu-
mended for air velocities between 0.05 to 0.15 m/s. rated with air, thus becomes lighter, and moves upwards carry-
ing the evaporated water with it. The heavier and drier room air
E = 0.000144  p w – p a  (2) moves down to take its place and this cycle continues. Equa-
tion 3 gives the evaporation due to this effect.
SHAH FORMULA FOR UNOCCUPIED POOLS Evaporation Due to Air flow by Ventilation System:
By the use of the well known relationship known as the Typical ventilation systems consist of supply diffusers on the
analogy between heat and mass transfer during natural ceiling near the wall on one side of the pool and return diffus-
convection, Shah (2002, 2008) derived the following formula ers on the opposite side of pool. These produce air currents
for evaporation from unoccupied pools. above the pool.
When the density of air at the pool surface is greater than
E 0 = 35 w   r –  w  1  3  W w – W r  (3) the density of room air, natural convection essentially ceases.
Then all evaporation will be due to these ventilating air
For (  r –  w) < 0.02, predictions were found to be some- currents. Thus by analyzing the data at negative density differ-
what low. This was attributed to stray air currents and edge ences, the formula for evaporation due to air currents can be
effects which become significant as the natural convection found.
currents become weak. To account for these effects, a 15% The available data for negative density difference are
increase in E0 was specified for (  r –  w) < 0.02. plotted in Figure 1. The following equation is fitted to these
This formula was found to be in good agreement with data.
virtually all published data for field and laboratory tests. The
shortcoming is that it cannot make any prediction when (  r – E 0 = 0.00005  p w – p r  (4)
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Table 1. Activity Factors FA for Various Pools to Be Used with the Carrier Equation,
as Given in ASHRAE Handbook (2007)
Type of Pool Air Temperature, C Water Temperature, C Relative Humidity,% FA
Baseline (unoccupied pool of any type) Any Any Any 0.5
Residential pool 24 to 29 24 to 29 0.5
Condominium 24 to 29 24 to 29 0.65
Hotel 28 to 29 28 to 30 0.8
50 to 60
Public, school, competition 24 to 29 24 to 29 1.0
Therapeutic 27 to 29 29-35 0.65
Whirlpool/spa 27 to 29 36 to 40 1.0
©2012 ASHRAE
461
Calculation Procedure velocity above the surface of a public pool and found it to aver-
Calculate evaporation by Equations 3 and 4. Use the age 0.15 m/s. ASHRAE Handbook recommends Equation 2
larger of the two. This is the evaporation rate from unoccupied for velocities of 0.05 to 0.15 m/s. So Equation 4 may be consid-
pools. ered to be applicable for velocities up to 0.15 m/s. At this veloc-
ity, air flow over the pool will be turbulent. Thus for higher
DEVELOPMENT OF FORMULA velocities, we may write in accordance with Equation 8:
FOR FORCED AIR FLOW OVER POOLS
E 0 = 0.00005  V  0.15  0.8  p w – p r  (9)
Evaporation rate can be expressed by the following equa-
tion:
The air velocities used for removing off-gases are likely
E0 = hM w  pw – pr   pa (5) to be less than 0.5 m/s. Analyzable data at such low velocities
could not be found. Many formulas have been proposed for
For turbulent flow over a flat surface, using the analogy calculating evaporation with forced air flow. They take the
between heat and mass transfer, we may write (Shah 1981): form:
hM L VL 0.8
---------- = Sc  0.036  ----------- – 836
13 E0 = C1 + C2 V (10)
(6)
D 
The values of the constants C1 and C2 in various formulas
The term ‘836’ is negligible as the Reynolds numbers in
are listed in Table 2. The Meyer formula was developed from
turbulent flows are very large. Variations of air density in the
data on large water reservoirs. The Carrier formula was based
typical range of conditions are small. Total air pressure varies
on data from a pool as noted earlier. The other formulas are
very little. Then combining Equations 5 and 6:
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
from laboratory tests. The tests of Powell were for air veloci-
E 0 = function  V 0.8 , L – 0.2 , air properties .  p w – p r  (7) ties from 0.5 to 2.8 m/s. The other tests were at velocities of 1
m/s and higher. Formulas of the form of Equation 10 are not
The air properties change vary little over the conditions of applicable at zero air velocity as then evaporation depends on
interest. Due to the small exponent of the pool width L and as density difference. Lurie and Michailoff (1936) noted that at
the variations in pool widths are rather limited, L can be zero air velocity, the formula of Himus and Hinchley (1924)
neglected as a variable. Then: for forced convection predicts three times the evaporation as
the Himus-Hinchley formula for natural convection.
E 0 = constant. V 0.8 .  p w – p r  (8)
Figure 2 shows the comparison of these formulas with
Smith et al. (1993) measured air velocity above a large Equation 9. It is seen that at 1 m/s velocity, the prediction of
indoor public swimming pool. The velocity varied from 0.035 Equation 9 are close to those by the Meyer, Carrier, and
to 0.05 m/s. Their test data agree well with the present author’s Powell-Griffiths equations. At 0.5 m/s, Equation 9 and the
method (Shah 2008). Hanssen and Mathisen (1990) measured Powell-Griffiths formula predictions are almost the same. At
higher velocities the agreement with the Meyer formula is
quite good.
Thus Equation 9 has been verified down to velocity of
0.5 m/s. It seems likely than that it will be good for lower
velocities also but it remains to be verified with test data.
OCCUPIED POOLS
Public Pools: The present author presented an empirical
correlation that fitted all available data for occupied and
Table 2. Values of Constants by Various

Researchers in the Equation E0 = C1 + C2 V
Author C1 C2
Carrier (1918) 0.000132 0.000116
Meyer (1942) 0.000116 0.000126
Powell and Griffith (1935) 0.000055 0.000136

Figure 1 Analysis of data for negative density difference,
i.e. (  a –  w) < 0. Lurie and Michiloff (1936) 0.000168 0.000128
462
ASHRAE Transactions
Table 3. Author’s Recommended Calculation
Method for Pools of Various Types
Type of Pool K =E/E0 Note
Baseline (unoccupied
1.0
pool of any type)
Based on FA
Residential pool 1.0 from ASHRAE
Handbook.
Condominium 1.3 See Table 1
Hotel 1.6
Public, school,
NA Use Equation 11
competition
Carrier formula does not have any basis for natural convection
which is the predominant mode and does not include param-
eters for it.
Figure 2 Comparison of the present formula, Equation
9, with various published formulas at 1 Pa
pressure difference. SUMMARY OF RECOMMENDED
CALCULATION PROCEDURE
partially occupied public pools with a mean deviation of • For pools with typical ventilation systems, calculate E0
16.2%. The ASHRAE Handbook method had a mean devia- by Equations 3 and 4 and use the larger of the two.
tion of 36.2%. For fully occupied pools, the Shah formula may (Table 5 provides calculated values of E0 at most com-
be written as: monly used conditions.)
E = 0.113 + 0.000059  p w – p r  (11) • For pools with forced air flow along the pool surface,
calculate E0 by Equations 3 and 9 and use the larger of
Various Other Pools: As noted earlier, test data are not the two values.
available for occupied pools other than public/competition • Fully occupied competition/public pools, use the
pools. Until such data become available, the best information author’s formula, Equation 11:
available are the ASHRAE Handbook activity factors listed in
Table 1. Using these, multiplication factors (K) listed in • For various other types of occupied pools, calculate as
Table 3 have been calculated. Using K, evaporation rate is stated in Table 3.
calculated as:
VALIDATION OF THE RECOMMENDED
E = K. E 0 (12)
CALCULATION PROCEDURE
These K factors have been arrived at as follows. For unoc-
The author had compared his original formulas for evap-
cupied pools, the author’s analytical procedure described
oration from occupied and unoccupied pools with a wide
above is very reliable as it is in good agreement with virtually
range of test data (Shah 2008). The data for unoccupied pools
all published data. In Table 1, FA for unoccupied pools is 0.5.
were from ten sources while those for occupied pools were
FA for hotel pool is 0.8. Thus experience has shown that evap-
from four sources. The range of those data is given in Table 4.
oration from hotel pools is 0.8/0.5 = 1.6 times that of an unoc-
No new analyzable data were found by literature search.
cupied pool. The K factors for residential and condominium
pools have been calculated in the same way as the water and All those data were compared with the modified proce-
room temperatures for all these pools are in the same range and dure recommended here. As the modifications affect mostly
the activities of occupants are similar. However, this procedure the data for air density difference less than 0.02 kg/m3, only a
may not be reliable for spas and therapeutic pools as the water small percentage of data were affected. Deviations of a few
temperatures are much higher and there is not much activity. data points increased while those of a few others decreased.
Therefore K factors have not been given for these pools. The overall deviations remained unchanged. All data for unoc-
While the K factors have been derived from the activity cupied pools were predicted with a mean deviation of 20.6%
factors, their use may be preferable as they are used with E0 by while the data for occupied pools were predicted with a mean
the author’s formula which has a firm theoretical basis and is deviation of 16.2%. Table 4 gives the range of data for which
thoroughly verified with test data. On the other hand, the the recommended calculation procedure has been validated.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
©2012 ASHRAE
463
Table 4. Verified Range of Author’s Formulas for Evaporation from Pools
UNOCCUPIED POOLS OCCUPIED POOLS
Pool area, m2 0.073 to 425 64 to 1209
Water temperature, C 7 to 94 25 to 30
Air temperature, C 6 to 35 27 to 32
Air relative humidity,% 28 to 98 33 to 72
(pw - pr), Pa 210 to 80,156 1,067 to 2,069
(  r –  w), kg/m 3
-0.004 to +1.002 0.0013 to 0.218
Number of occupants 0 8 to 180
Table 5. Evaporation from Unoccupied Pools in SI Units at Typical Design Conditions

Calculated from the Author’s Formulas
Evaporation from Unoccupied Pool (E0)

Space Air Temperature and Relative Humidity
25 C 26 C 27 C 28 C 29 C 30 C
Water
Temperature, C 50% 60% 50% 60% 50% 60% 50% 60% 50% 60% 50% 60%
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
25 0.1085 0.0809 0.0918 0.0636 0.0732 0.0515 0.0637 0.0450 0.0583 0.0382 0.0523 0.0311
26 0.1355 0.1042 0.1171 0.0872 0.0997 0.0693 0.0806 0.0547 0.0677 0.0479 0.0626 0.0479
27 0.1579 0.1290 0.1433 0.1118 0.1262 0.0941 0.1081 0.0753 0.0885 0.0582 0.0723 0.0510
28 0.1877 0.1556 0.1711 0.1380 0.1539 0.1200 0.1360 0.1014 0.1171 0.0818 0.0960 0.0618
29 0.2174 0.1841 0.2006 0.1661 0.1831 0.1477 0.1651 0.1287 0.1459 0.1092 0.1268 0.0888
30 0.2483 0.2146 0.2312 0.1962 0.2136 0.1773 0.1953 0.1579 0.1765 0.1380 0.1570 0.1176
31 0.2831 0.2474 0.2655 0.2285 0.2475 0.2091 0.2289 0.1892 0.2098 0.1688 0.1900 0.1480
32 0.3192 0.2825 0.3013 0.2631 0.2829 0.2432 0.2639 0.2228 0.2445 0.2019 0.2244 0.1805
39 0.6461 0.6035 0.6256 0.5808 0.6045 0.5575 0.5827 0.5334 0.5604 0.5087 0.5379 0.4833
40 0.7051 0.6617 0.6842 0.6386 0.6627 0.6148 0.6405 0.5903 0.6177 0.5655 0.5943 0.5391
APPLICATION TO OUTDOOR POOLS extended formula has been verified by comparison with all
All discussions till now have been for indoor pools. As published test data from ten sources.
seen in Figure 2, the proposed formula for evaporation by A tentative formula has been developed for the case when
forced convection agrees well with many correlations for air is blown along the surface of pool to remove off gases. This
forced flow at air velocities of 1 m/s and higher. The Meyer formula agrees well with several published correlations for
formula is based on extensive measurements on wind exposed velocities of 0.5 m/s and higher. It remains to be verified for
reservoirs. It therefore appears likely that the calculation lower velocities as analyzable data were not available.
procedure recommended here for indoor pools with forced air A method has been developed for calculating evaporation
flow may also be applicable to outdoor pools. This remains to from various types of occupied pools for which no test data are
be checked by comparison with test data. available. This consists of multiplication factors derived from
the experience-based activity factors in the ASHRAE Hand-
SUMMARY AND DISCUSSION book (2007), to be used with the author’s formula for unoccu-
The author’s formula for unoccupied indoor pools has pied pools. This method may be more reliable as the
been extended to include conditions when the density of air in unoccupied pool formula is well-verified and has a firm theo-
contact with water surface is higher than that of room air. The retical basis.
464
ASHRAE Transactions
Tables have been provided that give rate of evaporation Shah, M. M. 2002. Rate of Evaporation from Undisturbed
from unoccupied pools at typical design conditions. These Water Pools: Evaluation of Available Correlations.
make it easier to perform design calculations. International J. HVAC&R Research, 8 (1): 125-132
Shah, M. M. 2003. Prediction of Evaporation from Occupied
NOMENCLATURE Indoor Swimming Pools. Energy & Buildings, 35 (7):
707-713
The SI units given below apply for all equations in this
paper except for the Appendix. Shah, M. M. 2008. Analytical Formulas for Evaporation
from Water Pools. ASHRAE Transactions,
D = Coefficient of molecular diffusivity, m2/h
Smith, C. C., Jones, R. W., and Lof, G. O. G. 1993. Energy
E = Rate of evaporation from occupied pools, kg/m2h Requirements and Potential Savings for Heated Indoor
E0 = Rate of evaporation from un-occupied pools, kg/ Swimming Pools. ASHRAE Trans. 99(2):864 – 874.
m2h
APPENDIX
ifg = Latent heat of vaporization of water, kJ/kg
hM = Mass transfer coefficient, m/h RECOMMENDED CALCULATION METHOD
K = Multiplication factor in Equation 12, dimensionless IN I-P UNITS
L = Characteristic length (i. e. width) of water pool, m
All equations in the paper are given in SI units. In the
p = Partial pressure of water vapor in air, Pascals following, the recommended calculation procedure is given in
pa = Total pressure of air, Pascal Inch-Pound units.
Sc = Schmidt number = /D, dimensionless For unoccupied pools of all types with typical ventilation
V = Air velocity, m/s systems, calculate evaporation by Equations A1 and A2 and
use the larger of the two.
W = specific humidity of air, kg of moisture/kg of air
 = dynamic viscosity of air, kg/m.h E 0 = 290 w   r –  w  1  3  W w – W r  (A1)
 = Density of air, mass of dry air per unit volume of
moist air, kg/m3 (This is the density in E 0 = 0.0346  p w – p r  (A2)
psychrometric charts and tables)
Calculated values of E0 at commonly used design condi-
Subscripts tions are listed in Table A1.
w = Saturated at water surface temperature For unoccupied pools with forced air flow over the pool,
r = At room temperature and humidity calculate evaporation with Equations A1 and A3 and use the
larger of the two.
REFERENCES
E 0 = 0.0346  V  30  0.8  p w – p r  (A3)
ASHRAE. 2007. ASHRAE Handbook HVAC Applications.
ASHRAE, Atlanta, Georgia. For fully occupied public, school, and competition pools,
Carrier, W. H. 1918. The Temperature of Evaporation, use Equation A4.
ASHVE Trans., 24: 25-50.
Hanssen, S. A. and Mathisen, H. M. 1990. Evaporation from E = 0.023 – 0.0000162 + 0.041  p w – p r  (A4)
Swimming Pools. Roomvent 90:2nd International Con-
ference, Oslo, Norway. For various types of pools listed in Table 3, calculate by
Himus, G. W. and J. W. Hinchley. 1924. The Effect of A Equation A4, E0 calculated as described above or read from
Current of Air on the Rate of Evaporation of Water Table A1.
Below the Boiling Point. Chemistry and Industry,
August 22: 840-845. E = K. E 0 (A5)
Lurie, M. and N. Michailoff. 1936. Evaporation from Free
The units for the above equations are:
Water Surfaces. Industrial and Engineering Chemistry,
28(3): 345-349. E, E0 = lb/hr. ft2
Meyer, A. F. 1942. Evaporation from Lakes and Reservoirs. K = Dimensionless
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Minnesota Resources Commission.

pw, pr = inches of Hg
Powell, R. W. and Griffiths, E. 1935. The Evaporation of
Water from Plane and Cylindrical Surfaces. Trans. Insti- V = ft/min
tution of Chemical Engineers, 13: 175-198.
W = lb/lb
Shah, M. M. 1981. Estimation of Evaporation from Horizon-
tal Surfaces. ASHRAE Trans., 87(1).  r ,  w = lb/ft3
©2012 ASHRAE
465
Table A1. Calculated Evaporation Rate from Unoccupied Pools
at Typical Design Conditions By the Author’s Method in Inch-Pound Units
Evaporation from Unoccupied Pool (E0)

Space Air Temperature and Relative Humidity
76 F 78 F 80 F 82 F 84 F 86 F
Water Temperature,
F 50% 60% 50% 60% 50% 60% 50% 60% 50% 60% 50% 60%
76 0.0211 0.0159 0.0176 0.0120 0.0135 0.0099 0.0123 0.0085 0.0110 0.0069 0.0097 0.0053
78 0.0270 0.0211 0.0232 0.0173 0.0193 0.0133 0.0149 0.0106 0.0132 0.0091 0.0119 0.0075
80 0.0328 0.0266 0.0291 0.0228 0.0252 0.0188 0.0212 0.0146 0.0166 0.0114 0.0141 0.0098
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
82 0.0390 0.0326 0.0353 0.0287 0.0315 0.0245 0.0274 0.0204 0.0233 0.0160 0.0185 0.0110
84 0.0458 0.0391 0.0420 0.0351 0.0381 0.0309 0.0340 0.0266 0.0298 0.0222 0.0253 0.0176
86 0.0530 0.0461 0.0492 0.0419 0.0451 0.0377 0.0410 0.0333 0.0368 0.0288 0.0323 0.0241
88 0.0608 0.0536 0.0569 0.0494 0.0528 0.0450 0.0486 0.0405 0.0442 0.0358 0.0397 0.0311
90 0.0692 0.0618 0.0651 0.0574 0.0610 0.0529 0.0567 0.0482 0.0522 0.0435 0.0476 0.0386
102 0.1335 0.1250 0.1289 0.1199 0.1241 0.1147 0.1192 0.1093 0.1142 0.1037 0.1088 0.0979
104 0.1469 0.1383 0.1422 0.1333 0.1374 0.1279 0.1324 0.1222 0.1272 0.1165 0.1219 0.1106
466
ASHRAE Transactions
SA-12-012 (RP-1390)
Short-Term Curtailment of HVAC Loads

in Buildings
Lixing Gu, PhD, PE Richard Raustad

Member ASHRAE
Rapidly increasing requirements placed on utilities to With the rapidly increasing demands placed on utilities,
reduce peak loads has led to utility customer incentives to shift exacerbated by the increasing cost of generating energy,
peak demand to non-peak times or reduce peak loads when reducing peak loads and minimizing energy use assume
notified by the utility that the grid is close to capacity. This greater importance than ever before. It is well documented that
paper investigates methods used to reduce building demand building sub-systems (e.g. HVAC, lighting, equipment) can
during a fixed time window near a utilities on-peak period in use as much as 40% (PNNL, 2008) of the nation’s total energy
response to short-term notification (a few hours advanced resources. To combat these increasing demands, utilities offer
notice) through computer simulations using EnergyPlus. The incentives to customers that can shift peak demand to non-
work presented in this paper was performed under the peak times or reduce peak loads when notified by the utility
ASHRAE research project, 1390-RP. Two building types (office that the demand on the grid is close to capacity. Utilities have
and retail) were defined using two equally viable construction developed the concept of critical peak pricing (CPP) (PG&E
types (light and heavy mass). The four resulting building 2007, SCE 2007a, and SCE 2007b) and real time pricing
construction models, combining building types and construc- (RTP) (Gabel et al. 1998) wherein the utility provides custom-
tions, were simulated in five selected cities which represent a ers with a pricing structure typically one day in advance. Then
diverse range of climate conditions. The control strategies the utility offers a monetary reward to customers who can shift
investigated in the project include lighting power density or reduce their electric demand from the grid during peak
reductions, global thermostat setpoint setback control, chilled hours.
water temperature reset, and supply air temperature adjust- Building owners and operators have responded to time-
ment. Results from computer modeling were analyzed. of-day pricing by resorting to traditional methods, such as
Demand reductions associated with changes in the thermal load-shedding (i.e. by turning off non-essential components
mass were insignificant compared to those achieved with that use electricity during peak hours), or by using non-tradi-
control strategies. The lighting power density reduction is tional methods, such as Demand Response (DR) as docu-
effective and easily implemented. Higher percent reductions in mented by various publications including Motegi et al (2007).
demand were achieved for the smaller buildings and medium Demand response consists of demand control algorithms that,
retail building which used the packaged HVAC system type along with a host of other strategies (e.g. lighting load control)
with global thermostat setpoint setback control. The peak reduce peak load or shift demand and use the thermal mass of
demand reductions that are achievable when two or more the building (e.g., HVAC load control) to achieve these
control strategies are combined vary between 12% and 34% demand reductions during peak hours (Xu, 2009). The build-
and are dependent on building and HVAC system types. ing can be pre-cooled during non-peak hours using electricity
Lixing Gu is a principal research engineer and Richard Raustad is a senior research engineer in the Building Research Division at the Florida
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Solar Energy Center in Cocoa, FL, which is associated with the University of Central Florida.
©2012 ASHRAE
467
at lower than normal prices, which allows for an increase of strategies under the ASHRAE research project, 1390-RP. The
thermal set-points (i.e. zones and/or HVAC components) to objective of 1390-RP was to identify and evaluate peak
reduce cooling demands during peak hours. The building’s demand reduction strategies formulated in response to short
normal operation can be restored when the grid demand is notification of critical peak rates and time-of-use charges from
sufficiently reduced. Lee and Braun (2006a and 2006b) have utilities, for use in commercial buildings. The term “demand
conducted research to assess what they term ‘demand limit- reduction” is used throughout this paper to represent the aver-
ing’ by using the thermal mass for small commercial build- age electrical energy savings during the peak demand window.
ings. In an experimental setting and in some parametric The short notification time may be as short as 15 to 30 minutes,
simulations of zone set-point manipulation before and during and as long as a few hours. Ever increasing pressure on utilities
peak hours, the use of three thermostat control profiles in has resulted in pricing forecasts that are on the order of hours
conjunction with pre-conditioning the building thermal mass instead of days, and previously researched strategies like pre-
lead to a reduction in peak cooling loads of approximately cooling of building thermal mass will not be as useful in reduc-
20% to 40%. Research by Xing (2004) and Reddy and Norford ing peak load when the demand on the grid is at its highest. To
(2002) in the area of building load control and optimization for
overcome this limitation, strategies specific to short-term elec-
demand limiting and peak load shedding present some prelim-
trical price forecasts have been investigated and identified.
inary results of simulation modeling in which thermostat and
The peak demand window was chosen to be 5-9 A.M. in winter
economizer setpoints are varied for a simple VAV system and
and 2-5 P.M. in summer.
a three-zone building model. This review of the literature
reveals that little work has been done to develop, test and
implement effective load-reduction strategies when the utility BUILDING AND HVAC SYSTEM SPECIFICATIONS
notification is not provided a day ahead but rather on the same
day. Computational simulations are needed to provide a direc- The buildings selected for this study were obtained from
tion for further experiments since computer simulation the computer simulation reference buildings provided by the
enables a more comprehensive consideration of various United States Department of Energy (DOE) (Deru et al. 2009).
control strategies and the influence of building type, climate, This study included small, medium, and large representations
etc. of the office and retail building types. The building character-
In order to develop effective control strategies to reduce istics were selected to meet the minimum requirements for
peak demand in response to short notification (only a few those building types according to ASHRAE Standards 90.1-
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
hours advanced notice) of critical peak rates and time-of-use 2004 and 62.1-2004. The buildings have cooling setpoint at
charges from utilities, computational simulations were 24oC (75.2oF) and heating setpoint at 21oC (69.8oF) during
performed to examine the effectiveness of various control occupied hours. Table 1 summarizes the selected building
Table 1. Selected HVAC System Types and Associated Building Types
Office Retail
Building Type
Small Medium Large Small Medium Large
Area (m2) 511 4,982 46,320 46,320 174 348 2294 2294
(ft2) 5,500 53,626 498,588 498,588 1,873 3,745 24,693 24,693
HVAC PSZ MZ-VAV MZ-VAV Dual Duct PSZ PSZ MZ-VAV Dual Duct
Chilled Chilled
Cooling Type DX DX Chilled water DX DX Chilled water
water water
COP 3 3 4.45 4.45 3 3 4.45 4.45
Heating Type Gas furnace Gas furnace Hot water Hot water Gas furnace Gas furnace Hot water Hot water
Eff. (%) 80 80 80 80 80 80 80 80
Fan control Constant Variable Variable Constant Constant Constant Variable Constant
Note:
PSZ - packaged single zone system
MZ-VAV - multizone variable air volume fan system
Dual Duct - constant volume dual duct system
DX – direct expansion refrigeration system
468
ASHRAE Transactions (RP-1390)
types and associated HVAC systems. The table also provides period for each prototype day, instead of a full annual simu-
building floor areas, cooling and heating types and efficien- lation, dramatically reduced the time required to compute the
cies, and fan control modes for each HVAC system. results. For each simulation, the last day in the 10 consecutive
Two types of building constructions were chosen to repre- day period provides the results for the specific prototype day.
sent light and heavy thermal mass buildings to examine the The selected prototype days are Summer Peak, Summer Mid,
impact that thermal mass would have on a building’s peak Summer Low, Fall Cool High, Winter Peak, Winter Mid,
demand reduction potential. Five geographical regions were Winter Low, and Fall Heat High (Table 2).
chosen to study climate specific variations in the results. The Eleven different thermostat temperature profiles were
cities selected and corresponding climate types are: Miami - used to represent different approaches, or control profiles, for
hot and humid, Baltimore - mixed humid, Albuquerque - controlling the HVAC system. Linear, exponential, and nega-
mixed dry, Phoenix - hot dry, and Minneapolis - cold. tive exponential equations were used to calculate the eleven
different thermostat temperature profiles. Each profile repre-
SIMULATION PROTOCOL
sents a thermostat schedule based on inputs to the mathemat-
The development of simulation protocol consists of ical model. In the same manner, we also developed setpoint
selecting control strategies and developing the simulation profiles of chilled water temperature reset, and supply air
approach, including selection of prototype days. temperature adjustment. The detailed descriptions are
provided in the final report (Gu et al. 2011).
Selected Control Strategies
When the demand window closed, the thermostat sched-
Many control strategies were identified to reduce building ule was gradually restored to the original thermostat setpoint
peak demand. After careful selections and suggestions from temperature in a linear manner over a 1-hour time period for
the Project Monitoring Subcommittee (PMS) (Gu et al. 2011), the global thermostat setpoint setback control. This recovery
four control strategies were selected and investigated thor- strategy was used to minimize the spike in building demand
oughly in the ASHRAE project: lighting power density reduc- that would occur if the thermostat setpoint temperature was
tion, global thermostat setpoint setback control, chilled water immediately returned to the setpoint that was in use prior to the
temperature reset, and supply air temperature adjustment. The implementation of the demand reduction strategy. For the
last two control strategies are applied only to central plant chilled water temperature reset and supply air temperature
systems used for large office and large retail buildings and are adjustment control, both temperatures were gradually restored
more practical for use with building automation systems. to the original setpoint temperatures in the same manner as the
global thermostat control. For the lighting power reduction
Simulation Approach
control, the lighting power density was set to the original value
Over 30,000 EnergyPlus computer simulations were used in the base case.
performed. These simulations reported building average
energy savings potential over the peak demand window by SIMULATION RESULTS ANALYSIS
building type, geographical location, and day type (e.g.,
summer peak, winter peak, etc.) for individual control strate- This section analyzes the impact of thermal mass, four
gies as well as combinations of strategies where two or more individual control strategies and combinations of control strat-
individual strategies were used. To reduce the time required egies using simulation results. The relationship between indi-
for the simulations, prototype days were selected to represent vidual control strategies and combined control strategies are
typical working days in different seasons. Choosing these discussed. Thermal comfort was also simulated to investigate
prototype days and simulating only a 10 consecutive day the impact of the control strategies on occupant comfort. The
Table 2. Prototype Day Description
Summer Peak A day to present a summer day with peak cooling energy use during peak demand period in a season
Summer Mid A day to present a summer day with average cooling energy use during peak demand period in a season
Summer Low A day to present a summer day with low cooling energy use during peak demand period in a season
Fall Cool High A day to present a fall day with peak cooling energy use during peak demand period in a season
Winter Peak A day to present a winter day with peak cooling energy use during peak demand period in a season
Winter Mid A day to present a winter day with average heating energy use during peak demand period in a season
Winter Low A day to present a winter day with low heating energy use during peak demand period in a season
Fall Heat High A day to present a fall day with peak heating energy use during peak demand period in a season
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
©2012 ASHRAE
469
detailed analysis may be found in the project final report (Gu The LPD reduction may be achieved by a building auto-
et al. 2011). mation system or manual switches. Multiple stages of LPD
reductions are recommended to smooth out the changes in
Thermal Mass Impact building energy use during the peak demand window.
Simulation results show that the impact of the building

Global Thermostat Temperature Setpoint Setback
thermal mass on the building peak demand is relatively small
for the building types selected for this study. In this study it
The use of a global thermostat temperature (TST) setpoint
was assumed that utility notification of a demand reduction
setback schedule achieved average energy savings of 40% for
event would occur only a few hours in advance and insufficient
small buildings, 30% for medium buildings and 20% for large
time was available to pre-condition the building. The rela-
buildings over the peak demand windows in this study using
tively small average energy savings over the peak demand
a maximum thermostat setback temperature differential.
window associated with thermal mass were found to be insig-
nificant (<2%) compared to the average energy savings Office buildings generally showed 5% - 10% less savings than
obtained from other control strategies investigated during the found for retail building types. These savings are dependent on
course of this project. Therefore, the impacts of thermal mass geographical location and HVAC system types. For this proj-
on peak demand reduction need not be considered a significant ect, the maximum thermostat setback temperature differential
contributor to savings found for control strategies investigated is 3.3oC (6oF). Using this control strategy it is possible to
during this project. achieve these savings even for the simplest of HVAC systems
(e.g., PSZ and packaged MZ VAV). This strategy applies to
Lighting Power Density Reduction those buildings using a zone thermostat (analog or digital) and
to those using more complex building automation systems.
The lighting power density (LPD) reduction strategy is an This study also found that the global thermostat strategy had
effective way to meet a peak demand reduction requirement. less of an impact on buildings served by a central plant using
This strategy is easy to implement by simply turning off non- dual duct systems since this system type would respond to a
essential lights. Figure 1 shows the percent peak reductions of zone setpoint temperature increase by reducing the cold deck
facility electricity averaged over all locations and prototypes air flow rate. This in turn would increase the hot deck air flow
days. Lighting power density varied between 70% and 90% of rate for this constant volume system and increase heating
the base case value during the peak demand windows. The line energy required to maintain the hot deck supply air tempera-
drawings are based on 3 discrete points at 70%, 80% and 90%. ture setpoint.
Since the percent peak demand reductions vary linearly with
the percent changes of lighting power density, it is very conve- The selection of the maximum setback differential will
nient for building operators or control engineers to decide how impact the degree to which occupant comfort is sacrificed
much lighting power density should be reduced given a during the peak demand window. Since the shape of the
specific peak demand reduction target. It should be pointed out setback curve can impact the average energy savings over the
that any LPD reduction will increase heating energy use in peak demand window, it is a good practice to test different
winter and reduce cooling energy use in summer. shapes to maximize savings.
Figure 1 Average percent peak reduction with different light power density changes over all prototype days and geographical
locations.
470 ASHRAE Transactions (RP-1390)

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---

Chilled Water Temperature Reset demand window. In practice, fan speed regulation is strongly
Resetting the chilled water control temperature (CWT) recommended when resetting the supply air temperature
for chilled water plant loops without addressing pump speed setpoint.
control can reduce building peak demand by up to 7%. For this
Combined Control Strategies
project, the chilled water temperature (6.7oC [44.1oF] for the
baseline) was increased by a maximum of 5oC (9oF). Build- As previously described, demand savings due to each
ings with dual duct systems achieved higher percent reduc- individual strategy provided a reasonable savings in building
tions than buildings with VAV systems. Although an increase peak demand for specific building and HVAC system types. In
in chilled water temperature will improve the efficiency of general, a single control strategy did not provide the maximum
chillers, the increase in pump and fan energy could offset these possible savings and various combinations of these strategies
savings in certain instances. For this reason, additional simu- were investigated to determine how these control strategies
lations were performed to limit the chilled water pump flow worked in combination. Computer simulations also showed
rate to be no greater than the flow rate at the beginning of the that combining thermostat reset strategies with chilled water
peak demand window. When including pump speed control, or supply air temperature reset strategies did not provide
savings up to 15% can be achieved. Therefore, it is important savings equal to the sum of the savings for the individual strat-
to prevent the chilled water flow rate from increasing at the egies. In other words, relationships of peak demand reductions
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
same time when chilled water temperature is reset. between individual strategies and the corresponding
Selection of the maximum setback differential should combined strategy are not linear. This result applies whether
take into consideration the impact it will have on dehumidifi- or not a speed control strategy is used since a reset in thermo-
cation capacity during the peak demand window. Pump speed stat temperature reduces zone loads, and therefore the required
control can be extremely important for VAV systems. The supply air or chilled water flow rate, which eliminated savings
pump speed was limited to the pump flow rate simulated just due to the water or air reset strategies.
prior to the start of the peak demand window. The combined control strategies of lighting power density
reduction and global thermostat setpoint setback control can
Supply Air Temperature Adjustment be implemented in small office, small retail, medium office
Supply air temperature (SAT) adjustment does have the and medium retail buildings. Control strategies combining
potential to reduce building peak demand, but only for specific two or more of the individual control strategies of lighting
HVAC system types. For this project, the maximum reset power density reduction, global thermostat setpoint setback
supply air temperature differential is 5oC (9oF). For dual duct control, chilled water temperature reset, and supply air
systems, results showed up to a 16% savings in building temperature adjustment can be implemented in large office
demand for cooling operation. For VAV systems, results and retail buildings.
generally showed up to a 5% savings, however, building peak The percent reductions of electrical demand savings
demand could actually increase up to 10% for the Summer- during peak demand periods in a summer peak day using
Peak or SummerMid day types for certain climate locations. combined control strategies (i.e., utilizing all available strate-
VAV systems will increase fan speed in response to an increase gies mentioned above) are shown in Figure 2. The savings with
in supply air temperature which can increase energy use. For light and heavy mass constructions are also provided in the
this reason, additional simulations were performed to limit the figure. The peak demand reductions vary between 10% and
VAV fan flow rate to be no greater than the flow rate at the 40%. Figure 2 also indicates the insignificant impact of ther-
beginning of the peak demand window. When implementing mal mass by comparing demand reductions between light and
a fan speed control strategy for VAV systems, building peak heavy construction types.
demand reductions up to 12% were possible. Figure 3 presents averaged percent reductions using all
Dual duct systems mix the cool and hot air upstream of the available control strategies with maximum savings for each
individual zones before entering the spaces. The sum of the individual control strategy for different building and HVAC
cold and hot deck air stream flow rates is constant. When the system types. The values are averaged over all prototype days
cold deck supply air temperature is increased, the cold deck air and geographical locations. The difference between Figures 2
flow rate also increases to meet the same load. As a result, the and 3 is that Fig. 2 presents average results from a summer
hot deck air flow rate decreases and energy savings is primar- peak day, while Fig. 3 presents average results from all proto-
ily due to a reduction in the hot deck heating requirement. type days. Although the same combined control strategies are
Since this is a constant volume system, fan speed control is not applied in both Figures 2 and 3, different setpoint profiles may
applicable. have been selected for different building and HVAC system
The selection of the maximum setback differential will types. The selection criteria chose specific setpoint profiles
impact the dehumidification capacity of the system during the based on the maximum savings. These specific setpoint
peak demand window. Fan speed control can be extremely profiles were then combined to simulate the overall savings.
important for VAV systems. The fan speed was limited to the To compare the impact different control strategies had on
fan flow rate simulated just prior to the start of the peak peak electricity use, the electricity use for individual and
©2012 ASHRAE
471
Figure 2 Electrical demand window saving potential using combined strategies in a summer peak cooling day.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 3 Averaged percent reductions using combined control strategies over all prototype days and geographical locations.
combined control strategies in a large office building for a strategies were based on the best performance of each individ-
summer peak cooling day in Miami are presented in Figure 4. ual control strategy. If savings from individual strategies were
The selected cases are mentioned in the legend using the previ- found to be additive, percent reductions from combined
ously defined abbreviations. The “+” sign indicates combined control strategies could be derived from the reductions of
control strategies. It is observed that the maximum saving single measures. After performing a statistical analysis for
occurs in the case with combined strategies individual control strategies and combinations of these strat-
(LPD+TST+SAT+CWT, lowest line in the demand window). egies, the results for combined control strategies may be calcu-
For each individual control strategy, the performance order lated from the individual savings found for small and medium
with the best first is CWT, SAT, LPD and TST (upper most office building types where only lighting and thermostat
four lines below the base case). For the cases with LPD plus control strategies were used. However, for large office and
one of the other 3 control strategies, the performance order retail building types, the savings for combinations of control
with the best first is LPD+CWT, LPD+SAT and LPD+TST. strategies should not be calculated using the results for indi-
vidual strategies (i.e., individual results are not additive when
Relationship Between Individual Control Strategies thermostat reset is combined with air or water reset strategies).
and Combined Control Strategies
In order to reach the maximum peak demand reductions, Thermal Comfort Check
it was essential to combine different control strategy types. Understanding that the use of thermostat setback can
Since there are many possible combinations, only a select few adversely affect occupant comfort, a summary of simulated
variations were simulated. The selected combined control occupant comfort for the same summer demand simulations
472
Figure 4 Electricity use history in a large office building with MZ-VAV system for a peak summer cooling prototype day with
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
different control strategies.
Figure 5 Fanger predicted mean vote value comparison to baseline values in a summer peak cooling day.
presented in Figure 2 was also compiled and is shown in to 0.1 for all building types (each “Baseline” result in the
Figure 5. For this study, a setback temperature difference of figure). This range is between a slightly cool (slightly cool =
3.3oC (6oF) was used to limit the maximum possible offset -1) perception to just barely warmer than a neutral perception
from the original thermostat temperature schedule. The simu- (neutral = 0) of how occupants perceive the indoor environ-
lation results from the summer peak day are selected to present ment. Comparing the baseline thermal comfort to that simu-
thermal comfort check, because the summer peak day with the lated when all control strategies were combined shows an
setback control strategies may represent possible severe increase in the PMV value towards the warm comfort region
scenarios. (i.e., comparing the “Baseline” results to the “All” results for
The Fanger comfort model is used to describe occupant each building type). The Fanger PMV values range from -0.3
thermal comfort as a people-weighted average for the entire to 0.6 for the combined simulations (each “* - All” result in the
building. ASHRAE Standard 55-2010 specifies the acceptable figure) and is primarily due to higher indoor temperatures.
predicted mean vote (PMV) range is between -0.5 and +0.5. Since these values do not exceed the slightly warm criteria
Figure 5 shows that the comfort levels for the Summer Peak (slightly warm = 1), and only a few are slightly higher than the
light construction baseline simulations are in the range of -0.5 maximum value specified by ASHRAE Standard 55-2010, it
©2012 ASHRAE
473
is assumed that the control strategies selected for this study are REFERENCES
considered feasible for use when implemented as building
demand reductions strategies. However, note that these Deru, M., K. Field, D. Studer, K. Benne, B. Griffith, P. Tor-
summary values are averages over the entire building (i.e., cellini, M. Halverson, D. Winiarski, B. Liu, M. Rosen-
multiple zones) and specific zones may have lower or higher berg, J. Huang and M. Yazdanian & D. Crawley. 2009
values than reflected in these average data. DOE Commercial Building Benchmark Models for
Energy Simulation, National Renewable Energy Labo-
CONCLUSIONS AND RECOMMENDATIONS ratory, Golden, Colorado
The results of this study provided the following conclu- Gabel, S.D., L. Carmichael and G. Shavit. 1998. Automated
sions: control in response to real-time pricing of electricity.
ASHRAE Journal, November, 26-29.
• Higher percent reductions were achieved for the smaller Gu, L., R. Raustad & M. Basarkar, 2011, “Short-Term Cur-
buildings and medium retail building which used the tailment of HVAC Loads in Buildings,” Final report for
PSZ HVAC system type, as shown in Figure 3. the ASHRAE research project 1390-RP, FSEC-CR-
• The average values of percent reductions shown in Fig- 1881-11, Florida Solar Energy Center, Cocoa, FL 32922
ure 3 may be generally applied to small office, small
retail, medium office, medium retail and large office Lee, K.-h. and J.E. Broun. 2006a. Evaluation of methods for
with dual duct systems. The values for each of these determining demand-limiting set point trajectories in
building and HVAC system types varied in a narrow commercial buildings using short-term data analysis.
range for all prototype days and locations simulated. Proc. SimBuild 2006. 2nd National Conference of
Results for large retail VAV and dual duct systems were IBPSA-USA, Cambridge, MA
more dependent on prototype days and geographical Lee, K.-h. and J.E. Broun. 2006b. Development of methods
location. for determining demand-limiting set point trajectories in
• Linear superposition of combined control strategies may commercial buildings using short-term data analysis.
be applied to small and medium building types with DX Proc. SimBuild 2006. 2nd National
cooling coils (i.e., the savings of individual strategies
may be added together to find the combined savings). Conference of IBPSA-USA, Cambridge, MA
However, linear superposition of combined control strat- Motegi, N., M.A. Piettie, D.S. Watson, S. Kiliccote, P. Xu.
egies may not be applied to large office and retail build- 2007. Introduction to commercial building control strat-
ing types served by a central plant. egies and techniques for demand response. Lawrence
Berkeley National Laboratory report LBNL-59975.
The following recommendations for future work are May. Available from http://drrc.lbl.gov/pubs/59975.pdf
provided below:
PG&E. 2007. PG&E commercial building critical-peak-pric-
• More building types are needed to develop comprehen- ing tariff. http://www.pge.com/tariffs/pdf/E-CPP.pdf
sive control algorithms for real building controllers. It is PNNL, 2008, http://eere.pnl.gov/building-technologies
recommended to select all 16 reference buildings pro-
vided by the US DOE. Reddy, T.A. and L.K. Norford. 2002. Building operation and
dynamics within an aggregated load. Final Report for
• The recommendations from the present study provide a
ASHRAE Research Project 1146-RP. ASHRAE, Atlanta
direction to implement peak control demand strategies
in real buildings. Formal testing of existing buildings SCE. 2007a. Southern California Edison critical peak pricing
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
will verify whether the recommendations are valid or rate schedules http://www.sce.com/NR/rdonlyres/
not. Feedback from building testing is expected to pro- B73F4175-162B-4C4F-B953-4E0A94863390/0/
vide more requirements for modeling. Several cycles CPPFactSheet0407.pdf
from building testing and modeling are necessary to cre-
SCE. 2007b. Southern California Edison large-commercial-
ate more effective and realistic control strategies.
building interruptible-service tarriff. http://
www.sce.com/NR/sc3/tm2/pdf/ce76-12.pdf
ACKNOWLEDGMENTS
Xing, H.-Y. 2004. Building load control and optimization.
The authors thank the American Society of Heating,
Dept. of Architecture Ph.D. theses, Massachusetts Insti-
Refrigerating and Air-conditioning Engineers, Inc.
tute of Technology, Cambridge MA
(ASHRAE) and the research manager, Mike Vaughn, for fund-
ing this work. Sincere appreciation is expressed to members of Xu, P. 2009, “Case Study of Demand Shifting with Thermal
the ASHRAE Project Monitoring Subcommittee: John House, Mass in Two Large Commercial Buildings,” ASHRAE
Steve Blanc, Carlos Haiad, Srinivas Katipamula, and Peng Xu. Transactions, vol. 115, pt. 2, Louisville 2009
474
SA-12-013
Development of a Tool to Improve

the Energy Efficiency of Existing
Commercial and Institutional Buildings
Angela Lewis, PE
ABSTRACT cial buildings in the United States consumed 5.8 quadrillion

British thermal units (BTUs) (6.1 x 1018 J) of energy. This
Many commercial and institutional buildings have signif-
icant opportunities to improve energy efficiency. However, resulted in 10.6 quadrillion BTUs (1.1x1025 MJ) of primary
determining which strategies to use can be challenging due to energy (site use, production and delivery losses) (EIA 2003).
limited funds, time and staff available to research and imple- Seventy-five percent of commercial building energy
ment new strategies. To help facility managers and building consumption is from lighting, space heating, cooling and
operators determine and prioritize energy efficiency strate- water heating. The remaining 25 percent of energy consump-
gies, a question driven decision support system tool was devel- tion is from ventilation, cooking, refrigeration and office
oped. This paper summarizes the development, testing and equipment (including elevators and telephones) (Swenson
evaluation of the tool. Results of the evaluation by 38 industry 1998).
participants found that decision support systems that combine As stated above, the overall economic and environmental
energy and maintenance decision making processes can be impacts of building operations has been quantified at a high
helpful and useful to support decision making processes to level. However, understanding building performance and how
transform existing buildings into proactively operated and to transform current operations practices into more proactive
maintained, energy efficient buildings. practices at the building, systems and equipment level is a
current industry challenge. Hinge et al. (2009) states that
INTRODUCTION “getting energy performance data should be easy … but what
Over the life cycle of commercial and institutional build- we’ve learned is that energy consumption can be a complex
ings, design and construction costs are between 5 to 10 percent issue, and the data that are available need some analysis.” This
of the total cost. Acquisition, renewal and disposal costs are challenge is compounded by current operational conditions of
about 5 to 35 percent of the life cycle cost, while operations building control systems. Rios (2005) finds that for every
and maintenance costs are between 60 to 85 percent (Christian building control system that is operated successfully, there are
et al. 1997). A study by the National Institute of Building hundreds that are underutilized and incapable of achieving
Sciences (NIBS 1998) found that employee salaries and bene- basic energy savings. Brambley, et al. (2005) states that many
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
fits are 88 percent of the building life cycle cost, operations control systems are “essentially ‘broken’” and are often
and maintenance are 11 percent and design and construction “malfunctioning or disabled”. Brambley et al. (2005) also
are 1 percent. Thus, operation and maintenance costs are a found that a lack of systematic information exists to address
significant component of the life cycle cost of commercial and the causes, find solutions and determine financial paybacks of
institutional buildings. possible solutions. These statements are further supported
Environmentally, buildings in the United States account quantitatively by a study of 60 commercial buildings by Piette
for 71 percent of electricity consumption and 38 percent of et al. (1996): more than half of the buildings studied had
carbon dioxide emissions (USGBC 2007). In 2003, commer- temperature control problems, 40 percent had heating, venti-
Angela Lewis is a PhD student in the School of Construction Management and Engineering, University of Reading, Reading, United Kingdom,
and a project professional at Facility Engineering Associates in Fairfax, Virginia, USA.
©2012 ASHRAE
475
lating and air conditioning equipment problems and about 33 However, without maintenance, even if the most energy effi-
percent had improperly operating sensors. cient HVAC systems and equipment are installed in a building,
The challenge is further compounded by a lack of funding over time the energy efficiency will decrease as belts start to
for effective maintenance planning, the underutilization of slip, bearings wear and sensors drift out of calibration.
computerized maintenance management systems (CMMS)
and facility information management systems (Joudah 1996; DEVELOPMENT OF THE DECISION
Sapp 2008). As well as a common view that maintenance does SUPPORT SYSTEM
not save money because buildings are a liability, not an asset A decision support system is a model developed using
(Alexander 1996). Even if an effective energy management management science concepts that can be used to analyze past
program is put in place, without proper maintenance, over activities and help make decisions that require assessment,
time the performance of equipment degrades and sensors and analysis and/or planning (Kroenke et al. 2010) to help make
meters drift out of calibration. This results in energy perfor- timely and relevant decisions (Markland 1989). The purpose
mance data that is of little value. In fact, a building with good of the decision support system described in this paper is to
operations and maintenance practices that is poorly designed align current practices and goals of facility managers for
often will outperform a well-designed building with poor energy, maintenance and human factors. Within this paper,
operations and maintenance practices (ASHRAE 2009). human factors are defined as the interactions of humans with
Three of the last four ASHRAE presidential themes have energy and maintenance management systems and processes.
demonstrated that existing buildings and sustainability are As defined by Wickens et al. (1998), human factors includes
important focus areas to ASHRAE and its members. During both brain and body interaction with a system. Examples of
2008-2009, Bill Harrison’s theme, Maintain to Sustain – human factors include facility management and organiza-
Delivering ASHRAE’s Sustainability Promise, emphasized tional strategy and training of building operators.
the role of energy efficiency and maintenance to meet sustain- The concept for the decision support system was devel-
ability goals (Harrison 2008). In 2009-2010, Gordon Holness’ oped through observations from three case studies (Lewis et
theme was Sustaining our Future by Rebuilding our Past, al. 2011) and further exploration of the case study observa-
Energy Efficiency in Existing Buildings – Our Greatest tions through rigorous analysis of the results of a questionnaire
Opportunities for a Sustainable Future (Holness 2009). Lynn to study the current practices, challenges and needs of energy
Bellenger, 2010-2011 President, emphasized sustainability and maintenance management practices in North America
within her theme, Modeling a Sustainable World. Bellenger’s (Lewis et al. 2010). The purpose of the case studies was to
speech referenced Holness’ speech, stating that energy effi- understand the challenges faced by facility management teams
ciency of existing buildings remains “our strongest opportu- while also providing a case study report that could benefit each
nity to make an immediate impact on worldwide energy facility management organization. The three facility manage-
consumption” (Bellenger 2010). The increasing importance of ment teams managed different building types: a community
operations and maintenance to ASHRAE members is also college district, a laboratory building on a college campus and
demonstrated by the 2008 publication of ASHRAE/ACCA a hospital. The three case studies were selected because they
Standard 180, Standard Practice for Inspection and Mainte- represented forward thinking facility management teams
nance of Commercial Building HVAC Systems, and the devel- committed to improving building operations practices. It
opment of ASHRAE Guideline 32 Sustainable, High should be noted that although an office building was not
Performance Operation and Maintenance. specifically selected, each of the facilities included office
Given the challenges identified, as well as the cost and space. To compare and synthesize the findings of the three
environmental impacts of buildings, new tools are needed to
help meet energy efficiency goals. This need will increase as
more goals are set to design, construct and operate net zero,
high performance buildings. This paper discusses the devel-
opment of a question-driven decision support system tool to
help facility managers and other members of the building
industry to transform existing building energy and mainte-
nance management processes into more proactive, data-driven
processes.
FOUNDATIONAL CONCEPTS OF THE DECISION

SUPPORT SYSTEM
It is well recognized that energy efficiency is a core
component of sustainability. However, the role of mainte-
nance is often not included in discussions of energy efficiency, Figure 1 Interdependent relationship between energy
even though the concepts are interrelated (Figure 1). efficiency and maintenance (Lewis et al. 2011).
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
476
ASHRAE Transactions
case studies, content analysis was performed. To complete the Coding Rules
content analysis, the Integrated Building Process Model As stated previously, the results of the questionnaire test-
(IBPM) for building operations (Sanvido et al. 1990; Guvenis ing were used to develop a set of coding rules for the decision
1989) was adapted to classify the content and determine what support system. The coding rules were used to develop the
energy and maintenance practices within the three case studies processing algorithms. The algorithms are used to link the user
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
could be considered standard practices. Classifying the data responses to the question set to the Best Practice Recommen-
using the IBPM allowed the data from the case studies to be dation Sheets (BPRs) and to determine the Proactive/Reactive
systematically and consistently organized to support the Score.
development of ten research propositions. A web-based ques-
tionnaire was developed to test the propositions. Further Best Practice Recommendation Sheets
discussion of the development and results of the questionnaire, A Best Practice Recommendation Sheet is a one-page
including a table of the propositions and hypothesis can be summary of a best practice that could be implemented at a
found in Lewis et al. (2010). The results of the proposition facility. The 35 BPRs developed provide a structured,
testing were used to develop the decision support system economical and quantifiable summary of a practice that can be
coding rules. used to meet the needs identified by the decision support
system. The goal of the BPRs is to provide an overview of an
COMPONENTS OF THE DECISION action an in-house facility management or consultant team
SUPPORT SYSTEM could implement to transition current practices to more proac-
tive energy, maintenance and human factors management
The decision support system includes three parts: a ques- practices. The BPRs include:
tion set, processor and a Recommendations Report (Figure 2).
By answering a set of multiple choice questions, the user • Tool format: Checklist, flow chart, score card, planning
receives a Recommendations Report that includes three Best guide, spreadsheet or informational brochure
Practice Recommendation Sheets and a Proactive/Reactive • Description of a specific best practice, highlighting key
Score to help the facility management team improve their components
energy, maintenance and human factors management prac- • Benefits of the specific best practice
tices. The question set includes topics such as energy and • Tips to help develop a business case for the use of the
maintenance goals, the use of building automation systems best practice
and computerized maintenance management systems, the
existence of a maintenance plan, data collection and bench- Figure 3 provides an example of a BPR, the energy perfor-
marking strategies used and types of maintenance activities mance indicator selection worksheet. The purpose of this
worksheet is to help a facility management team determine
performed on a route basis. The question set was developed so
what metrics, measurements and benchmarks to use to quan-
that all of the questions could be answered by someone famil-
tify building energy performance. As shown in the figure, the
iar with the day to day practices of the facility, without the need BPR includes a description of the tool, a brief summary of why
to look up information or input data from utility bills, main- to use the process described on the BPR, how to use the energy
tenance reports or other documents. performance indicator worksheets and screen shots of the
worksheets. Although further information could be provided
on the BPR, it was necessary to limit the development of each
BPR to one page for the current scope of research.
Proactive/Reactive Score
As the researcher developed the concepts for the decision
support system, insight was gathered from industry practitio-
ners about what would be useful and helpful for inclusion
within the tool. One repeated request was a simple graph that
quickly identified how proactive or reactive the practices at a
facility were. Thus, the researcher developed a bar graph from
negative five to positive five for each of the three categories,
called the Proactive/Reactive Score (Figure 4). Negative five
indicates the facility is nearly to entirely reactive and positive
five indicates the facility is proactive at a world class level. To
understand what each numeric rating indicates, a qualitative
Figure 2 Components of the decision support system. value is assigned at each 1.0 interval. The score for a specific
©2012 ASHRAE
477
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 3 Example of a best practice recommendation sheet.
478
ASHRAE Transactions
Facility management teams with multiple buildings and
multiple campuses of buildings were considered significantly
during the decision support system development. Organiza-
tions with multiple buildings, especially those with multiple
campuses with different energy and maintenance management
practices, have a greater opportunity to benefit from the use of
the decision support system. The larger the number of build-
ings that the recommended practices can be applied in a stan-
dardized manner, the greater the potential for cost and time
savings.
Figure 4 Example of proactive/reactive score.
TESTING AND EVALUATION OF THE DECISION
SUPPORT SYSTEM
facility is determined from the responses to the question set After the programming and debugging were complete,
provided by the user. Many of the question choices are the decision support system was tested by 56 industry partic-
assigned a numeric value within the processing algorithms ipants. Thirty-one of the test participants completed an eval-
that calculate the total score. uation of the decision support system through a web-based
questionnaire.
Decision Support System Coding
To recruit testing participants, the researcher established
Nine different software options were evaluated to deter- a contacts database over the duration of the research by inter-
mine what software to use to create the processing algorithms. acting with industry professionals at conferences and posting
The evaluation determined that the use of LAMP (Linux, information on the research website. Additionally, the
Apache, MySQL and PHP) used with a commercially avail- researcher posted information about the testing process on
able survey tool as the frontend would best meet the develop- multiple professional networking sites and listservs. This
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
ment needs for the prototype tool. Using a commercially effort resulted in 20 test participants (34%) from the contacts
available survey tool for the question set (frontend) was valu- database and 38 test participants (66%) from listservs and
able because most computer users are generally familiar with networking sites.
survey tools. LAMP was selected because it provided the flex- The testing process found that it took an average of one
ibility necessary to support a web-based output (the Recom- hour and 21 minutes to complete the question set. This is a
mendations Report), supports commonly used databases significant finding because the use of benchmarking tools and
(such as MySQL), and PHP (hypertext preprocessor) is a condition assessments can take between eight hours and one
commonly known computer language used by software devel- week to gather the data before a recommendation can be made.
opers for web-based applications. Thus, the decision support system can provide a recommen-
dation more quickly, allowing planning and implementation to
USERS OF THE DECISION SUPPORT SYSTEM start sooner. This may be especially to helpful to facility
Through the foundational components of the research, the management teams who have very limited staff and funding to
three case studies and the questionnaire, as well as multiple support process improvements.
discussions with industry members, the researcher found that at
a high level, many of the operations and maintenance challenges Sample Decision Support System Output
in commercial and institutional buildings were similar. Thus, Each user who answers the question set receives a
the decision support system was developed to be useful to many Recommendations Report. The Recommendations Report
commercial and institutional building types, including: includes an overview of how the Recommendations Report
was generated, a list and brief descriptions of the three recom-
• Dormitory/Barracks mended BPRs, web links to the BPRs and the Proactive/Reac-
• Hospital tive Score. The BPRs are provided in a prioritized (numbered)
• Laboratory order to suggest which BPR to review, and if appropriate,
• Lodging/Hotel implement first. The prioritized order is determined based on
• Manufacturing the user’s responses to the question set. Two examples of the
type of user specific content within the Recommendations
• Multi-family residential
Report is provided below. The first example is for a university
• Office, both medical and non-medical and the second is for a hospital.
• Public assembly, such as library, museum and church A university facility manager completed the question set
• Restaurant for a single building built before 1969. Maintenance goals,
• Retail such as moving from reactive to proactive practices when
• Schools, including K-12 and college/university economically justifiable, were set before the question set was
©2012 ASHRAE
479
answered. The facility manager was interested in using the compared to buildings considered best in class, the energy and
tool to help determine energy efficiency goals. The resulting maintenance management practices are somewhat proactive.
Recommendation Report suggested the following BPRs: The human factors management practices are very proactive.
The largest opportunities for improvement are maintenance
1. Energy Program Planning Pyramid: A set of bounded and energy management practices. Human factors manage-
steps to identify goals and determine criteria to implement practices are the most proactive and nearing proactive at
ment an energy efficiency program. Bounded steps are a world class level. As the Proactive/Reactive Score for main-
summarized to provide high level guidance for teams tenance is the lowest, it is not surprising that the recommended
who have limited time to develop a program. BPRs are for maintenance.
2. Energy Efficiency Incentives and Justification Planning
Guide: A list of databases and resources to find tax Evaluation of the Decision Support System
rebates and incentives to support energy efficiency proj- A web-based questionnaire with 30 questions was sent to
ects, as well as guidance to help justify energy efficiency the 56 test participants. Thirty-one (55 %) of the test partici-
to financial decision makers. pants completed the evaluation. Given the web-based struc-
3. Energy Indicator Selection Worksheet: A structured ture of the questionnaire, it was not possible to determine the
multi-tab spreadsheet to help facility management teams reason for the 45 percent non-response rate. Thus, it should be
determine what measurements, metrics and benchmarks noted that the data likely reflects a non-response bias, as well
to use to quantify energy consumption at the building, as an optimism bias. An optimism bias likely occurred because
system and equipment level. an evaluator, regardless of the topic of study, who likes what
he/she is evaluating is more likely to participate in the evalu-
The Proactive/Reactive Scores were -0.98 for energy
ation process.
management, 0.42 for maintenance and 1.13 for human
The goal of the evaluation was to determine how helpful
factors. This suggests that compared to buildings considered
and useful the decision support system was to industry partic-
to be best in class, the energy management practices are reac-
ipants. Within the evaluation, helpful was defined as the
tive, but nearly proactive. The maintenance practices consist
amount of assistance the system provided to the decision
of both proactive and reactive practices and the human factors
making process and useful was defined as the amount of prac-
management practices are proactive, but nearly reactive. As
tical and/or beneficial use of the decision support system. The
scores can range from negative five to positive five, there are
evaluation was limited to the use of a questionnaire because
multiple opportunities to move towards more proactive
the quantification of measurable energy and/or cost savings
energy, maintenance and human factors management prac-
would take several years of data to validate. Wireman (2010)
tices.
finds that it can take three to five years for a reactive mainte-
A hospital facility manager completed the question set for
nance organization to transform into a proactive organization
a campus of hospital buildings, many built before 1969.
that utilizes best practices. The collection of quantitative
Although maintenance goals had been set, 60 percent of main-
energy and/or cost savings data would be further complicated
tenance practices were reactive. Although a maintenance plan
because it would be difficult to differentiate changes in energy
is used, it includes the minimum amount of information and
and/or maintenance management practices as a result of using
does not document how to verify if work was completed as
the decision support system, compared to other organized
required. The resulting Recommendation Report suggested
factors.
the following BPRs:
Demographics of Evaluation Participants
1. Maintenance Approach Evaluator: A spreadsheet to
determine the capital and life cycle costs of using reac- Four questions within the evaluation were used to collect
tive, preventive, predictive and reliability centered main- demographic information: the role of the participant within the
tenance for different types of mechanical equipment. industry, type of service(s) provided, if the testing process was
2. Work Scheduling Prioritization Calculator: A spread- completed for a single building or campus of buildings, and the
sheet to plan, schedule and evaluate completed work type(s) of building(s) the testing process was completed for.
orders to determine how to more effectively prioritize The participants represented various industry roles: facil-
maintenance, given limited budgets and staff time. ity managers, mechanical engineers, energy engineers, facility
3. Critical Equipment Selector: A set of criteria to help management consultants, maintenance technicians, commis-
prioritize maintenance needs using the ranking index for sioning agents, architects and maintenance service providers.
maintenance (RIME) method. Fifty percent of the participants were facility managers. The
second and third largest roles represented were mechanical
In contrast to the university example, the hospital team’s engineers (13%) and energy engineers (10%). The services
goals focus more strongly on maintenance management. The provided by the participants included energy management or
scores were 2.43 for energy management, 1.65 for mainte- commissioning (36%), maintenance (33%), construction
nance and 4.06 for human factors. This suggests that (15%), design (15%) and various types of consulting services
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
480
ASHRAE Transactions
(1%). Fifty percent completed the testing process for a single only be somewhat useful. None of the evaluators stated that the
building, while the other fifty percent of the participants tool was not at all useful.
completed the testing for a campus of buildings. Fifteen differ- The evaluators were asked what the most useful part of the
ent building types, including office, public assembly, college/ decision support system was (Figure 7). Evaluators were to
university, laboratory, medical office building, manufacturing select one of the following: full Recommendations Report,
and dormitory/barracks were included. Offices (not medical) Proactive/Reactive Scores, rank ordering of the BPRs, infor-
represented 31 percent of the test population, while the other mation contained on the BPRs, user guide or question set. The
types of buildings each represented between two and 11 full Recommendations Report received the highest percentage
percent of the total. of responses, 39 percent. This demonstrates the decision
support system as a whole is useful to industry practitioners
RESULTS because the full Recommendations Report includes the rank
Overall, the results of the testing process revealed that the ordered list of BPRs, information on the BPRs and Proactive/
decision support system was rated as helpful to very helpful Reactive Scores. The set of questions and user guide provide
and useful to very useful. Additionally, the information the structure for the user to receive and understand how to use
conveyed by the decision support system was generally the full Recommendations Report, so it is not surprising these
communicated clearly or very clearly. components were ranked the lowest.
To evaluate how helpful the decision support system was Although the information on the BPRs were not selected
for combining energy and maintenance management deci- as the most useful component of the decision support system,
sions, a question with a five point scale ranging from not at all most of the evaluators found the BPRs to be either useful or
helpful to extremely helpful was used. As shown in Figure 5, very useful (Figure 8). As shown in the figure, between 31 to
all participants found the decision support system to be at least 53 percent of evaluators found the BPRs to be useful, while 26
helpful (53 percent), 40 percent found the decision support to 54 percent found them very useful. One area of improve-
system very helpful and seven percent found the decision ment suggested by evaluators was to include more content on
support system to be extremely helpful when combining the BPRs.
energy and maintenance management decisions within the
A five point scale was provided for the evaluators to indi-
same process.
cate how clear the information of various components of the
The decision support system was evaluated to determine Recommendation Report was communicated (Figure 9). The
how useful it was for moving from reactive to proactive evaluators generally found the decision support system to
energy, maintenance and human factors management prac- convey the information presented clearly (between 43 to 46%)
tices, using a five point scale. As shown in Figure 6, the deci- or very clearly (36 to 44%) for all four of the components eval-
sion support system was found to be either useful or very uated. None of the evaluators found the four components eval-
useful for moving from reactive to proactive energy, mainte- uated to be very unclearly communicated.
nance and human factors management practices by most of the
evaluators. Between 7 percent to 11 percent found the decision Evaluators were asked about the likelihood they would
support system to be extremely useful to transition from reac- use the decision support system again and if they would
tive to proactive practices. Thirty-six to 57 percent found the recommend the use of the tool to a colleague. Forty-eight
tool to be very useful, while 26 to 39 percent found the tool to
be useful. In contrast, only 7 to 14 percent found the tool to
Figure 6 Usefulness of decision support system to

Figure 5 Helpfulness of the decision support system to transition from reactive to proactive energy,
support combined energy and maintenance maintenance and human factors management
management decisions. practices.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
©2012 ASHRAE
481
mend the decision support system to a colleague and 21
percent would very likely recommend it. In contrast, only 4
percent would unlikely recommend the decision support
system to colleague and seven percent would very unlikely
recommend it.
Qualitative, open-ended feedback was requested as part
of the evaluation. Some of the comments received included:
• “The [decision support system] takes advantage of eas-

ily accessible technology to acquire, contain and ana-
lyze information about a particular [facility]... thus joins
the user to colleagues in helping to set priorities.”
• It “can provide some valuable guidance for facility man-
Figure 7 Usefulness of decision support system agement groups that utilize” it
components.
• “A fee or donation or donation for use could certainly be
acceptable” for use
Overall, the qualitative comments received were support-

ive and multiple comments suggested that value could be
added to the decision support system through further develop-
ment of the BPRs and refinement of the question set.
SUMMARY AND CONCLUSIONS
A decision support system was developed to align two

interdependent concepts that are often researched and applied
separately. By answering a set of multiple choice questions
about the energy and maintenance management practices for
a facility, the tool users receives a Recommendations Report
that includes three Best Practice Recommendation Sheets and
Figure 8 Usefulness of BPRs. a Proactive/Reactive Score. The testing and evaluation of the
decision support system by industry participants found that for
the test population, the tool was helpful for facility manage-
ment teams to combine energy and maintenance management
decisions when seeking to move from reactive to proactive
energy and maintenance management practices for heating,
ventilating and air-conditioning systems. It is concluded that
to transform existing buildings into energy efficient, high
performance, green buildings, new tools that include both
energy and maintenance management in the decision making
process, such as the tool discussed within this paper, are
needed. To further quantify and solidify the findings described
in this paper, the decision support system should be further
developed and tested by a larger sample using a longitudinal
Figure 9 Clarity of contents of Recommendations Report. case study approach.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
ACKNOWLEDGMENTS
percent of respondents would likely use the decision support
A portion of this research was completed using funding
system again and 10 percent would very likely use the decision
from an ASHRAE Graduate Student Grant-In Aid with the
support system again. In contrast, only 3 percent would be distinction of “Life Member Club Grant Recipient.” The
unlikely to use the decision support system again, while 10 author also thanks all those who participated in various stages
percent would very unlikely use the decision support system of the research, including the case studies, questionnaire, deci-
again. Forty-three percent of respondents would likely recom- sion support system development and evaluation.
482
ASHRAE Transactions
REFERENCES Lewis, A., A. Elmualim,D. Riley. 2011. Linking energy and
maintenance management for sustainability through
Alexander, K. 1996. Chapter 1: Facilities management Facil-
three American case studies. Facilities 29(5/6): 243.
ities Management Theory and Practice. K. Alexander.
Lewis, A., A. Elmualim, D. Riley. 2010. Development of a
London, E &FN Spon.
framework for improving building operating decisions:
ASHRAE. 2009. The Decision-maker's guide to energy effi- findings from the questionnaire. International Confer-
ciency in existing buildings. Atlanta, GA, ASHRAE. ence on Innovation in Architectural Engineering and
ASHRAE. 2008. ASHRAE/ACCA Standard 180-2008: Construction Proceedings, State College.
Standard practice for inspection and maintenance of Markland, R. E. 1989. Topics in management science. Hobo-
commercial building HVAC systems. Atlanta, GA. ken, NJ, John Wiley and Sons
Bellenger, L.G. 2010. Modeling a sustainable world. NIBS. 1998. Excellence in facility management. National
ASHRAE Presidential speech, 2010-2011. Institute of Building Sciences, Washington DC.
Joudah, A. 1996. Developing facilities information manage-
Brambley, M., P. Haves, et al. 2005. Department of energy
ment systems. Facilities Management Theory and Prac-
advanced controls R & D planning workshop. June 11,
tice. K. Alexander. London, E &FN Spon.
2003. Washington DC.
Piette, M. A. and B. Nordman. 1996. Costs and benefits of
Christian, J. and A. Pandeya (1997). Cost predictions of utility funded commissioning of energy-efficiency mea-
facilities. Journal of Management in Engineering. 13(1): sures in 16 buildings.” ASHRAE Transactions 102(Part
52-61. 1).
EIA. 2003. Commercial buildings energy consumption sur- Rios, J. A. 2005. Building controls and green buildings.
vey. www.eia.doe.gov/emeu/cbecs/contents.html. HPAC Engineering 77(9):9,12.
Guvenis, M. M. 1989. Integrated facility operation process Sanvido, V., S. A. Khayyal, et al. 1990. An Integrated build-
model, Technical report #6. University Park, The Penn- ing process model, Technical report #1. V. Sanvido. Uni-
sylvania State University. versity Park, The Pennsylvania State University.
Harrison, W.A. 2008. Maintain to Sustain – Delivering Sapp, D. 2008. Computerized maintenance management sys-
ASHRAE’s sustainability promise. ASHRAE Presiden- tems (CMMS). www.wbdg.org/om/cmms.php.
tial speech, 2008-2009. Swenson, A. 1998. A Look at commercial buildings in 1995:
Characteristics, energy consumption and energy expen-
Hinge, A. W. and D. J. Winston. 2009. Documenting perfor-
ditures, Department of Energy, Energy Information
mance does it need to be so hard? High Performing
Administration, Office of Energy Markets and End Use.
Buildings. Winter 18-23.
USGBC. 2007. A National green building research agenda.
Holness, G. 2009. Sustaining our future by rebuilding our United States Green Building Council. Washington,
past, Energy efficiency in existing buildings – Our great- D.C.
est opportunity for a sustainable future. ASHRAE Presi- Wickens, C. D., S. E. Gordon, et al. 1998. An Introduction to
dential speech 2009-2010. human factors engineering. New York, Longman.
Kroenke, D. M. and D. J. Auer. 2010. Database processing Wireman, T. 2010. Benchmarking Best Practice in Mainte-
fundamentals, design and implementation. Boston, nance Management. Second Edition. New York, Indus-
Prentice Hall trial Press.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,
©2012 ASHRAE
483
SA-12-014
Laboratory Testing of a
Fabric Air Dispersion System
D. Kulkarni A.N. Nalla S. Idem, PhD K. Gebke

Student Member ASHRAE Member ASHRAE Member ASHRAE
ABSTRACT piezometric rings) be located upstream and downstream of a

flexible/fabric duct test section, in order to measure the result-
Friction factors for a 368 mm (14.5 in.) diameter round non- ing pressure loss. Because of concerns about the disparity in
porous fabric air dispersion system (fabric duct and fabric the fabric duct/steel duct diameter at the terminal connections,
diffuser combination product) were measured per ASHRAE it was suggested that additional pressure loss tests be
Standard 120 (2008). Two configurations were tested. In one performed on fabric ducts using Pitot-static tubes mounted at
instance an internal frame (to maintain a round cross section the duct centerline, at axial locations prescribed by ASHRAE
when no air pressure was in the system) was absent. For other Standard 120. The results of the revised test program based on
tests a skeletal metal frame was inserted into the fabric tube to that recommendation are reported herein.
maintain the tube shape in the absence of static air pressure.
Fabric ducts are frequently used to supply and disperse air
The resulting pressure loss data suggest that 368 mm (14.5 in.)
in open ceiling architecture. Common applications include
diameter round fabric air dispersion systems lacking an internal
factories, warehouses, and gymnasiums, etc. Fabric ducts are
support frame should be categorized as "MEDIUM SMOOTH"
made from specific blends of fabrics rather than from metal or
in the Duct Design chapter of the ASHRAE Handbook of Funda-
plastic. In order to ensure safety, the material may be treated
mentals. Round fabric air dispersion systems possessing a skel-
to be both heat-resistant and flame-retardant. Air passing
etal metal internal support frame can be categorized as
through the fabric walls eliminates the risk of condensation,
"MEDIUM ROUGH" in the Handbook.
and minimizes dust accumulation on the duct surfaces.
Compared to sheet metal duct and diffuser systems, fabric air
INTRODUCTION dispersion systems are lighter and less expensive. Simple
This paper presents pressure loss data for a 368 mm suspension systems reduce installation time. Hence fabric
(14.5 in.) diameter non-porous polyester fabric duct with an duct systems are cost effective, and can be an aesthetically
attractive alternative to metal ductwork. However, there is a
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
acrylic/urethane coating. The presence of the coating ensures

that air leakage through the fabric is negligible when the lack of published performance data for fabric duct system
system is pressurized within the limits of manufacturer’s spec- components available to design engineers through such refer-
ifications. The measurements were performed in accordance ences as the ASHRAE Duct Fitting Database (2008) and the
with ASHRAE Standard 120 (2008). Previous pressure loss Duct Design chapter of the ASHRAE Handbook of Funda-
tests performed on a similar fabric duct were reported in Idem mentals (2009). Hence this project will contribute to the
et al. (2011). They concluded that the friction factor data were improved designs of fabric duct systems.
dependent on the particular type of connection used to join the
fabric ducts to the 356 mm (14.0 in.) diameter rigid steel ducts EXPERIMENTAL PROGRAM
at the terminal ends of the test section. That test arrangement The test apparatus shown in Figure 1, which depicts the
was mandated by requirements in ASHRAE Standard 120, measurement planes employed in this study, was in compli-
which stipulate that static wall pressure taps (arranged to form ance with ASHRAE Standard 120. The test duct was
D. Kulkarni is a graduate student and S. Idem is a professor in the Department of Mechanical Engineering, Tennessee Tech University,
Cookeville, TN. A.N. Nalla is an aftermarket engineer with Cummins Filtration, Stoughton, WI. K. Gebke is a new product development engi-
neer with DuctSox Corporation, Peosta, IA.
484
ASHRAE Transactions
Figure 1 Fabric duct test setup.
comprised of a single piece of fabric, i.e., there were no trans-

verse joints, and one longitudinal seam was present running
down the entire length of the specimen. From plane 1 to
plane 2, the duct had an overall length of 30.5 m (100 ft). It was
suspended from the laboratory ceiling by means of a hangar
system per the manufacturer’s recommendations; refer to
Figure 2. Two configurations were tested. In one instance there
was no internal frame to maintain a round cross section when
no static air pressure was in the system. In order to support the
structure of the fabric section in that case, the minimum inter-
nal static pressure maintained in the experiments was 125 Pa
(0.5 in. wg). For other tests a skeletal metal frame was inserted
into the duct to maintain the duct shape in the absence of static
air pressure. A schematic diagram of the frame is provided in
Figure 3. The ring and spoke components of the frame
consisted of 6.4 mm (0.25 in.) diameter steel rod, and the
diameter of central aluminum tube was 20.3 mm (0.8 in.). The
separation distance L' between the rings equaled 1.88 m
(74 in). There were six equally-spaced spokes mounted to the
terminal rings of the skeletal metal frame, whereas the interior
rings had four spokes. The outer diameter of the ring elements
closely matched the inner duct diameter. Hence the maximum
Figure 2 Fabric duct suspension system.
reduction of the air flow cross section due to the presence of
the skeletal frame was 13.5%.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
The test setup included a 20 hp centrifugal fan, followed outlet; refer to Figure 4. The open area of the orifice was
by a cylindrical nozzle chamber which was used for flow adjustable. Screens mounted upstream and downstream of the
measurement. Primary control of air flow through the test duct nozzle board inside the chamber were used to settle the flow.
was accomplished by means of a variable frequency drive, The system was blow-through in nature. The nozzle board
which was used to control the fan speed. Additional control of contained three long-radius spun aluminum flow nozzles
air flow through the test duct was achieved using a porous having throat diameters of 89 mm (3.5 in.), 152 mm (6 in.) and
fabric orifice mounted at the downstream metal test duct 254 mm (10 in.). The nozzles were mounted on a 25 mm (1 in.)
©2012 ASHRAE
485
Figure 3 Skeletal metal frame.
Table 1. Test Set-Up Dimensions
LZ – 1 L1 – A LA – C LC – D LD – B LB – 2
------------
- -------------- -------------- --------------
- --------------- --------------
D D D D D D
12.9 18.0 3.3 35.2 11.2 15.0
measurement locations in the fabric duct provided an entrance

duct section to achieve fully developed turbulent flow
(upstream length), the test section, and a tail portion (down-
stream length). The dimensions conformed to the require-
ments of ASHRAE Standard 120. Pressure taps arrayed in a
piezometric ring were mounted on the rigid steel ducts in order
to obtain an auxiliary measure of the pressure loss across the
entire length of the fabric duct. When an internal support was
inserted into the test apparatus, the pressure loss over the
length of the frame was measured using the static pressure
ports of the Pitot-static tubes. Under those circumstances the
pressure taps mounted on the rigid steel ducts were not used to
measure the pressure loss. Pertinent test section dimensions
are provided in Table 1. The downstream Pitot-static tube was
Figure 4 Fabric orifice. situated after the end of the support structure, so that wake
effects impacting that sensor were minimized. As outlined
subsequently, the tare pressure loss of the fabric duct was
thick plywood board. Various combinations of flow nozzles subtracted from the pressure drop measured using the Pitot-
were employed, depending on the desired flow rate. Unused static tubes in order to evaluate pressure loss per unit length
nozzles were blocked using smooth vinyl balls. The pressure over the internal metal frame, i.e., from plane C to plane D.
drop was measured by two piezometer rings located 38 mm Pressure drop measurements over the pressure taps
(1.5 in.) on each side of the nozzle board, with both sides mounted on the rigid steel ducts and across the nozzle board
connected to a single manometer to measure differential pres- were performed using liquid-filled micromanometers having
sure. a measurement accuracy of 0.025 mm (0.001 in.). Like-
Tests were performed with the fabric duct connected to wise, the static pressure upstream and downstream of the test
356 mm (14 in.) diameter steel ducts located upstream and section and the pressure loss indicated by the Pitot-static
downstream of the fabric duct test section. The terminal ends tubes was measured by means of electronic manometers
of the fabric duct were attached to the rigid steel ducts with having a readability of 0.25 mm (0.01 in.). Similarly static
minimum overlaps of 50 mm (2 in.), and clamped tightly using pressure in the nozzle chamber was measured using an elec-
straps in order to make an airtight connection. Pitot-static tronic manometer having the scale readability of 0.25 mm
tubes aligned precisely with the air flow were mounted (0.01 in.). The air temperature in the nozzle chamber was
upstream and downstream of the fabric duct test section to measured using a mercury thermometer having a scale read-
measure the pressure loss per unit length. The Pitot-static tube ability of  0.5C (1.0F). The dry-bulb and wet-bulb
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
486
ASHRAE Transactions
temperatures of the ambient air were measured using an aspi- Additional equations necessary to support the flow calcu-
rated psychrometer, with an accuracy of 0.5C (1.0F). The lation per Equation 4 can be found in ASHRAE Standard 120.
test section temperature was not measured directly, but was Determination of the flow rate required the measurement of
assumed to be the same as the temperature of the air inside the pressure drop across the nozzle board, the static pressure
the nozzle chamber. Ambient pressure was measured with a of the chamber, and the temperature inside the chamber. The
Fortin-type barometer, with an accuracy of 0.25 mm (0.01 density of air in the test section was calculated by means of the
in.) of mercury. All measurements of temperature and pres- correlations presented in ASHRAE Standard 120, based on
sure in this project were in compliance with ASHRAE Stan- measurements of the ambient dry-bulb and wet-bulb temper-
dard 120. atures, barometric pressure, test section temperature, and aver-
age static pressure.
DATA REDUCTION The Reynolds number in the test section was determined
Referring to Figure 1, when the internal support frame by Equation 5.
was not present the pressure loss p s 1 – 2 was measured  1 V 1 /  D 1 /1000 
directly by means of the pressure taps mounted on the rigid Re 1 = ---------------------------------------
- (SI) (5)
1
steel ducts. Likewise the pressure loss p s A – B was evaluated
using the static pressure ports of the Pitot-static tubes.  1  V 1 /60 /  D 1 /12 
However, when the internal frame was inserted into the appa- Re 1 = ----------------------------------------------
- (I-P) (5)
1
ratus the pressure loss was calculated by Equation 1:
The average velocity in the test section ‘V’ was defined by
p f
p s C – D = p s A – B –  L A – C + L D – B   -------- (1) the continuity equation as
 L 
Q 1 /1000
The terms LA-C and LD-B represent the separation distance V 1 =  -------------------- (SI) (6)
A
between the upstream Pitot tube static pressure ports and the
entrance plane of the frame, and the exit plane of the frame and Q1
the downstream Pitot tube static ppressure ports, respectively. V 1 =  ------ (I-P) (6)
A
The friction loss per unit length --------f is the duct tare pressure
L
loss per unit length, as calculated by Equation 2: The measured pressure loss data were plotted on a Moody
diagram in terms of friction factor ‘f’ as a function of relative
p f b roughness ‘/D’ and Reynolds number. These quantities are
-------- = aV (2)
L related by the Colebrook equation
The coefficients ‘a’ and ‘b’ were determined by a least 1 /D 2.51
squares curve fit, based on Pitot tube pressure loss data ----- = – 2 log -----------1 + --------------- (SI) (7)
f 3.7 Re f
obtained in the absence of the support frame. 1
The Darcy friction factor was calculated by Equation 3.

1 = – 2 log 12/D 2.51 -
-----------------1- + --------------
----- (I-P) (7)
p s i – j /L i – j f 3.7 Re 1 f
f = ----------------------------------------
- (SI) (3)
1--- 2
 1 V 1 /  D/1000 
2 The relative roughness was determined iteratively by
fitting the experimentally determined friction factors to the
p s i – j /L i – j Colebrook equation, using the least squares method. Guessed
f = ---------------------------------------------------------
- (I-P) (3)
1--- 2 relative roughness values were substituted successively into
 1  V 1 \/1097  /  D/12 
2 the Colebrook equation until the square of the difference of the
calculated and experimentally determined friction factors
In Equation 3 the evaluation planes i-j refer to 1-2, A-B,
approached zero.
or C-D, depending on the particular circumstances of the
measurement. The flow rate for each test point was calculated All dimensional measurements were assumed to have an
by Equation 4, where 5 denotes the section upstream of the accuracy of ±1%. In some instances, the measurement uncer-
nozzle and 6 indicates the nozzle throat. tainty of a parameter exceeded the basic scale readability of a
particular instrument. For example, that occurred when
2  p s 5 – p s 6  random fluctuations in the system static pressure were present,
-   Cn An 
Q 1 = 1000Y n -------------------------------- (SI) (4) and those fluctuations exceeded the scale readability of the
5
manometer. Estimates of the measurement uncertainty of
several quantities are presented in Table 2, for the conditions
2  p s 5 – p s 6 
-   Cn An 
Q 1 = 1098Y n -------------------------------- (I-P) (4) typically encountered in the experiments. The friction factor
5 measurements were subjected to an uncertainty analysis based
©2012 ASHRAE 487

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---

on the method of Kline and McClintock (1953), as prescribed range 2 × 105 to 6 × 105. The results are plotted in Figure 5.
by ASHRAE Standard 120 for random variations of the Two configurations were tested. In one instance an internal
measurands. In every instance the measurement uncertainty frame to maintain a round cross section when no static air pres-
estimates were performed with a 95% confidence level. sure was in the system was absent. Under these circumstances
the measured duct relative roughness based on Pitot tube pres-
RESULTS sure loss data equaled 0.0003. Likewise, the relative rough-
Friction factors for a 368 mm (14.5 in.) diameter non- ness determined using pressure taps mounted on the rigid steel
porous fabric duct were measured over the Reynolds number ducts located at each end of the fabric duct yielded a relative
roughness of 0.0005. For the case where a skeletal metal frame
was inserted into the fabric duct, the pressure loss measure-
Table 2. Uncertainties in Measured Parameters ments over the test section performed using Pitot-static tubes
yielded a relative roughness of 0.0046. In every instance when
0.6C
Dry Bulb Temperature plotted on a Moody diagram, the data closely followed a single
(1F)
relative roughness curve. The horizontal bars through the data
0.6C points represent the range of expected uncertainty in the
Wet Bulb Temperature
(1F) measured Reynolds number, with a 95% confidence limit, and
the vertical bars through each point depict the range of
0.6C
Plenum Chamber Temperature expected uncertainty in the measured friction factor.
(1F)
0.6C The absolute roughness values are summarized in Table 3

Test Section Temperature in terms of the terminal fabric duct/steel duct connection, and
(1F)
the presence/absence of an internal skeletal metal support
25 Pa frame. For cases where an internal support frame was absent
Plenum Chamber Static Pressure
(0.1 in. wg) the absolute roughness of the ducts based on test data obtained
5 Pa using Pitot-static tubes was 0.11 mm (0.0004 ft). This value lies
Pressure Drop Across Nozzle Chamber
(0.02 in. wg) within the duct roughness range described as "MEDIUM
SMOOTH" in the Duct Design chapter of the Handbook. When
2.5 Pa
Test Section Static Pressure an internal support frame was present the absolute roughness of
(0.01 in. wg)
the ducts based on Pitot tube test data was 1.69 mm (0.0056 ft).
Barometric Pressure
0.25 mm Hg This conforms closely to the category of "MEDIUM ROUGH"
(0.01 in. Hg) in the Duct Design chapter of the Handbook.
Figure 5 Moody diagram for 368 mm (14.5 in.) diameter fabric duct.
488
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
ASHRAE Transactions
Table 3. Test Results for Fabric Ducts
Internal Frame Absent Internal Frame Present

Pressure Loss
Measurement Relative Roughness Absolute Roughness , Relative Roughness Absolute Roughness ,
Method /D mm (ft) /D mm (ft)
0.11 1.69
Pitot-Static Tubes 0.0003 0.0046
(0.0004) (0.0056)
0.18
Pressure Taps 0.0005 — —
(0.0006)
PREDICTED DUCT PRESSURE LOSS Table 4. Predicted Pressure Loss Pa/m (in. water/100
A parametric study was performed to predict the pressure ft) for Fabric Ducts
loss performance of 368 mm (14.5 in.) diameter fabric ducts
Q,
over a range of typical volumetric flow rates. Two cases were
L/s (ft3/min)
considered, namely (i) the internal frame was absent and (ii)
the internal frame was present. In the former case the absolute Test 236 472 708 944 1180
roughness was assumed to equal 0.11 mm (0.0004 ft), whereas Condition (500) (1000) (1500) (2000) (2500)
in the latter case the absolute roughness was taken to be 1.69 mm
(0.0056 ft). The pressure loss for a prescribed duct length was Internal
0.172 0.621 1.323 2.271 3.471
Frame
obtained by means of Equation 8, which is referred to as the (0.021) (0.076) (0.162) (0.278) (0.425)
Absent
Darcy equation.
Internal
1000fL V
2 0.253 0.980 2.181 3.863 6.012
p f = ----------------- --------- (SI) (8) Frame
D 2 (0.031) (0.120) (0.267) (0.473) (0.736)
Present
V 2
p f = -----------   ------------
12fL
(I-P) (8)
D  1097 presence of the internal skeletal metal frame increased the
measured relative roughness in the duct, relative to the case
The pressure loss was calculated in units of Pa/m (in. of where no frame was used. In addition, accurate results were
water per 100 feet of duct length) for flow rates ranging from 236 critically dependant on where the friction factor data were
L/s (500 ft3/min) to 1180 L/s (2500 ft3/min). The Colebrook taken and whether pressure taps mounted on the terminal rigid
equation was used to solve iteratively for the friction factor by steel ducts or Pitot-static tubes were used to measure the pres-
means of a standard root-solving procedure. In every instance sure loss in the test section. The disparity between the resulting
standard conditions of temperature and pressure were assumed relative roughness values is consistent with test data reported
when calculating air thermal properties. The resulting predicted in Idem et al. (2011). They attributed this variation to the pres-
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
pressure loses are presented in Table 4. For the specified duct ence of transitions in the fabric duct adjacent to the pressure
diameter the flow rates correspond to air velocities ranging from taps mounted on the steel ducts, which were caused by the
2.21 m/s (436 ft/min) to 11.1 m/s (2180 ft/min). Current fabric different diameters of the fabric and steel ducts. It was postu-
air dispersion system design guidelines dictate that maximum lated this affected the measured pressure loss in the test
airflow velocities should be limited to approximately 8.13 m/ section. It is standard practice in the fabric air dispersion
s (1600 ft/min), in order to reduce stress and noise, and to system industry to manufacture the fabric duct larger than the
obtain a better balanced system. When the fabric air dispersion rigid supply duct to which it is attached, in order to facilitate
system has an internal skeletal metal frame that tensions the the mounting of the fabric duct. Typical fabric duct installa-
walls of the system, the maximum can rise to 12.7 m/s tions do not employ a connection to a steel duct at the fabric
(2500 ft/min) or higher, depending on the internal static air system outlet. It is suggested that the static pressure ports of
pressure of the system. The system design guidelines also Pitot tubes mounted at the centerline of the test duct be
show that increased duct roughness and pressure losses assist employed to measure the pressure loss, provided that other
in providing uniform air dispersion from the inlet to the stipulations of ASHRAE Standard 120 are satisfied. Since
extremities of the system. only one diameter of fabric air dispersion system was studied
in this test program, it is suggested that additional pressure loss
CONCLUSIONS tests be conducted over a wider range of duct diameters, with
Two trends in the data were readily apparent from the and without the skeletal metal frame being present, in order to
measured pressure loss data for non-porous fabric ducts. The establish whether the duct roughness is significantly affected
©2012 ASHRAE
489
by the duct diameter. If duct roughness is a function of duct C plane C
diameter, it is necessary to determine whether the roughness D plane D
increases or decreases with increasing duct diameter.
Z plane Z
NOMENCLATURE 1 plane 1
A = cross-sectional duct area, m2 (ft2) 2 plane 2
An = nozzle throat area, m2 (ft2) 5 plane 5
Cn = nozzle discharge coefficient, dimensionless 6 plane 6
D = diameter, mm (in.)
REFERENCES
f = friction factor, dimensionless
L = length, m (ft) ASHRAE. 2008. ANSI/ASHRAE Standard 120-2008,
Method of Testing to Determine Flow Resistance of
L' = separation distance, m (ft)
HVAC Ducts and Fittings. Atlanta: American Society of
Q = flow rate, L/s (ft3/min) Heating, Refrigerating and Air-Conditioning Engineers,
Re = Reynolds number, dimensionless Inc.
V = velocity, m/s (ft/min) ASHRAE. 2009. Duct Fitting Database, Version 5.00.08.
Yn = nozzle expansion coefficient, dimensionless Atlanta: American Society of Heating, Refrigerating
p = pressure loss, Pa (in. wg) and Air-Conditioning Engineers, Inc.
ASHRAE, 2009. I-P & SI Handbook-Fundamentals, Table 1,
 = absolute roughness, mm (ft)
p. 35.7. Atlanta: American Society of Heating, Refriger-
 = air density, kg/m3 (lbm/ft3) ating and Air-Conditioning Engineers, Inc.
 = dynamic viscosity, N·s/m2 (lbm/ft·s) Idem, S., A.N. Nalla, and K. Gebke. 2011. Laboratory Test-
ing of Fabric Air Dispersion System Friction Loss, Final
Subscripts
Report, DuctSox Corp.
A plane A Kline, S.J. and F.A. McClintock. 1953. Uncertainty in Single
B plane B Sample Experiments. Mechanical Engineering 75:3-8.

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---

SA-12-015 (RP-1448)
Ventilation Requirements for Refrigerating

Machinery Rooms
Scot K. Waye, PhD, PE Ronald L. Petersen, PhD Anke Beyer-Lout

ABSTRACT Q = 100 G
ANSI/ASHRAE Standard 15-2010 Section 8.11.5 supplies
a formula to determine the ventilation rate requirement for  Q = 70 G 
refrigerating machinery rooms. The current formula suggests where Q = the airflow in cubic feet per minute (liters per
that the appropriate ventilation rate to flush a refrigerant second) and G = the mass of refrigerant in pounds (kilograms)
during a leak is proportional to the square root of the mass of in the largest system, any part of which is located in the
the refrigerant in the room. A search for the historical roots of machinery room.”
the basis of this approach indicates that this formula was likely The Standard 15 User’s Manual (Fenton and Richards
developed before 1930. A more rigorous and scientifically 2002) states that the table of ventilation requirements, on
sound approach yields a technically defensible rationale for which the formula is thought to be based, was likely developed
ventilation requirements in refrigeration machinery rooms to before 1930. In fact, Brown (2005) shows that this table
maintain safety. This paper examines recorded accident appeared first in the Explosive and Hazardous Trades of the
scenarios to develop an appropriate leakage scenario and a New York Municipal Code, 1927 (Section 220 of Article 18)
ventilation requirement that takes into account several impor- and also gives some other previous history in developing
tant parameters: the mass of the refrigerant, refrigerant prop- ventilation requirements. This table did not appear in the
erties, the volume of the room, the delay in detection, and the ASRE handbook until 1939 when it was adopted by the then
recommended concentration limit suggested for health and B9 committee with modifications. The modifications
safety purposes. After determining an appropriate, time-vary- consisted of higher ventilation rates for systems with greater
ing emission rate, a mass balance equation is used to determine than 1000 lbs of refrigerant (Brown 2005).
the concentration of the refrigerant in the room. The refriger- The Standard 15 equation has several fundamental flaws:
ant concentration in the room is kept below the recommended 1) the equation does not account for variations in the maxi-
concentration limit by adjusting the ventilation rate. mum acceptable refrigerant concentration (or recommended
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
concentration limit – RCL); 2) the equation does not account
INTRODUCTION
for differences in refrigerant properties such as boiling point,
Section 8.11.5 of ASHRAE Standard 15 (ASHRAE vapor pressure, and molecular weight; and 3) it does not
2010) states: account for different room sizes.
“The mechanical ventilation required to exhaust an accu- In an effort to improve the Standard 15 equation, this
mulation of refrigerant due to leaks of a rupture of the system research program was carried out. First, the history of acci-
will be capable or removing air from the machinery room in dents related to machinery room refrigerant leaks was
not less than the following quantity: reviewed to help determine conservative, but reasonable
Scot K. Waye is a senior engineer, Ronald L. Petersen is a principal and vice president, and Anke Beyer-Lout is a project scientist at CPP,
Inc. in Fort Collins, CO.
©2012 ASHRAE
491
scenarios for future accidents. Next, a differential equation Table 1. Reported Fixed Source Releases (NRC)
was developed to estimate the ventilation rate needed to limit from Jan 1, 2004 through Dec 31, 2009
the concentration of the refrigerant to the maximum limits for
the proposed accident scenario. Finally, a simplified equation Refrigerant
Number Incidents Reportable
was derived for easy application and implementation into the Reported Quantity (lb)
standard.
R-11 1 5000
The objective was to establish a technical rationale for
specifying ventilation requirements in refrigeration machin- R-12 2 5000
ery rooms to maintain safety and to develop equations or other R-22 13 *(1)
methods to calculate required ventilation rates. To meet this
objective, ventilation strategies and rates required to minimize R-123 0 *(1)
health risks resulting from a leak were based on a realistic acci- R-134a 2 not listed
dent scenario, reflect the fundamentals of dilution, account for
the recommended concentration limits and physical properties R-245f 0 not listed
of each refrigerant, and are simple enough to use in standards R-407a 0 not listed
and building codes for determining refrigerating machinery
room ventilation requirements. R-410a 0 not listed
R-417a 0 not listed

BACKGROUND
R-422a 0 not listed
An overview of several databases reveals release scenar-
ios, including quantities, duration, location in the system, and R-290 (propane) 0 10,000
the cause of the release. Several have suggested their opinions
R-601 0 not listed
on realistic release rates to be used to calculate the ventilation
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
requirement. R-601a 0 not listed
Reported Accidents R-717 (Ammonia) 6010 100
Searching for historical refrigerant leaks yielded direc- R-744 146 not listed
tion to define realistic and appropriate accident scenarios upon *(1) chemical subject to reporting under section 313 and section 6607 of the Pollution
Prevention Act.
which ventilation requirements can be based. All accidents in
North America and Europe on record were considered with
accidents related to machinery room refrigerant leaks, includ- Therefore, almost all accidents are Ammonia. Some of
ing the magnitude of the leak, the cause, and any related the NRC reported accidental releases are also included
damage to people or property included. Four main databases in the ARIP database.
were searched:
• ARIP (The Accidental Release Information Program)
• MARS (Major Accident Reporting System) (MAHB (US EPA 2010): ARIP is a database compiled by the
2001): database on “major accidents” reported under EPA that focuses on accidental releases at fixed facilities
Seveso, OECD and UN-ECE Managed by the Major that resulted in off-site consequences or environmental
Accident Hazards Bureau (MAHB). The current MARS damage. All non-routine releases of oil and chemicals
database contains over 700 accidents and near misses are required to be reported (to the NRC, Coast Guard or
collected since 1982 from the Member States of the EPA regional offices). The US Environmental Protec-
European Union. Seventeen accidents were considered tion Agency (EPA) compiles these reports into the
from examined between May 1991 and December 2001. Emergency Response Notification System (ERNS) data-
• NRC (The National Response Center) (NRC 2010): base. Significant accidents are selected from the ERNS
NRC has made available yearly data files containing database under the Accidental Release Information Pro-
data related to incidents (oil, chemical, radiological, bio- gram (ARIP) and a questionnaire is sent out to the facil-
logical, and etiological releases anywhere in the United ity involved, consisting of 23 questions about the
States and its territories). Table 1 shows the number of facility, the circumstances and causes of the incident,
reports made to the NRC (2010) concerning the refriger- and the accidental release prevention practices and tech-
ants suggested above as likely being in machinery nologies in place prior to, and added or changed as a
rooms. The reportable quantity (RQ) (US EPA 2001) is result of, the event. The ARIP database was searched for
the threshold quantity where the accumulated loss of a accidents involving indoor refrigerant releases for the
refrigerant within a 24 hour period must be reported to timeframe of 1990 to 1999. Since the EPA draws infor-
the NRC by law. Ammonia is the only refrigerant with a mation from the NRC database, most of the accidents
low enough reportable quantity to be meaningful. found in the ARIP database also appear in the NRC
492
ASHRAE Transactions
database. However, the ARIP database provides more listed for most accidents reported. Unfortunately, mostly
detailed information on the cause of the release as well Ammonia accidents are recorded (see Table 2.1 for explana-
as the release quantity and release duration. Therefore, tion). The most frequent release quantifies vary between 100
the information extracted from the ARIP database was and 10,000 lbs. The released quantities were binned and
supplemented with the NRC data, if possible, to provide counted. The emission rate probability distribution is shown in
as much information as possible for each incident. In Figure 1.
case the NRC database conflicts with the ARIP data- From the databases, accidents seem more frequent before
base, it is assumed that the ARIP database is more accu- 1993, as shown in Figure 2. In 1992, OSHA issued the Process
rate. 223 accidents were recorded that met the criteria. Safety Management (PSM) of Highly Hazardous Chemicals
• OSHA (The Occupational Safety and Health Adminis- standard (OSHA 1992), which contains requirements for the
tration) (OSHA 2010): The OSHA database focuses on management of hazards associated with processes using
accidents affecting workers and included accident inves- highly hazardous chemicals. For refrigeration systems the
tigation summaries. The Accident Investigation Sum- standard applies to systems that contain 10,000 pounds of
mary database details accidents affecting workers and ammonia. Any facility having the threshold quantity of ammo-
has limited information on the cause of a chemical nia must have a PSM program in place. All new facilities are
release, release duration or release quantity. A search required to develop a plan and have it in place prior to intro-
was performed and accidents involving indoor refriger- ducing ammonia above the threshold quantity.
ant releases between January 1990 and February 2010
were extracted, yielding 94 applicable accidents. Eighty Table 2 lists the total emission rate for specific probabil-
of the reported releases were for Ammonia. Limited ities of exceedance. These emission rates were obtained by
information is provided on the quantities released. linear interpolation. One percent of all accidents evaluated had
a total emission rate greater than 2.99 kg/s (395.9 lb/min). Five
• NIOSH (National Institute for Occupational Safety and
percent of the accidents had a total emission rate greater than
Health) (NIOSH 2010): The center for disease control
1.13 kg/s (149.7 lb/min).
(CDC) and NIOSH report deaths from some accidents
(mostly workplace) through the Fatality Assessment and So far, no differentiation has been made depending on the
Control Evaluation (FACE) program. Only one fatality release type (gas, liquid or both). Figure 3 shows the fraction
could be found relating to a refrigeration leak. In 1992 of the different release types from the ARIP/NRC database.
in Alaska, an assistant ice rink manager died of asphyxi- For 75% of the reported accidents, gaseous ammonia was
ation while trying to stop a CFC-22 leak inside of a released and only 13% of the accidents involved a pure liquid
compressor room. release. However, the boiling point of ammonia is 240K (-
28°F). A liquid pool of ammonia would generate vapor due to
All of the databases listed were analyzed to determine heat transfer to the liquid from the floor in most machinery
accident scenarios that are conservative, yet reasonably likely room settings. Therefore, even a purely liquid release would
to occur in standard practice. The ARIP/NRC databases are have a gaseous component. On the other hand, a liquid release
the most comprehensive. The release quantity, release dura- could be mistaken for a gaseous release due to the fact that the
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
tion, as well as the release parameters (gas, liquid, etc.) are sudden pressure drop at the release point causes a portion of
Figure 1 Emission rate probability distribution from ARIP/NRC database (Ammonia).
©2012 ASHRAE
493
Figure 2 Emission rates from accidents by date from ARIP/NRC database (Ammonia).
Table 2. Emission Rate Probability Distribution from ARIP/NRC Database (Ammonia)
Probability of Exceedance 1% 2% 5% 10%
Total Emission Rate (metric) 2.99 kg/s 2.29 kg/s 1.13 kg/s 0.61 kg/s
Total Emission Rate (english) 395.9 lb/min 302.4 lb/min 149.7 lb/min 80.4 lb/min
The ARIP database also details the release location. The

MARS and OSHA databases were also evaluated. Both the
OSHA and the MARS databases have a large percentage of
unknown locations. Even though no emission rate information
is given, release locations are listed Figure 5 shows the fraction
of occurrence of different release locations. The most common
leak location is at a valve. The second most probable location
is the piping.
Although the total mass of the refrigerant is not as impor-
tant as the emission rate, it plays a role in the analysis because
it provides a bounding limit of the amount of refrigerant that
can be released. This leads to an emission time, where
increased ventilation may slow the increase in refrigerant
concentration in the room, but until the leak stops or all the
mass is released, the ventilation will not clear the room.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
The upper limit is generally defined by the catastrophic
Figure 3 Release types from ARIP/NRC database. failure of the largest vessel in the machinery room. The quan-
tity of refrigerant stored in this vessel could all be released if
a large rupture occurred at the bottom. The released refrigerant
the liquid jet to flash into vapor and the remaining liquid may would consist of the initial flash gas followed by the remaining
evaporate almost immediately. liquid. The liquid pool the machinery room floor would gener-
Figure 4 shows the percentages for different causes of the ate vapor due to heat transfer to the liquid from the floor. After
releases reported. Equipment failure is the major cause for a time, as the floor cools and approaches the equilibrium
accidents. Calculating the average emission rate for all release temperature at atmospheric pressure for the refrigerant, the
causes resulted in 0.27 kg/s (35.7 lb/min) for operator error, vapor generation rate would decrease dependent on heat trans-
0.27 kg/s (35.7 lb/min) for equipment failure and 0.16 kg/s fer from the air above the liquid surface. Another failure that
(21.2 lb/min) for other. Therefore, no distinction can be made could cause a large release would be a large rupture in a liquid
between the sizes of the release due to equipment failure or due supply line. If the flow is not stopped, a very large quantity of
to operator error. liquid could be expelled by the line. Several incidents have
494
ASHRAE Transactions
occurred in this manner resulting in the release of large quan- of a ½” high pressure, high temperature liquid line is the most
tities of ammonia. probable design worst-case scenario. Depending on the refrig-
In commercial buildings with refrigerating machinery eration system’s piping and valves, the leak rate can be
rooms containing water chillers, the smaller release incidents computed by applying the fluid flow principles in conjunction
are similar to those that occur with ammonia in industrial with the thermodynamic properties of the refrigerant.
systems. These leaks are caused by valve packing failure, Seidl and Taylor (2005) examine a leak from a 0.25 in.
mechanical seal leakage, pipe and tube fitting failure, corro- (6.35 mm) hole. Using R-22 and temperatures and pressures of
sion failure, or incorrect operation or maintenance procedures. 40ºF (4.4ºC) and 83 psia (572 kPa) and 100ºF (37.8ºC) and 210
Refrigerant storage vessels are not generally used in conjunc- psia (1448 kPa), the leak rate was calculated to be 3.5 lb/min
tion with water chillers and so the quantity of refrigerant in (0.026 kg/s) and 8.5 lb/min (0.064 kg/s), respectively. For a
these systems (R-11, R-22, R-123, R-134a, etc.) is consider- half in hole at the higher temperature and pressure, the leak
ably less. Consequently, the largest release of refrigerant from rate was calculated to be 34 lb/min (0.257 kg/s). They recom-
water chillers is significantly smaller than the potential large mend a leak rate of 15 lb/min (0.11 kg/s), arguing that similar
release from an industrial facility. non-toxic refrigerants will have similar leak rates and that a
0.25 in. (6.35 mm) hole is not an unreasonable estimate of a
Refrigerant Release Types puncture or accidental drilling.
General. Brown (2005), Stoecker (1998), Richards Refrigerant releases are generally jet flows that consist of
(1986), and others suggest that a failure consisting of a rupture only vapor, only liquid, or a mixture of vapor and liquid (i.e.,
two-phases). Ventilation systems are designed to dilute refrig-
erant vapor concentration to a selected value, so the antici-
pated refrigerant vapor leak rate selected by the designer
determines the ventilation rate. The four variables that influ-
ence the vapor leak rate are: pressure, temperature, state of the
refrigerant (vapor or liquid), and physical size of the hole or
orifice. The variables that usually cause higher leak rates are
high pressure, high temperature, large orifices, and/or a liquid
release. Low pressure, low temperature, small orifices, and/or
a vapor release usually are characteristic of lower leak rates.
This is not completely the case in all situations, but it does
correctly show general trends. For example, a leak of high
pressure, high temperature liquid through a large orifice would
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
be large. In contrast, a leak of low pressure, low temperature
vapor through a small orifice would be relatively small.
Liquids under high pressure provide a greater leak rate due to
high density and the greater velocity of flow through the
orifice. High temperature liquids also augment the leak rate
Figure 4 Cause of release from ARIP/NRC database. due to the higher fraction of flash vapor generated. From
Figure 5 Release location from (a) ARIP/NRC, (b) MARS, and (c) OSHA databases. A: Valves; B: Piping, C: Compressor/
Process Vessel, D: Pump, E: Other, and F: unknown.
©2012 ASHRAE
495
another perspective, refrigerant liquid leaks are greater and releases are not considered except to predict the liquid flow
thus more difficult to handle than vapor leaks through the same approaching the hole or rupture.
orifice, same pressure, and same temperature. Therefore, high The flow of liquid from a vessel, such as a reservoir tank,
pressure, high temperature, liquid leaks produce the largest through a hole depends on the pressure difference between the
vapor quantity (volume) and thus pose the greatest safety risk inside and outside of the vessel and on the gravity head devel-
in machinery rooms. oped by the liquid above the hole (AIChE 1996):
Liquid leaks are possible with the following combination p – po 12
of temperatures and pressures: high pressure and high temper- E l = c o A h  l 2  -------------- + 2gH l (1)
 l 
ature, high pressure and low temperature, and low pressure
and low temperature. With high pressure and high tempera- The shape of the vessel also influences the flow rate of a
ture, the flash vapor comprises approximately 20% - 40% of liquid through a hole. For a vertically oriented vessel, the
the total leaked mass flow rate depending on the actual pres- liquid height may be computed using (AIChE 1996).
sure and temperature. Flashing of the vapor cools the remain-
ing liquid to its thermodynamic equilibrium temperature at 4V
atmospheric pressure. The refrigerant supplying the leak also H l = ---------l (2)
d 2
drops in temperature according to its pressure. When the
upstream pressure drops to atmospheric pressure, the leak rate When the vessel is horizontal, then (AIChE 1996),
through the orifice is nearly zero. However, the refrigerant
d
may still exist in the source vessel or pipe depending on the H l = ---  1 – cos  l  (3)
2
location of the leak. The liquid that has leaked out will absorb
heat from surfaces that it comes in contact with and conse-
Ld 2 sin  2 l 
quently vaporize while cooling that surface. If sufficient liquid V l = ---------   – -------------------- (4)
4  2 
is leaked, it may pool on the floor of the machinery room and
after producing large vapor quantities cooling the floor, Reservoir tanks are commonly located outside refrigerant
produce only a small quantity of vapor. At that point, the heat machinery rooms. Therefore, the liquid jet release out of a pipe
vaporizing the refrigerant liquid only comes from small is a more appropriate scenario. The liquid flow out of a pipe
amount of thermal energy still left in the floor and the depends on the difference between the liquid pressure and the
surrounding air. ambient pressure, as well as the head loss due to fittings (such
If the liquid is at a low temperature, then the flash vapor as elbows, tees and valves) and wall friction (ASHRAE 2009).
quantity is considerably less (on the order of 10% by mass) and Thus,
so the vapor produced is less. But if the pressure is high, the 12
leaked liquid is rapidly replenished and the leaving liquid  
 p – po 
produces vapor cooling the surfaces it comes in contact with E l = A h  l 2  --------------------------------- (5)
  ff L 
  l  ------ + K
D  
in the machinery room. Low pressure liquid will also produce
flash vapor at a similarly low fraction but will not be vigor-
ously replenished as with a high pressure liquid. The friction loss depends on the Reynolds number Re of
The quantity of vapor that will leak from an orifice the flow and on the roughness height,  , of the pipe wall
depends only on the orifice size and the upstream pressure surface (ASHRAE 2009):
assuming the flow is not choked. The upstream temperature
8 12 1  12
f f = 8  ------ + -----------------------
only influences the leak rate by how it affects the density of the 1 -
(6)
upstream vapor. However, if the upstream pressure is greater Re  A + B  1.5
than the critical pressure (and depending on the hole size), then
the flow is choked and is not dependent on the upstream pres-   16
 1 
sure. Under chocked flow conditions, upstream temperature A h = 2.457  ln  -------------------------------------------
- (7)
  -----
7 0.9
 0.27- 
and pressure changes will influence the flow rate in the same   Re- +  ------------ D 
manner that it influences the upstream vapor density.
Liquid Jet Releases. When the hole causing the release 37530 16
B =  --------------- (8)
is at or below the liquid vapor interface, then liquid will exit.  Re 
The pressure inside the vessel or pipe resulting from the ther-
modynamic equilibrium condition pushes the liquid through 4E l
Re = ---------------
- (9)
the hole. With refrigerants, the sudden pressure drop causes a  l Dv
portion of the liquid jet to flash into a vapor. Consequently, this
is a two-phase flow. Because flashing of liquid to vapor will Because the friction loss f depends on the mass emission
generally occur to some extent with refrigerants, purely liquid rate, Ql, the mass emission rate has to be calculated iteratively.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
496
ASHRAE Transactions
For high Reynolds numbers, i.e. fully turbulent flow, the fric- must be appropriate allowing for an all liquid flow equation
tion factor becomes independent of the Reynolds number. similar to the Bernoulli equation.
Two-Phase Jet Releases. The release of liquid refriger- Evaporating Pool. For a two-phase release, liquid can
ant involving a significant pressure drop will result in a portion form a pool on the floor while at the same time a super cooled
of the liquid flashing to vapor. If it is assumed that the expan- vapor is released. The emission rate for the two mechanisms
sion of the liquid is adiabatic, that air does not mix with the will be additive. One equation that is considered for the liquid
expanding fluid, and that the expansion process is reversible, spill is provided below (US EPA 1992):
then the amount of flashing liquid can be calculated by an isen-
tropic energy balance (API 1996): 6.94  10 – 7  1 + 0.0043  T 2 – 273.15  2 
Q = (12)
Ef S l1 – S l3  U 0.75  A s  M w  V p  V Ph
----- = -------------------
- (10)
El S vap
exp(76.858 – 7245.2  T 2 – 8.22ln  T 2 
V ph = (13)
If the kinetic energy of the fluid expansion is neglected, + 0.0061557 T 2
the isenthalpic balance can be used to calculate the amount of
flashing liquid. The heat of vaporization is obtained entirely The vapor pressure is calculated using the Clausius Clap-
from the enthalpy of the liquid being released. It can be shown eyron equation:
that the fraction of liquid flashed is (API 1996):
 H vap M w 1 1 
cp  T – Tb  V p = 101325  exp  --------------------  ----- – -----
Ef (14)
----- = ------------------------
- (11)  R T b T 2 
El H vap
The release of liquid refrigerant from high pressures The surface area and surface velocity of the liquid pool
involves a substantial change in kinetic energy, therefore the can all be adjusted to obtain specific evaporation rates for
isenthalpic balance approach might be inappropriate for the alternative emission scenarios and room sizes.
calculation of the amount of flashing liquid. Neglecting the The evaporation of a liquid pool is a time varying process.
kinetic energy of the fluid expansion causes an over prediction Initially, the pool is very small; therefore the evaporation rate
of flashing liquid. According to the Manual for Modeling is insignificant. However, as time progresses, the liquid pool
Hypothetical Accidental Releases to the Atmosphere (API, becomes larger, increasing the evaporation rate. Eventually,
1996) Equation 11 should be used in cases where a saturated the pool is large enough that equilibrium is reached and the
liquid is released. For refrigerant blends, the flashing process evaporation rate approaches the liquid release rate (total mass
is much more complex and calculations must include infor- emission rate minus emission rate of flashing liquid). The
mation concerning the vapor-liquid equilibrium conditions of limiting size of the pool is the room floor area or area of any
the blend. dike that is used.
Releases of two-phase flow from vessels and pipes Vapor Jet Releases. The release of vapor jets can be
involving choked flow remain an area of research. While approximated well assuming an isentropic expansion (revers-
empirical relationships have been proposed, they only apply to ible and adiabatic). The estimates provided yields reasonable
simple situations: refrigerants composed of single compo- temperatures of the initial vapor cloud and velocity of the
nents. One approach that has been used for single component vapor entering the cloud. For vapor releases that result in a
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
two-phase flows is the homogeneous equilibrium flow model large change in velocity, the isentropic approximation has
(HEM) which is based on the assumptions: been shown to be preferred over the isenthalpic approximation
(Moran and Shapiro, 2000).
• Homogeneous liquid-vapor mixture
A vapor jet forms upon the development of a small hole
• Thermal equilibrium between liquid and vapor phases
generating a path for the refrigerant to flow through and enter
• No slip between liquid and vapor phases
the surrounding air in the machinery room. When the pressure
• Isentropic expansion process
in the vessel or pipe containing the refrigerant vapor exceeds
The HEM has been applied in determining the choked the critical pressure, the flow velocity is sonic (equals the
flow rate of flashing liquid releases, but refrigerant releases speed of sound) through the hole and the flow is defined as
have not been found in the literature. Several researchers have “choked.” Regardless of how much greater the refrigerant
used this approach to predict laboratory results, but only pressure is above the critical pressure the flow rate is constant.
(Sallet 1990) has predicted flow rates of several refrigerants Only the density inside the vessel or pipe (through temperature
including: R-290 (propane), ammonia, carbon dioxide, and R- and pressure) influences the flow rate. Assuming the refriger-
22. ant vapor behaves as an ideal gas, the critical pressure ratio is
The monograph (AIChE 1996) reports another approach k
for predicting two-phase choked flows where analytical + 1- ----------
p-  k----------- k–1
-
----  (15)
expressions are employed. The thermodynamic conditions po  2 
©2012 ASHRAE
497
Values for k generally range from 1.2 to about 1.5 and the The exponential terms are actually time constants that depend
pressure ratio defined by Equation 16 is approximately 2 for on the physical size of the pipe and the size of the rupture.
most gases and vapors.
The choked flow condition for an ideal gas or vapor based RELEASE SCENARIOS
on isentropic flow is given by The purpose of this section is to present the method used
to develop a realistic vapor release rate was for a machinery
k + 1 1/2
2 ------------ room based on the type of refrigerant and upon historical
E = c o A h k  ------------ k – 1 (16)
k+1 information on accidental releases.
There are five factors that affect the vapor release rates as
Values for the discharge coefficient vary from 0.6 to previously discussed: size of the release orifice, state of refrig-
almost one and are due to non-ideal flow affects. Observe that erant (gas or liquid); properties of the refrigerant; pressure and
if the pressure is above the critical pressure in a vessel, the temperature. Different combinations of these factors will
temperature will decrease somewhat causing the density to produce different vapor release rates. Below are the combina-
decrease in turn decreasing the flow rate. tions ordered from high to low vapor release potential as indi-
When pressures are less than the critical pressure ratio, cated by Brown (2005):
the condition of choked flow no longer holds and the mass
flow rate is dependent on the pressure inside the vessel of pipe. • Release Scenario 1: Liquid, high pressure, high temper-
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
The flow rate is given by ature

• Release Scenario 2: Gas, high pressure, high tempera-
+ 1- 1 / 2
k----------- ture
k  po 2/k
 p----o- k 
E = c o A h 2p  -----------   ----- –  (17)
 k – 1   p   p 
• Release Scenario 3: Liquid, high pressure, low tempera-
ture
Note that the mass flow rate is less than that for choked • Release Scenario 4: Liquid, low pressure, low tempera-
flow. ture
For a given release, the rate of change of pressure inside • Release Scenario 5: Gas, low pressure, low temperature
the vessel or pipe depends on the specific situation governing
the release. Either an adiabatic or an isothermal flow process Of most interest is the worst-case realistic scenario and
may be appropriate. An example of an isothermal flow would provide background information for use in selecting a realistic
result when a small leak occurs in a pipeline because the cool- release rate to be used for ventilation room design. Figure 6
ing that occurs by the expansion is counteracted by the friction shows a schematic of a typical refrigeration cycle identifying
heating and heat transfer through the pipe from the outside. where these release scenarios could occur in the process.
Thus, the temperature in the pipe would remain constant;
however, with relatively large vapor release flow rates, the Release Scenario 1: Liquid, High Pressure,
expansion process cools the internal volume of the pipe or High Temperature
vessel and the adiabatic process is the appropriate choice. A liquid, high pressure, high temperature release is a two-
Moreover, if the pipe or vessel is insulated, then the adiabatic phase flow. The pressure inside the vessel or pipe resulting
process is again the appropriate choice. In those situations from the thermodynamic equilibrium condition pushes the
when the isothermal process holds, the inside temperature is liquid through the hole. The sudden pressure drop at the
constant. With the adiabatic flow process, the pressure release point causes a portion of the liquid jet to flash into a
decreases as the refrigerant is released. The pressure may be vapor. The remainder of the liquid pools on the machinery
estimated by using conveniently small time increments and room floor at the boiling point temperature (Brown, 2005)
calculating the mass released over the each time step. For each after which it slowly warms to ambient temperature. The vapor
subsequent time step, subtract the mass that escaped from the emission rate is therefore calculated in 3 steps: 1) the total
mass inside the vessel or pipe recalculating the pressure. In liquid mass emission rate out of the hole is determined; 2) the
this manner, even the case of isothermal flow can be accom- fraction of flashing liquid is computed; and 3) the evaporation
modated as well as heat transfer to or from the vessel or pipe rate from the liquid pool surface is calculated.
applying reasonable values for the convection heat transfer Several assumptions have to be made concerning the
coefficient. release size, release area, release temperature and air velocity
Further elaboration on calculation methods for vapor over refrigerant pool. The liquid refrigerant pool size was
releases are presented in the monograph (AIChE 1996) where taken to be the equilibrium pool size (evaporation rate was
procedures first developed by Wilson (1981) for releases from limited to be not more than liquid release rate minus emission
pipe ruptures are discussed. The concept developed involves rate of flashing liquid), with a liquid depth of 0.394 in. (10
defining the initial vapor mass in the pipe and treating the mm) (US EPA 1992). During the release out of the pipe, the
release as an isothermal vapor release characterized by an liquid will expand and cool to the boiling point. For refriger-
expression consisting of the sum of two exponential terms. ants with boiling points below ambient conditions, the liquid
498
ASHRAE Transactions
A simplified short pipe release without wall friction,
valve loss coefficients and gravity head minimizes required
assumptions about the refrigeration system. This release
scenario is solely dependent on the diameter of the hole and the
difference in pressure inside and outside of the pipe. There-
fore, this scenario is simple but still realistic.
Equation 2 simplifies to:
p – po 1/2
E l = c o A h  l 2  -------------- (18)
 l 
The EPA recommends a discharge coefficient co = 0.6,

however, this coefficient depends on the flow conditions out of
the orifice, with co = 1 representing flow that is not hindered
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
by the shape of the orifice (such as a severed pipe).
During the course of the release the pressure and temper-
ature inside the storage vessel are kept constant (US EPA
1992). In reality, the loss of liquid refrigerant out of the valve
causes evaporation inside the vessel and therefore a reduction
of the temperature and pressure of the system. The resulting
total liquid emission rate would therefore decrease slightly
faster over time, than the one calculated using the assumption
of constant pressure and temperature, which is therefore the
simpler and more conservative approach.
As described above, the flashing emission rate of the R-
134a release is calculated using the isentropic method. The
remaining liquid R-134a forms a pool. The pool size is time
dependent and calculated using an air velocity of 50 ft/min
Figure 6 Refrigeration cycle with locations of high and low (0.254 m/s).
pressure gas and liquid. The rate of flashing liquid is proportional to the total
liquid release rate and decreases slowly over time. Shortly
after the release, the pool size is small and therefore the pool
pool will gradually become warmer but will most likely never evaporation rate is insignificant. However, over time the pool
reach ambient temperature before it is entirely evaporated. grows until equilibrium is reached, where the amount of liquid
The liquid pool will also cool the surrounding air. Therefore, refrigerant supplying the pool is equal to the evaporation rate,
the pool temperature, T2, in Equation 13 was taken to be the resulting in a total vapor emission rate that is equal to the total
greater of 32ºF (0ºC) or the boiling point temperature of the liquid emission rate.
specific refrigerant. This is a conservative assumption for The above described calculations were repeated for other
those chemicals that have boiling points lower than 32ºF (0ºC) hole diameters. Figures 7 and 8 show the dependence of the
in that the liquid will remain cooler than 32ºF (0ºC) for some various calculated emission rates on the diameter of the
period of time. For example calculation purposes, the air release orifice for R-134a and Ammonia, respectively. The
velocity over the spill is taken to be 50 ft/min (0.254 m/s). This total liquid mass emission rate, the vapor emission rate due to
would correspond to a ventilation rate of 20,000 cfm, which is flashing, the equilibrium liquid pool evaporation rate and the
recommended by Brown (2005) for all machinery rooms, and total vapor emission rate are plotted with the hole diameter on
a room cross-sectional area of 400 ft2 (37.2 m2). A lower venti- a log-log scale for two examples: Ammonia (R-717) and
lation rate and a smaller room size would give a similar air 1,1,1,2-Tetrafluoroethane (R-134a). The vapor released from
velocity which is the only important input for the evaporation flashing and evaporation of a 0.25 in. (6.35 mm) hole repre-
calculation. sents approximately 95% (5% would exceed this) of the acci-
To determine the total vapor emission rate, the mass flow dents found in the databases. A 0.5 in. (12.7 mm) hole exceeds
of the flashing liquid (using the isentropic balance) is added to 99% of the accidents.
the evaporation rate.
A potential failure location of a valve is directly behind a Release Scenario 2: Gas, High Pressure,
storage vessel of liquid high pressure and high temperature High Temperature
refrigerant, for example a defective liquid line valve behind This was initially rated as the second worst vapor release
the receiver tank. potential. A high pressure, high temperature gas release is
©2012 ASHRAE
499
Figure 7 Emission rate for Ammonia as a function of hole diameter for release scenario 1: high temperature, high pressure.
Figure 8 Emission rate for R-134a as a function of hole diameter for release scenario 1: high temperature, high pressure.
possible in the piping between the compressor and the All equations used in Scenario 1 can be applied to this scenario
condenser (see Figure 6). If the pressure in the pipe exceeds as well. Insignificant differences in the resulting emission
the critical pressure, the flow velocity equals the speed of rates are expected for this scenario compared to Scenario 1,
sound through the hole and the flow is “choked.” If that is the because the total liquid emission rate is solely driven by the
case, the vapor emission rate is constant regardless of how pressure difference between the inside and outside of the
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
much greater the pressure inside the pipe is above the critical vessel. The lower temperature of the refrigerant causes a
pressure. The choked condition is therefore the limiting case slightly lower flashing rate and a lower evaporation rate.
for this scenario. This scenario has significantly lower emis- However, once the system reaches equilibrium (i.e. the evap-
sion rates than observed (see Table 3.2) and will not be consid- oration rate equals the total liquid emission rate minus the
ered as a reasonable case from which to base the room flashing rate) the temperature difference between the two
ventilation rate. scenarios will have no effect and this scenario will provide the
same emission rates as Scenario 1.
Release Scenario 3: Liquid, High Pressure,
Low Temperature Release Scenario 4: Liquid, Low Pressure,
Low Temperature
This scenario is very similar to Scenario 1 described. A
high pressure, low temperature liquid refrigerant release is A low pressure, low temperature liquid release of refriger-
possible between the heat exchanger and the expansion valve. ant is possible between the expansion valve and the evaporator.
500
ASHRAE Transactions
If the pressure inside the pipe or vessel drops below the ambient actual leak rate, ventilation may, or may not, dilute the refrig-
pressure, the release is governed by the weight of liquid above erant concentration to the RCL or lower using the current Stan-
the orifice and no release may occur until the pressure is equal- dard 15 method. This section will presents methods whereby
ized. Due to the lower temperature compared to Scenario 1, the a designer can specify a ventilation rate to ensure the concen-
rate of flashing liquid is significantly lower. The equilibrium tration will be below the RCL value. In subsequent para-
pool size is smaller than that for Scenario 1 because the overall graphs, a derivation of the equations needed to estimate the
liquid emission rate is lower. Since the calculated emission rates concentration in a machinery room due to an accidental
are significantly lower than scenario 1, this scenario has also be release of any refrigerant is presented.
ruled out for use in determining the ventilation rate.
Detector Activation of Ventilation System
Release Scenario 5: Gas, Low Pressure,
Standard 15 (ASHRAE 2010) outlines the requirement
Low Temperature
that every refrigerating machinery room contains a detector
This scenario is very similar to Scenario 2. A low pres- that activates an alarm and mechanical ventilation in case of a
sure, low temperature gaseous refrigerant release is possible refrigerant leak. The alarm is to be activated at a refrigerant
between the evaporator and the compressor. All equations concentration level no higher than the threshold limit value
used in Scenario 2 can be applied to this scenario as well. The (TLV) time weighted average (TWA). The threshold limit
resulting vapor emission rate is expected to be lower than that value-time-weighted-average (TLV-TWA) is the average
calculated in Scenario 2, because the total vapor emission rate concentration for a normal workday exposure (40 hours a
is solely driven by the pressure difference between the inside week, 8 hours a day) without any adverse effect. Table 3 lists
and outside of the vessel. A reduction in that pressure differ- the TLV-TWA for different refrigerants. In the following anal-
ence causes a reduction in the emission rate. If the pressure yses and examples the TLV-TWA shall be used as the thresh-
inside the pipe or vessel falls below the critical pressure, the old for the activation of the ventilation. For those chemicals
condition of choked flow no longer holds, so the vapor flow where a TLV-TWA is not available, 10% of the RCL is used as
rate is dependent on the pressure inside the vessel or pipe. The the detector threshold value.
total mass flow rate for un-choked flows is less than that for
choked flows. Since the calculated emission rates are signifi- Refrigerant Concentration in a Room
cantly less than Scenario 1, this scenario will also be ruled out
A mass balance equation is used to solve for the concen-
for use in determining the ventilation rate.
tration in a room as a function of the sources of the contami-
Recommended Release Scenario for Ventilation nant and the losses, mainly ventilation. Several assumptions
Room Design are made while deriving and solving this mass balance:
To account for a likely worst-case scenario, it is recom- 1. The room is well-mixed. When the contaminant refriger-
mended that the high temperature, high pressure release be ant is released into the enclosed room and is gas, the emit-
used. Although the two-phase release is the most realistic, ted mass is instantaneously diluted in the volume of the
accounting for all the variables for the evaporating pool, room. Without much more sophisticated tools to account
including surface velocity, pool depth, pool size (floor area), for densities, mixing rates, and other parameters to deter-
and heat transfer considerations becomes very difficult and mine stratification and the concentration gradient, this
full of ambiguous assumptions. In addition, the emission rate assumption will underestimate the concentration in part
for the evaporating pool can become very large. Hence, it is of the room and overestimate it in others. For example, if
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
assumed that a drain is provided to capture any liquid that the refrigerant is heavier than air and the ventilation rate
would accumulate. All vapor emissions will be created by the is initially small (low mixing), the concentration would
flashed vapor from a 0.5 in. (12.7 mm) hole. Only taking into be higher at the floor level and lower near the ceiling. For
account the flashing vapor (no evaporation from potential this reason, many emergency exhaust intakes are located
liquid), for Ammonia, this represents an accident scenario nearer to the floor. Once the ventilation rate is increased,
with only a 5% probability of exceedance which seem like a better mixing will most likely occur.
reasonable probability for design purposes (including evapo- 2. The contaminant is inert. There are no reactions, either
ration, this scenario would yield a 1% probability of exceed- homogenous or heterogeneous, in the volume of the
ance). It is further assumed that similar probabilities would room.
occur for other refrigerants. 3. There is no nucleation or condensation. Once the contam-
inant refrigerant is gas, no nucleation or condensation
VENTILATION RATE
occurs.
The purpose of ventilating machinery rooms is to dilute
the air in the room such that the recommended concentration For a volume of space, V, the change in concentration, C,
limit (RCL) in Tables 1 and 2 of ANSI/ASHRAE Standard 34 is equal to the emission rate, E, minus the ventilation, Q, of the
(ASHRAE 2010) is not exceeded. However, depending on the contaminant out of the room:
©2012 ASHRAE
501
Table 3. Refrigerant Quantities to Calculate Ventilation Rate
Refrigerant RCLa Concentration Setpointb m  Qmax
ppm (v/v) g/m3 lb/Mcf ppm (v/v) mg/m3 lb/Mcf sec cfm L/s
R-12 18,000 90 5.6 1,000 4,945 0.31 39 0.33 44,400 21,000
R-22 59,000 210 13 1,000 3,537 0.23 32 0.34 23,400 11,100
R-23 41,000 120 7.3 4,100 12,000 0.75 24 0.61 95,700 45,200
R-32 36,000 77 4.8 3,600 7,700 0.49 28 0.34 73,600 34,800
R-123 9,100 57 3.5 910 5,700 0.36 161 0.049 2,550 1,210
R-124 10,000 56 3.5 1,000 5,600 0.35 51 0.3 48,600 23,000
R-125 75,000 370 23 1,000 4,909 0.31 28 0.54 24,700 11,700
R-134a 50,000 210 13 1,000 4,173 0.27 40 0.35 19,800 9,310
R-143a 21,000 70 4.5 1,000 3,437 0.22 33 0.46 94,400 44,600
R-152a 12,000 32 2.0 1,200 3,200 0.2 50 0.29 85,200 40,200
R-170 7,000 9 0.54 1,000 1,230 0.077 39 0.64 873,000 412,000
R-245fa 34,000 190 12 3,400 19,000 1.2 98 0.14 3,580 1,690
R-290 5,300 10 0.56 1,000 1,804 0.12 53 0.4 372,000 176,000
R-404A 130,000 500 31 1,000 3,992 0.25 32 0.5 14,600 6,860
R-407C 76,000 270 18 7,600 27,000 1.7 31 0.4 22,200 10,500
R-410A 130,000 390 25 13,000 39,000 2.5 27 0.42 18,300 8,640
R-507A 130,000 520 32 1,000 4,043 0.26 31 0.5 14,500 6,820
R-600a 4,000 10 0.6 400 960 0.06 87 0.29 164,000 77,200
R-717c 320 0 0.014 25 17 0.0011 45 0.21 10,100,000 4,730,000
R-1270 1,000 2 0.1 500 861 0.054 48 0.4 2,310,000 1,090,000

a
From ANSI/ASHRAE Standard 34-2010 (ASHRAE 2010). If disputes arise, Standard 34 takes precedent.
b
TLV/TWA value or 1/10 RCL if TLV/TWA not available
c
R-717 see IIAR Standard 2 (IIAR 2008) for required ventilation rates
dC  t  E t
V -------------- = E  t  – QC  t 
dt
(19) C  t  = e – t  e t ---------
V
- dt (23)
Dividing all terms by the volume yields a first order Assuming a constant emission rate, E  t  = E 0 , and an
differential equation. initial condition of C(0) = C0, the analytical solution for C(t)
is:
dC  t - + C  t  = E  t -
------------- --------- (20)
dt V E0
C  t  = ------  1 – e –t  + C 0 e – t (24)
By multiplying all terms by et and using the reverse chain Q
rule, an intermediate solution for C(t) is found: For most other emission rates, a numerical approach is
dC  t  E t
needed to find a solution.
e t -------------- + e t C  t  = e t ---------- (21) The analytical solutions consist of two fundamental
dt V
mechanisms. The decay of the concentration due to ventilation
d- Et (the only loss considered in this approach) is given by the
----  C  t e t  = e t ---------- (22)
dt V C 0 e –t term. The second term(s), with the  1 – e – t  term or
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
502
ASHRAE Transactions
variants of it, represent the approach of the concentration lating the concentration as a function of time, these factors can
towards an asymptote that represents a steady-state concen- be considered.
tration. This term is multiplied by the source emission to deter- Although a numerical integration of the Equation 25
mine the magnitude. through various implicit or explicit methods would yield a
For each unique set of source and loss terms (emission pure numerical solution, it can be used for each time step,
rate and ventilation rate), the equation should be reconsidered resetting the value for the initial concentration, C0, and the
with a new C0 and E0 and any other parameter that may have emission rate, E0, correlated to each time step, n. The super-
changed. Each time the emission or ventilation rates change, script n denotes the time step.
the system attempts to reach a new equilibrium. For example,
En
C n = ------n  1 – e –  t – t   + C n – 1 e –  t – t 
n n n–1 n n n–1
for a short time at the beginning of the leak, the emergency (27)
exhaust ventilation has not yet started, so the concentration Q
rises in the room. When the ventilation rate increases, the The release duration is defined by the total mass of the
steady-state concentration value decreases. refrigerant. The emitted liquid, which either flashed into vapor
Assuming a constant emission rate, at some point a or spilled as liquid and then evaporated, is limited by the reser-
steady-state concentration is reached. This also constitutes the voir of available refrigerant. Therefore, the flashed vapor ends
worst case scenario. For a constant emission rate, the concen- when the liquid emission is done and the liquid pool stops
tration of the refrigerant in the room is independent of the growing and diminishes in size while the refrigerant evapo-
room volume and is only a function the emission rate and the rates. Figure 9 shows a sample plot of the emission rate, show-
RCL as shown below: ing the total liquid emission rate, the flashing emission rate,
and the evaporation emission rate. The figure demonstrates the
E E
C  t  = ------0- = -----0- (25) relationship between the various mechanisms contributing to
V Q refrigerant vapor in the room.
Therefore, in order to calculate the ventilation rate for the The evaporation from the pool creates difficult non-
worst case scenario (steady-state for a given leak), the formula linearities in the solution because it introduces additional vari-
will be: ables that must be resolved, such as the depth of the pool, air
velocity at the surface of the pool, and the size of the pool,
E0 which necessitates a floor area. Comparing a leak from a 0.25
Q = ----------
- (26) in. (6.35 mm) hole with pool evaporation and from a 0.5 in.
RCL
(12.7 mm) hole without evaporation yielded similar gas emis-
where RCL is the recommended concentration limit given by sion rates for most refrigerants. Therefore, the simplification
Standard 34 and E0 is determined by the accident scenario. to increase the hole size but ignore the evaporation from the
The steady state solution is conservative because it only liquid pool was made for the selected accident scenario.
examines the maximum emission rate while not being able to Further, it is recommended that a drain be present in the room
consider such factors as the room volume, time varying emis- to remove liquid from the floor.
sion rate and variable ventilation (i.e., low ventilation when For each time step, the liquid emission rate is broken up into
concentration is below the TLV and high thereafter). By calcu- the portion that flashes and the liquid pool, which evaporates,
Figure 9 Example emission rate for R-134a for a high pressure, high temperature release scenario.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
©2012 ASHRAE
503
but is ignored. This total vapor emission rate is used to calculate dimensional variable incorporates the refrigerant mass, room
the concentration in the room. The maximum concentration is volume, RCL, and accident scenario. If the safe room volume
found and a routine solves for the ventilation rate such that the fraction, f, is greater than unity, no additional ventilation will
maximum concentration does not exceed the RCL. Several of be needed if there is an accidental release.
the variables that can be changed are: the total refrigerant mass,
the hole diameter for the accident scenario, the temperature and Detector Delay Time
pressure of the refrigerant in the vessel, the volume of the room,
and the RCL and TLV-TWA values (which are dependent on the The time between the start of the refrigerant release and
refrigerant). initiation of the ventilation fan may have a big impact on the
The steady-state solution yields the easiest calculation for ventilation rate needed. Increased ventilation is needed to slow
the required ventilation rate, but also is the most conservative. down the increase in concentration in the room. When the
It may be a starting point for the calculation, but a time varying reservoir of available refrigerant is exhausted, the ventilation
analysis that accounts for the emission rate not reaching equi- will continue to flush the room. An example of this effect on
librium gives a much more accurate solution. The transient concentration is shown in Figure 10, which uses the same
analysis ends up calculating the required ventilation rate by parameters for R-134a as the accident scenario. The concen-
using the concentration in the room at the maximum emission tration rises until all the available refrigerant has leaked out,
rate. Depending on a number of variables, including the mass whereupon the ventilation purges the room of the refrigerant.
of the refrigerant, the room size, and how the liquid pool is The goal is keep the maximum concentration below the RCL
contained, the ventilation requirement may be higher than at all times. If left too long, the concentration will exceed the
previously presented in Standard 15 (ASHRAE 2010) by RCL. Ventilation will slow the increase in concentration and
using the 100 G equation (for lbs of refrigerant). For lower eventually purge the room, but the RCL will have already been
masses and larger rooms, the ventilation rate may be lower and exceeded.
the RCL may not even be reached.
The maximum delay time was found for which some
Safe Room Volume amount of ventilation could prevent the concentration from
exceeding the RCL. This ventilation rate is excessively high.
The safe room volume, as defined here, is the volume of Figure 11 shows that delaying ventilation increases the venti-
room that is needed in order for the concentration of the leaked lation requirement dramatically as it gets to the maximum
refrigerant to remain below the RCL if all the mass was leaked delay time. To limit the amount of ventilation required, a
and mixed instantaneously. It also takes into account the detector delay time has been provided such that it is 50% of the
flashed portion of the liquid release from the accident maximum detector delay time. All maximum ventilation rates
scenario. The safe room volume may be defined as: (Qmax) are calculated at this point.
G- For each refrigerant and machinery room, the allowable
V s = ---------- (28)
RCL detector delay time, in seconds, is calculated using m (Table 3)
The values for  are found in Table 3. and f:
The safe room volume fraction, f, is defined as the ratio of
the room volume to the safe room volume (V/VS). This non- t d  mf for f  1 (29)
Figure 10 Example time history of concentration of R-134a for high pressure, high temperature release scenario.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
504
ASHRAE Transactions
Figure 11 Sample required ventilation rate required as a Figure 12 Ventilation requirement for all refrigerants as a
function of detector delay time. Values and units function of safe room volume fraction.
omitted on abscissa due to generalization of trend.
Table 4. Comparison of Proposed Ventilation Requirement with Current Standard 15
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Q Q
Chiller Room Volume Refrigerant Mass  RCL f Qmax
[this work] [Standard 15]
ft3 lb lb/Mcf cfm cfm cfm
3300 134a 124 0.35 13 0.99 19,800 521 1114
7956 134a 400 0.35 13 0.74 19,800 10,140 2000
4730 123 750 0.049 3.5 0.45 2550 2222 2739
33,895 123 1050 0.049 3.5 2.31 2550 0 3240
5299 134a 355 0.35 13 0.55 19,800 15,181 1884
36,229 134a 2300 0.35 13 0.59 19,800 14,464 4796

8788 134a 760 0.35 13 0.43 19,800 17,603 2757
20,925 134a 760 0.35 13 1.02 19,800 0 2757

29,700 134a 625 0.35 13 1.77 19,800 0 2500
Recommended Ventilation Rate (ASHRAE 2010). Since the approach in this work contains
By calculating the required ventilation rate as a function many more variables than just the refrigerant mass, the trend
of the safe room volume fraction, f, the required ventilation of the ventilation requirement is not quite as predictable as the
rate fitted to the data (see Figure 12) becomes: square root of the mass. In general, the ventilation requirement
goes down when the mass of the refrigerant is lower, when the
Q = 0 for f  1 (30) volume of the room is larger, when the refrigerant is not as
Q = Q max  1 + 0.3f – 1.3f 2  for f  1 (31) volatile, and when the RCL is higher. To compare the two
methods, a comparison has been calculated using chiller
The values for Qmax are provided in Table 3. rooms that have been designed, as listed in Table 4. In some
cases, the ventilation requirement goes up. In others, it
DISCUSSION decreases. For a few rooms, the safe volume fraction is above
It is useful to compare this method of calculating the unity, so no additional ventilation is needed to keep a leak of
ventilation requirement with that of the Standard 15 the entire mass from exceeding the RCL in the room.
©2012 ASHRAE
505
CONCLUSIONS U = air velocity
The current recommendation for ventilation in refrigera- V = volume
tion machinery rooms found in Standard 15 only accounts for Vp = vapor pressure (at T2)
the mass of the refrigerant in the room and does not have a Vph = vapor pressure of hydrazine (at T2)
strong scientific rationale for its existence. This work presents VS = safe room volume, M/RCL
a ventilation requirement for refrigerant machinery rooms in
the case of a release. The approach is simple enough to be Greek
codified or put in a standard and includes the major variables
 = roughness height
that would affect the ventilation requirement: mass of the
refrigerant in the room, volume of the room, refrigerant prop-  = air exchange rate, Q/V
erties, a concentration limit, and an appropriately conservative  = kinematic viscosity
and realistic accident scenario. Further work is needed to  = angle relative to the upward vertical direction
expand the refrigerants with RCL values as well as better formed by the liquid surface remaining around the
defining some of the refrigerant properties in Standard 34 to circumference of the tank
complete this work for all refrigerants.  = density
 = percentage of the liquid that flashes, found in Table
ACKNOWLEDGMENTS
3
This work was funded by ASHRAE research project
1448-RP through TC 4.3 – Ventilation Requirements & Infil- Subscripts
tration. b = boiling
NOMENCLATURE d = detector delay
f = flashing
A = area
h = hole
co = discharge coefficient
l = liquid
cp = specific heat at constant pressure
v = vapor
cv = specific heat at constant volume
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
s = surface
d = diameter
0 = ambient or initial
E = mass flow rate
1,2,3 = states
ff = friction factor
max = maximum
f = safe volume fraction, V/VS
G = mass REFERENCES
g = acceleration due to gravity AIChE (American Institute of Chemical Engineers). 1996.
H = height Guidelines of Vapor Cloud Dispersion Models, 2nd Edi-
Hvap = enthalpy of vaporization (the difference between Hv tion. New York: Center for Chemical Process Safety.
and Hl at the boiling point and ambient pressure) API (American Petroleum Institute). 1996. Manual for Mod-
eling Hypothetical Accidental Releases to the Atmo-
p = pressure of the liquid inside the vessel,
sphere. API PUBL 4628.
K = fitting loss coefficient ASHRAE. 2009. 2009 ASHRAE Handbook--Fundamental.
k = ratio of specific heats cp/cv Atlanta: American Society of Heating, Air-Conditioning
L = length and Refrigeration Engineers, Inc.
M = mass ASHRAE. 2010. ANSI/ASHRAE Standard 15-2010, Safety
Standard for Refrigeration Systems. Atlanta: American
Mw = molecular weight
Society of Heating, Air-Conditioning and Refrigeration
m = detector delay time influence factor (Table 3) Engineers, Inc.
Re = Reynolds number ASHRAE. 2010. ANSI/ASHRAE Standard 34-2010, Desig-
Q = volume flow rate nation and Safety Classification of Refrigerants.
R = universal gas constant Atlanta: American Society of Heating, Air-Conditioning
and Refrigeration Engineers, Inc.
S = entropy
Brown, R. 2005. Machinery Room Ventilation for Industrial
 Svap = entropy of vaporization (difference between Sv and Refrigeration Systems: A Rational Engineering
Sl at the boiling point and ambient pressure) Approach. Technical Paper No. 5. IIAR Ammonia
T = temperature Refrigeration Conference and Exhibition. Acapulco,
t = time Mexico.
506
ASHRAE Transactions
Fenton, D. L. and W.V. Richards. 2003. User’s Manual for Wilson, D.M. 1981. Along-wind diffusion and source tran-
ANSI/ASHRAE Standard 15-2001 Safety Standard for sients. Atmospheric Environment, 15:489-495.
Refrigeration Systems. ASHRAE Special Project SP-93.
Atlanta: American Society of Heating, Air-Conditioning DISCUSSION
and Refrigeration Engineers, Inc.
IIAR. 2008. ANSI/IIAR 2-2008 (Addendum A), Americation Joy Kohler, Engineering Manager, Johnson Controls,
National Standard for Equipment, Design and Installa- York, PA: The report cited only one example of fluorocarbon
tion of Closed-Circuit Ammonia Mechanical Refrigerat- leaks in machinery rooms: a fatality accident in Alaska. Most
ing Systems. Alexandria: International Institute of of the data is from ammonia systems, which are significantly
Ammonia Refrigeration. different from fluorocarbon systems due to the self-alarming
MAHB (Major Accident Hazard Bureau). 2001. eMARS – nature of ammonia and the typical design of ammonia
Major Accident Reporting System. systems. What effort was made to locate such data? Certainly
www.emars.jrc.ec.europa.eu/?id=4. there must be accident data for such systems.
Moran, M.J. and H.N. Shapiro. 2000. Fundamentals of Engi-
Scot K. Waye: The major databases that were searched for
neering Thermodynamics, 4th Ed. New York: John
refrigerant leaks in machinery rooms are the following:
Wiley and Sons.
MARS, NRC, ARIP, OSHA, and NIOSH. Table 1 of the paper
NIOSH (National Institute for Occupational Safety and
lists the number of accidents reported to the NRC for different
Health). 2010. Fatality Assessment and Control Evalua-
refrigerants. The number of fluorocarbon refrigerant releases
tion (FACE) Program. www.cdc.gov/niosh/face/.
reported for the indicated time period are very few, especially
NRC (National Response Center). 2010. www.nrc.uscg.mil/
as compared to ammonia. The same is true for the OSHA data-
download.html
base. No accidents involving fluorocarbon refrigerants were
OSHA (Occupational Safety & Health Administration).
found in the MARS database, and only one incident was found
1992. Process Safety Management (PSM) of Highly
in the NIOSH database (the Alaska incident). Summaries of
Hazardous Chemicals Standard. 29 CFR 1910.119.
the incidents reported in the NRC database are in the appendix
OSHA (Occupational Safety & Health Administration).
of the full report for ASHRAE RP-1448. [Editor’s note: final
2010. Fatality and Catastrophe Investigation Summa-
reports of ASHRAE Research Projects are available for free
ries. www.osha.gov/pls/imis/accidentsearch.html
download to members at www.ashrae.org.] Efforts were made
Richards, W.V. 1986. How Codes and Regulations Affect the
to find additional databases and sources for more reported
Design of Your Plant. IIAR Ammonia Refrigeration
leaks involving non-ammonia refrigerants, but these failed to
Conference and Exhibition. Innisbrook, FL.
provide additional incidents.
Sallet, D.W. 1990. Critical two-phase mass flow rates of liq-
uefied gases. Journal of Loss Prevention in the Process For many refrigerants, apart from Ammonia, the quantity
Industries 3:38-42. released that triggers a mandatory report is high, which we
Seidl, R. and S.T. Taylor. 2005. Sizing Exhaust System for believe causes an underreporting of incidents for fluorocarbon
Refrigerating Machinery Rooms. Taylor Engineering, leaks. For some refrigerants, no reportable quantity is listed at
Alameda, CA. all. The report for the Alaska incident was due to a fatality,
Stoecker, W. F. 1998. Industrial Refrigeration Handbook. causing an investigation and a report. For some cases, it is
New York: McGraw-Hill Companies, Inc. presumed that without injury or fatality a report of the leak was
US EPA. 1992. Workbook of Screening Techniques for not mandatory under the current requirements; this also causes
Assessing Impacts of Toxic Air Pollutants (Revised). underreporting. We acknowledge that better reporting and
Office of Air Quality Planning and Standards. more data representing fluorocarbon refrigerant leaks would
US EPA. 2001. Consolidated List of Chemicals Subject to enhance understanding of the issue, but this data was not avail-
the Emergency Planning and Community Right-To- able to us through public and open sources. Although much of
Know Act (EPCRA) and Section 112® of the Clean Air the data from actual releases was for ammonia, trends of
Act. EPA 500-B-01-003. Office of Solid Waste and potential leak locations and causes may be examined. Addi-
Emergency Response (5104). US EPA, USA. tionally, a threshold probability for emission rate exceedance
US EPA. 2010. www.epa.gov/emergencies/tools.htm#arip can be estimated for worst-case emission scenarios.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
©2012 ASHRAE
507
SA-12-016
Secondary Control Strategies for Cycling of

Air Conditioning and Refrigeration Systems
Swarooph Nirmal Seshadri Bryan P. Rasmussen, PhD

Member ASHRAE
ABSTRACT their speed to the cooling load, achieving higher efficiencies at

part load conditions. For both standard and rapid cycling there
A large percentage of air conditioning and refrigeration
is a need to evaluate the impact of the secondary fan and valve
systems rely on compressor cycling strategies to meet cooling
control strategies on system efficiency.
loads. This paper examines the impact of expansion valve and
heat exchanger fan control strategies on system efficiency, Existing research regarding optimal cycling focuses
assuming an on-off cycle of the compressor. The first step is the almost exclusively on optimizing the duty cycle and schedul-
development of a start-up/shut-down dynamic simulation ing of the compressors corresponding to a cooling load profile.
model that is used as a tool to understand the effect that valve/ (Chang, Lin et al. 2005) and (Rampazzo 2010) presented
fan cycling has on the system’s cyclic efficiency. Refrigerant methods for determining optimal chiller sequencing, consid-
migration due to valve strategies and fan delay at shutdown are ering the effects of loading limit and minimum down time.
determined as crucial issues that have an effect on system’s (Honglian, Larsen et al. 2009) used a simple model of an AC
cyclic efficiency. Optimization approaches are then used to system to determine the optimal compressor cycle input by
search for the optimal cycle strategies. The simulation studies optimizing a function including thermal comfort, energy effi-
are followed by extensive experimental tests conducted on a 3- ciency and compressor wear (number of switches). More
ton residential air conditioner, confirming the major conclu- recently, (Li and Alleyne 2010) considered the same optimi-
sions of the simulation study, namely that the valve and fan zation variables and used relay feedback to determine optimal
control strategies during compressor shutdown can result in compressor switching. (Jian and Zaheeruddin 1998) calcu-
significant variations in cyclic COP. lated the optimal lead time for the compressor switching given
a cooling load profile, while using an optimization function
INTRODUCTION based on temperature performance and energy performance.
Despite all these studies, the role of expansion valve and the
Despite the prevalence of modern controls systems,
heat exchanger fan control during cycling has yet to be inves-
control of vapor compression based Air Conditioning and
tigated.
Refrigeration (AC&R) systems remains fairly simple. The
compressor and fans are cycled on and off to meet cooling The remainder of the paper is organized as follows. First
loads, and the expansion valve uses mechanical or electronic a brief summary of dynamic modeling of vapor compression
feedback control to regulate evaporator superheat. In this system components will be presented. These models are then
study, conventional cycling refers to HVAC&R systems that used to conduct simulation studies and evaluate control strat-
compressor on/off cycle period is on the order of minutes. egies, and reveal the critical underlying physical phenomena
Rapid compressor cycling refers to cycle periods on the order that lead to cyclic efficiency gains/losses. This is followed by
of seconds. In both cases, the compressor duty cycle is varied a brief description of the experimental system, and detailed
to meet the desired cooling load. Variable speed compressors experimental results that verify the major inferences from the
are an alternative to compressor cycling strategies, and vary simulation study.
B.P. Rasmussen is an assistant professor of mechanical engineering at Texas A&M University in College Station, TX. S. Seshadri is a mechan-
ical engineer with The Mathworks Inc.
508
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
ASHRAE Transactions
DYNAMIC MODELING OF h out,isentropic – h in
VAPOR COMPRESSION SYSTEMS  a = ---------------------------------------------- (3)
h out – h in
This section describes the modeling of the various
P out
components in the vapor compression based AC&R systems:  a = f 2   k , ---------- (4)
 P in 
compressor, electronic expansion valve, the evaporator and
the condenser. The models were drawn from (Rasmussen
2005) and (Gupta 2007) and implemented in the MATLAB/ Electronic Expansion Valve
Simulink environment. The heat exchangers were modeled The mass flow rate through the electronic expansion valve
using the Finite Control Volume (FCV) approach similar to is based on the standard orifice equation,
(Gruhle and Isermann 1985) (MacArthur and Grald 1987).
·
This modeling approach uses a spatially dependent approach m v = C d A v   P in – P out  (5)
where the heat exchanger is divided into a number of constant
volumes. A lumped parameter approach is then applied to each C d A v = f 3  u v  P  (6)
of the control volumes, thus producing a discretized distrib-
uted parameter approach. Another popular technique is the The valve area and discharge coefficient are modeled using a
Moving Boundary (MB) technique that estimates the point in semi-empirical map that depended on the valve command and
a heat exchanger at which phase change occurs (Wedekind, the pressure differential across the valve (Equation 6). The
Bhatt et al. 1978). Here the heat exchanger length is divided valve is assumed to be isenthalpic.
dynamically depending on the refrigerant phase. (Bendapudi,
h v,out = h v,in (7)
Braun et al. 2008) compared the two modeling techniques and
concluded that the main advantage of the MB models is their
speed of execution with predictions similar to the FCV Heat Exchangers
models. Although FCV models are computationally slower, As noted before, FCV based modeling approach was
they are better suited to capture the complex start-up and shut- adopted for the simulation model. The governing differential
down transients. Additional details on dynamic modeling of equations were obtained from conservation of refrigerant
vapor compression systems can be found in the review papers energy, mass and heat exchanger wall energy. A detailed deri-
(Bendapudi and Braun 2002) (Rasmussen 2012; Rasmussen vation can be found in (Gupta 2007). In the FCV evaporator
and Shenoy 2012). model, the state of the refrigerant in a particular control
volume is determined by the enthalpy of the refrigerant at the
Modeling Assumptions exit of that control volume. Figure 1 depicts the FCV heat
exchanger model. In the figure m· in , h in and m· out , h out are the
• The mechanical dynamics of the compressor and the mass flow rates and the enthalpies at the inlet and the exit of
electronic expansion valve are significantly faster than the heat exchanger. The average enthalpy in the k th control
the two-phase heat exchanger dynamics and can be region, h k , is assumed to the same as the enthalpy leaving the
modeled with algebraic relationships. region. The mass flow rate exiting each region is denoted m· k .
• Heat exchangers are modeled as horizontal tubes with A nonlinear state space of form in Equation 8 is used to
one dimensional fluid flow. describe the heat exchanger dynamics, based on conservation
of mass and energy. This contains 2n+1 dynamic states
• The heat transfer occurring at the heat exchangers was
(enthalpy of n regions + wall temperature of n regions + heat
assumed to be isobaric.
exchanger pressure). These are contained in the dynamic state
vector x, as expressed in Equation 9. The input vector u (Equa-
Compressor tion 10) and consists of external inputs to the heat exchanger
The compressor was modeled using two relationships. such as inlet air temperature and air mass flow rate. For addi-
The first determined the mass flow rate of the refrigerant at the tional modeling details, see (Rasmussen 2012).
compressor outlet involving an empirically derived volumet- ·
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Z  x,u x = f  x,u  (8)

ric efficiency (Equation 1).
T
·
m k = V k  k  k (1) x = [P e ... h e,1 ... h e,k ... h e,n T w,1 ... T w,k ... T w,n ] (9)
· · ·
u =  m in m out h in T a,in m air  T (10)
P out
 k = f 1   k , ---------- (2)
 P in 
SIMULATION
The second expression calculates the outlet enthalpy based on In this section the simulation approach for optimizing the
the assumption that compression was adiabatic with an isen- expansion valve and the evaporator fan cycles and the associ-
tropic efficiency (Equations 3 and 4). ated results are discussed in detail. A cyclic efficiency quantity
©2012 ASHRAE
509
Figure 1 FCV heat exchanger model.
denoted ICOPCycle is defined for this purpose. This is Parameterization

followed by the presentation of the efficiency trends with
respect to the expansion valve and evaporator fan profile To find an optimal expansion valve and evaporator fan
parameters. Refrigeration migration is identified as a primary profile given a particular compressor cycle requires parame-
cause for changes in system efficiency, as these directly affect terization of the valve and fan control signals. In place of the
the time required for pulling down evaporator pressure and infeasible problem of optimizing the valve position/evapora-
achieving the desired level of cooling. Fan control strategies tor fan speed profiles at each sampling point, the actuator
affect both start-up and shut-down transients, and directly profiles were parameterized using a generalized sigmoid-
account for changes in fan power consumption. These results curve model given in Equation 12 and depicted in Figure 2.
are consistent with earlier work where refrigerant migration Here, K and A are the values for the upper and lower asymp-
was noted as a major factor affecting cyclic efficiency (Wang totes, and M and B dictate the time and slope of the transition.
and Wu 1990). Similarly, the effect of fan delay has also been The values Q and v are governed by the initial condition of the
explored in (Mulroy 1986). An optimization approach is them curve, and are constant in this study.
employed to determine the best sequencing strategies for fan
and expansion valve. K–A
Y  t  = A + ------------------------------------------------
- (12)
 1 + Qe – B  t – M   1  v
Objective Function
Coefficient of Performance (COP) is the measure of Trends
energy efficiency of an air conditioning system. For optimi-
zation routines the inverse COP (ICOP) was selected for mini- To visualize the impact on efficiency changes, a normal-
mization. A moving average of instantaneous cooling and ized COP Cycle was used for comparing different strategies
power allows the comparison of what energy efficiency over (Equation 13). This performance metric is similar to the one
the entire cycle, as opposed to the instantaneous efficiency at used in (Mulroy 1986). The simulations were run for two
particular points in time during the cycle. The resulting different cycle lengths to see the effect cycle length has on
cycling efficiency is calculated continuously as per Equation COP Cycle .
11, where t cycle is the cycle period, and t c = t – t cycle .
t COP Cycle
 Q    d Normalized COP Cycle = ------------------------------------------------
 0
- (13)
COP BenchmarkCycle
 ------------------------- for t  t cycle
 t
  W    d Short Cycle. Simulation results obtained for a short cycle
 tc
COP Cycle  t  =  (11) length of 200 seconds are presented here. The benchmark
 t
  Q    d cycling strategy assumes the expansion valve and the evapo-
 ------------------------
0
- for t  t cycle rator fan are cycled simultaneously with the compressor. The
 t
  W    d rest of the cycling strategies are described in Table 1, with
 tc performance metrics shown in Figure 3. Case A leaves the
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
510
ASHRAE Transactions
Figure 2 Sigmoid parameter variations.
Table 1. Basic Cycling Strategies
Expansion Valve Strategy Evaporator Fan Strategy

Cycling
Scheme Compressor Compressor Compressor Compressor
Off On Off On
Benchmark CLOSED OPEN OFF ON
A OPEN OPEN ON ON
B CLOSED OPEN ON ON
C OPEN OPEN OFF ON
OPEN for 20 [s] then

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
D OPEN OFF ON
CLOSED
E CLOSED OPEN ON for 20 [s] then OFF ON
valve open and the fan on during compressor shutdown. Cases during shutdown. This is confirmed by Figure 4(c) where, the
B and C evaluates each of these strategies separately. Cases D amount of refrigerant mass in the evaporator during the off
and E assess the impact of delayed shutdown of the valve and period is almost twice in Case A as in the benchmark. Since the
fan respectively. refrigerant flow to the evaporator is effectively cut off in the
The data show a potential change of 45% in COP due to benchmark condition, the refrigerant is completely vaporized
secondary control strategies, with A having the worst perfor- and superheated. The effect of valve cycling on superheat is
mance and E with an efficiency that similar to the benchmark. evident from Figure 4(b). Here the superheat is lost during the
Figure 4(a) shows simulation details for the benchmark, and off period, and the evaporator is flooded as the valve remains
Cases A and E. For Case A, since the valve is on during the off open. The instantaneous cooling and work are as shown in
cycle, the pressures equalized, whereas the valve is closed in Figures 4(d) and 4(e), and the startup cooling for Case A is
the benchmark so the pressures do not equalize. This influ- significantly less than the benchmark as it takes longer to
ences the startup efficiency as it avoids refrigerant migration remove the excess refrigerant and pull down the evaporator
©2012 ASHRAE
511
Figure 3 Comparison of basic cycling strategies, simulation.
pressure. As expected COP Cycle for the A is much less when the range in Table 3, the other parameters were held constant
compared to the benchmark as seen in Figure 4(f). at their benchmark values.
In Case E, the pressures almost equalize because of the Figure 5(a) and 5(b) show the relatively negligible effect
faster heat transfer during shutdown since the fan is on for 20 of shutdown slope and startup slope respectively on, assuming
seconds after shutdown. Case E loses some superheat at shut- the response is sufficiently fast. Figure 5(e) shows that the
down because of the extended operation of the fan leading to valve has to be strictly closed during the off cycle for the best
heat transfer. Loss in cyclic efficiency for Case E can be attrib- cyclic efficiency of the system. Leakage in the valve means
uted to the evaporator fan power during the off period. Switch-
loss of system efficiency, as this delays evaporator pressure
ing the fan on during the compressor off period did result in
pull-down, and thus cooling. Figure 5(c) and Figure 5(d) pres-
additional cooling, but the trade-off between cooling and fan
ent the effect of valve shutdown time and startup time on. In
power did not work in favor of the cyclic efficiency.
both the cases, large lead/delay operation of the valve did not
Overall, the various cycling schemes show the important show increase in cyclic efficiency. The conclusions that were
dynamics that affects the cyclic efficiency of the system. drawn from the trend analysis for the valve are thus summa-
However, there is still a need to investigate the expansion valve rized in
and the evaporator fan cycling sigmoid parameters individu-
ally. Variations in expansion valve profile were considered The effects of evaporator fan signal parameters on
first, while the evaporator fan was assumed to be on for the COP Cycle is also shown in Figure 5. For these analyses the
entire cycle. The values of the sigmoid parameters used for valve was cycled in sync with the compressor. This approach
this simulation are specified in Table 2. Higher values of slope was justified since the previous discussion has already estab-
correspond to steeper transients in the control signal, and lished that valve cycling is preferred. Figure 5(a) and 5(b)
faster actuator response. Negative time values mean a lead show the relatively negligible effect of shutdown slope and
operation (open/close action before the compressor cycles on/ startup slope respectively on COP Cycle , assuming the fan
off) whereas positive value means a delayed operation (action response is sufficiently fast. Figure 5(f) shows that leaving the
after the compressor cycles). fan on at the specified speed throughout the off cycle is detri-
The detailed trend analysis was one dimensional. mental to system efficiency. Figure 5(c) also indicates that
(Figure 5 presents the results of the detailed trend analysis for delayed fan shutdown may not result in significant efficiency
the individual sigmoid parameters for the expansion valve gains, as this generally consumes more power than justified by
profile.) That is, when one parameter was varied according to the additional cooling obtained. Figure 5(d) shows that a large
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
512
ASHRAE Transactions
Figure 4 Short cycle: detailed comparison of select cases, simulation.
Table 2. Simulation Sigmoid Parameters Range for Detailed Trends
Shutdown Slope Startup Slope Shutdown Time [s] Startup Time [s]
0.5 to 32 0.5 to 32 -20 to 20 -20 to 20
©2012 ASHRAE 513

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 5 Short cycle: detailed trends for parameter variations, simulation.
Table 3. Trend Analysis Summary for Expansion Valve – Short Cycle
Expansion Valve Cycle Recommendations for Better

Representative Figures
Sigmoid Parameter Cycle Efficiency
Shutdown Slope Fast Response Figure 5(a)
Startup Slope Fast Response Figure 5(b)
Shutdown Time With Compressor Figure 5(c)
Startup Time With Compressor Figure 5(d)
Valve position during OFF cycle Closed Figure 5(e)
514
ASHRAE Transactions
Table 4. Trend Analysis Summary for Evaporator Fan – Short Cycle
Evaporator Fan Cycle Recommendations

Sigmoid Parameter for Better Cycle Efficiency

Shutdown Time With Compressor/Delay Figure 5(c)
Air mass flow rate during OFF cycle Zero (Fan Off) Figure 5(f)
lead or lag in fan startup does not significantly improve effi- compressor is turned off, the system’s pressure differential is
ciency. The conclusions that were drawn from the trend anal- also maintained higher during the off period, allowing the
ysis are summarized in Table 4. system to reach full capacity almost instantaneously upon
Long Cycle. Simulation results obtained for a relatively start-up. The increase in efficiency (~2% compared to bench-
long cycle length of 1000 seconds are discussed here. Figure mark) is also affected by the minor decrease in fan power
3 shows the trends for the various cycling strategies for both consumption during the cycle. By slightly delaying the fan
long and short cycles. A quick glance at Figure 3 indicates that during startup and leading it during shutdown, the fan is oper-
the cycling schemes exhibit more modest changes in cycle ating for a less time. Although the simulation assumes nearly
efficiency for long cycle periods. This is expected as the start- instantaneous response in expansion valve opening/closing,
up and shut-down transients occupy a relatively smaller the fan is modeled with inertial fan effects that keep some air
percentage of the cycle period, and thus the valve and fan strat- flowing across the evaporator as the fan is turned off. Although
egies have a proportionally smaller effect. Although a detailed the efficiency gains are minor, the results indicate that
trend analysis was conducted for the valve and fan control improvements are possible over the completely synchronous
signal parameters, the results are not significantly different approach. These strategies will also have a more dramatic
from expectations, and thus they are not included in this paper. effect for systems of significantly different size, or systems
employing rapid cycling.
Optimization
In this section key parameters are varied to determine Controller Evaluation
optimal values for the valve and fan control signals. Although The benchmark set in this paper cycles the valve and fan
the results of such an optimization will bew highly dependent synchronously with the compressor. However, in practice the
on the system parameters, the results are indicative of expansion valve is electronically or mechanically controlled
approaches that may be adopted as good practices. Based on to regulate evaporator superheat. In this section, two simple
the trend analysis, the parameters selected for optimization control strategies are evaluated to illustrate the potential
are: effects on efficiency.
• Expansion valve startup and shutdown time • Control Scheme 1: Valve position is determined by the
• Evaporator fan startup and shutdown time superheat controller during the complete on-and-off
cycle. This represents electronic valve controllers that
The Nelder Mead simplex algorithm (Nelder and Mead operate independently, or thermostatic expansion valves
1965) was used to determine the minimum. Since the short that operate using mechanical feedback.
cycle displayed significant changes in efficiency with respect • Control Scheme 2: Valve position is determined by the
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
to the valve and the fan sigmoid parameters, the optimization superheat controller through the on cycle. During the off
was carried out only for this cycle length. The results of opti- cycle the control signal is disregarded and the valve is
mization are shown in Table 5, and the resulting valve and fan completely shut off to avoid refrigerant migration. Note
singals are shown in Figure 6. that this requires coordination between compressor and
The ‘optimized’ condition achieves a faster pressure pull- valve controllers.
down at startup because of the slight delay in start-up, which
allows the compressor to begin evacuating the evaporator For simplicity and a uniform comparison, the controller
before the valve allows additional refrigerant to enter, and used in the both cases is a simple PI controller, and control
before the fan begins to extract heat. This also allows the scheme 1 is selected as the benchmark for comparison. The
system to reach the desired level of superheat sooner. By simulation results shown in Figure 7 illustrate large changes in
delaying the valve closing and fan shut down slightly after the efficiency due to valve control strategy. The main source of
©2012 ASHRAE
515
Table 5. Optimization Results
Cycle Sigmoid Parameter Solution
Shutdown time (seconds) -1.0

Expansion Valve Startup time (seconds) 1.4
Shutdown time (seconds) -2.9

Evaporator Fan Startup time (seconds) 3.4
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 6 Short cycle: optimization results, simulation.
inefficiency is the valve controller remaining open during part Figure 8. The system was instrumented with a full array of
of the off period. As discussed earlier, this results in refrigerant pressure sensors, thermocouples, flow rate sensors, humidity
migration, and the subsequent efficiency losses. In this case sensors etc. The actuators used in the systems are an electronic
this is the result of integrator windup in the controller. expansion valve and two variable speed fans. The system was
However, similar effects would be observed in electronic or charged with R-410A refrigerant, used a two stage fixed speed
thermostatic expansion valves that continue to attempt to regu- scroll compressor.
late superheat during the off period.
The various valve command signals are shown in Trends
Figure 7(a), and the resulting superheat response is given in
Figure 7(b), and the average cyclic efficiency in Figure 7(c). Similar to the simulation studies, a series of experiments
The strategies that coordinate the valve action with the were conducted to assess the effects of basic secondary control
compressor control result in 30% improvement in efficiency strategies on cyclic efficiencies. This included experiments for
due to improved startup performance, and minimized refrig- both relatively short and long cycle periods.
erant migration during the off period. Short Cycle. Figure 9 shows the general trends for the
various cycling strategies defined in Table 1, with results
EXPERIMENTAL EVALUATION comparable to the simulation study. Cases C and E were
selected for detailed presentation Figure 10. For case C, since
Experimental System the valve is open during the compressor on period the pres-
To verify the conclusions drawn from the simulation sures equalized (Figure 10(a)), whereas in benchmark the
study, similar tests were conducted on a 3-ton residential air valve is closed and a pressure differential is maintained. The
conditioner. Photos of the system components are shown in subsequent flooding of the evaporator results in a slower, and
516
ASHRAE Transactions
Figure 7 Short cycle: comparison of valve control strategies, simulation.
larger, superheat response upon startup (Figure 10(b)), and a simulation study. Specifically, valve and fan responses should
decrease in cooling capacity (Figure 10(c)). Although refrig- be sufficiently fast, the valve should be closed during the
erant charge cannot be measured directly, as in the simulation compressor off period, and some modest improvements in
studies, the results clearly indicate a problem of refrigerant efficiency can be realized with delayed fan shutdown. These
redistribution. Figure 10(d) shows the instantaneous work and conclusions are summarized in Table 6 and 7.
the cyclic efficiency in Figure 10(e). Case E gives more cool-
ing per cycle because of extended fan operation (Figure 10(d)) Long Cycle. Results obtained for a long cycle length of
during the initial part of the off cycle shown in the instanta- 1000 seconds are presented in Figure 9. To determine whether
neous cooling in Figure 10(c). This results in a modest the changes in efficiency are significant enough to merit a
increase in efficiency despite the increased power consump- further detailed investigation, a quick comparison of some
tion as shown in Figure 10(e). possible cycling schemes is presented. Similar to the long
Similar to the simulation study, the detailed trend analysis cycle simulations, the discerning characteristic in the long
was conducted for expansion valve and evaporator fan sepa- cycle is the major effect of fan cycling on cyclic efficiency. The
rately. The results in Figure 11 confirm the conclusions of the physical reasons for this are same as noted in the simulation
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
©2012 ASHRAE
517
Figure 8 Residential air conditioning system.
Figure 9 Comparison of basic cycling strategies, experiment.
section. Given these results, a detailed parameter analysis was ciency differences are even more pronounced, with significant
not conducted. flooding of the evaporator, and the overcompensation of the
valve resulting in large overshoot in the superheat response
Control Strategies during the compressor on period.
Real-time optimization of the actuator control signals on
the experimental system is not feasible, as maintaining condi- CONCLUSION
tions environmental conditions throughout the many iterations This paper makes specific contributions to the field of air
is not practical. However, an evaluation of the basic valve conditioning cycling. To the author’s knowledge, no previous
control strategies confirm the results of the simulation study. research evaluated the effects of specific expansion valve and
Again, there is a significant change in efficiency due to the evaporator fan cycling schemes had on an air conditioning
coordination of the valve and compressor controllers, as seen system’s cyclic efficiency. Through extensive simulation anal-
in Figure 12. However, in the experimental study the effi- ysis, several key vapor compression system dynamics were
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
518
ASHRAE Transactions
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 10 Short cycle: detailed comparison of select cases, experiment.
identified as causes for increasing/decreasing system’s cyclic strategies can result in significant degradations is achievable
efficiency. Start-up dynamics, particularly evaporator pres- cyclic efficiencies. These results are confirmed with experi-
sure pull-down, are critical indicators of cyclic efficiencies. mental tests using a residential air conditioning system. Eval-
Thus refrigerant migration and heat transfer during the uation of typical expansion valve control strategies indicate
compressor off period will strongly affect efficiency. Recom- that coordination between compressor cycling and expansion
mendations for fast-acting expansion valve and evaporator valve is critical to avoid refrigerant migration during the off
fan, optimized valve and fan delays to improve start-up tran- cycle, and improve the response of cooling capacity upon
sients and extract additional cooling during shutdown, and start-up.
complete closing of the expansion valve during the compres- This work also reveals areas for further study. These
sor off period will result in modest improvements in efficien- include a more detailed evaluation of strategies for systems
cies compared to synchronous operation. However, other with rapid compressor cycling, optimal secondary cycling
©2012 ASHRAE
519
Figure 11 Short cycle: detailed trends for parameter variations, experiment.
·
strategies for multi-evaporator systems, and the impact of m = Mass Flow Rate
these strategies on condenser performance and connected  = Efficiency
systems (e.g. reheat, or waste heat reclamation systems).
 = Density
Nomenclature A = Area
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
P = Pressure Cd = Discharge Coefficient

T' = Temperature V = Volume
Q = Heat  = Rotational Velocity
W = Work Z = Matrix
h = Specific Enthalpy f. = Continuous Func tions
520
ASHRAE Transactions
Table 6. Trend Analysis Summary for Expansion Valve – Experiment
Expansion Valve Cycle Recommendations


Shutdown Time With Compressor Figure 11(c)
Valve position during OFF cycle Closed Figure 11(e)
Table 7. Trend Analysis Summary for Evaporator Fan – Experiment
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Evaporator Fan Cycle Recommendations
Shutdown Time Delay Figure 11(c)
Air mass flow rate during OFF cycle Zero (Fan Off) Figure 11(f)
u = Input Signals Chang, Y.-C., F.-A. Lin, et al. (2005). "Optimal chiller
x = Dynamic States sequencing by branch and bound method for saving
energy." Energy Conversion and Management 46(13-
Subscripts 14): 2158-2172.
Gruhle, W.-D. and R. Isermann (1985). "Modeling and Con-
1,2, ... j, ... n = First, Second, etc. region
trol of a Refrigerant Evaporator." Journal of Dynamic
air = Air Systems, Measurement, and Control 107(4): 235-240.
in = In Gupta, A. (2007). Reduced Order Modeling of Heat
out = Out Exchangers using High Order Finite Control Volume
e = Evaporator Models. Department of Mechanical Engineering. Col-
lege Station, TX, Texas A&M University. Master of Sci-
k = Compressor ence.
v = Valve Honglian, D., L. Larsen, et al. (2009). Control of systems
with costs related to switching: applications to air-con-
ACKNOWLEDGMENTS dition systems. Control Applications, (CCA) & Intelli-
The authors gratefully acknowledge the assistance of gent Control, (ISIC), 2009 IEEE.
anonymous reviewers in improving the manuscript. Dr. Jian, W.-L. and M. Zaheeruddin (1998). "Sub-optimal on-off
Rasmussen’s efforts were supported by the National Science switching control strategies for chilled water cooling
Foundation under grant CMMI-0644363. systems with storage." Applied Thermal Engineering
18(6): 369-386.
REFERENCES Li, B. and A. G. Alleyne (2010). Optimal on-off control of
Bendapudi, S. and J. E. Braun (2002). A review of literature an air conditioning and refrigeration system. American
on dynamic models of vapor compression equipment. Control Conference (ACC).
S.l., s.n., American Society of Heating, Refrigerating, MacArthur, J. W. and E. W. Grald (1987). Prediction of
Air-Conditioning Engineers. cyclic heat pump performance with a fully distributed
Bendapudi, S., J. E. Braun, et al. (2008). "A comparison of model and a comparison with experimental data.
moving-boundary and finite-volume formulations for Mulroy, W. J. (1986). Effect of Short Cycling and Fan Delay
transients in centrifugal chillers." International Journal on the Efficiency of a Modified Residential Heat Pump.
of Refrigeration 31(8): 1437-1452. ASHRAE Transactions.
©2012 ASHRAE
521
Figure 12 Short cycle: comparison of valve control strategies, experiment.
Nelder, J. A. and R. Mead (1965). "A Simplex-Method for Rasmussen, B. P. and B. M. Shenoy (2012). "Review Article:
Function Minimization." Computer Journal 7(4): 308- Dynamic Modeling of Vapor Compression Systems -
313. Part II: Simulation Tutorial." HVAC&R Research.
Rampazzo, M. (2010). Efficient Management of HVAC Sys-
Wang, J. and Y. Wu (1990). "Start-up and shut-down opera-
tems. Dipartimento di Ingegneria dell'Informazione,
tion in a reciprocating compressor refrigeration system
Università Degli Studi Di Padova.
with capillary tubes." International Journal of Refrigera-
Rasmussen, B. (2005). Dynamic Modeling and Advanced
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
tion 13(3): 187-190.

Control of Air-Conditioning and Refrigeration Systems.
Department of Mechanical Engineering. Urbana, IL, Wedekind, G. L., B. L. Bhatt, et al. (1978). "A System Mean
University of Illinois. PhD. Void Fraction Model for Predicting Various Transient
Rasmussen, B. P. (2012). "Review Article: Dynamic Model- Phenomena Associated With Two-Phase Evaporating
ing of Vapor Compression Systems - Part I: Literature and Condensing Flows." International Journal of Multi-
Review." HVAC&R Research. phase Flow 4(1): 97-114.
522
ASHRAE Transactions
SA-12-018
Development of High Efficiency Home

Heating and Humidification Technology
Dexin Wang, PhD Ainan Bao, PhD William Liss
ABSTRACT equipment. ASHRAE Standard 62-1989, states, “relative

humidity in habitable space preferable should be maintained
A novel humidification technology has been developed,
between 30% and 60%... to minimize growth of allergenic and
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
which can boost the efficiency of a mid-efficiency furnace from
pathogenic organisms” (see Fig.1). Notably, the lack of proper
around 80% to more than 95%, and at the same time provides
whole house humidification without requiring an external space humidification enhances the rate of influenza virus,
water supply. The technology uses a nanoporous membrane resulting in a significant number of illnesses and deaths each
capillary condensation separation mechanism to transport
water vapor only from furnace combustion flue gas to safely
humidify the building air. The water transported through the
membrane is clean and mineral free, avoiding maintenance
issues commonly seen with conventional humidifiers. The
advanced technology has been developed and evaluated for
long term operation at a laboratory environment, to prove its
feasibility for a wide range operating conditions. Additionally,
it has been demonstrated in two occupied single family homes
to capture real world performance for a whole heating season.
The results showed both homes achieved comfortable whole
house humidification and significant furnace efficiency
improvement. We foresee this technology can be also applied
to commercial and industrial needs, with the benefit of no pota-
ble water consumption and significant energy savings.
INTRODUCTION
It is generally accepted, and has been confirmed by stud-
ies, that humidification of dry indoor air to raise relative
humidity (RH) during the heating season is beneficial to the
comfort and health of building occupants. There are also
significant energy savings possible due to the “apparent
temperature” phenomenon that allows people to feel more
comfortable (i.e., warmer) at higher RH. Humidification also
prevents adverse effects on wood floors and furniture and Figure 1 ASHRAE indoor relative humidity
reduces static electricity buildup which can harm electronic recommendations.
Dexin Wang is a senior engineer, Ainan Bao is an engineer, and William Liss is managing director of end use solutions at Gas Technology
Institute in Des Plaines, IL.
©2012 ASHRAE
523
year. Humidity control is important in commercial buildings – design make it promising for a commercial product. To our
including hospitals – as well as many industrial processes, knowledge, no practical technology has ever been developed
such as electronic and semiconductor manufacturing, medical for humidifying room air with flue gas moisture for residential
supply, printing application, woodworking and storage, and use. TMH technology can reduce fuel use, eliminate city water
textile industries. consumption, completely avoid mineral deposition and white
Currently, the most widely used residential humidification dust, and avoid microbial growth, improving both the physical
technologies are forced air furnace-mounted bypass wetted and financial health of the homeowners.
media, spray mist, and steam humidifiers. These use city water For a typical 3,000 square foot (278 square meter) home,
as a water source and require additional furnace heat or elec- 12 gallons (45.4 liter) per day of water is generally needed to
tricity to evaporate the water. Mineral deposition, white dust and add into the air to achieve a satisfactory relative humidity
microbial growth problems are associated with most of these level. But using a furnace humidifier has adverse effects,
humidifiers. For commercial building humidification, deminer- including the consumption of water (up to 20 gallons, or 75
alized water is typically used for humidification equipment like liter per day considering loss through drain) and need for 4%
steam heat exchangers, electric and ultrasonic humidifiers, or more of the furnace fuel input to evaporate the water into
compressed air atomizers, and high pressure cold water foggers. vapor. This TMH concept can be developed into humidifica-
In addition to the energy consumption for the water evaporation, tion technology for residential, commercial, and industrial
energy is also needed to produce high-quality demineralized needs, with the benefit of no potable water consumption and
water through a reverse osmosis process. significant energy savings.
A Transport Membrane Humidifier (TMH) concept was This paper describes the TMH technology development
developed by using a nanoporous membrane that facilitates a details for mid-efficiency residential furnaces (around 80%
capillary condensation separation mechanism which trans- AFUE), which includes TMH concept technical evaluation,
ports water vapor only from furnace combustion flue gas to prototype TMH laboratory investigation, and two TMH field
humidify building air. The capillary condensation action trials to show its real world performance.
enables high water transport rates while also blocking non-
condensable gases from transporting across the membrane. TMH CONCEPT AND DESIGN
There are other research efforts aimed at using
membranes to separate and transport water vapor for gas TMH Concept
stream dehydration, humidity control, and energy recovery in
commercial HVAC systems. None of these applications, A TMH unit is made from many nanoporous membrane
however, has attempted to extract water vapor from a flue gas tubes. Figure 2 shows the photomicrograph of a typical
stream to humidify air. For all these and similar applications, membrane tube cross section. It consists of a top layer with a
only very small trans-membrane total pressure is available. pore size of 60 to 80 Å (about 2 to 4 µm thick), a 500 Å pore
The driving force for water vapor to transport from one side of size intermediate layer (typically 20 to 50 µm thick), and a 0.4
the membrane to the other side relies mainly on the water µm pore size substrate (about 1 mm thick). This structure is
vapor partial pressure difference between the two gas streams.
For all these reported applications, they are dealing with trans-
porting moisture from a high humidity air stream to a low
humidity air stream, the water partial pressure difference is
relatively small, less than 0.4 psi (2,760 pascal).
A flue gas stream typically has a dew point of 120 to
136°F (49 to 58°C). This high temperature high humidity level
can create a greater than 2 psi (13,800 pascal) water vapor
partial pressure difference with the circulating room air, which
usually has a dew point of 50˚F (10°C) or lower. Using flue gas
moisture to humidify the room air can provide five times larger
driving force across the membrane, therefore substantially
less membrane surface area is needed. The reduced surface
area greatly lowers the cost and improves the prospect for a
cost effective commercial application using the TMH. In addi-
tion, since the flue gas is typically at much higher temperature
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
(over 250˚F, or 121°C), the TMH functions as a heat

exchanger to preheat the air stream to save energy.
The combined energy saving and humidification function
with no potable water consumption makes this technology Figure 2 Photomicrograph of a nanoporous ceramic
unique. The reduced membrane surface area and simple membrane tube cross-section.
524
ASHRAE Transactions
used to achieve high separation ratio with minimal resistance living area. Importantly, the capillary condensation separation
to flux of the permeating species and very low pressure drop. mode of the membrane prohibits the transfer of non-conden-
sible gases such as nitrogen, carbon dioxide, etc.
Figure 3 is one arrangement to show the TMH working
concept for a residential furnace, where the TMH module is
TMH Design and Optimization
installed in the furnace air inlet ductwork. In the TMH,
moisture-laden flue gas flows from the membrane feed side, A CFD (Computational Fluid Dynamics) software
while the building air that requires heating and humidifica- FLUENT was used to optimize the ceramic membrane tube
tion flows on the permeate side. The low-temperature, high- dimensions and the TMH module design. The objective of
flow-rate, dry room air passing over the membrane surfaces using CFD modeling is to achieve the following results:
provides adequate membrane cooling to facilitate the high-
performance capillary condensation water vapor separation • Optimize designs for good heat transfer between the
mode. Water from the flue gas is transported to the air side, building air and flue gas
simultaneously heating and humidifying the air. • Minimize air side and flue gas side pressure drops
• Minimize the overall membrane surface area to reduce
It is important to emphasize that moisture transported the TMH module cost
through the membrane is mineral free, since it comes from the
flue gas in vapor phase and passes through the nano-scale Figure 4 shows simulation results of membrane tube inner
membrane pores, and emerges as vapor phase in the air stream. surface temperature distributions for a TMH module, and the
This avoids a common failure mode in home humidifiers due detailed temperature profile along one of the tube. CFD results
to mineral deposition, and related white dust carryover to the prove that, even with very low air and flue gas side pressure
drops, the TMH design still can achieve very good heat trans-
fer and the flue gas can be cooled down from 280°F (138°C)
to 70-80°F (21-26°C).
Another factor for the TMH design is the air ductwork
dimensions, and specific designs are required for different
kinds of residential furnaces. For assembling the TMH
module, appropriate adhesive and tube holder materials have
been selected to provide strong bonding strength and avoid
any leaking problem.
TECHNICAL EVALUATION
A laboratory TMH unit was designed and installed into a
residential furnace, with a firing capacity of 120,000 Btu/hr.
This is a typical mid-efficiency home furnace with rated annual
fuel utilization efficiency (AFUE) at 80%. The TMH module
Figure 3 TMH working concept for a gas-fired residential was installed in the furnace air inlet ductwork, with flue gas
furnace. flowing inside TMH membrane tubes and room air flowing
Figure 4 Temperature distributions for a TMH module and along one membrane tube by CFD modeling.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
©2012 ASHRAE
525
outside to pick up both heat and moisture from the flue gas. outlet temperature and vapor mass fraction, respectively.  fg
Table 1 lists the TMH module characteristic dimensions. is the flue gas density, Cp fg is the average flue gas specific
Figures 5 show photos of the TMH module and the installation. heat, and  is the water vapor latent heat.
Figure 6 shows the P&ID of the TMH laboratory test setup. The moisture transport rate is calculated by the following
Furnace natural gas is measured by a natural gas flow meter, and formula:
the room air flow rate is calculated based on air velocities inside
the ductwork. Air and flue gas inlet/outlet temperatures are m = FR air   air   wt % in – wt % out  (2)
measured by thermocouples. Air and flue gas inlet/outlet dew
points are measured by hygrometers. An ID fan was installed to
where  air is air density, FRair is the air flow rate, and wt%in
overcome the TMH additional pressure resistance in the flue gas
and wt%out are air inlet and outlet moisture mass fractions,
side, and its electricity usage is measured by a power meter. The
respectively.
furnace stack O2 level is measured by a portable gas analyzer
HORIBA. All experimental data are collected by a data acqui- Table 2 shows typical TMH performance details. From
sition system for post-processing. the experimental results, the TMH was able to transfer 63%
moisture (5.97 pounds per hour, or 2.71 kg per hour) from the
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
The furnace instant efficiency is calculated by the follow-

ing formula: flue gas side to the air side, and improve the furnace instant
efficiency by about 15%. Pressure drops are 0.23 inches of
FR NG  HHV – FR fg   fg  Cp fg   T fg – T ref  water (57 pascal) for the air side, and 0.51 inches of water (127
out
pascal) for the flue gas side. The low pressure drops allows the
– FR fg   fg  wt% vapor  Y TMH be installed inside the furnace air ductwork without
 = ----------------------------------------------------------------------------------------------------------------------
out
- (1)
FR NG  HHV causing problem of the furnace operation. We also monitored
both air and flue gas side gas components using the gas
Where, HHV is natural gas higher heating value, and analyzer, and Table 3 shows the typical gas component data,
FRNG and FRFG are volume flow rates for natural gas and flue which indicates no hazardous gas leaking or transporting
gas. The flue gas flow rate is calculated by the excess air ratio across the TMH to room air side.
and natural gas flow rate, and excess air ratios are calculated
The technical evaluation shows the TMH technology can
from the stack O2 level. Tfgout and wt%vaporout are flue gas
provide home air humidification without additional city water
and energy consumption, and also significantly increase the
Table 1. TMH Modules Characteristic Dimensions furnace efficiency.
Overall TMH dimension, inch 9.3 x 8.2 x 24.5 PROTOTYPE TMH LABARATORY INVESTIGATION
Membrane tube length, inch 23.75 A prototype TMH unit was built for a 120,000 Btu/hr
(3.51 kW) residential furnace, to verify the TMH perfor-
Membrane tube OD/ID, inch 0.315/0.236
mances in a wide operation range, and test out the TMH long
Total Tube number 130 term performance.
Figure 5 Pictures of the TMH module and installation with Figure 6 P&ID of the TMH test setup.
a furnace setup.
526
ASHRAE Transactions
Table 2. TMH Performance Details
Unit without TMH with TMH

Air  P through TMH inch-water 0.23
FG  P through TMH inch-water 0.51
NG flow rate SCFH 108 108
Air flow rate SCFM 2000 1850
Air inlet temperature F 69.0

Air inlet dew point F 31.8
Air outlet temperature F 75.7
Air outlet dew point F 37.3
FG inlet temperature to TMH F 293.7

FG inlet dew point F 122.0
FG outlet temperature F 94.9
FG outlet dew point F 88.9
Moisture transport rate % 63.45

Moisture transport rate lb/hr 5.97
Furnace efficiency (HHV) % 77.43 93.80
Table 3. Typical Gas Components Before and After the TMH (FG: flue gas)
Air inlet Air outlet FG inlet FG outlet

NOx ppm 0.5 0.5 27.6 27.6
CO ppm 0.5 0.6 7.5 7.2
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
CO2 % 0.08 0.08 4.54 4.51

O2 % 13.4 13.5 13.4 13.5
Parametric Study of TMH Performance increases to 22°F (-5.6°C) and RH increases to 16%; and at a
Figure 7 shows TMH moisture transport rates at different high humidity day (Dp=63°F, or 17.2°C, and RH=83%), the
furnace air flow rates, which indicates air flow rate has no air outlet dew point only increases to 65°F (18.3°C) and RH
significant effect on the water transport rate. to 87%. Therefore, the TMH will humidify the air less at a
The furnace flue gas O2 level effect was also studied. high humidity condition. Considering the fact that limited
Figure 8 clearly shows both the moisture transport rate moisture available from flue gas, it eases concerns of over-
through the TMH and the furnace efficiency increase, when humidification.
furnace flue gas O2 level decreases. This is mainly due to the Furnace operation mode also affects the TMH perfor-
higher flue gas inlet dew point to the TMH at lower O2 level, mance, and Figure 10 shows the moisture transport rate at
which makes flue gas moisture capillary condensation inside different furnace operation cycle conditions, where furnace
the membrane pores easier. operation cycle percentage here is defined as the percentage of
The air inlet dew point is another factor to affect the TMH furnace ON time in the overall time. Smaller operation cycle
performance. Figure 9 shows air side dew point and relative percentage has higher moisture transfer rate.
humidity changes through the TMH. At a very dry day Parametric study shows the TMH works well in a wide
(Dp=10°F, or -12.2°C, and RH=10%), the air outlet dew point operation range for a residential furnace, therefore a TMH can
©2012 ASHRAE
527
work at different room air conditions, which may be affected
by local weather conditions and other factors.
Prototype TMH Long Term Testing

For the prototype TMH unit, a long term testing has been
carried out to verify its long term performance and membrane
fouling effect. Figure 11 shows the TMH performance for a
5,000 hour furnace operating time, which is equivalent to
about 4 year operation of a typical furnace. No cleaning or
other maintenance had been conducted in this time frame. It
can be seen the TMH performance reached a stable condition
after an initial performance drop during the first 100 hour’s
operation. At the end of the testing period, the furnace effi-
ciency was about 91% and the moisture transport rate was 5.0
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
lb/hr (2.27 kg/hr). Results indicate that the TMH performance

Figure 7 Air flow rate effect on TMH moisture transfer is stable and high over a long period, without any maintenance
rate. need. This significantly distinguishes this technology from
Figure 8 TMH moisture transport rate and furnace efficiency at different flue gas O2 levels.
Figure 9 Air dew point and RH change through TMH at different air inlet dew points.
528
ASHRAE Transactions
Figure 10 TMH moisture transport rate at different furnace operating cycles.
Figure 11 Long term test results for furnace efficiency and TMH moisture transport rate.
conventional humidification methods, most of them requiring ft2 (251m2) living space, and home 2 has 1,350 ft2 (125m2). The
wet media replacements every 2-3 months or at least every TMH modules were installed into the return air plenum going
heating season due to mineral deposition clogging the wet into the furnaces, and the flue gas heat and water were simul-
media, which is highly related to the water supply quality. taneously recovered in the TMH and distributed into the homes
after being further heated by the furnaces. Flue gas inside the
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
FIELD TRIALS TMH was maintained at a slightly negative pressure by a draft

Two occupied single family homes were selected to fan, so flue gas will not pass to the air side in case of any poten-
demonstrate the whole house TMH heat recovery and humid- tial leaks. Also a pressure switch was installed at the TMH flue
ification technology for residential furnaces, to verify their gas outlet in addition to the furnace draft fan pressure switch,
real world performance on furnace efficiency improvement this switch will shutdown the whole furnace if the flue gas side
and whole house humidification. pressure after the TMH reaches positive.
Based on previous TMH design and assembling experience, Furnace efficiency calculation results show that the home
we re-designed two TMH modules with even lower air and flue 1 furnace efficiency increases from 81.5% without the TMH
gas pressure drops, and the TMH modules overall dimensions to 95.5% with the TMH, and the home 2 furnace efficiency
are based on the corresponding furnace air ductwork cross increases from 80.6% to 96.9%. The average moisture trans-
sections. Table 4 listed both TMH modules dimensions. port rates are 2.7-6.2 gallon per day (10-23 L per day) for
Figure 12 shows the installation pictures. Furnace in home home 1, and 1.5-4.8 gallon per day (5.7-18 L per day) for
1 is 110,000 BTU/hr (3.22 kW) fuel input, and furnace in home home 2, depending on different room air temperatures and
2 is 90,000 BTU/hr (2.63 kW) fuel input. Home 1 has a 2,700 dew point conditions. Humidity levels for both homes have
©2012 ASHRAE
529
Table 4. TMH Module Characteristic Dimensions for Field Trials
Home 1 Home 2
Tube number 130 130
OD/ID, inch 0.315/0.236 0.315/0.236
Membrane tube length, inch 23.5 20.25
Air ductwork cross section, inch 24 x 10 20 x 20
Overall TMH dimensions, inch 24.3 x 8.2 x 9.3 18.9 x 21.3 x 5.5
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
humidification level to the home owners and also greatly
increased the furnace efficiencies.
REFERENCES
Asaeda, M., L. Du and M. Ushijima. 1985. Feasibility study on
dehumidification of air by thin porous alumina gel mem-
brane p472-478; Drying’ 85. R. Toei and A.S. Mujumkar,
ed. Washington D.C.: Hemisphere Publishing Corp.
ASHRAE. 1989. ANSI/ASHRAE Standard 62.1989,Ventila-
tion for Acceptable Indoor Air Quality. Atlanta: Ameri-
can Society of Heating, Refrigerating and Air-
ASHRAE. 1992. ASHRAE Standard 55.1992,Thermal Envi-
ronmental Conditions for Human Occupancy. Atlanta:
American Society of Heating, Refrigerating and Air-
Figure 12 Two home demos, left for Home1 and right Ray, R., D. D. Newbold, S.B. McCray, and D. T. Frlesen.
Home2. 1992. A novel membrane device for the removal of
water vapor and water droplets from air. 22nd interna-
tional conf. on environmental system, Seattle, Washing-
been maintained in a comfortable range of 40-55% relative ton, SAE 921322.
humidity. The flue gas side pressure drops through the TMH Scovazzo, P., A. Hoehn, and P. Todd. 2000. Membrane
were measured as, 0.35-0.4 inches of water (87-99 pascal) for Porosity and Hydrophilic Membrane-based Dehumidifi-
home 1 TMH module, and 0.2-0.25 inches of water (50-62 cation Performance. J. Membrane Science 167:217-225.
pascal) for home 2. Scovazzo, P., J. Burgos, A. Hoehn, P. Todd. 1998. Hydro-
philic Membrane-based Humidity Control. J. Membrane
SUMMARY Science 149:69-81.
Sterling, E.M., Arundel, A., and Sterling, T.D. 1985. Criteria
A TMH concept has been developed into practical designs, for Human Exposure to Humidity in Occupied Build-
for providing home air humidification with no external water ings. ASHRAE Transactions 91:611-622.
supply need, and improving furnace efficiency simultaneously. Zhang, L. and Y. Jiang. 1999. Heat and Mass Transfer in a
Technical evaluation testing shows that the TMH can transfer up Membrane-based Energy Recovery Ventilator. J. Mem-
to 6 lb/hr (2.7 kg/hr) moisture from the flue gas to the air side brane Science 163:29-38.
to meet the humidification demand without additional water and
energy consumption. This design also drastically increases DISCUSSION
furnace efficiency by more than 15%. Parametric study shows Bill Roy, Manager of Advanced Combustion, Carrier
the TMH works well for a wide operation range for residential Corporation, Indianapolis, IN: Do you have a plot of the
furnaces. Long term performance study with more than 5,000 home relative humidity versus furnace efficiency?
hour’s operating time has proved that the TMH performance is Dexin Wang: No.
stable for long term operation without maintenance need. Two Richard Ferguson, Engineer, Toyota, Erlanger, KY: Could
field trail TMH units were designed and tested for two typical the nano-membrane technology be adapted to large industrial
mid-efficiency home furnaces in two occupied single family indirect-fire HVAC systems?
homes, which proved the TMH can provide a comfortable Wang: Yes.
530
ASHRAE Transactions
SA-12-019 (RP-1339)
Selection of Desiccant Equipment at Altitude
Nelson Fumo, PhD Pedro J. Mago, PhD Eric Kozubal

Associate Member ASHRAE Member ASHRAE Associate Member ASHRAE
ABSTRACT turing, food storage and packaging, hospitals, ice rinks, librar-
ASHRAE Standard 139 provides parameters which are ies, and residences among others. These applications
used in the desiccant industry to calculate the performance of generally involve thermal comfort, but also include places
a desiccant dehumidifier. These performance parameters can where moisture in the air can lead to safety and equipment/
be obtained from any manufacturer by means of performance process efficiency issues. The correct selection of equipment
curves or selection software. However, solid desiccant dehu- is essential to the proper dehumidification for the specific
midifiers are generally rated at standard (sea level) conditions. application and location. Some companies provide selection
Although some manufacturers provide a means for estimating software to aide customers. Other companies provide techni-
the performance at altitude based on standard conditions, cal resources on different scenarios where dehumidification
there is no methodology that is accepted by all. The scope of may be necessary, with examples for each scenario on how to
this paper involves investigating how performance parameters size the necessary equipment. All companies will provide the
such as the moisture removal capacity (MRC), regeneration ratings of their equipment based on specific design conditions.
specific heat input (RSHI), process outlet temperature, and There are several design parameters of a solid desiccant
pressure drop through the wheel are affected by altitude and wheel that affect its performance. These parameters include
developing a ‘simple methodology’ that can be applied to any process and regeneration air velocity (mass flow rate), rota-
desiccant wheel to estimate these performance parameters. tional speed of the desiccant wheel, process and regeneration
For development and validation of the methodology, data was air temperature, process and regeneration air relative humid-
obtained by testing a test cassette at Mississippi State Univer- ity, and bed length. Out of these parameters, only some are
sity and the National Renewable Energy Laboratory. affected by altitude due to the variation in atmospheric pres-
sure. As a comment, in the industry it is common to find the
INTRODUCTION use of the word reactivation instead of regeneration.
A desiccant dehumidifier can either be composed of a The National Renewable Energy Laboratory (NREL) has
solid or a liquid desiccant. Liquid desiccants work by absorp- performed tests to investigate the effect of altitude on the
tion, which changes the state of the desiccant as it absorbs performance of a solid desiccant system. The tests concluded
moisture. A solid desiccant uses the process of adsorption, that if the number of transfer units (NTU) and inlet humidity
which does not alter the chemical state of the desiccant, but ratios were kept constant between altitude and sea level, the
merely adds additional mass to the desiccant from the water difference in performance between the two locations will be
vapor that is adsorbed. Desiccant dehumidification can be negligible. However, these tests were performed using
used for a wide variety of applications, including breweries, mechanical means to set the pressures. Therefore, empirical
candy manufacturing, corrosion control, electronics manufac- data at two actual altitudes is required in order to fully under-
Nelson Fumo is an assistant professor in the Department of Mechanical Engineering at The University of Texas at Tyler, Tyler, TX. Pedro J.
Mago is an associate professor in the Mechanical Engineering Department at Mississippi State University, Mississippi State, MS. Eric
Kozubal is a senior mechanical engineer with Electricity, Resources & Building Systems Integration, National Renewable Energy Laboratory,
Golden, CO.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
©2012 ASHRAE
531
stand how altitude affects the moisture capacity of the desic- et al. [4]. They stated that for a constant inlet temperature and
cant. humidity ratio, the pressure will decrease the relative humidity
For this reason, the main objective of the research conducted of the process and regeneration airflows. By examining a sorp-
was to perform tests on a solid desiccant dehumidifier in order to tion isotherm for any arbitrary solid desiccant, it can be seen
develop a ‘simple methodology’ or procedure that designers, that the adsorbed water content of the desiccant will decrease
engineers/architects, or end-users can use to estimate the mois- with decreasing relative humidity [1, 5], decreasing the mass
ture removal capacity (MRC), regeneration specific heat input transfer rate.
(RSHI), process outlet temperature, and pressure drop through By applying these concepts to the two main figures of
the wheel of a desiccant dehumidifier at any altitude using the merit for solid desiccant dehumidifiers, the MRC and RSHI,
manufacturers' published performance data for sea level. Data it can be seen that for a fixed inlet temperature, inlet humidity
for this research was obtained by testing a test cassette at two ratio, and mass flow rate, the MRC will decline with altitude
different locations: Mississippi State University (MSU), Missis- due to the change in relative humidity (vapor pressure). The
sippi State, Mississippi (elevation » 100 m or 330 feet) and RSHI will change inversely proportional to the MRC due to its
NREL, Golden, Colorado (elevation » 1,783 m or 5,850 feet). definition, since the energy used for regeneration heat is
The testing conditions were decided based on information gath- assumed to remain constant due to the nature of the selection
ered from previous research, manufactures experience, informa- process, which assumes constant inlet temperatures and mass
tion regarding the concept behind the adsorption process of a flow rates. Therefore, with the change in the specific heat of air
desiccant, and to keep the number of tests in a reasonable being negligible below 3,000 m (10,000 feet), and the method
amount. of regeneration heating being the same, the energy consump-
tion of the regeneration will not change. Research by Slayzak
LITERATURE REVIEW et al. [6] suggested that effect of atmospheric pressure on the
MRC due to altitude can be considered negligible when using
In the field of desiccant dehumidification at altitude, very
constant inlet humidity ratios and mass flow rates. This
few studies have been performed empirically. Pesaran and
discrepancy between the previously discussed studies could
Heiden [1] stated that the performance of a desiccant dehu-
be due to the nature of the experiment, which varied the pres-
midifier is related to the number of heat transfer units (NTUh)
sure in the test loop by keeping it sealed from the atmosphere
and the number of mass transfer units (NTUm). NTUh is a
and mechanically altering the pressure.
function of geometric properties, mass flow rate, Nusselt
number, and the specific heat of air. For a specific unit the Another important design parameter that can be affected
geometric properties will remain constant with altitude. The by altitude is the pressure drop through the wheel. The fan
change in specific heat of air can be considered negligible laws state that for a constant mass flow rate and wheel face
below 10,000 feet [1, 2], and the Nusselt number, which is a area, the face velocity of the air must increase with decreasing
function of the heat transfer coefficient, is independent of air density. Harshe et al. [3] presented a general equation for
density for fully developed laminar flow, and therefore pressure drop in their mathematical models that is a function
constant with altitude [1]. Therefore, in order to keep NTUh of velocity which is similar in form to the pressure drop equa-
constant, the mass flow rate must also remain a constant. Pesa- tion used in various other applications relating to the flow of
ran and Heiden [1] also concluded that, similar to the change air through a channel [1, 6, 7, 8], differing in the fact that some
in viscosity due to altitude, NTUm is not affected by pressure. studies neglect entrance and exit losses. Also, experimental
In addition they stated that: “The moisture capacity [of a desic- data by Tretiak and Abdallah [9] showed that pressure drop
cant] depends on [total] pressure at a fixed humidity ratio.” through a packed bed is a second order polynomial as a func-
Their study showed that the moisture capacity of a desiccant tion of velocity and Reynolds number. In pressure drop equa-
is inversely proportional to the total pressure, meaning that, for tions the friction factor is considered as the ratio of a parameter
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
a fixed humidity ratio, the moisture capacity of the desiccant that depends on geometric properties of the channel and Reyn-
decreases with increasing altitude. This statement is reflected olds number [10]. Belady [8] points out that many of the vari-
in models by Harshe et al. [3], which show that the water ables in pressure drop equations will remain constant with
content of a desiccant is a function of wheel speed, geometric altitude (e.g. terms related to geometry).
properties of the desiccant, and air humidity ratio (which all
UNCERTAINTY ANALYSIS
stay constant with altitude), as well as a mass transfer coeffi-
cient and the equilibrium humidity at the surface of the desic- The uncertainty associated with the measured variables
cant. They discussed that the equilibrium humidity at the has been considered through the accuracy of the measurement
surface of the desiccant was a function of humidity ratio, and systems. This is particularly the case for the inlet conditions.
therefore declines with altitude. This shows that, for a given However, for other parameters of interest, such as the pressure
mass flow rate and inlet humidity ratio, the outlet humidity drop through the wheel (process and regeneration) and the
ratio will increase with altitude, decreasing the grain depres- process outlet temperature, additional analysis is required
sion and in turn decreasing the MRC. This decrease in the because their results are a function of other variables. This is
moisture capacity is also shown in numerical models by Ruivo also the case for the most important variable to be considered,
532
ASHRAE Transactions
the MRC, which is computed as a function of other variables. GPI = Humidity ratio at the process inlet, [Gr/lbm]
In this paper, the Taylor Series Method for propagation of GPO = Humidity ratio at the process outlet, [Gr/lbm]
uncertainties was applied to obtain the equations describing Using Equation (2) along with the Taylor Series Method
how the uncertainties of the independent variables define the for propagation of uncertainties [11-13], the uncertainty for
uncertainties of the resulting variables. In this section the the MRC can be estimated as
equations derived from the application of the Taylor Series
Method are presented. These equations were implemented in MRC   MRC 2 1/2
U MRC =  ---------------
- U · + -------------------- U GPP (3)
a MathCad code and the actual data acquired from the tests  m· m   GPP 
were used to estimate the uncertainties. The magnitude of the
uncertainties is presented in the figures in the Test Results where:
section through error bars, and in table format in the appen- UMRC = Uncertainty for MRC
dixes. U· = Uncertainty for mass flow rate
m
UGPP= Uncertainty for grain depression
Uncertainty In MRC
The main figure of merit used to determine the perfor- Methods for Calculating Humidity Ratio
mance of a desiccant dehumidifier is the moisture removal
There are three accepted methods for calculating the
capacity. ASHRAE Standard 139 defines the MRC as
humidity ratio, each using different measured variables [11,
MRC =  Scmh P  1.2 kg/m 3   G i – G o    1000 (1a, SI) 12]. They are:
where: • Dry-bulb temperature and relative humidity (Relative

MRC = Moisture removal capacity, [kg/h] Humidity Method)
• Dry-bulb temperature and wet-bulb temperature (Wet-
ScmhP = Standard volumetric flow rate of process air, [cmh]
Bulb Temperature Method)
Gi = Inlet humidity ratio, [g/kg] • Dry-bulb temperature and dew-point temperature (Dew-
Go = Outlet humidity ratio, [g/kg] Point Temperature Method)
Relative Humidity Method. This method uses the rela-

MRC =  Scmh P   60 min/h  0.075 lb/ft 3    G i – G o    7000 tive humidity, atmospheric pressure, and saturation vapor
(1b, I-P) pressure (function of dry-bulb temperature) to calculate the
humidity ratio using Equation (4) [14, Chapter 1]:
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
where:
MRC = moisture removal capacity, [lb/h] p ws
w = 0.621945 ---------------------------- (4)
p atm – p ws
ScfmP = Standard volumetric flow rate of process air, [cfm]
Gi = Inlet humidity ratio, [Gr/lbm] where:
Go = Outlet humidity ratio, [Gr/lbm] w = Humidity ratio, [kg/kg] [lbm/lbm]
Equation (1) can also be written as  = Relative humidity
pws = Saturation pressure of water vapor at a given
· ·
MRC =  m  G PI – G PO    1000 = m GPP  1000 (2a, SI) temperature, [kPa] [psia]
patm = Atmospheric pressure, [kPa] [psia]
where:
· The uncertainty in humidity ratio for the relative humidity
m = Mass flow rate of the process air stream, [kg/h]
GPI = Humidity ratio at the process inlet, [g/kg] method can be estimated as:
GPO = Humidity ratio at the process outlet, [g/kg] w 2 w w 1/2
U w =  ------U  +  ----------- U p  +  -------------U p 
2 2
(5)
GPP= Grain depression through the process side of the  p ws ws p atm atm
desiccant wheel, defined as GPI – GPO, [g/kg]
where:
60 · 60 ·
MRC = ------------ m  G PI – G PO  = ------------ m GPP (2b, I-P) Uw = Uncertainty in humidity ratio
7000 7000
U = Uncertainty in relative humidity
where: Up = Uncertainty in saturation pressure of water vapor
· ws
m = Mass flow rate of the process air stream, [lbm/ Up = Uncertainty in atmospheric pressure
min] atm
GPP= Grain depression through the process side of the Wet-Bulb Temperature Method. This method takes
desiccant wheel, defined as GPI – GPO, [Gr/lbm] humidity ratio as a function of dry- and wet-bulb temperatures,
©2012 ASHRAE
533
and the saturation humidity ratio. The saturation humidity Next the humidity ratio is calculated using Equation (4) or
ratio is calculated using Equation (4), taking  = 1 and the Equation (6), depending on the method used. With the atmo-
saturation pressure of water vapor using the wet-bulb temper- spheric pressure, dry-bulb temperature, and humidity ratio
ature. This can be seen in Equation (6) [14, Chapter 1]: known, the density of the moist air can now be calculated as
[14, Chapter 1]:
 2501 – 2.326T wb w s – 1.006  T db – T wb 
w = ----------------------------------------------------------------------------------------------------
- (6a, SI) p atm 1 + w
2501 + 1.86T db – 4.186T wb
 = ------------ --------------------------- (9)
R da T 1 + 1.608w
 1093 – 0.556T wb w s – 0.240  T db – T wb 
w = ----------------------------------------------------------------------------------------------------
- (6b, I-P) where:
1093 + 0.44T db – T wb
 = Density of moist air, [kg/m3] [lbm/ft3]
where: Rda = Gas constant for dry air, [287.042 J/(kgda×K]
Twb = Wet-bulb temperature, [°C] [°F] [53.350 ftlbf/lbmda°R]
ws = Saturation humidity ratio, [kg/kg] [lbm/lbm]
The uncertainty for density can be estimated as:
Tdb = Dry-bulb temperature, [°C] [°F]
w w 2 w 2 1/2
U  =  -------------U p  +  ------U w +  ------U T
2
The uncertainty for humidity ratio for the wet-bulb  p atm atm  w   T 
(10)
temperature method can be estimated as:
where:
w w w
U w =  ------------ U T  +  --------- U w  +  ----------- U T 
2 2 2 1/2
(7) U = Uncertainty in density
T wb wb w s s T db db
UT = Uncertainty in absolute temperature
where:
It should be noted that it is assumed that there is no uncer-
UT = Uncertainty in wet-bulb temperature
wb tainty involved with the gas constant.
Uw
s
= Uncertainty in saturation humidity ratio By measuring the volumetric flow rate, the mass flow rate
UT
db
= Uncertainty in dry-bulb temperature can be calculated using the previously calculated density. The
mass flow rate is found using Equation (11):
Dew-Point Temperature Method. This method uses the
· ·
same equations as the wet-bulb temperature method, but m = V (11)
replaces all measured temperatures (dry- and wet-bulb) with
the dew-point temperature. where:
·
m = Mass flow rate, [kg/h] [lbm/min]
Equations Used to Calculate MRC ·
V = Volumetric flow rate, [cmh] [cfm]
From the three methods previously outlined, MSU took
the relative humidity approach while NREL took the dew- The uncertainty for the mass flow rate is estimated using
point temperature approach. The following equations show the Taylor Series Method as:
how the measured variables are used to calculate the depen- w w  2 2 1/2
U · =  ------U  +  ------U (12)
dent variables needed to find the MRC. m     · V· 
V
For both methods the saturation pressure of water vapor
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
must be found using Equation (8) [14, Chapter 1]: where:
U· = Uncertainty in volumetric flow rate
ln p ws = C 8  T + C 9 + C 10 T + C 11 T 2 + C 12 T 3 + C 13 ln  T  (8) V
Once the mass flow rate and the humidity ratios are
where
known, Equation (2) can be used to calculate the MRC. Using
SI Units I-P Units the instrument accuracies for each location, the Taylor Series
Method can be applied to each of these equations to find the
C8 = –5.8000 220 6 E+03 –1.044 039 7 E+04
uncertainty involved with the MRC in Equation (3).
C9 = 1.391 499 3 E+00 –1.129 465 0 E+01
C10 = –4.864 023 9 E–02 –2.702 235 5 E–02 Uncertainty In Process Outlet Temperature
C11 = 4.176 476 8 E–05 1.289 036 0 E–05 The process outlet temperature cannot be exactly defined
C12 = –1.445 509 3 E–08 –2.478 068 1 E–09 because the temperature is the result of heat and mass transfer
C13 = 6.545 967 3 E+00 6.545 967 3 E+00 processes that are not known. However, if it is assumed that the
temperature will depend strongly on the mass transfer, the
Pws = Saturation pressure, [Pa] [psia] latent heat released to the air due to the dehumidification of the
T = Absolute temperature, [K] [°R] air stream can be defined using Equation (13).
534
ASHRAE Transactions
· GPP Uncertainty In Pressure Drop Through The Wheel
Q G = m ---------------- h fg (13a, SI)
1000
Pressure drop equations proposed in other studies
· GPP accounts for geometric parameters and fluid properties that
Q G = m ---------------- h fg (13b, I-P) will make the uncertainty analysis quite laborious in order to
7000
consider the exact uncertainties for pressure drop in each test.
where: Therefore, in the rationality for pressure drop through the
QG = Heat produced by dehumidification, [kJ/h] [BTU/ wheel (Section 7.4), Equation (17) is proposed as a general
min] equation representing the pressure drop equations suggested
hfg = Specific enthalpy of evaporation, [kJ/kg] [BTU/ in other studies.
lbm]
P = C l V + C k t V 2 (17)
Neglecting the uncertainty for hfg, the uncertainty for the
heat rate can be estimated as: where:
Cl = Coefficient related to laminar flow through the
Q G  2  Q G 2 1/2
channel
UQ =  -------------
- U · + ------------------- U GPP (14)
G  m· m  GPP  Ck,t = Coefficient related to entrance and exit losses and
turbulent flow
where:
Although geometric parameters in these coefficients are
UQ = Uncertainty in heat rate
G constants, fluid parameters change with the air properties for
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Once the uncertainty for the heat rate is known, Equation each test. However, analysis of some cases showed that the
(15) can be used to find the estimated process outlet temper- uncertainty for the pressure drop can be conservatively
ature. approximated as equal to the percentage uncertainty of the
mass flow rate.
Q G
T PO = ----------
· - + T PI (15)
m cp EXPERIMENTAL SETUPS
Experimental setups at MSU and NREL were used to
where: accommodate a test cassette. The test cassette was assembled
TPO = Process outlet temperature, [°C] [°F] at NREL with 50.8 mm (2") Isocyanurate foam for good heat
cp = Specific heat of air, [kJ/kg°C] [BTU/lbm°F] transfer retention and reinforced with aluminum angle. The
TPI = Process inlet temperature, [°C] [°F] test cassette contained the 300 mm (11”) diameter and 100 mm
(4”) depth solid desiccant wheel made of silica gel set for a
The uncertainty for the process outlet temperature can be geometric configuration of 50/50 split and rotation at 14 rph.
estimated using the Taylor Series Method as: The estimated effective area for the process and regeneration
sides is 3.8555E-2 m2 (0.415 ft2). Both experimental setups
T PO 2 T PO  2  T PO 2 1/2
=  -------------- U Q  +  ------------ 
UT - U + ------------
- U (16) were made to comply with the requirements/specifications of
m m  T PI PI
·
PO Q G G · T
ASHRAE Standard 139 [15]. The accuracies of the measure-
ment systems at both locations, as well as the accuracies
where: required by the standard, are shown in Table 1. It can be seen
UT
PO
= Uncertainty in process outlet temperature from this table that all of the instrument accuracies fall within
UT
PI
= Uncertainty in process inlet temperature the required limits of the ASHRAE Standard 139.
Table 1. Measurement System Accuracies
Measurement MSU NREL ASHRAE
Mass Flow ±2.4% of reading ±2% of reading ±3% of reading

±0.18°F (Humidity Probe)
Dry-Bulb Temperaturea ±0.3°F ±0.5°F
±0.27°F (RTDa)
Dew-Point Temperature — ±0.3°F ±0.5°F
Relative Humidity ±1% RH — ±3%
Pressure Measurements ±0.5% FS +/– 0.5% of FS ±1%
a. RTD used for high temperature at regeneration inlet.
©2012 ASHRAE
535
The requirements for the testing method state that the where:
tests must be performed under steady-state conditions, which patm = Atmospheric pressure at altitude, [kPa] [psia]
have been held for at least fifteen minutes. The standard also
states that the mass flow rates must be calculated using the air Z = Altitude, [m] [ft]
flow at the flow meter, and also the density calculated at the Since the actual atmospheric pressure was used to calcu-
flow meter by the temperature and humidity probes. late the humidity ratios during the tests, the relative humidity
One important consideration that was taken during the (MSU) and dew-point temperature (NREL) were adjusted
experiments was the pressure differential between the differ- accordingly so that the resulting humidity ratios were as
ent test plenum chambers. In order to maintain the required defined in the test conditions.
balances, the pressure differential between the Process Outlet I-P units were preferred at MSU, while SI units were
and Regeneration Inlet chambers was set to zero. This assured preferred at NREL. Results are presented in both unit systems
that no mixing of the air occurred between the process out and through the paper but figures are given in I-P units since it is
regeneration inlet air streams, leading to false readings of the the common unit system for the industry. For the tests, the
actual process outlet and regeneration inlet conditions and mass balance, moisture balance, and enthalpy balance were
flow rates. Also, the pressure differential between the Process computed showing agreement with the limits recommended
Inlet and the atmosphere was set to be equal-but-opposite of by ASHRAE Standard 139 [15]. Regarding the uncertainty,
the pressure differential between the Regeneration Outlet and Appendix A presents the uncertainties associated to the exper-
the atmosphere. This ensured that the system was properly iments computed as outlined in Section 3.
balanced, allowing for minimal leakage between the different
Tables 2 and 3 outline the results from MSU for the
chambers of the desiccant apparatus and the atmosphere.
process and regeneration air streams, respectively. Tables 4
and 5 outline the results from NREL for the process and regen-
TEST RESULTS
eration air streams, respectively. For Tables 2 and 4 it is impor-
Test conditions for the experiments were decided based tant to notice that the actual process inlet mass flow rate
on information gathered from manufactures as to how they corresponds to the mass flow rate reported in the column
have handled the selection process at altitude in the past, infor- ‘Process Out’ and not to the mass flow rate reported in the
mation regarding the concept behind the adsorption process of ‘Process In’. As mentioned in Section 4, this can be explained
a desiccant, and previous research. Tests were performed for since the difference in pressure between the process outlet
three different mass flow rates, defining the identification plenum chamber and regeneration inlet plenum chamber was
letter of the test (A, B, and C). The flow rates were decided set as zero. Therefore, for the process side, the actual mass
based on common face standard velocities used in the industry, flow rate crossing the desiccant wheel is the mass flow rate
and capacities of the testing apparatuses. For all tests, the face measured at the outlet.
standard velocity for the regeneration side was set to be half of
the standard velocity on the process side. The flow rate for ANALYSIS OF RESULTS
Tests C was set so that there would be a set of matching actual
face velocities between both locations (MSU-C matches This section present an analysis of the results obtained
actual face velocity at NREL-B, NREL-C matches actual face from the experiments performed at MSU and NREL. The
velocity at MSU-A). For the process side, five combinations of results focus on the MRC, the process outlet temperature, and
temperature and humidity ratios were decided to cover general the pressure drop through the wheel.
cases for common operation of dehumidifiers and to identify
trends based on temperature and humidity ratio variation. Moisture Removal Capacity (MRC)
These combinations correspond to the number associated to Since MRC is a function of mass flow rate and grain
the identification of the test (1, 2, 3, 4, and 5). In order to elim- depression, and because mass flow rate was a controlled vari-
inate the regeneration inlet conditions as a variable in the tests, able, it is important to illustrate the behavior of grain depres-
for all the tests the regeneration temperature and relative sion in order to explain the behavior of the MRC. Figure 1
humidity before the heater were set as 32.2 °C (90°F) and 60% shows the grain depression as a function of mass flow rate for
with a regeneration temperature of 93.3 °C (200°F). The rela- Tests 1 (1A, 1B, and 1C). The results presented in this figure
tive humidity of 60% was fixed at both locations since it was are for three different mass flow rates but for the same process
considered as a measured variable defining the driving force inlet temperature and humidity ratio. The other test conditions,
for mass transfer. To define the test conditions at both loca- Tests 2, 3, 4, an 5, present similar trends to the one obtained for
tions (MSU and NREL), the atmospheric pressure at each Test 1. Figure 1 illustrates that the grain depression decreases
location was calculated using Equation (18) [14, Chapter 1]: with mass flow rate. This can be explained since, for lower
velocities through the wheel, the air is dried more deeply
p atm = 101.325  1 – 2.25577  10 –5 Z  5.2559 (18a, SI) because the air is in contact with the desiccant longer and can
therefore come closer to equilibrium [16, Chapter 23], i.e. to
p atm = 14.696  1 – 6.8754  10 – 6 Z  5.2559 (18b, I-P) the point of maximum possible dehumidification.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
536
ASHRAE Transactions
Table 2a. MSU Test Results for Process Air Stream (SI Units)
Process In Process Out Process Atmospheric

Test Mass Tdb W Mass Tdb W P Pressure
[kg/s] [°C] [kg/kg] [kg/s] [°C] [kg/kg] [cm WC] [kPa]
1A 0.158 23.9 0.0113 0.142 39.7 0.0066 1.438 102.0
2A 0.156 23.8 0.0131 0.141 41.0 0.0078 1.450 100.6
3A 0.155 23.9 0.0150 0.141 42.2 0.0092 1.466 100.6
4A 0.157 18.4 0.0130 0.141 37.9 0.0070 1.458 102.0
5A 0.158 12.9 0.0091 0.142 30.8 0.0041 1.389 101.3
1B 0.105 23.9 0.0112 0.095 42.2 0.0054 0.970 101.3
2B 0.104 23.9 0.0131 0.094 43.9 0.0066 0.955 101.3
3B 0.106 23.9 0.0152 0.095 45.0 0.0082 0.991 101.3
4B 0.105 18.3 0.0131 0.094 42.0 0.0059 0.965 99.9
5B 0.104 12.9 0.0090 0.094 34.5 0.0030 0.950 102.0
1C 0.128 23.9 0.0112 0.115 41.4 0.0060 1.186 100.6
2C 0.128 23.9 0.0131 0.115 42.4 0.0072 1.179 100.6
3C 0.128 23.9 0.0151 0.115 43.7 0.0085 1.227 101.3
4C 0.128 18.4 0.0131 0.115 39.9 0.0064 1.179 101.3
5C 0.127 12.9 0.0090 0.115 32.7 0.0034 1.158 101.3
Table 2b. MSU Test Results for Process Air Stream (I-P Units)
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
[lbm/min] [°F] [Gr/lbm] [lbm/min] [°F] [Gr/lbm] [in WC] [psia]
1A 20.9 75.0 79 18.7 103.4 46 0.566 14.8
2A 20.7 74.9 92 18.6 105.8 55 0.571 14.6
3A 20.5 75.1 105 18.6 107.9 64 0.577 14.6
4A 20.7 65.1 91 18.6 100.3 49 0.574 14.8
5A 20.9 55.1 63 18.7 87.4 29 0.547 14.7
1B 13.9 75.0 78 12.5 107.9 38 0.382 14.7
2B 13.8 75.0 92 12.4 111.0 46 0.376 14.7
3B 14.1 75.0 107 12.5 112.9 57 0.390 14.7
4B 13.8 65.0 91 12.5 107.6 41 0.380 14.5
5B 13.8 55.2 63 12.4 94.0 21 0.374 14.8
1C 16.9 75.1 79 15.2 106.4 42 0.467 14.6
2C 16.9 75.0 92 15.2 108.3 51 0.464 14.6
3C 16.9 74.9 106 15.2 110.6 59 0.483 14.7
4C 17.0 65.0 92 15.2 103.9 45 0.464 14.7
5C 16.8 55.2 63 15.3 90.8 24 0.456 14.7
©2012 ASHRAE
537
Table 3a. MSU Test Results for Regeneration Air Stream (SI Units)
Regeneration In Regeneration Out Regen

Test Mass Tdb W Mass Tdb W P
[kg/s] [°C] [kg/kg] [kg/s] [°C] [kg/kg] [cm WC]
1A 0.071 93.2 0.0186 0.082 57.0 0.0260 0.975
2A 0.070 93.4 0.0184 0.081 55.2 0.0272 0.973
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
3A 0.071 93.2 0.0182 0.083 53.4 0.0283 0.983
4A 0.071 93.0 0.0184 0.083 52.1 0.0282 0.960
5A 0.071 93.2 0.0186 0.083 54.3 0.0262 0.980
1B 0.048 93.3 0.0183 0.057 50.9 0.0264 0.663
2B 0.047 92.9 0.0183 0.056 49.3 0.0276 0.643
3B 0.047 93.2 0.0182 0.058 47.4 0.0292 0.653
4B 0.047 93.4 0.0184 0.056 45.0 0.0289 0.648
5B 0.047 93.7 0.0183 0.056 45.8 0.0265 0.650
1C 0.057 93.3 0.0183 0.068 55.3 0.0247 0.795
2C 0.057 93.2 0.0183 0.068 53.6 0.0273 0.798
3C 0.057 93.5 0.0183 0.068 51.2 0.0287 0.792
4C 0.057 93.3 0.0182 0.067 47.6 0.0288 0.782
5C 0.071 93.2 0.0186 0.082 57.0 0.0260 0.790
Table 3b. MSU Test Results for Regeneration Air Stream (I-P Units)

[lbm/min] [°F] [Gr/lbm] [lbm/min] [°F] [Gr/lbm] [in WC]
1A 9.3 199.7 130 10.9 134.7 182 0.384
2A 9.3 200.1 129 10.8 131.4 190 0.383
3A 9.4 199.8 128 10.9 128.1 198 0.387
4A 9.4 199.3 129 11.0 125.8 198 0.378
5A 9.3 199.8 130 10.9 129.7 184 0.386
1B 6.3 199.9 128 7.6 123.6 185 0.261
2B 6.2 199.2 128 7.4 120.8 193 0.253
3B 6.3 199.8 128 7.6 117.2 205 0.257
4B 6.2 200.1 128 7.4 113.0 202 0.255
5B 6.3 200.6 128 7.5 114.4 186 0.256
1C 7.6 200.0 128 9.0 131.5 173 0.313
2C 7.6 199.8 128 9.0 128.5 191 0.314
3C 7.6 200.3 128 9.0 124.2 201 0.312
4C 7.6 200.0 128 8.9 117.8 202 0.308
5C 7.6 200.0 128 9.0 120.0 187 0.311
538
ASHRAE Transactions
Table 4a. NREL Test Results for Process Air Stream (SI Units)

[kg/s] [°C] [kg/kg] [kg/s] [°C] [kg/kg] [cm WC] [kPa]
1A 0.158 23.9 0.0113 0.141 39.5 0.0068 1.753 82.7
2A 0.157 23.9 0.0132 0.141 40.8 0.0081 1.803 82.7
3A 0.158 23.9 0.0152 0.141 42.0 0.0096 1.803 82.7
4A 0.158 18.3 0.0132 0.141 37.4 0.0074 1.702 82.7
5A 0.158 12.8 0.0091 0.141 30.5 0.0042 1.676 81.3
1B 0.105 23.9 0.0113 0.094 43.6 0.0058 1.194 82.7
2B 0.105 23.9 0.0131 0.094 45.1 0.0070 1.219 82.7
3B 0.104 23.9 0.0152 0.094 46.4 0.0084 1.219 82.7
4B 0.105 18.3 0.0132 0.094 41.7 0.0063 1.143 82.7
5B 0.105 12.8 0.0093 0.094 35.0 0.0033 1.041 81.3
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
1C 0.129 23.9 0.0113 0.116 41.4 0.0063 1.473 82.7

2C 0.129 23.9 0.0131 0.116 42.8 0.0076 1.473 82.7
3C 0.129 23.9 0.0153 0.116 44.4 0.0090 1.372 81.3
4C 0.129 18.3 0.0132 0.116 39.6 0.0069 1.422 82.7
5C 0.129 12.8 0.0092 0.116 32.7 0.0037 1.321 81.3
Table 4b. NREL Test Results for Process Air Stream (I-P Units)

[lbm/min] [°F] [Gr/lbm] [lbm/min] [°F] [Gr/lbm] [in WC] [psia]
1A 20.8 75.0 79 18.7 103.1 48 0.69 12.0
2A 20.8 75.0 92 18.7 105.5 57 0.71 12.0
3A 20.8 75.0 106 18.7 107.6 67 0.71 12.0
4A 20.8 65.0 92 18.7 99.3 52 0.67 12.0
5A 20.8 55.0 64 18.7 87.0 29 0.66 11.8
1B 13.9 75.0 79 12.5 110.5 41 0.47 12.0
2B 13.8 75.0 92 12.4 113.2 49 0.48 12.0
3B 13.8 75.0 106 12.4 115.6 59 0.48 12.0
4B 13.9 65.0 92 12.5 107.1 44 0.45 12.0
5B 13.8 55.1 65 12.4 94.9 23 0.41 11.8
1C 17.1 75.0 79 15.3 106.6 44 0.58 12.0
2C 17.1 75.0 92 15.3 109.0 53 0.58 12.0
3C 17.0 75.0 107 15.3 111.9 63 0.54 11.8
4C 17.0 65.0 92 15.3 103.3 48 0.56 12.0
5C 17.0 55.0 64 15.3 90.8 26 0.52 11.8
©2012 ASHRAE
539
Table 5a. NREL Test Results for Regeneration Air Stream (SI Units)

[kg/s] [°C] [kg/kg] [kg/s] [°C] [kg/kg] [cm WC]
1A 0.071 93.3 0.0228 0.082 60.0 0.0284 1.194
2A 0.071 93.3 0.0227 0.082 58.0 0.0288 1.168
3A 0.071 93.3 0.0227 0.082 55.9 0.0298 1.168
4A 0.071 93.4 0.0226 0.082 53.9 0.0302 1.168
5A 0.071 93.3 0.0227 0.082 55.9 0.0286 1.168
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
1B 0.047 93.3 0.0228 0.057 53.1 0.0292 0.787

2B 0.047 93.3 0.0228 0.057 50.9 0.0303 0.762
3B 0.047 93.3 0.0227 0.057 48.7 0.0315 0.762
4B 0.047 93.3 0.0227 0.056 46.6 0.0316 0.737
5B 0.047 93.3 0.0227 0.056 48.6 0.0298 0.737
1C 0.058 93.3 0.0227 0.069 56.7 0.0285 0.965
2C 0.058 93.3 0.0226 0.069 54.5 0.0295 0.965
3C 0.058 93.3 0.0227 0.068 52.3 0.0309 0.940
4C 0.058 93.3 0.0227 0.068 50.3 0.0309 0.940
5C 0.058 93.3 0.0227 0.068 52.4 0.0293 0.940
Table 5b. NREL Test Results for Regeneration Air Stream (I-P Units)

[lbm/min] [°F] [Gr/lbm] [lbm/min] [°F] [Gr/lbm] [in WC]
1A 9.3 200.0 160 10.9 140.0 199 0.47
2A 9.3 200.0 159 10.9 136.4 202 0.46
3A 9.3 200.0 159 10.9 132.7 209 0.46
4A 9.3 200.0 158 10.9 129.1 211 0.46
5A 9.3 200.0 159 10.9 132.6 200 0.46
1B 6.2 200.0 159 7.5 127.6 204 0.31
2B 6.2 200.0 159 7.5 123.6 212 0.30
3B 6.2 200.0 159 7.5 119.7 220 0.30
4B 6.2 200.0 159 7.4 115.9 221 0.29
5B 6.2 200.0 159 7.4 119.5 209 0.29
1C 7.7 200.0 159 9.1 134.1 200 0.38
2C 7.7 200.0 158 9.1 130.2 206 0.38
3C 7.7 200.0 159 9.0 126.1 217 0.37
4C 7.7 200.0 159 9.0 122.6 216 0.37
5C 7.7 200.0 159 9.0 126.3 205 0.37
540
ASHRAE Transactions
Figure 1 Grain depression as a function of mass flow rate for results of tests 1A, 1B, and 1C at MSU and NREL.
To study the behavior of the grain depression as a function MRC increases with the humidity ratio. This behavior is
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
of humidity ratio, tests having the same mass flow rate (tests explained by the fact that if the dry-bulb temperature is kept
A, B, or C) and dry-bulb temperature of 23.9 °C (75°F) were constant, a higher humidity means a higher partial pressure of
considered. Figure 2 shows the grain depression as a function water vapor, which is the driving force for desiccant dehumid-
of process inlet humidity ratio for test A (1A, 2A, and 3A). For ification.
tests B (1B, 2B, and 3B) and C (1C, 2C, and 3C), the behavior The comparison of the MRC at both locations is presented
and trend are similar to the ones obtained for test A. From in Table 6. It can be seen that the variation is about 5% for each
Figure 2, it can be observed that for the same process mass test, with the exception of Tests 5A, 5B, and 5C, which have
flow rate and dry-bulb temperature, the grain depression a lower variation. For the same mass flow rate (Tests A, B, or
increases with inlet humidity ratio. This can be explained C), the variations are due to the effect of the inlet temperature
since, keeping all other inlet conditions constant, a higher and humidity ratio on the heat and mass transfer processes. It
humidity ratio implies a higher partial pressure differential should be kept in mind that the effect of humidity ratio is actu-
between the air and the desiccant surface, leading to a larger ally due to the partial pressure of water vapor. If the mass flow
capacity for dehumidification. rate, dry-bulb temperature, and humidity ratio are kept
Figure 3 shows the MRC as a function of mass flow rate constant when comparing desiccant performance at altitude,
for Tests 1 (1A, 1B, and 1C). For the other test conditions, by analyzing the equation for the MRC, it can be seen that the
Tests 2, 3, 4, and 5, similar trends are obtained when the data only variable producing a change in the MRC is the outlet
is plotted. Figure 3 illustrates that the MRC increases with the humidity ratio, Go. Therefore, the difference in inlet partial
mass flow rate. By analyzing the equation for MRC, it seems pressure of water vapor is defining the variation in Go and
that this behavior should be obvious. However, since the grain therefore in the MRC. To explain the lower variation in MRC
depression decreases with mass flow rate, the results show that of Tests 5A, 5B, and 5C with respect to the other tests, it is
the effect of the increase of mass flow rate on the MRC has a necessary to recall that the water capacity of a given desiccant
larger impact than the decrease of grain depression as a conse- decreases at altitude at a constant humidity ratio. This is the
quence of the increase in mass flow rate. consequence of a lower partial pressure of water vapor (or
Figure 4 illustrates the MRC as a function of process inlet relative humidity) at altitude for the same temperature and
humidity ratio for tests A (1A, 2A, and 3A). For tests B (1B, humidity ratio.
2,B, and 3B) and C (1C, 2C, and 3C), the behavior and trend As mentioned in Section 5, since the relative humidity
is similar to tests A. From Figure 4, it can be noticed that for before the heater was set equal at both locations instead of
the same process mass flow rate and dry-bulb temperature the humidity ratio, the regeneration inlet humidity ratio at MSU
©2012 ASHRAE
541
Figure 2 Grain depression as a function of inlet humidity ratio for results of tests 1A, 2A, and 3A at MSU and NREL.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 3 MRC as a function of mass flow rate for results of tests 1A, 1B, and 1C at MSU and NREL.
542
ASHRAE Transactions
Figure 4 MRC as a function of inlet humidity ratio for results of tests 1A, 2A, and 3A at MSU and NREL.
(18.4 g/kg or 129 Gr/lb) is lower than the value at NREL (22.7 mass transfer process is inconclusive without a mathematical
g/kg or 159 Gr/lb). Based on the results, it can be said that this model. However, it should be noticed that since the variation
condition may cause a slightly greater dehumidification in MRC would have been lower if test at MSU would have
(MRC) at MSU compared with tests that could be performed been conducted at the same NREL’s regeneration humidity
with the same humidity ratio of NREL. By using performance ratio, the variations in process outlet temperature should also
data from two manufacturers, analysis of the tests conducted be lower than the values reported in Table 7. Therefore, for the
at MSU suggests that the variation in MRC between locations same mass flow rate, dry-bulb temperature, and humidity
would be lower than half of the differences reported in Table 6 ratio, the variation in process outlet temperature between loca-
if the tests at MSU would have been conducted with a regen- tions can be considered negligible.
eration humidity ratio of 22.7 g/kg (159 Gr/lbm). Therefore,
for the same mass flow rate, dry-bulb temperature, and humid- Pressure Drop through the Wheel
ity ratio, the variation in MRC between locations can be As mentioned in Section 3.4, the pressure drop is a func-
considered negligible. tion of face velocity. Since mass flow rate is directly related to
face velocity, the pressure drop is presented here as a function
Process Outlet Temperature of mass flow rate. Besides, mass flow rate is a variable to be
The Process Outlet Temperature at MSU and NREL with kept constant when comparing performance at different alti-
the respective temperature variation for each test are presented tudes. Figure 5 and 6 present the pressure drop as a function
in Table 7. Results indicate that the maximum variation for the of mass flow rate for the process streams and regeneration
process outlet temperature is 2.4% with most of the variations streams for tests 1 (1A, 1B, 1C), respectively. For tests 2 (2A,
below 1%. If it is said that the MRC summarizes the perfor- 2B, 2C), 3 (3A, 3B, 3C), 4 (4A, 4B, 4C), and 5 (5A, 5B, 5C)
mance of the desiccant system, then a low variation in MRC the behavior and trend are similar for both air streams (process
between sites suggests that the variation in atmospheric pres- and regeneration respectively).
sure does not have a significant impact on the heat and mass Analysis of pressure drop (Figures 5 and 6) shows a
transfer processes, and therefore on the process temperature constant trend between MSU and NREL: for a given mass
rise. Results suggest that for the different tests and humidity flow rate, pressure drop at NREL is always higher than MSU.
conditions (Tests 1, 2, 3, 4, and 5) there is not a clear trend. This can be explained since for the same mass flow rate, NREL
Therefore the results show that the influence of temperature will always have a higher velocity than MSU to compensate
and humidity ratio in the energy balance from the heat and for the lower density at altitude. This higher velocity implies
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
©2012 ASHRAE
543
Table 6a. Comparison of MRC at MSU and NREL Table 7a. Comparison of Process Outlet Tempera-
(SI Units) ture at MSU and NREL (SI Units)
MSU NREL Variation MSU NREL Variation

Test Test
[kg/hr] [kg/hr] [%] [°C] [°C] [%]
1A 2.43 2.29 0.13 5.4 1A 39.7 39.5 0.2 0.4
2A 2.70 2.57 0.13 4.7 2A 41.0 40.8 0.2 0.4
3A 2.98 2.84 0.14 4.7 3A 42.2 42.0 0.2 0.4
4A 3.06 2.93 0.13 4.4 4A 37.9 37.4 0.6 1.5
5A 2.55 2.50 0.05 1.8 5A 30.8 30.6 0.2 0.7
1B 1.96 1.86 0.10 5.3 1B 42.2 43.6 –1.4 –3.4
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
2B 2.19 2.08 0.11 5.2 2B 43.9 45.1 –1.2 –2.8

3B 2.39 2.29 0.11 4.5 3B 44.9 46.4 –1.5 –3.3
4B 2.43 2.32 0.12 4.7 4B 42.0 41.7 0.3 0.7
5B 2.04 2.02 0.02 0.9 5B 34.4 34.9 –0.5 –1.5
1C 2.17 2.07 0.10 4.6 1C 41.3 41.4 –0.1 –0.3
2C 2.42 2.32 0.10 3.9 2C 42.4 42.8 –0.4 –0.9
3C 2.73 2.60 0.13 5.0 3C 43.7 44.4 –0.7 –1.7
4C 2.77 2.63 0.14 4.9 4C 39.9 39.6 0.3 0.8
5C 2.33 2.28 0.05 2.1 5C 32.7 32.7 0.0 0.0
Table 6b. Comparison of MRC at MSU and NREL Table 7b. Comparison of Process Outlet Tempera-
(I-P Units) ture at MSU and NREL (I-P Units)
MSU NREL Variation MSU NREL Variation

Test Test
[lbm/hr] [lbm/hr] [%] [°F] [°F] [%]
1A 5.35 5.06 0.29 5.4 1A 103.4 103.1 0.3 0.3
2A 5.95 5.67 0.29 4.7 2A 105.8 105.5 0.4 0.3
3A 6.57 6.26 0.31 4.7 3A 107.9 107.6 0.4 0.3
4A 6.75 6.45 0.29 4.4 4A 100.3 99.3 1.0 1.0
5A 5.62 5.52 0.10 1.8 5A 87.4 87.0 0.5 0.6
1B 4.33 4.10 0.23 5.3 1B 107.9 110.5 –2.6 –2.4
2B 4.83 4.58 0.25 5.2 2B 111.0 113.2 –2.2 –2.0
3B 5.28 5.04 0.25 4.5 3B 112.9 115.6 –2.6 –2.3
4B 5.36 5.11 0.26 4.7 4B 107.6 107.1 0.5 0.5
5B 4.49 4.45 0.05 0.9 5B 94.0 94.9 –0.9 –0.9
1C 4.79 4.57 0.22 4.6 1C 106.4 106.6 –0.1 –0.1
2C 5.33 5.12 0.22 3.9 2C 108.3 109.0 –0.7 –0.7
3C 6.03 5.73 0.29 5.0 3C 110.6 111.9 –1.3 –1.2
4C 6.11 5.81 0.30 4.9 4C 103.9 103.3 0.6 0.6
5C 5.14 5.03 0.11 2.1 5C 90.8 90.8 0.0 0.0
544
ASHRAE Transactions
Figure 5 Pressure drop versus mass flow rate of the process stream for tests 1A, 1B, and 1C at MSU and NREL.
Figure 6 Pressure drop versus mass flow rate of the regeneration stream for Tests 1A, 1B, and 1C at MSU and NREL.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
©2012 ASHRAE
545
a higher pressure drop according to the equations proposed to humidity at altitude is used to define the humidity ratio at sea
describe the pressure drop. The other trend that can be noticed level to obtain the performance of the wheel using manufac-
from the analysis of the figures is that, as mass flow rate turer’s performance data. Then, the humidity ratio change
increases, the difference in pressure drop between MSU and (grain depression) at sea level is used to obtain the outlet
NREL also increases. This trend can also be explained through humidity ratio at altitude by subtracting the humidity ratio
the equations describing pressure drop since pressure drop change from the inlet humidity ratio at altitude.
follows a second-order polynomial behavior with respect to
the velocity. Although the second-order term is considerably Rationality for Regeneration Specific Heat Input
smaller than the first-order term, with increasing velocity By definition, the RSHI is the ratio of the energy input of
comes a larger variation due to the velocity-squared term. the regeneration heater to the MRC. Since the focus of the
methodology is to examine the same dehumidification unit
METHODOLOGY between sea level and altitude, the method used for the regen-
eration heat should also remain the same. As discussed, the
Rationality for Moisture Removal Capacity specific heat of air changes negligibly at altitudes less than
To compare performance between locations, the mass 3,048 m (10,000 feet). With the mass flow rate and air temper-
flow rate, dry-bulb temperature, and humidity ratio are the atures at the inlet and exit of the heater constant, the energy
variables to be considered constants. From the experimental input to the heater (Q) will also remain a constant. Therefore,
data, it was found that the variations in MRC between MSU since from the previous section the MRC at altitude (up to
(100 m or 330 ft » sea level) and NREL’s altitude (1,783 m or 1,524 m or 5,000 ft) is approximated to the MRC at sea level,
5,850 ft) fall between 4.1% and 5.4% for most of the tests, and the RSHI at altitude is approximated to the RSHI at sea level
between 1% and 2.2% for the tests at low temperature and high as shown in Equation (19).
relative humidity (pre-cooling), with the MRC decreasing Q Q
RSHI z = ---------------  --------------- = RSHI o (19)
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
with altitude. These variations are within the experimental MRC z MRC o
uncertainties of both MSU and NREL. On the other hand, as
discussed previously, it is important to recall that these varia-
Rationality for Process Outlet Temperature
tions would have been less than half if tests at MSU would
have been conducted with the same regeneration inlet humid- As discussed in Section 6.2, the process outlet tempera-
ity ratio of NREL. Therefore, due to (a) the small variation in ture did not follow any specific trend between the two loca-
MRC obtained from experimental data, (b) its comparison tions, and the variation was quite small, varying between -1.44
with the uncertainty associated with the experimental data, (c) °C (-2.6 °F) (» -2.4%) and 0.56 °C (1.0°F) (» 1.0%). Since the
the fact that the results were obtained from a specific desiccant process temperature rise is due to the amount of water vapor
wheel with specific regeneration parameters, (d) the tendency adsorbed by the desiccant (latent heat) plus an additional
of a lower variations in MRC for common pre-cooling condi- amount of heat that is equal to between 5 and 25% of the latent
tions, and (e) results from previous studies conducted at heat [14, Chapter 32], it can be concluded that the variation in
NREL, for altitude up to 1,524 m (5,000 ft) (84.3 kPa or 12.2 process outlet temperature between altitude and sea level can
psia atmospheric pressure), it seems to be unwise to propose be neglected because (a) since from Section 7.1 the MRC at
a simple methodology other than a series of steps that leads to altitude (up to 1,524 m or 5,000 ft) is approximated to the
a MRC at altitude equal to the MRC obtained from manufac- MRC at sea level, there is no variation associated to the water
turer's data at sea level. In practical applications, the common vapor adsorbed by the desiccant, and (b) since the inlet
safety design factor used in engineering can be defined to take temperature for process and regeneration are kept constant at
care of the small decrease of MRC with altitude that arises as altitude when compared to sea level, the difference in temper-
consequence of the lower water vapor partial pressure or rela- ature that is the driving force for heat transfer is the same
tive humidity found at altitude for the same humidity ratio regarding of altitude. This reasoning, supported by the exper-
when compared to sea level. imental results, leads to the conclusion that for altitude up to
It is important to point out that if the process inlet humid- 1,524 m (5,000 ft) the process outlet temperature at altitude
ity ratio is kept constant between altitude and sea level, for can be approximated to the process outlet temperature at sea
high relative humidity conditions at altitude, the humidity level.
ratio at sea level may not be defined since it would be greater
Rationality for Pressure Drop through the Wheel
than the humidity ratio at saturation. Therefore, since the
process inlet state of the air stream can be defined with dry- The pressure drop was found to hold the most significant
bulb temperature and relative humidity, and because it was difference with respect to altitude. To propose a methodology
found that the performance at altitude can be approximated to some analysis of how pressure drop has been considered in
the performance at sea level, an approach based on relative previous studies is of great importance.
humidity can be used to convert humidity ratio from altitude Pesaran and Heiden [1] suggested Equation (20) to esti-
to sea level and from sea level to altitude. That is, the relative mate the pressure drop through the wheel for fully developed
546
ASHRAE Transactions
laminar flow assuming negligible entrance, exit, and acceler- where the coefficients Cl (related to laminar flow) and Ck,t
ation effects. (related to entrance and exit losses and turbulent flow) are
specific for each desiccant wheel, which account for geomet-
·
1 m l L 1 l L ric parameters of the desiccant wheel, thermophysical prop-
P = --- --------------
- = --- ---------
-V (20)
2 D h2 A 2 D h2 erties of the air, and units. By knowing two different sets of
actual face velocity and pressure drop, the coefficients Cl and
where: Ck,t can be found for a specific desiccant dehumidifier. When
l = Geometric factor keeping the mass flow rate constant, and with the coefficients
Slayzak et al. [6] used Equation (21) for laminar flow and known, the pressure drop at altitude can be calculated as
Equation (22) for turbulent flow when analyzing the effect of
altitude on pressure drop, with the assumption that the loss P z = C l V z + C k t  z V z2 (26)
coefficient for turbulent entrance effects quantitatively
follows this relationship. where Vz and z are the actual face velocity and air density
at altitude for the design mass flow rate, respectively. Due to
64 L V 2 L viscous effects of temperature, the coefficients Cl and Ck,t
P lam =  ------ ------ ------ = 32 ------2- V (21)
Re D h 2 Dh must be found for process (low temperatures) and regenera-
tion (high temperatures) air streams independently.
c L V2
P turb =  --------------
- ------ ------ (22) Since to estimate pressure drop at altitude the mass flow
Re 0.25 D h 2
rate is kept constant, the face velocity at altitude can be defined
Harshe et al. [3] presented Equation (23) as a general as
equation to describe the pressure drop through a desiccant
wheel in their mathematical models. In Equation (23), f repre- o
V z = -----V 0 (27)
sents the friction factor which is a function of the inverse of z
Reynolds number. Therefore, the first term in Equation (23)
has the same form as Equations (20) and (21). where subscript ‘o’ refers to standard conditions. By substi-
tuting Equation (27) into Equation (26) and comparing with
2 fV 2 L 1 Equation (25), the pressure drop at altitude can be estimated
P = ------------------- + K --- V 2 (23)
Dh 2 from pressure drop at standard conditions as
For all the equations Reynolds number can be defined as 0
P z = -----P 0 (28)
· z
VD mD
Re = -------------h- = ----------h- (24)
 A
It should be noted that, when calculating the density at
Since all geometric parameter are fixed, by analyzing the altitude, the actual atmospheric pressure should be used;
previous equations it can be observed that if the inlet condi- however, although weather conditions may cause small vari-
tions and the mass flow rate are kept constant between sea ations with negligible effects compared with the uncertainty of
level (standard conditions) and altitude, the pressure drop due the measurements, the standard atmospheric pressure found
to laminar, turbulent, entrance, and exit effect can be described using Equation (18) can be used. Since the methodology
by two terms, one depending on velocity and the other depend- compares densities at the same temperature and humidity
ing on the square of the velocity. Although air properties will ratio, as defined by equations from ASHRAE [14, Chapter 1],
change as the air moves through the desiccant wheel, the tests the ratio of the densities can be defined as the ratio of standard
results suggest no significant variation in temperature between atmospheric pressures (Equation (18)). Therefore, Equation
sea level and altitude. Therefore, effects due to variation of (29) is equivalent to Equation (28).
temperature can be neglected and only the effect of variation
in density due to altitude becomes important. 1
P z = ------------------------------------------------------------------
-P (29a, SI)
Although pressure drop should be computed using the  1 – 2.25577  10 – 5 Z  5.2559 0
flute velocity, since geometric parameters remain constant, the
pressure drop can related to face velocity. Based on this 1
P z = ---------------------------------------------------------------
-P (29b, I-P)
reasoning and the previous discussion, the pressure drop  1 – 6.8754  10 –6 Z  5.2559 0
through a honeycomb matrix as a function of actual face veloc-
ity and density can be well represented as a second order poly- By using Equation (28), the variations between the estimated
nomial of the form pressure drops and actual pressure drops for NREL showed
appropriate agreement as illustrated in Tables 8 and 9 for
P = C l V + C k t V 2 (25) process and regeneration, respectively.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
©2012 ASHRAE
547
Table 8a. Comparison of Actual and Estimated Table 9a. Comparison of Actual and Estimated
Process Pressure Drop at NREL (SI Units) Regeneration Pressure Drop at NREL (SI Units)
Actual Estimated Error Actual Estimated Error

Test Test
[cm WC] [cm WC] [%] [cm WC] [cm WC] [%]
1A 1.750 1.773 0.023 1.3 1A 1.199 1.212 0.013 1.1
1B 1.207 1.186 –0.020 –1.7 1B 0.785 0.818 0.033 4.2
1C 1.471 1.448 –0.023 –1.6 1C 0.968 0.978 0.010 1.0
2A 1.796 1.773 –0.023 –1.3 2A 1.176 1.199 0.023 1.9
2B 1.207 1.173 –0.033 –2.7 2B 0.772 0.795 0.023 3.0
2C 1.478 1.438 –0.041 –2.7 2C 0.970 0.983 0.013 1.3
3A 1.801 1.791 –0.010 –0.6 3A 1.176 1.209 0.033 2.8
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
3B 1.209 1.212 0.003 0.2 3B 0.765 0.803 0.038 5.0
3C 1.384 1.519 0.135 9.7 3C 0.935 0.986 0.051 5.4
4A 1.715 1.801 0.086 5.0 4A 1.158 1.196 0.038 3.3
4B 1.146 1.171 0.025 2.2 4B 0.737 0.792 0.056 7.6
4C 1.425 1.443 0.018 1.2 4C 0.935 0.968 0.033 3.5
5A 1.671 1.730 0.058 3.5 5A 1.179 1.229 0.051 4.3
5B 1.039 1.186 0.147 14.2 5B 0.742 0.818 0.076 10.3
5C 1.328 1.440 0.112 8.4 5C 0.940 0.993 0.053 5.7
Table 8b. Comparison of Actual and Estimated Table 9b. Comparison of Actual and Estimated
Process Pressure Drop at NREL (I-P Units) Regeneration Pressure Drop at NREL (I-P Units)
Actual Estimated Error Actual Estimated Error

Test Test
[in WC] [in WC] [%] [in WC] [in WC] [%]
1A 0.689 0.698 0.009 1.3 1A 0.472 0.477 0.005 1.1
1B 0.475 0.467 –0.008 –1.7 1B 0.309 0.322 0.013 4.2
1C 0.579 0.570 –0.009 –1.6 1C 0.381 0.385 0.004 1.0
2A 0.707 0.698 –0.009 –1.3 2A 0.463 0.472 0.009 1.9
2B 0.475 0.462 –0.013 –2.7 2B 0.304 0.313 0.009 3.0
2C 0.582 0.566 –0.016 –2.7 2C 0.382 0.387 0.005 1.3
3A 0.709 0.705 –0.004 –0.6 3A 0.463 0.476 0.013 2.8
3B 0.476 0.477 0.001 0.2 3B 0.301 0.316 0.015 5.0
3C 0.545 0.598 0.053 9.7 3C 0.368 0.388 0.020 5.4
4A 0.675 0.709 0.034 5.0 4A 0.456 0.471 0.015 3.3
4B 0.451 0.461 0.010 2.2 4B 0.290 0.312 0.022 7.6
4C 0.561 0.568 0.007 1.2 4C 0.368 0.381 0.013 3.5
5A 0.658 0.681 0.023 3.5 5A 0.464 0.484 0.020 4.3
5B 0.409 0.467 0.058 14.2 5B 0.292 0.322 0.030 10.3
5C 0.523 0.567 0.044 8.4 5C 0.370 0.391 0.021 5.7
548
ASHRAE Transactions
Proposed Methodology 50 split and rotation at 14 rph using set design inlet conditions.
As result from this research project, the following infor- Test conditions for the experiments were decided based on
mation was proposed for dissemination as “Methodology for information about industry practices, information regarding
Solid Desiccant Dehumidifier Correction for Altitude.” the concept behind the adsorption process of a desiccant, and
previous research. The desiccant dehumidifier was tested by
This methodology is proposed as a simple methodology
MSU and NREL, and the results of the tests were compared to
to correct manufacturer’s performance data for solid desiccant
find the relationship between the performance at sea level and
dehumidifiers at sea level for altitudes up to 1,524 m (5,000 ft)
at altitude. With the exception in the pressure drop through the
(83.3 kPa or 12.2 psia atmospheric pressure). The methodol-
wheel, results from the experimental tests suggested that if the
ogy was developed based on experimental tests conducted at
mass flow rate, dry-bulb temperature, and humidity ratio at the
Mississippi State University and the National Renewable
process and regeneration inlet are kept constant, the perfor-
Energy Laboratory. It should be understood that the method-
mance of a desiccant dehumidifier at altitude up to 1,524 m
ology was developed within the limits of experimental accu-
(5,000 ft) can be approximated to the manufacturers’ perfor-
racy.
mance at sea level. Regarding the pressure drop through the
For a given process and reactivation inlet air mass flow
wheel, thee drop at altitude is greater than pressure drop at sea
rate, temperature, and humidity ratio, the performance of a
level by the ratio of the densities or atmospheric pressures.
desiccant dehumidifier is constant regardless of altitude, with
the exception for the pressure drop through the wheel. NOMENCLATURE
When the relative humidity of the process inlet air is near
saturation at altitude and the humidity ratio at altitude is above English Symbols
saturation at sea level, the following method can be used to
A Area, [m2] [ft2]
determine the dehumidifier performance at altitude:
cp Specific heat of air, [kJ/kg °C] [BTU/lbm °F]
1. Determine the air mass flow rate, inlet temperature and G Humidity ratio, [g/kg] [Gr/lbm]
relative humidity of both process and reactivation air Gr Grains of water
streams at altitude. If the reactivation air is direct gas hfg Specific enthalpy of evaporation, [kJ/kg] [BTU/lbm]
fired, its inlet air humidity ratio should be corrected for ·
m Mass flow rate, [kg/s] [lbm/min]
the water vapor from combustion, which can be estimated
as 0.0437 and 0.036 grams per kilogram of air per degree MRC Moisture Removal Capacity, [kg/h] [lbm/hr]
Celcius (0.17 and 0.14 grains per pound of air per degree N Fan speed, [rpm]
Fahrenheit) of temperature rise for natural gas and L.P. p Pressure, [kPa] [psia] [in W.C.]
gas, respectively. Q Heat, [kJ/s] [BTU/hr]
2. Use the values from Step 1 (inlet mass flow rate, temper- RSHI Regeneration Specific Heat Input, [kJ/kg] [BTU/hr]
ature, and relative humidity) and the manufacturer's sea
Scmh Standard volumetric flow rate, [cmh]
level performance data to determine the discharge
temperature and humidity ratio change (grain depression) Scfm Standard volumetric flow rate, [cfm]
of the process air (and also reactivation air, if it is of inter- T Temperature, [K] [°C] [°F] [°R]
est). V Face velocity, [m/s] [fpm]
3. Use the discharge temperature and humidity ratio ·
V Volumetric flow rate, [cmh] [cfm]
changes from Step 2 and the inlet temperature and rela- W Humidity ratio, [kg/kg] [lbm/lbm]
tive humidity values at altitude to determine the dehumid-
ws Saturation humidity ratio, [kg/kg] [lbm/lbm]
ifier performance at altitude.
Z Altitude, [m] [ft]
The pressure drop through the desiccant wheel (process
and reactivation sides) obtained from manufacturer’s perfor- Greek Symbols
mance data at sea level must be corrected by the ratio of atmo-  Variation (grain depression, pressure drop)
spheric pressure at sea level (101.325 kPa or 14.696 psia) to  Dynamic viscosity of air, [lbm/s ft]
altitude.”  Density of moist air, [lbm/ft3]
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
CONCLUSION  Relative humidity
This paper presented the research performed to develop a Subscripts

simple methodology to be used for the estimation of desiccant
atm Atmospheric
dehumidifiers performance at altitude. The research involved
taking performance data from a test cassette containing a 300 db Dry-bulb temperature
mm (11”) diameter and 100 mm (4”) depth solid desiccant dp Dew-point temperature
wheel made of silica gel set for a geometric configuration of 50/ f Fan
©2012 ASHRAE
549
h Heat Niu, J.L., Zhang, L.Z. 2002. Heat Transfer and Friction
I Inlet Coefficients in Corrugated Ducts Confined by Sinusoi-
m Mass dal and Arc Curves. International Journal of Heat and
Mass Transfer Vol. 45, pp. 571-578.
o Sea level
Pesaran, A.A., Heiden, R. 1994. Influence of Altitude on the
O Outlet
Performance of Desiccant-Cooling Systems. Energy Vol.
P Process 19, No. 11, pp. 1165-1179.
R Regeneration Ruivo, C.R., Costa, J.J., Figueiredo, A.R. 2011. Influence of
z Altitude the Atmospheric Pressure on the Mass Transfer Rate of
ws Saturation pressure of water vapor at a given Desiccant Wheels. International Journal of Refrigera-
temperature tion Vol. 34, No. 3, pp. 707-718.
wb Wet-bulb temperature Ruivo, C.R., Costa, J.J., Figueiredo, A.R. 2011. Numerical
Study of the Influence of the Atmospheric Pressure on
ACKNOWLEDGMENTS the Heat and Mass Transfer Rates of Desiccant Wheels.
Support from ASHRAE and the National Renewable International Journal of Heat and Mass Transfer Vol.
Energy Laboratory (NREL) is gratefully acknowledged. The 54, pp. 1331-1339.
authors also want to recognize the contributions of Kenneth Schultz, C.C. 2006. Engineering for High Altitudes. Engi-
Jacobs as graduate student participating in this research proj- neered Systems, pp. 26-33.
ect. Slayzak, S.J., Ryan, J.P. 1998. Instrument Uncertainty Effect
on Calculation of Absolute Humidity Using Dewpoint,
REFERENCES
Wet-Bulb, and Relative Humidity Sensors. Prepared for:
ASHRAE. 2007. Method of Testing for Rating Desiccant Solar 98: Renewable Energy for the Americas Confer-
Dehumidifiers Utilizing Heat for the Regeneration Pro- ence: Albuquerque, New Mexico. NREL/CP-550-
cess. ASHRAE: Atlanta, GA. ANSI/ASHRAE Standard 24523.
139-2007. Slayzak, S.J., Ryan, J.P. 2000. Desiccant Dehumidification
ASHRAE. 2008. 2008 ASHRAE Handbook – HVAC Systems Wheel Test Guide. National Renewable Energy Labora-
and Equipment. ASHRAE: Atlanta, GA. tory: Golden, Colorado. NREL/TP-550-26131.
ASHRAE. 2009. 2009 ASHRAE Handbook – Fundamentals.
American Society of Heating, Refrigeration, and Air- Slayzak, S.J., Ryan, J.P., Jalalzadeh-Azar, A.A. 2002. Mea-
Conditioning Engineers, Inc.: Atlanta, GA. sured Effect of Altitude on the Performance of a Regen-
Belady, C.L. 1996. Design Considerations for Air Cooling erated Desiccant Matrix. ASHRAE Transactions:
Electronic Systems in High Altitude Conditions. IEEE Symposia. Vol. 108(2), pp. 556-562. (NREL Report No.
Transactions on Components, Packaging, and Manufac- 33844).
turing Technology – Part A Vol. 19, No. 4, pp. 495-500. Tretiak, C.S., Abdallah, N.B. 2009. Sorption and Desorption
Coleman, H.W., Steele, W.G. 2009. Experimentation, Vali- Characteristics of a Packed Bed of Clay-CaCl2 Desic-
dation, and Uncertainty Analysis for Engineers. Hobo- cant Particles. Solar Energy Vol. 83, pp. 1861-1870.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
ken, NJ: John Wiley & Sons. Print. Wen, T., Tian, J., Lu, T.J., Queheillalt, D.T., Wadley, H.N.G.
Harshe, Y.M., Utikar, R.P., Ranade, V.V., Pahwa, D. 2005. 2006. Forced Convection in Metallic Honeycomb Struc-
Modeling of Rotary Desiccant Wheels. Chemical Engi- tures. International Journal of Heat and Mass Transfer
neering & Technology Vol. 28, No. 12, pp. 1473-1479. Vol. 49, pp. 3313-3324.
550
ASHRAE Transactions
APPENDIX A – UNCERTAINTIES
Table A1a. Data Uncertainties for Process Stream at MSU (SI Units)
Mass Flow Humidity Ratio In Humidity Ratio Out Grain Depression MRC
Test
[kg/s] [%] [g/kg] [%] [g/kg] [%] [g/kg] [%] [kg/hr] [%]
1A 0.0029 2.1 0.20 1.8 0.46 6.9 0.50 10.5 0.26 10.7
2A 0.0029 2.1 0.20 1.6 0.50 6.4 0.53 10 0.28 10.2
3A 0.0029 2.1 0.21 1.4 0.53 5.8 0.57 9.7 0.29 9.9
4A 0.0029 2.1 0.21 1.7 0.41 5.9 0.44 7.4 0.24 8.1
5A 0.0029 2.1 0.16 1.7 0.27 6.8 0.30 6 0.17 6.7
1B 0.0022 2.3 0.20 1.8 0.51 9.5 0.56 9.6 0.20 9.9
2B 0.0022 2.3 0.20 1.6 0.57 8.6 0.60 9.4 0.21 9.7
3B 0.0022 2.3 0.21 1.4 0.60 7.4 0.64 9.1 0.23 9.4
4B 0.0022 2.3 0.23 1.7 0.51 8.8 0.54 7.6 0.20 8.3
5B 0.0022 2.4 0.16 1.7 0.34 11.2 0.37 6.2 0.14 6.6
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
1C 0.0025 2.2 0.20 1.8 0.50 8.4 0.59 10.9 0.23 10.4
2C 0.0025 2.2 0.21 1.6 0.53 7.3 0.61 10.4 0.24 10.0
3C 0.0025 2.2 0.21 1.4 0.57 6.7 0.64 9.8 0.26 9.5
4C 0.0025 2.2 0.21 1.7 0.46 7.2 0.51 7.7 0.22 8.0
5C 0.0025 2.2 0.16 1.7 0.30 9.2 0.34 6.1 0.15 6.5
Table A1b. Data Uncertainties for Process Stream at MSU (I-P Units)
Test
[lbm/min] [%] [Gr/lbm] [%] [Gr/lbm] [%] [Gr/lbm] [%] [lbm/hr] [%]
1A 0.39 2.1 1.4 1.8 3.2 6.9 3.5 10.5 0.57 10.7
2A 0.39 2.1 1.4 1.6 3.5 6.4 3.7 10.0 0.61 10.2
3A 0.39 2.1 1.5 1.4 3.7 5.8 4.0 9.7 0.65 9.9
4A 0.39 2.1 1.5 1.7 2.9 5.9 3.1 7.4 0.54 8.1
5A 0.39 2.1 1.1 1.7 1.9 6.8 2.1 6.0 0.37 6.7
1B 0.29 2.3 1.4 1.8 3.6 9.5 3.9 9.6 0.43 9.9
2B 0.29 2.3 1.4 1.6 4.0 8.6 4.2 9.4 0.46 9.7
3B 0.29 2.3 1.5 1.4 4.2 7.4 4.5 9.1 0.50 9.4
4B 0.29 2.3 1.6 1.7 3.6 8.8 3.8 7.6 0.44 8.3
5B 0.29 2.4 1.1 1.7 2.4 11.2 2.6 6.2 0.30 6.6
1C 0.33 2.2 1.4 1.8 3.5 8.4 4.1 10.9 0.50 10.4
2C 0.33 2.2 1.5 1.6 3.7 7.3 4.3 10.4 0.53 10.0
3C 0.33 2.2 1.5 1.4 4.0 6.7 4.5 9.8 0.57 9.5
4C 0.33 2.2 1.5 1.7 3.2 7.2 3.6 7.7 0.49 8.0
5C 0.33 2.2 1.1 1.7 2.1 9.2 2.4 6.1 0.33 6.5
©2012 ASHRAE
551
Table A2a. Data Uncertainties for Regeneration Table A2b. Data Uncertainties for Regeneration
Stream at MSU (SI Units) Stream at MSU (I-P Units)
Mass Flow Inlet Humidity Ratio Mass Flow Inlet Humidity Ratio
Test Test
[kg/s] [%] [g/kg] [%] [lbm/min] [%] [Gr/lbm] [%]
1A 0.0017 2.0 0.40 2.5 1A 0.22 2.0 2.8 2.5
2A 0.0017 2.0 0.40 2.6 2A 0.22 2.0 2.8 2.6
3A 0.0017 2.0 0.41 2.7 3A 0.22 2.0 2.9 2.7
4A 0.0017 2.0 0.40 2.6 4A 0.23 2.0 2.8 2.6
5A 0.0017 2.0 0.39 2.5 5A 0.22 2.0 2.7 2.5
1B 0.0012 2.1 0.40 2.6 1B 0.16 2.1 2.8 2.6
2B 0.0012 2.1 0.43 2.7 2B 0.16 2.1 3.0 2.7
3B 0.0012 2.1 0.37 2.5 3B 0.16 2.1 2.6 2.5
4B 0.0012 2.1 0.31 2.0 4B 0.16 2.1 2.2 2.0
5B 0.0012 2.1 0.31 2.1 5B 0.16 2.1 2.2 2.1
1C 0.0014 2.1 0.46 2.9 1C 0.18 2.1 3.2 2.9
2C 0.0014 2.1 0.46 3.0 2C 0.18 2.1 3.2 3.0
3C 0.0014 2.1 0.37 2.4 3C 0.19 2.1 2.6 2.4
4C 0.0014 2.1 0.39 2.5 4C 0.18 2.1 2.7 2.5
5C 0.0014 2.0 0.24 1.6 5C 0.19 2.0 1.7 1.6
Table A3a. Data Uncertainties for Process Stream at NREL (SI Units)
Test
[kg/s] [%] [g/kg] [%] [g/kg] [%] [g/kg] [%] [kg/hr] [%]
1A 0.0028 2.0 0.13 1.1 0.09 1.2 0.14 3.3 0.09 3.9
2A 0.0028 2.0 0.14 1.1 0.10 1.2 0.17 3.4 0.10 3.9
3A 0.0029 2.0 0.16 1.1 0.11 1.1 0.20 3.5 0.11 4.0
4A 0.0029 2.0 0.14 1.1 0.09 1.2 0.17 2.9 0.10 3.5
5A 0.0028 2.0 0.10 1.1 0.06 1.2 0.11 2.4 0.08 3.1
1B 0.0019 2.0 0.13 1.1 0.07 1.2 0.14 2.6 0.06 3.3
2B 0.0019 2.0 0.14 1.1 0.09 1.2 0.17 2.7 0.07 3.4
3B 0.0019 2.0 0.16 1.1 0.10 1.1 0.19 2.8 0.08 3.5
4B 0.0019 2.0 0.14 1.1 0.07 1.2 0.16 2.4 0.07 3.1
5B 0.0019 2.0 0.10 1.1 0.04 1.3 0.11 2.0 0.05 2.8
1C 0.0023 2.0 0.13 1.1 0.07 1.2 0.14 2.9 0.07 3.6
2C 0.0023 2.0 0.14 1.1 0.09 1.2 0.17 3.0 0.09 3.6
3C 0.0023 2.0 0.17 1.1 0.10 1.1 0.20 3.1 0.10 3.7
4C 0.0023 2.0 0.14 1.1 0.09 1.2 0.17 2.6 0.09 3.3
5C 0.0023 2.0 0.10 1.1 0.04 1.2 0.11 2.1 0.07 2.9
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
552
ASHRAE Transactions
Table A3b. Data Uncertainties for Process Stream at NREL (I-P Units)
Test
[lbm/min] [%] [Gr/lbm] [%] [Gr/lbm] [%] [Gr/lbm] [%] [lbm/hr] [%]
1A 0.37 2.0 0.9 1.1 0.6 1.2 1.0 3.3 0.20 3.9
2A 0.37 2.0 1.0 1.1 0.7 1.2 1.2 3.4 0.22 3.9
3A 0.38 2.0 1.1 1.1 0.8 1.1 1.4 3.5 0.25 4.0
4A 0.38 2.0 1.0 1.1 0.6 1.2 1.2 2.9 0.23 3.5
5A 0.37 2.0 0.7 1.1 0.4 1.2 0.8 2.4 0.17 3.1
1B 0.25 2.0 0.9 1.1 0.5 1.2 1.0 2.6 0.14 3.3
2B 0.25 2.0 1.0 1.1 0.6 1.2 1.2 2.7 0.15 3.4
3B 0.25 2.0 1.1 1.1 0.7 1.1 1.3 2.8 0.17 3.5
4B 0.25 2.0 1.0 1.1 0.5 1.2 1.1 2.4 0.16 3.1
5B 0.25 2.0 0.7 1.1 0.3 1.3 0.8 2.0 0.12 2.8
1C 0.31 2.0 0.9 1.1 0.5 1.2 1.0 2.9 0.16 3.6
2C 0.31 2.0 1.0 1.1 0.6 1.2 1.2 3.0 0.19 3.6
3C 0.31 2.0 1.2 1.1 0.7 1.1 1.4 3.1 0.21 3.7
4C 0.31 2.0 1.0 1.1 0.6 1.2 1.2 2.6 0.19 3.3
5C 0.31 2.0 0.7 1.1 0.3 1.2 0.8 2.1 0.15 2.9
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Table A4a. Data Uncertainties for Regeneration Table A4b. Data Uncertainties for Regeneration
Stream at NREL (SI Units) Stream at NREL (I-P Units)
Mass Flow Inlet Humidity Ratio Mass Flow Inlet Humidity Ratio
Test Test
[kg/s] [%] [g/kg] [%] [lbm/min] [%] [Gr/lbm] [%]
1A 0.0014 2.0 0.24 1.0 1A 0.19 2.0 1.7 1.0

2A 0.0014 2.0 0.24 1.0 2A 0.19 2.0 1.7 1.0
3A 0.0014 2.0 0.24 1.0 3A 0.19 2.0 1.7 1.0
4A 0.0014 2.0 0.23 1.0 4A 0.19 2.0 1.6 1.0
5A 0.0014 2.0 0.24 1.0 5A 0.19 2.0 1.7 1.0
1B 0.0009 2.0 0.24 1.0 1B 0.12 2.0 1.7 1.0
2B 0.0010 2.0 0.24 1.0 2B 0.13 2.0 1.7 1.0
3B 0.0010 2.0 0.24 1.0 3B 0.13 2.0 1.7 1.0
4B 0.0010 2.0 0.23 1.0 4B 0.13 2.0 1.6 1.0
5B 0.0009 2.0 0.24 1.0 5B 0.12 2.0 1.7 1.0
1C 0.0011 2.0 0.24 1.0 1C 0.15 2.0 1.7 1.0
2C 0.0011 2.0 0.23 1.0 2C 0.15 2.0 1.6 1.0
3C 0.0011 2.0 0.24 1.0 3C 0.15 2.0 1.7 1.0
4C 0.0011 2.0 0.23 1.0 4C 0.15 2.0 1.6 1.0
5C 0.0011 2.0 0.24 1.0 5C 0.15 2.0 1.7 1.0
©2012 ASHRAE
553
Table A5a. Wheel Pressure Drop Uncertainties for MSU and NREL (SI Units)
MSU NREL
Test Process Regeneration Process Regeneration
[cm WC] [%] [cm WC] [%] [cm WC] [%] [cm WC] [%]
1A 0.0305 2.1 0.0229 2.0 0.0356 2.0 0.0229 2.0
2A 0.0305 2.1 0.0229 2.0 0.0356 2.0 0.0229 2.0
3A 0.0305 2.1 0.0229 2.1 0.0356 2.0 0.0229 2.0
4A 0.0305 2.1 0.0229 2.0 0.0356 2.0 0.0229 2.0
5A 0.0279 2.1 0.0229 2.0 0.0330 2.0 0.0229 2.0
1B 0.0229 2.3 0.0178 2.0 0.0229 2.0 0.0152 2.0
2B 0.0229 2.3 0.0152 2.1 0.0254 2.0 0.0152 2.0
3B 0.0229 2.3 0.0152 2.1 0.0254 2.0 0.0152 2.0
4B 0.0229 2.3 0.0152 2.1 0.0229 2.0 0.0152 2.0
5B 0.0229 2.4 0.0152 2.1 0.0203 2.0 0.0152 2.0
1C 0.0254 2.2 0.0203 2.1 0.0305 2.0 0.0203 2.0
2C 0.0254 2.2 0.0203 2.1 0.0305 2.0 0.0203 2.0
3C 0.0254 2.2 0.0203 2.0 0.0279 2.0 0.0178 2.0
4C 0.0254 2.2 0.0203 2.1 0.0279 2.0 0.0178 2.0
5C 0.0254 2.2 0.0203 2.1 0.0254 2.0 0.0178 2.0
Table A5b. Wheel Pressure Drop Uncertainties for MSU and NREL (I-P Units)
MSU NREL
Test Process Regeneration Process Regeneration
[in WC] [%] [in WC] [%] [in WC] [%] [in WC] [%]
1A 0.012 2.1 0.009 2.0 0.014 2.0 0.009 2.0
2A 0.012 2.1 0.009 2.0 0.014 2.0 0.009 2.0
3A 0.012 2.1 0.009 2.1 0.014 2.0 0.009 2.0
4A 0.012 2.1 0.009 2.0 0.014 2.0 0.009 2.0
5A 0.011 2.1 0.009 2.0 0.013 2.0 0.009 2.0
1B 0.009 2.3 0.007 2.0 0.009 2.0 0.006 2.0
2B 0.009 2.3 0.006 2.1 0.010 2.0 0.006 2.0
3B 0.009 2.3 0.006 2.1 0.010 2.0 0.006 2.0
4B 0.009 2.3 0.006 2.1 0.009 2.0 0.006 2.0
5B 0.009 2.4 0.006 2.1 0.008 2.0 0.006 2.0
1C 0.010 2.2 0.008 2.1 0.012 2.0 0.008 2.0
2C 0.010 2.2 0.008 2.1 0.012 2.0 0.008 2.0
3C 0.010 2.2 0.008 2.0 0.011 2.0 0.007 2.0
4C 0.010 2.2 0.008 2.1 0.011 2.0 0.007 2.0
5C 0.010 2.2 0.008 2.1 0.010 2.0 0.007 2.0
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
554
ASHRAE Transactions
Table A6. Process Outlet Temperature Uncertainties for MSU and NREL (SI and I-P Units)
MSU NREL
Test
[°C] [°F] [%] [°C] [°F] [%]
1A 1.3 2.4 2.3 0.4 0.8 0.8
2A 1.3 2.3 2.2 0.5 0.9 0.9
3A 1.2 2.1 1.9 0.5 0.9 0.8
4A 1.4 2.6 2.6 0.5 0.9 0.9
5A 1.3 2.4 2.7 0.6 1.0 1.1
1B 1.3 2.3 2.1 0.6 1.0 0.9
2B 1.6 2.8 2.5 0.6 1.0 0.9
3B 1.4 2.6 2.3 0.6 1.1 1.0
4B 1.3 2.4 2.2 0.6 1.1 1.0
5B 1.4 2.5 2.7 0.7 1.2 1.3
1C 1.2 2.2 2.1 0.6 1.0 0.9
2C 1.1 2.0 1.8 0.6 1.0 0.9
3C 0.9 1.7 1.5 0.4 0.8 0.7
4C 0.9 1.6 1.5 0.4 0.8 0.8
5C 0.8 1.4 1.5 0.4 0.8 0.9
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
©2012 ASHRAE
555
SA-12-020
Thermal Performance and Charge Control

Strategy of a Ventilated Concrete Slab (VCS)
with Active Cooling Using Outdoor Air
Yuxiang Chen A.K. Athienitis, PhD, PE Khaled Galal, PhD, PE

ABSTRACT 1990, Morgan and Krarti 2007). This is mainly due to the fact
that their large exposed surface areas provide good thermal
This paper presents full-scale measurements and simula-
tion results of the thermal performance of a ventilated concrete coupling between them and the room. Improvements include
slab (VCS) with active cooling using cool outdoor air. A meth- enhanced thermal comfort by reducing room temperature
odology for the charge control and sizing of building-inte- fluctuations, shifting and decreasing peak energy demands,
grated thermal energy storage (BITES) systems that can be higher efficiency and smaller required capacity of mechanical
used for active slab cooling is developed. This VCS, which has heating/cooling equipment, and offsetting the mismatch of
a floor area of 30 m2 (323 ft2) and a storage volume of about energy supply/availability and demand. Concrete is a common
5 m3 (177 ft3), is located in a near net-zero energy solar house. and effective building material used as thermal mass (ACI
It serves as an active-charge/passive-discharge thermal mass, committee 122 2002). Concrete slabs are effective TES
as well as the basement slab – it is a BITES system. An on-site systems because of their distributed form, which ensures a
test under summer nighttime outdoor condition was performed large contact area with room air (Shaw et al. 1994) and
and the dynamic temperature distribution of the slab was moni- provides better local thermal comfort (Feustel and Stetiu
tored. The measured data is used for the validation of a thermal 1995, Inard et al. 1998). Slabs are also an inherent part of the
model under cooling operation. The validated thermal model building fabric. Therefore, making effective use of concrete
is then used to predict the thermal performance of the slab slabs as thermal storage systems will avoid extra occupation of
under different operating conditions. The simulation results living space and material cost. Various hydronic slab systems
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
show that under typical operating conditions, roughly 6.3 kWh that store thermal energy and provide space thermal condition-
(21.5 MBTU) of heat (1.26 kWh per m3 of concrete (4.3 MBTU/ ing continue to attract significant research interest. These
m3)) can be extracted by the flowing air from the VCS in 6 systems provide energy saving potentials and better thermal
hours. A 3-parameter (temperature and velocity of inlet air, comfort as compared to traditional forced-air space condition-
and cooling duration) correlation is developed to approximate ing systems (Henze et al. 2008, Lehmann et al. 2007, Xu et al.
the thermal energy storage under different operating condi- 2010). Air-core systems (Howard 1999) have been shown to
tions. Similar correlation can be developed and used for design provide stronger thermal coupling between high-mass slabs
and control of such systems. The methodology for design and and room air (e.g. higher convective heat transfer rate) and
charge control presented in this paper is applicable to other hence to significantly reduce space cooling load (Braham
types of BITES systems in different types of buildings. 2000). Research on the efficient utilization of ventilated
concrete slabs (VCS) in cooling applications for commercial
INTRODUCTION buildings has been conducted by several groups (Ren and
Distributed thermal energy storage (TES) systems have Wright 1998, Winwood et al. 1997). VCS systems may satisfy
been shown to improve the thermal performance of buildings the thermal conditioning load by using the precast concrete
significantly (Dincer 2002, Hadorn 2005, Howard and Fraker slab with ventilated cavities as the thermal energy storage
Yuxiang Chen is a doctoral candidate, A.K. Athienitis is a professor, and Khaled Galal is an associate professor in the Department of Build-
ing, Civil and Environmental Engineering at Concordia University in Montreal, Quebec.
556
ASHRAE Transactions
system and the ventilation system (Winwood et al. 1997). For a storage volume of about 5 m3 (177 ft3). Chen et al. (2010b)
example, slabs may be pre-cooled during nighttime and presented the design, construction, and monitoring of this
become heat sinks during daytime. VCS. The temperature distribution of the slab is being moni-
Concrete slabs are common in the basements of Canadian tored since 2009. The locations of the thermocouples are
houses. Utilization of basement slabs as TES systems with shown in Figure 2. The ductwork configuration in the house
proper control can improve the thermal performance of the enables both active slab heating and cooling. When used in
house; however, the thermal behavior of the slabs in heating heating, the VCS stores heat recovered from a roof-mounted
and/or cooling applications have to be well understood and building-integrated photovoltaic/thermal (BIPV/T) system
quantified in order to realize successful designs and opera- (Candanedo et al. 2010). At the BIPV/T system, outdoor air is
tions. Furthermore, easy application of control is critical in drawn with a variable speed fan underneath the warm/hot PV
promoting building-integrated thermal energy storage, espe- panels which are amorphous laminates integrated into the
cially for houses in terms of effort and cost. To serve these
metal roof. The air extracts the heat from the PV panels and
purposes, a prototype VCS was constructed in a demonstra-
then passes through the VCS as shown in Figure 1 (see Chen
tion house for in-depth study. The house was built under the
et al. (2010a) for more details about the BIPV/T and VCS
Canadian EQuilibrium Housing demonstration competition,
which started in 2006. The house is located near Eastman, active heating operation). In the active slab cooling operation,
Quebec, Canada. Detailed information on the house can be which normally happens at night, cool outdoor air passes
found in work by Chen et al. (2010a). Implementing TES is through the same path as the warm BIPV/T air. Better arrange-
one of the techniques employed in this house to reduce annual ment for shorter ductwork for the cooling air intake is possible
space heating energy consumption and peak power demand. in future similar applications. The stored heat or coolness is
The VCS is located in the basement of the house released passively to the basement space through the top
(Figure 1). It serves as a TES system, as well as the basement surface of the VCS by radiation and natural convection. The
slab – it is a building-integrated thermal energy storage VCS slab considered in this paper is an active-charge/passive-
(BITES) system. The VCS is a ribbed concrete slab with voids discharge BITES system, meaning it is charged by the air pass-
that act as air channels (Figure 2) – concrete is poured on top ing through it, but the release of thermal energy is passive –
of standard steel decks. It has a floor area of 30 m2 (323 ft2) and through radiation and natural convection from its top surface.
Figure 1 Schematic of BIPV/T system and thermal mass (direct gain system in main living area and VCS in basement).
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
©2012 ASHRAE
557
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 2 Plan view, longitudinal cross section (a), and traversal cross section (b) of VCS with monitoring points indicated.
558
ASHRAE Transactions
The site measurements and simulation for the thermal even less in cool floor situations (i.e. when the floor cooler
performance of the slab in the active slab heating operation than the air). For the bottom layer “Y-9” (Figure 3), the thick-
mode were reported by Chen et al. (2010b). In this paper, the ness is 2.7 mm (0.106 in.). It is 1/16 of the bottom width (38
thermal performance of the active slab cooling is studied and mm (1.5 in.)) of the flute (the strut of the steel deck). Assuming
the methodology for its sizing and charge control is presented. equal to 15 W/m2/K (2.64 BTU/hr/ft2/°F) for the air channel,
the calculated Biot number is about 0.02. From exterior layer
NUMERICAL MODEL AND VALIDATION to interior layer, the thickness of the layer increases with a
Concrete is used as the storage medium because it also has scale factor of two. The calculation of effective conductance in
good structural properties for basement floors. The thermo- Y direction between nodes with irregular shapes uses area-
physical properties of concrete are assumed as follows: weighted equivalent distance.
specific heat is 900 J/kg/K (0.215 BTU/lb/°F); density is 2300 Upwind differencing scheme method (Patankar 1980) is
kg/m3 (144 lb/ft3); conductivity is 1.9 W/m/K (1.1 BTU/hr/ft/ used for the control volume of the air inside the channel, as
°F). Small variations (of the order of ±10%) of the physical shown in Equation (1).
properties do not affect the slab’s thermal behavior to a signif- ·
icant extent (Bilgen and Richard 2002). This statement is also m c p.air  T air i – 1 +  Tj
j surface  U c 
j
h cr = ---------------------------------------------------------------------------------
- (1)
validated by the preliminary design simulations for the VCS · j
m c p.air +  j U c
(Chen et al. 2010b).
A simplified three-dimensional, explicit finite difference where m· c p.air is the mass flow rate of the air. T airi – 1 is the
thermal model was developed for the VCS. The 9-layer temperature of the air in the previous control volume. U c is
discretization scheme of the typical section of the VCS (not the convective conductance between current air node and the
the manifold portions) is shown in Figure 3. This discretiza- surface in contact in direction j (i.e. either the bottom nodes of
tion scheme is based on the typical pattern of temperature the slab or the ground). T surface is the surface temperature in
distribution on the cross section simulated with regular struc- direction j.
tured discretization (Chen et al. 2010b). Dense discretization Equation (2) developed by Chen et al. (2010b) is used as
at the top and bottom layers (i.e. Biot number smaller than 0.1) the correlation of the convective heat transfer coefficient
enhances the accuracy of the heat transfer modeling at the (CHTC) inside the air channel for air velocities ranging from
boundaries. The thickness of the control volumes is 2.4 mm 0.67 to 1.68 m/s (132 and 330 FPM).
(0.094 in.) at the top layer “Y-1”. The corresponding Biot
number (Kreith and Bohn 2001) is 0.015, assuming the surface h c.vcs = 3.94  V air + 5.45 [W/m 2 /K]  V air is in m/s  (2)
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
combined convective and radiative heat transfer coefficient is

10 W/m2/K (1.76 BTU/hr/ft2/°F). In natural convection condi- In order to determine the adequate grid size of “Z-Step” in
tions, the value of is less than 10 W/m2/K (1.76 BTU/hr/ft2/°F) the Z direction (Figure 3) and time step, three preliminary simu-
when the floor is less than 10°C (18°F) warmer than the room lations were performed in which these two discretization
air (ASHRAE 2008, 2009a, Karadag and Akgobek 2008). It is parameters were varied while all other parameters and boundary
Figure 3 Schematic of the 9-layer 3D discretization of the VCS.
©2012 ASHRAE
559
conditions are kept the same. Table 1 compares the different comparison indicated a good agreement between the
outputs from these three simulations. It can be seen that, 300 measured and simulation results.
mm (12 in.) for “Z-Step” and time step of 10 s is adequate to In the present paper, site-measured thermal behavior of
obtain a sufficiently converged accurate solution. 150 mm (6 the slab in cooling mode is used to validate the thermal model.
in.) “X-Step” in X direction (i.e. the center-to-center distance of A site test of active slab cooling using outdoor cool air was
the channels) is used for all three simulations. carried out on July 15th and 16th, 2009. Cool outdoor air was
For the radiative heat transfer inside the air channels, it is drawn at night into the VCS at 200 L/s (424 CFM) for 7 hours
assumed that the concrete nodes only exchange radiative heat - from 11:00pm of the 15th to 6:00 am of the 16th. The period
with the insulation surrounding the same air control volume. between midnight and 6:00am is discussed below.
This assumption is based on two factors. First, the length of the Figures 4 and 5 show the measured temperature profiles
air control volume stream-wise is 300 mm (12 in.) in this of the test. During the data sampling period, the forced-air
discretization scheme (same length as the “Z-step”). This is ventilation system was continuously under fan mode (i.e. there
four times the height of the air control volume (i.e. the height was continuous air circulation between different floors with-
of the air channel). Second, the temperature difference out mechanical cooling). The measured variables (Figure 4),
between two adjacent concrete nodes or two adjacent insula- such as the temperature and the flow rate of inlet air, soil
tion nodes is negligibly small. temperatures underneath the slab insulation, and the room air
Validation of the thermal model for active slab heating temperature were used as the boundary conditions for the
was previously performed by Chen et al. (2010b). The simulation. The measured soil temperatures based on nine-
measured temperatures from site tests was compared with point monitoring setup (Figure 2) were extrapolated and inter-
those from simulations that used the corresponding measured polated to obtain the soil temperatures under each of the
boundary conditions (i.e. the flow rate, the temperatures of the bottom grid of the slab model (Chen et al. 2010b). These
room air, the ground soil, and the inlet air) as inputs. This values are kept constant in the simulation. This is a reasonable
Table 1. Comparison of Outputs Under Different Configurations
Model Configuration Thermal Energy, kWh (MBTU)
Z-Step, mm (inches) Time Step (s) Recovered from Air Stored in VCS
150 (6) 6 -8.378 (-28.59) -5.540 (-18.90)
300 (12) 6 -8.273 (-28.23) -5.589 (-19.07)
300 (12) 10 -8.275 (-28.24) -5.591 (-19.08)

Note: negative values in energy mean loss of heat.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 4 Measured temperature profiles of site test on July 16th (family room is on top of basement).
560
ASHRAE Transactions
Figure 5 Measured temperature profiles and simulated outlet air temperature profile (July 16th).
Table 2. Measured and Simulated Temperature Values for July 16th, °C (°F)
Whole Slab Average Temperature Slab Top Surface Average Temperature Initial Room
Data Source
Temperature
Beginning End Change Beginning End Change
Measurements 19.3 (66.7) 18.4 (65.1) -0.9 (-1.6) 19.5 (67.1) 18.6 (65.5) -0.9 (-1.6) 20.3 (68.5)
Simulation 19.3 (66.7) 18.2 (64.7) -1.1 (-2.0) 19.5 (67.1) 18.5 (65.3) -1.0 (-1.8) 20.4 (68.7)
Note: the accuracy of the measurement is ± 0.5°C (0.9°F).
approach as shown by the measured data in Figure 4. Equation ence is about 0.3°C (0.5°F). The simulated profile follows the
(3) (Nu is the Nusselt number) (ASHRAE 2009a) is used in the measurement closely. The data from Table 2 and Figure 5 indi-
model to calculate the natural convective heat transfer cate good agreement between the measured and the simulated
between the cool floor and the warmer space above. results. Factors such as the accuracy of the thermocouples (±
0.5°C (0.9°F)), the difference of boundary conditions between
Nu = 0.27  Ra 1  4 (3) reality and simulation setup, the uncertainties in the exact slab
concrete thickness and evenness, the measured flow rate, the
The room temperature in the simulation was set 0.9°C CHTC at the inlet and the outlet, and other minor factors,
(1.6°F) constantly higher than that of the varying slab top contribute to the discrepancy between the simulated and
surface temperature (denoted as T clg.rm.slab ). It is based on measured values (see Chen et al. (2010b).
the measured data shown in Figure 5 – from 0.8°C (1.4°F) at
the beginning of the on-site test to 1.0°C (1.8°F) at the end.
VCS THERMAL PERFORMANCE
The 0.9°C (1.6°F) average temperature difference between the
surface and air was observed in other long-term data recorded The thermal performance based on measurements will be
as well. Measured data also indicate that the average slab presented first, followed by simulated performance. At the end
surface temperature is typically about 0.5°C (0.9°F) lower of the cooling period, the temperature readings at location B2
than the room air temperature outside active slab cooling peri- (center of the VCS, Figure 2-a) in the Y direction were 19.0,
ods (denoted as T off.rm.slab ). These two just denoted vari- 18.4 and 18.3°C (66.2, 65.1, and 65.0°F) for TC-1, TC-2, and
ables are important and used for analysis in later sub-sections. TC-5 (Figure 2-b), respectively. Therefore, the difference
Table 2 tabulates the slab temperatures from the simula- between the bottom and top surfaces was 0.7°C (1.3°F). For
tion and the measurement. The differences between the end locations at A2 and C2 (Figure 2-a), the differences were 0.5
values are small. Figure 5 compares the outlet air temperatures and 0.6°C (0.9 and 1.1°F) respectively. Figure 6 shows the
between measurement and simulation. The maximum differ- temperature profiles of the middle layer at those four locations
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
©2012 ASHRAE
561
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 6 Measured temperature profiles at different locations in the middle layer (“TC-2” in Figure 2(b)) (July 16th).
Table 3. Thermal Energy Shares Based on Simulations

[16°C (60.8°F) air at 200 L/s (424 CFM) for 6 hrs] kWh (MBTU)
Average Slab Recovered From Slab to From Slab to From Slab to From Air to
Date Stored in Slab
Temperature from Air Upper Space Soil Air Soil
July 16th 19.3°C (66.7°F) -3.24 (-11.1) -2.94 (-10.0) -0.97 (-3.3) 0.46 (1.6) 3.45 (11.8) 0.21 (0.7)
August 10th 21.8°C (71.2°F) -6.15 (-21.0) -5.81 (-19.8) -1.00 (-3.4) 0.50 (1.7) 6.32 (21.6) 0.17 (0.6)
during the cooling period. The temperature differentials Numerical analysis of nighttime active slab cooling was
increased from about 0.5 to 1.2°C (0.9 to 2.2°F) longitudinally carried out using the thermal model discussed earlier. Table 3
(i.e. in Z direction, the temperature difference between B1 and lists some key results from two sets of simulations: July 16th
B3) and from about 0.2 to 0.5°C (0.4 to 0.9°F) transversely and August 10th of 2009. Their operating conditions are the
(i.e. in X direction, the temperature difference between A2 and same – 16°C (60.8°F) air at a flow rate of 200 L/s (424 CFM)
C2). Note that the temperatures of these locations at the begin- for 6 hours, but the initial average slab and soil temperatures
ning of the plotted period were not close to each other. That is were different - 19.3 and 13.2°C (66.7 and 55.8°F) for July
because the slab cooling began one hour prior to the period 16th, and 21.8 and 15.0°C (71.2 and 59.0°F) for August 10th,
plotted. From Figure 5, it can be seen that the difference respectively. The simulated cooling power of air (i.e. heat flux
between the average temperatures of the slab top and bottom removed from the slab) for July 16th was about 600 W (2.05
surfaces was about 0.4°C (0.7°F) throughout the cooling MBTU/hr or 0.17 Ton), while it reached a higher value of 1050
period. These measurements indicate that the temperature W (3.58 MBTU/hr or 0.30 Ton) for August 10th – a typical
distribution in the VCS is quite uniform – the maximum summer night in Montreal. In the simulation for August 10th,
temperature difference was about 1.2°C (2.2°F) longitudinally 6.15 kWh (21.0 MBTU) of heat was extracted by the cool air
over 7.6 m (25.0 ft) length, except the areas near the inlet and after 6 hours of slab cooling - 6.32 kWh (21.6 MBTU) of heat
outlet manifolds. extracted from the slab and 0.17 kWh (0.6 MBTU) lost to the
soil (the soil was cooler than the air). The amount of heat
At the end of the cooling period, the measured tempera-
removed from the slab was 5.81 kWh (19.8 MBTU), which
tures of the slab near the outlet were 18.8°C (65.8°F) at B3
included 1.00 kWh (3.4 MBTU) of heat gained from the living
(Figure 6), and 18.7°C (65.7°F) at C3; while the outlet air
space through the slab top surface, 0.50 kWh (1.7 MBTU) of
temperature was 17.8°C (64.0°F) (Figure 5). This temperature
heat lost to the soil, and 6.32 kWh (21.6 MBTU) of heat
difference of about 1.0°C (1.8°F) indicates effective heat
extracted by the air passed through its hollow channels.
exchange between the warm slab and the cool air. Note that
locations B3 and C3 are about 1.5 m (4.9 ft) and 1.2 m (3.9 ft) Figure 7 shows the contour plots of the slab top surface
upstream from the outlet, respectively. temperature distribution at the end of the simulations for
562
ASHRAE Transactions
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 7 Simulated VCS top surface temperature distribution at the end of the cooling period for August 10th ((a)
temperature in Celsius; (b) temperature in Fahrenheit).
August 10th. Small temperature gradients are also indicated the channel for 0.67 and 1.68 m/s (132 and 330 FPM) air
here, matching those from field measurements. The fact that velocities are about 3685 and 9240, respectively.
the slab is cooler at the inlet manifold than it is to the east of The measured pressure drop across the VCS is about 0.07
the manifold indicates that a considerable amount of heat kPa (0.281 in. of water) for a flow rate of 200 L/s (424 CFM).
exchange took place at the inlet manifold area. This is because The fan power consumption for this VCS is therefore about 28
the CHTC at the manifold area is about an order of magnitude W (0.038 hp), assuming a fan efficiency of 50%. The corre-
higher than that in the air channel due to the former’s higher sponding air velocity for a flow rate of 200 L/s (424 CFM) is
turbulent flows. The cooler temperatures of the inlet area of 1.35 m/s (265 FPM).
reveal the importance of proper distribution of air flow. An approximately linear relationship is observed between
Improper air path design will result in an uneven and ineffi- the inlet air temperature and E store , while exponential rela-
cient distribution of stored thermal energy. tionship exists between E store and air velocity (i.e. flow rate)
Figure 8 shows the amount of coolness that will be stored or the cooling duration. Equation (4) is the fitted correlation
in the slab, E store , for August 10th initial conditions with with a maximum discrepancy of 8.6% of the corresponding
different operating conditions. It is equivalent to the heat simulation result1. The standard deviation of the discrepancies
extracted from the slab by the channel air minus those transfer is 1.9%. This correlation could be further developed for better
from the room air to the slab. Inlet air temperatures from 16 to regression. Correlations similar to Equation (4) for other types
20°C (60.8 to 68°F) in 1.0°C (1.8°F) increments, cooling dura- of TES systems with active charge can be developed using
tions from 2 to 10 hours in 2-hour increments, and channel air their corresponding thermal models.
velocities of 0.67, 1.01, 1.35 and 1.68 m/s (132, 199, 265, and
 3.4  T inlet.air – 74    1 – --------------------------------------
-
1
331 FPM) are combined and used as the operating conditions.  exp  0.85  V  air
There are totally 180 different operating conditions. The inlet E store = (4)
  1 – ---------------------------
- + 0.065
air temperature is set constant in each simulation. Note that, an 1
 exp  0.1  t 
initial slab temperature of 21.8°C (71.2°F) (denoted as slab
reference temperature T ref.slab ), constant soil temperature of
15.0°C (59.0°F), and constant T cl.grm.slab of 0.9°C (1.6°F) 1.
Percentage discrepancy = (correlation value – simulation value)
are used in all these simulations. The Reynolds numbers inside / simulation value
©2012 ASHRAE
563
Figure 8 Predicted thermal energy transferred from air to slab for August 10th initial conditions. (Initial slab and soil
temperatures were 21.8 and 15.0  C (71.2 and 59.0  F), respectively. (a) 1.35 m/s (265 FPM) constant air velocity
with different inlet air temperatures; (b) 16  C (60.8  F) constant inlet air temperature with different air velocities.)
where E store is negative in kWh, V air is the air velocity inside When a floor slab is used as cool storage, the top surface
the VCS channel in meter per second, T inlet.air is the constant temperature of the slab will affect the thermal comfort of occu-
inlet air temperature in Celsius, and t is the time in hours. pants. For example, for people with normal indoor footwear in
office space, the floor temperature should not be lower than
CHARGE CONTROL AND SIZING METHODOLOGY 19°C (66.2°F) (ASHRAE 2004). In residential and commer-
In this section, a simple methodology for the charge cial/institutional buildings, the lower limit of the floor temper-
control and sizing of the VCS TES system will be presented. ature, Tmin.slab , will also be selected according to the
The methodology makes use of the thermal comfort require- occupants’ preference. The standard-complied or occupant-
ment and the observed thermal behavior of the VCS from preferred T min.slab and the initial slab temperature just before
long-term on-site measured data. 0
the active slab cooling, T slab , define the maximum allowable
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
564
ASHRAE Transactions
amount of thermal energy, E allow , that can be stored in the slab tural weight, which are important issues for high-rise build-
at current slab cooling operation (Equation (5)). C slab is the ings. Increasing the slab thickness will also increase the
0
thermal capacity of the slab. T slab 0 (initial
is related to T rm construction cost (materials and labour) and embodied energy
room air temperature). When active slab cooling is desired, the cost. The final slab thickness may be smaller than desired due
amount of the coolness to be stored should equal to E allow to architectural and structural requirements. Also note that
achieve maximum storage. general VCS design will, to a small degree, weaken the struc-
tural strength of the slab due to hollow voids and openings.
E allow =  T slab
0 – T min.slab   C slab (5) Larger C slab can also be achieved by using structural material
with higher volumetric heat capacity; though at the cost of
The observation of the thermal behavior of the VCS indi- adding weight. Adding phase-change materials (PCM) into
cates that the slab temperature distribution is almost even. The the concrete composite is possible, but the chemical deterio-
changes of the T clg.rm.slab (temperature difference between ration of PCM and the negative influence on the structural
room air and slab during active slab cooling operation) and the strength of the concrete composite remain challenges (Dincer
T off.rm.slab (temperature difference between room air and 2002, Zhu et al. 2009).
slab outside active slab cooling period) are slight, respectively.
This observed thermal behavior of the VCS can be used to When E allow is known from Equation (5) or (6), by
derive Equation (6) from Equation (5) for more convenient setting E store equal to E allow , Equations (4) can be used to
use. The average slab temperature can be estimated based on estimate the cooling duration t and V air (i.e. flow rate) with
room air temperatures instead of being measured using corresponding T inlet.air . Weather forecast data should be
sensors. Since slab temperature is usually T clg.rm.slab lower available so that the average outdoor air temperature can be
than the room air temperature during slab cooling operation, used as T inlet.air . Since fan power is proportional to the cube
T min.slab can be replaced with T min.rm – T clg.rm.slab . T min.rm of flow rate, smaller V air and longer t should be used if oper-
is the minimum allowable room air temperature. It can be set ating conditions allow. Based on the predicted T inlet.air , a suit-
by occupants but should comply with thermal comfort stan- able V air should be chosen along with a long t to meet the
dards. Similarly, T slab can be replaced with thermal energy storage requirement. Iterative calculations
0 – T
T rm . The values of T clg.rm.slab and with Equation (4) are needed since T inlet.air , t , and V air are
off.rm.slab
T off.rm.slab can be obtained through on-site measurements. interrelated. The maximum flow rate for fan capacity sizing
can be determined using Equation (4) with marginally favor-
[  T rm
0 – T
off.rm.slab 
able T inlet.air and by setting E store equal to E max . Sometimes,
E allow = (6) when E allow is small, the cooling period may be shortened
–  T min.rm – T clg.rm.slab ]  C slab
given minimum flow rate is reached. On the other hand, some-
In buildings with energy-saving measures such as night- times even with high flow rate and longest active cooling time
time set-up (i.e. mechanical cooling is shut off, and the room possible (e.g. from 10 pm at night to 8 am of the next morning,
air temperature is allowed to float up) and adaptive thermal depending on the outdoor air conditions), the largest possible
comfort (ASHRAE 2009b) manuals, T rm 0 can be higher than E store could still be less than E allow . Note that E allow will be
T sp.rm (mechanical space cooling setpoint). It is also possible constant if room air temperature is constant (i.e. mechanical
space cooling is continuous; T rm 0 = T
0 is lower than T
that T rm sp.rm in Equation (6)). If
sp.rm in mild weather conditions.
the cooling duration is set constant, Equation (4) can be further
In the preliminary design stage, optimal C slab can be
simplified, as can the control.
determined using Equation (7), which is derived from Equa-
0 replaced by T
tion (6) with T rm sp.rm , and E allow replaced by
E max (maximum storage of coolness). E max is less than or DISCUSSION
equal to the maximum 24-hr cooling load of the building,
which can be estimated through building thermal modeling The thermal behavior of the VCS under active slab cool-
under design weather conditions. ing is affected by the varying room air temperature, initial slab
temperature, inlet air temperature and flow rate, and cooling
E max duration. Without a complex thermal model of the whole
C slab = ---------------------------------------------------------------------------------------------------------------------------- (7)
 T sp.rm – T off.rm.slab  –  T min.rm – T clg.rm.slab  building and weather forecast data input, the prediction of the
thermal behavior, potential thermal storage capacity, and opti-
C slab is the product of the thickness and the volumetric mal control strategy development are difficult. ASHRAE
heat capacity (J/m3/K (BTU/ft3/°F)) of the slab. If larger slab (2007) reported that advanced optimal strategies often do not
thermal capacity C slab is desired, it can be obtained by result in significant cost-savings benefits as compared to well-
increasing the slab thickness. However, in commercial build- designed basic strategies that rely on simpler control routines.
ing applications, the slab thickness is normally determined by Furthermore, easy application of control is critical in promot-
the structural need and the architectural plan (e.g. the storey ing building-integrated thermal energy storage. In order to
height). Increasing the slab thickness for thermal storage simplify the thermal modeling and hence develop a simple
purpose will occupy extra living space and increase the struc- strategy for the control of the VCS in active slab cooling, the
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
©2012 ASHRAE
565
following two main approximations are adopted in the thermal (4)) with suitable T ref.slab , T clg.rm.slab , and T off.rm.slab .
model used in this paper: These three temperatures can be fine-tuned based on on-site
measurements. Furthermore, since the slab temperature can be
• The varying room air temperature is replaced by the slab estimated using T off.rm.slab and room air temperature, detec-
top surface temperature and the temperature difference tion of a problematic low room air temperature can be used to
between the room air and the slab top surface during the stop the cooling operation. If not enough coolness is stored by
slab cooling operation, T clg.rm.slab . the end of the cooling period, the only drawback is that some
• The average temperature of inlet air is used throughout potential energy saving will be lost. In any case, no thermal
the cooling period. discomfort will occur. The greatest benefit of using this meth-
odology for charge control is that frequent and complex ther-
Based on the simulation results from the simplified mal modeling can be avoided. The simplified thermal
model, a correlation that has only three unknown variables, modeling will be needed only twice: firstly in the preliminary
T inlet.air , t , and V air , is able to estimate E store with accept- design stage to determine the optimal thermal capacity of the
able accuracy. For the approximation of inlet air temperature slab and the fan capacity, and secondly after building construc-
being constant, results from simulations using constant and tion with fine-tuned parameters (e.g. T ref.slab and T ’s).
varying inlet air temperatures can be compared to quantify any
In the design and operation of the active slab cooling,
potential discrepancy. However, since E store is linear with
there are several other important considerations as follows:
constant inlet air temperature, it can be assumed that the
discrepancy will not be significant. Using the average of the • Active slab pre-cooling will not be needed if the weather
forecast air temperature as the constant value will lessen the condition of the following day does not create a cooling
error in the application of Equation (4). The observations of demand or allows free space cooling. Weather forecasts
T off.rm.slab , T clg.rm.slab , and an even temperature distribu- and finding the balance temperature are useful for con-
tion in the slab during and outside the slab cooling period trol.
simplify the estimation of the thermal energy allowed to be • If the outdoor air is not cool enough for efficient slab-
stored. On-site measured T ref.slab , T off.rm.slab , T clg.rm.slab cooling, chiller-assisted active slab cooling (i.e. cooling
and after the completion of construction will improve the the inlet air using chillers) is viable, especially with low-
accuracy of the correlation. The bottom boundary temperature cost, off-peak electrical energy (Braun 2003). Cost-
of the slab considered in this paper is almost constant during related efficiency should also be considered.
the cooling period; however in other cases, this temperature
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
• For buildings with multiple zones, the slab temperatures

may vary during the period, such as the case in which a living
can be different in different zones. The active slab cool-
space is underneath the slab. In this case, another constant
ing has to be controlled separately.
between the room air underneath and the slab can be used. If
• Low floor surface temperature will cause low air tem-
there is a layer of insulation between them, the heat transfer
peratures at the ankle, and hence can cause thermal dis-
can be ignored since the temperature difference between the
comfort (ASHRAE 2004). The lower limit of the surface
slab and the room underneath will not be significant.
temperature of the floor should comply with thermal
Equation (4) is generated with a constant initial slab
comfort standard or occupant’s choice.
temperature, which is referred to as slab reference tempera-
ture, T ref.slab . Meanwhile, the initial slab temperature T slab 0 • When the slabs are covered by indoor objects, such as
will vary depending on the building space cooling strategies furniture or carpets, the passive heat exchange (e.g. radi-
(e.g. setpoint set-up and chillers shutting down during unoc- ation and natural convection) between the room air and
cupied hours). The actual thermal energy that will be stored in the slabs will be reduced. Active discharge can be used
the slab will differ from the E store estimated by Equation (4) to make effective use of the VCS and other TES sys-
0
if T slab is different from T ref.slab . This discrepancy can be tems.
reduced by using average initial slab temperatures as T ref.slab , • Coolness from ground can be a thermal energy source
or creating a set of correlations for different values of T ref.slab . for cooling the thermal storage system through certain
Moreover, the discrepancy will be self-corrected by the slab kind of heat transfer fluid (e.g. air or water). Cost effec-
due to the energy-storing nature of the slab (Shaw et al. 1994). tiveness has to be considered as well.
For example, if T slab0 is higher than T ref.slab before the active
cooling operation, more coolness will be stored since the CONCLUSION
temperature difference between the actual slab temperature This paper presented full-scale measurements and simu-
and the inlet air temperature is larger. If T slab 0 is lower than lation results of the thermal performance of a building-inte-
T ref.slab , less coolness will be stored. grated thermal energy storage (BITES) system – a ventilated
All the errors introduced by the assumptions and simpli- concrete slab (VCS) – with active slab cooling using outdoor
fications will be self-corrected by the slab to a certain extent cool air in a near net-zero solar house. A methodology for
because of the slab’s high thermal capacity and energy-storing sizing and control of VCS that can be used for active slab cool-
nature. The key step is to develop correlations (e.g. Equations ing was presented as well.
566
ASHRAE Transactions
The long-term monitored data show that during the active NOMENCLATURE
slab cooling period, the room air temperature was about 0.9°C BITES = building-integrated thermal energy storage
(1.6°F) ( T clg.rm.slab ) higher than the average slab top surface
CFM = cubic feet per minute
temperature. This difference was about 0.5°C (0.9°F)
( T off.rm.slab ) outside the cooling period. These two impor- CHTC = convective heat transfer coefficient
tant observations are used in the development of a simple ther- FPM = feet per minute
mal model, and in the estimation of the building’s energy TES = thermal energy storage
storage requirement. T clg.rm.slab and T off.rm.slab can be
VCS = ventilated concrete slab
fine-tuned based on on-site measurements after construction
completion. An on-site test under an early-summer nighttime t = time or duration (hr)
outdoor condition was performed and the dynamic tempera- h cr = combined convective and radiative heat
ture distribution of the slab was monitored. The measured data transfer coefficient
are used to validate the thermal model for active slab cooling. C slab = thermal capacity of the slab
The validation shows good agreement between the measured E allow = thermal energy that is allowed to be stored in
and simulated results. The simulation outputs are adequate for the slab
design and control purposes. The validated thermal model is
E max = maximum storage of thermal energy (i.e.
then used to predict the thermal performance of the slab under
coolness in this study)
different operating conditions (e.g. various air flow rates,
temperatures of the inlet air and the soil underneath, and cool- E store = thermal energy (coolness) that will be stored
ing durations) with a initial slab temperature of 21.8°C in the slab
(71.2°F). The simulation for the typical operation summer T inlet.air = temperature of the inlet air entering the slab
condition (cool air at about 200 L/s (424 CFM) and 16°C T min.rm = minimum allowable room temperature
(60.8°F) for 6 hours) shows that about 6.3 kWh (21.5 MBTU) T min.slab = minimum allowable slab temperature (i.e.
(1.26 kWh/m3 of concrete (4.3 MBTU/m3)) of heat can be the lowest slab temperature that slab cooling
extracted by the cool air. Both on-site measurement and simu- wants to achieve)
lations show that the temperature distribution within a typical
T ref.slab = reference slab temperature, used as the initial
section of the slab is quite uniform.
slab temperature in the slab cooling
Based on the simulation results, an approximately linear simulation
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
relationship is observed between the inlet air temperature and 0

T rm = initial room air temperature
the coolness that will be stored in the slab ( E store ), while expo- 0
T slab = initial slab temperature
nential relationship exists between E store and air velocity (i.e.
T sp.rm = mechanical space cooling setpoint
flow rate) or the cooling duration. A correlation was developed
to predict E store under different operating conditions. V air = velocity of air inside the air channel of the
slab
The lower limit of the slab surface and room air temper-
T clg.rm.slab = temperature difference between room air
atures based on comfort requirements define the amount of
and slab during active slab cooling operation
heat can be extracted from the slab. The use of the approximate
temperature differences T clg.rm.slab and T off.rm.slab greatly T off.rm.slab = temperature difference between room air
simplifies the sizing of the BITES system and the estimation and slab outside active slab cooling period
procedure of the daily energy storage requirement to simple
equations. The correlation facilitates the selection of required REFERENCES
flow rate and duration period for the energy storage. The meth- ACI committee 122. 2002. 122R-02: Guide to thermal prop-
odology of sizing and control presented in this paper is appli- erties of concrete and masonry systems. In: American
cable to other types of BITES systems in different kinds of Concrete Institute (ACI), ed. Manual of Concrete Prac-
buildings with appropriate modifications. tice. Detroit, Mich, US: American Concrete Institute.
ASHRAE. 2004. Standard 55: Thermal environmental con-
ACKNOWLEDGMENT ditions for human occupancy, Atlanta, GA, USA, Amer-
ican Society of Heating Refrigerating and Air-
Canada Mortgage and Housing Corporation (CMHC) Conditioning Engineers (ASHRAE).
conducted the EQuilibrium demonstration program. The ASHRAE. 2007. Thermal Storage. ASHRAE Handbook -
house was built by Alouette Homes, an industrial partner of HVAC Applications. SI ed. Atlanta, GA, US: American
the Canadian Solar Buildings Research Network. Financial Society of Heating Refrigerating and Air-Conditioning
support of this work was provided by Natural Sciences and Engineers (ASHRAE).
Engineering Research Council (NSERC) of Canada through ASHRAE. 2008. Panel heating and cooling. ASHRAE Hand-
the Solar Buildings Research Network (SBRN). book - HVAC Systems and Equipment. SI ed. Atlanta,
©2012 ASHRAE
567
GA, US: American Society of Heating Refrigerating Howard, B. D. and H. Fraker. 1990. Thermal energy storage
and Air-Conditioning Engineers (ASHRAE). in building interiors. In: B. Anderson, ed. Solar Building
ASHRAE. 2009a. Heat transfer. ASHRAE Handbook - Fun- Architecture. Cambridge, MA, USA: MIT press.
damentals. SI ed. Atlanta, GA, USA: American Society Inard, C., A. Meslem and P. Depecker. 1998. Energy con-
of Heating, Refrigerating, and Air-Conditioning Engi- sumption and thermal comfort in dwelling-cells: A
neers (ASHRAE). zonal-model approach. Building and Environment,
ASHRAE. 2009b. Thermal comfort. ASHRAE Handbook - 33(5): 279-291.
Fundamentals. SI ed. Atlanta, GA, USA: American Karadag, R. and O. Akgobek. 2008. The prediction of con-
Society of Heating, Refrigerating, and Air-Conditioning vective heat transfer in floor-heating systems by artifi-
Engineers (ASHRAE). cial neural networks. International Communications in
Bilgen, E. and M. A. Richard. 2002. Horizontal concrete Heat and Mass Transfer, 35(3): 312-325.
slabs as passive solar collectors. Solar Energy, 72(5): Kreith, F. and M. Bohn. 2001. Principles of heat transfer,
405-413. Australia; Pacific Grove, Calif., Brooks/Cole Pub.
Braham, G. D. 2000. Mechanical ventilation and fabric ther- Lehmann, B., V. Dorer and M. Koschenz. 2007. Application
mal storage. Indoor and Built Environment, 9(2): 102- range of thermally activated building systems tabs.
110. Energy and Buildings, 39(5): 593-598.
Braun, J. E. 2003. Load control using building thermal mass. Morgan, S. and M. Krarti. 2007. Impact of electricity rate
Journal of Solar Energy Engineering-Transactions of structures on energy cost savings of pre-cooling controls
the ASME, 125(3): 292-301. for office buildings. Building and Environment, 42(8):
Candanedo, L. M., A. K. Athienitis, J. A. Candanedo, W. 2810-2818.
O'Brien and Y. X. Chen. 2010. Transient and Steady Patankar, S. V. 1980. Numerical heat transfer and fluid flow,
State Models for Open-Loop Air-Based BIPV/T Sys- New York, USA, Hemisphere Pub. Corp.
tems. ASHRAE Transactions, 116(1): 600-612.
Ren, M. J. and J. A. Wright. 1998. A ventilated slab thermal
Chen, Y., A. K. Athienitis and K. Galal. 2010a. Modeling, storage system model. Building and Environment, 33(1):
design and thermal performance of a BIPV/T system 43-52.
thermally coupled with a ventilated concrete slab in a
Shaw, M. R., K. W. Treadaway and S. T. P. Willis. 1994.
low energy solar house: Part 1, BIPV/T system and
Effective use of building mass. Renewable Energy, 5(2):
house energy concept. Solar Energy, 84(11): 1892-1907.
1028-1038.
Chen, Y., K. Galal and A. K. Athienitis. 2010b. Modeling,
Winwood, R., R. Benstead and R. Edwards. 1997. Advanced
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
design and thermal performance of a BIPV/T system
fabric energy storage I: review. Building Services Engi-
thermally coupled with a ventilated concrete slab in a
neering Research and Technology, 18: 1-6.
low energy solar house: Part 2, ventilated concrete slab.
Solar Energy, 84(11): 1908-1919. Xu, X., S. Wang, J. Wang and F. Xiao. 2010. Active pipe-
Dincer, I. 2002. On thermal energy storage systems and embedded structures in buildings for utilizing low-grade
applications in buildings. Energy and Buildings, 34(4): energy sources: A review. Energy and Buildings, 42(10):
377-388. 1567-1581.
Feustel, H. E. and C. Stetiu. 1995. Hydronic radiant cooling - Zhu, N., Z. J. Ma and S. W. Wang. 2009. Dynamic character-
preliminary assessment. Energy and Buildings, 22(3): istics and energy performance of buildings using phase
193-205. change materials: A review. Energy Conversion and
Hadorn, J.-C. (ed.) 2005. Thermal Energy Storage for Solar Management, 50(12): 3169-3181.
and Low Energy Buildings: International Energy
DISCUSSION
Agency (IEA) Solar Heating and Cooling Task 32 -
Advanced storage concepts for solar and low energy Hugo Hens, Professor Emeritus, University of Leuven,
buildings, IEA. and BPh Consult bvba, Leuven, Belgium: 1) Romans used
Henze, G. P., C. Felsmann, D. E. Kalz and S. Herkel. 2008. the smoke of the fire in the heating room to warm the floors,
Primary energy and comfort performance of ventilation not hot air (Hypocaustus). 2) Be careful with corrosion of the
assisted thermo-active building systems in continental corrugated steel plates. 3) What with feet comfort?
climates. Energy and Buildings, 40(2): 99-111. Yuxiang Chen: The slab is normally warmer than the air
Howard, B. 1999. The CMU air-core passive hybrid heat flowing through. There will be no condensation due to air
storage system. Proceedings of the Renewable and temperature drop. The surface temperature of the slab can be
Advanced Energy Systems for the 21st Century, April: controlled according to ASHRAE thermal comfort guide-
11-15. lines.
568
ASHRAE Transactions
SA-12-021 (RP-1365)
Thermal Performance of Building Envelope

Details for Mid- and High-Rise Buildings
Patrick Roppel, PEng Mark Lawton, PEng Neil Norris

Member ASHRAE
ABSTRACT quate condensation resistance at intersections of components

ASHRAE Research Project 1365 (RP-1365) had the and compromised occupant comfort.
objective to provide thermal performance data of 40 common ASHRAE Research Project 1365 (RP-1365) had the
building envelope details for mid- and high-rise construction. objective to provide thermal performance data of 40 common
The goal was to develop procedures and a catalogue that will building envelope details for mid- and high-rise building
allow designers quick and straightforward access to informa- construction. The goal of the project was to develop proce-
tion but with sufficient complexity and accuracy to reduce dures and a catalogue that will allow designers quick and
straightforward access to information but with sufficient
uncertainty in the thermal performance of building envelope
complexity and accuracy to reduce uncertainty in the thermal
components.
performance of building envelope components.
This paper provides an overview of RP-1365 and focuses
The principal audience for the project was the users of the
on how the procedures and building envelope thermal perfor-
ASHRAE Handbook – Fundamentals; however, the informa-
mance catalogue can be utilized in practice for users of the
tion advanced by the project should be of direct interest to the
ASHRAE Handbook – Fundamentals and developers of ANSI/
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
developers and users of energy standards, such as ANSI/

ASHRAE/IESNA Standard 90.1.
ASHRAE/IESNA Standard 90.1. The procedures and cata-
INTRODUCTION logue produced for RP-1365 provide simple and straightfor-
ward options to consider heat flow through major thermal
The reduced thermal resistance due to thermal bridging bridges such as floor edge conditions, intersections of assem-
through steel and concrete framing can have a significant blies, or structural penetrations, not currently addressed by
impact on the whole building energy performance. Designers Standard 90.1.
of buildings are starting to respond by trying to better predict A principal advancement of RP-1365 is that thermal
the heat loss through the building envelope by utilizing proce- performance data is available for a wide spectrum of common
dures outlined in the ASHRAE Handbooks and two-dimen- and generic building envelope details considering three-dimen-
sional steady-state heat transfer software to determine the sional heat flow; however, a significant benefit is the method-
effective thermal resistance of assemblies. However, these ology of calculating and reporting thermal transmittances.
methods require many assumptions to estimate the thermal Thermal transmittance was calculated and reported by three
performance of complex three-dimensional architectural categories of thermal anomalies: clear field, linear, and point
details at intersections of walls, windows and roofs. This leads anomalies. Linear and point anomalies are reported using linear
to potentially large uncertainty about the thermal performance and point transmittances. This is a departure from common
of the building envelope, which can lead to inefficient design practice in North America of using only an area weighted aver-
of HVAC systems, building operation inefficiencies, inade- age approach. Linear and point transmittances were utilized to
Patrick Roppel is an associate and building science specialist, Mark Lawton is a principal and senior building science specialist, and Neil
Norris is a building science consultant in the Buildings, Technology, and Energy division at Morrison Hershfield Limited in Vancouver, BC,
Canada.
©2012 ASHRAE
569
overcome drawbacks with the area weighted average method 1. To be relevant to ASHRAE/IES Standard 90.1;
for anomalies contained within opaque assemblies. 2. To be relevant to existing and future building stock and
This paper provides an overview of ASHRAE Research capture both retrofit and new construction details;
Project 1365 and focuses on how the procedures and building 3. Represent both high thermal performance envelopes and
envelope performance catalogue can be utilized in practice, standard building practice;
particularly the developers and users of Standard 90.1. 4. Represent typical interior finishing and cladding systems
and attachment methods.
HEAT TRANSFER MODEL AND CALIBRATION
Selection of details was based on the following method-
Modeling for this project was done using a three-dimen- ology.
sional finite element analysis heat transfer software package
by Siemens PLM Software; FEMAP and Nx, with Maya’s a. Highest priority details have thermal bridges in three-
TMG thermal solver. The ability of the software and tech- dimensions;
niques used to predict conductive thermal performance of b. Focus on details not already addressed in ASHRAE pub-
building envelope details containing high conductivity (non- lications (build on past work);
insulating) thermal bridges was demonstrated by calibrating c. Include clear wall and two-dimensional intersections
and benchmarking against measured public-domain thermal when not already effectively addressed by ASHRAE pub-
performance data and deterministic analytical solutions. lications;
d. Details should include intersections throughout a typical
Validation exercises in ISO standards were modeled to
building, i.e. at parapets, opaque walls to glazing, inter-
demonstrate that the TMG solver will yield high precision
secting walls, and balconies.
results for well-defined problems.
Numerous well documented cases with guarded hot-box The calibration exercises highlighted some important
test measurements from the literature, 29 in total, were simu- considerations for modeling generic assemblies and details; it
lated. Simulated results were compared to the guarded hot-box was deemed important to not get lost in the details when produc-
data sets to demonstrate that the techniques and procedures to ing the catalogue. Assemblies were modeled in a generic
produce the catalogue and the TMG model will accurately manner, when possible, that embodied small differences from
reproduce test measurements using measured (or typical) mate- less significant factors. Examples of factors deemed less signif-
rial properties and consistent boundary conditions. Good agree- icant are steel gauge thickness and fasteners into sheathings.
ment between simulated and measured thermal performance for These factors were incorporated into each assembly, but in a
both steady-state and transient conditions was found. Good consistent approach for every assembly without micro analysis.
agreement was defined by the simulated results being within Details were allocated by wall type: steel stud, poured-in-
±8% of the reported measured results, which is based on the place concrete, glazing spandrel section, precast concrete and
reported accuracy of the guarded hot-box method and previ- sandwich panels, and concrete block. Details varied by type of
ously used to demonstrate the accuracy of thermal modeling for intersection of components, variations in insulation strategy,
similar projects (Kosny and Desjarlais 1994, Kosny and Childs and cladding attachment and support.
2000, Kosny, Christian, Desjarlais 1998, Kosny 1995 & 1994, An example data sheet for one detail can be found in
Enermodal Engineering 1996). Nevertheless, the simulated Appendix A. The data sheets identify all materials, dimen-
thermal resistances were on average lower by 1% than the sions, and material properties necessary to evaluate the ther-
measured thermal resistances for 24 steel stud assemblies1. mal performance of the detail.
Sensitivity analysis showed the importance of assuming appro-
priate values for contact resistance particularly between steel MODELING PROCEDURES
flanges and sheathings when modeling steel framed assemblies.
The modeling procedures followed from the calibration
The final report for RP-1365 provides an extensive
exercises and are intended to produce thermal performance data
discussion and comparison between simulated results and
that will allow designers quick and straightforward access to
guarded hot-box data sets.
information, but with sufficient complexity and accuracy to
reduce uncertainty in the thermal performance of building enve-
BUILDING ENVELOPE DESIGN DETAIL
lope components. Moreover, the procedures are intended to
CATALOGUE
produce results that will answer the fundamental questions of
The catalogue of 40 building envelope design details was how overall geometry and materials affect the overall thermal
developed based on a set of generic and common interface performance.
details suitable for mid- and high-rise construction. The objec- The calculated thermal performance data includes thermal
tives of the catalogue are: transmittances and indexed surface temperatures. The focus of
the remainder of the paper is on the primary benefit of the proj-
1.
The thermal resistance was measured for two temperature differ- ect, which is the methodology of calculating and reporting ther-
ences for some assemblies. The higher reported thermal resistance mal transmittances. Refer to the RP-1365 final report for a
was selected for comparison and calculating percent difference. complete discussion of the general modeling principles (cut-off
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
570
ASHRAE Transactions
planes, contact resistance, material properties, air cavities, envelope, including multiple assemblies (walls, windows,
boundary conditions) and indexed surface temperatures. roofs, floors types), using an area weighted average approach.
The impact of a thermal bridge on thermal transmittance is
THERMAL TRANSMITTANCE assigned to an “area of influence”. The difficulty is that assign-
The basis of calculation and reporting of thermal trans- ing “effective” lengths (or area of influence) to complex three-
mittances is established by three categories of thermal anom- dimensional heat flow paths can be an arbitrary or complex
alies as summarized below and presented in Figure 1. process when there are strong lateral heat flow paths. An alter-
For each linear or point thermal anomaly (slab, shelf native is to assign the heat loss due to major thermal bridges
angle, parapet, beam, etc.) the thermal transmittance for the to a line or point of no width (i.e. of no area) and calculating
whole assembly (U-value) was compared to the thermal trans- overall heat loss by summing heat loss through the clear field
mittance without the thermal anomaly (clear field value, Uo). (area) to the linear thermal anomalies and the point thermal
Thermal transmittances were simulated using steady- anomalies. If needed the overall U-value of an enclosure area
state boundary conditions, constant material properties, and of interest can be calculated by dividing the heat loss for linear
contact resistances. Reported thermal transmittances are air- and point transmittances by the overall projected enclosure
to-air values in order to provide consistent values that are not area. Utilizing linear and point thermal transmittances is
dependent on three-dimensional surface temperatures, i.e.
uncomplicated and well suited to quantify all thermal anom-
surface resistances are included in the thermal transmittances
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
alies that interrupt clear opaque field assemblies.
as outlined in the previous section.
Currently, it is common practice in North America to Linear and point transmittances are discussed in more detail
calculate the overall thermal transmittance of the building later in this section. First, a discussion about the drawbacks of
Clear Field Anomalies – thermal bridges uniformly distributed by a sufficient

amount such that they can be assumed to modify the thermal transmittance of the
assembly (area based) and are considered not practical to account for on an individual
basis.
Examples are brick ties, girts supporting cladding, and structural framing. A steel
stud assembly with horizontal z-girts is shown to the left as an example.
Linear Anomalies – thermal bridges that are continuous and/or uniformly distrib-
uted typically along a considerable portion of a building perimeter or height in one
dimension.
Examples are shelf angles, slab edges, balconies, corner framing, parapets, and
window interfaces. A steel stud assembly with a structural concrete slab bypassing
the wall assembly is shown to the left as an example.
Point Anomalies – thermal bridges that are countable points and are considered
feasible to account for on an individual basis for whole building calculations.
Examples are three way corners, structural steel penetrations through insulation,
ducts. A steel stud assembly with a structural steel beam (to support a canopy, sign,
etc.) bypassing the wall assembly thermal insulation is shown to the left as example.
Figure 1 Illustrations of Example Thermal Anomalies.
©2012 ASHRAE
571
assigning and utilizing “effective” lengths (or area of influence)
for the thermal performance catalogue follows below.
The effective length is the distance from a thermal anomaly
where heat flow is affected by the thermal bridge. In other
words, it is the distance, on a section, between the points where
the clear field heat flow is not significantly affected by the
anomaly. Effective lengths are typically assessed by comparing
the heat flow through projected areas of the interior and exterior
surfaces at varying distances away from the thermal anomaly to
the clear field heat flow. The procedure of establishing effective
lengths using a thermal model is simple enough for straightfor-
ward section details. Heat flows can be evaluated at any surface
by a heat transfer model. The challenge is that the calculated
effective length is dependent on the location of insulation rela-
tive to strong lateral heat flow paths.
Figure 2 illustrates the detail of a concrete parapet with an
interior insulated wall and exterior insulated roof showing
“bands” of area or control areas where the heat flow can be eval- Figure 2 Illustration of hypothetical effective lengths and
uated and compared to the clear field heat flow. Figure 3 shows area of influence of a concrete parapet thermal
the heat flow through the interior surface and the exterior anomaly.
Figure 3 Heat flow through the interior surface and the exterior surface of control areas, dependent on the distance from the
roof/wall junction.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
572
ASHRAE Transactions
surface of control areas, dependent on the distance from the These drawbacks can be partially overcome by setting
roof/wall junction. standard effective lengths for groups of anomalies. However,
For the wall assembly, an effective length of approxi- this simplification leads to the assigned value being arbitrary
mately 200 mm (8 inches) is calculated using the interior without real significance. This leads to the concept of linear
surface heat flow compared to an effective length of approxi- and point transmittances, which avoids the complication of
mately 850 mm (33 inches) using the exterior surface heat assigning “affected area” by simply defining it as zero.
flow2. A similar, but reversed, difference is evident for the roof
portion which was calculated to be 1400 mm (55 inches) using
the interior surface heat flows. In contrast, the effective length
using the exterior surface heat flows is approximately 700 mm
(26 inches). The reason for this difference is that the concrete
is highly conductive compared to the thermal insulation;
therefore, there is significant lateral heat flow through the
concrete to/from the concrete parapet thermal anomaly.
While the interior surface typically defines the boundary
of zones for building load calculations, the parapet example
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
above shows how assigning the effective length based on the
interior surface heat flow alone does not provide a consistent
and usable definition of effective length in many common situ-
ations. Furthermore, this theoretical procedure does not work
well for assemblies where the clear field contains complex
three-dimensional heat flow paths. For a slab edge condition
(exposed concrete or shelf angle) with an exterior insulated
assembly and steel passing through the insulation parallel to
the slab (brick ties or horizontal girts); the heat flow through
the exterior surface of the wall cladding is not affected much
by the slab edge thermal anomaly. Figure 4 illustrates the ther-
mal gradients of a slab edge thermal anomaly.
Ultimately the logical location to assign an effective length
for assemblies with complex heat flow paths may be at the
midpoint between major thermal anomalies. When there are Figure 4 Slab edge (shelf angle) thermal anomaly with
multiple thermal bridges bypassing the thermal insulation, complex three-dimensional heat flow paths.
assigning effective lengths and utilizing the area-weighted aver-
age method can get very awkward. For example, defining the
effective area for a steel post and beam contained in a steel stud
assembly with horizontal z-girts supporting the metal cladding
(see Figure 5) is either a very challenging or arbitrary process
and using this information with the area-weighted average
method for whole building load calculations is tedious.
This work highlights a few major drawbacks for using the
area-weighted average method to account for thermal anom-
alies contained within opaque building envelope assemblies:
1. Assigning effective areas to three-dimensional heat flow

paths can be a complex and, frankly, arbitrary procedure.
2. A catalogue of effective areas will be specific to simu-
lated details and not easily transferable to other cases.
3. Utilizing the area-weighted average method to account
for all the significant thermal anomalies of typical build-
ings is a tedious process at best.
2.
Distance was determined where the thermal anomaly has a negli-
gible effect (less than 2% difference) on the assembly heat flow by
comparing the heat flow through sections away from the anomaly Figure 5 Steel Post and Beam Point Anomaly where
to the clear field heat flow. assigning an area of Influence is not practical.
©2012 ASHRAE
573
First, one has to accept that there cannot be effective is the total projected surface area (ft2 or m2). This step is
lengths specific to each simulated detail and simplifications not necessary if the thermal anomaly is contained within
are necessary for a generic thermal performance catalogue to a single assembly.
be useful for practitioners. The effective length might as well 4. Simulate a section that contains a length of the linear ther-
be zero. mal anomaly and a good portion of clear field assemblies
Second, one has to be open to the idea of letting go of the on each side.
area-weighted average method, at least for thermal anomalies 5. Determine the overall transmittance for the section with
contained within opaque building envelope assemblies, if the thermal anomaly, U (Btu/h•ft2•°F or W/m2 K)
another method provides the same end result but with a lot less 6. The difference between the simulated total heat flow and
effort. overall clear field heat flow for a set area is the heat flow
The concept of linear and point thermal transmittances ( that is ascribed to the linear transmittance. Equations
or  respectively) is quite simple. The heat flow through an summarizing the procedures follow.
assembly with a thermal anomaly is compared to the same
assembly without the thermal anomaly for the same gross The linear and point thermal transmittances are deter-
area. The difference in the heat flow is attributed to the linear mined by the following equations using data from the simu-
length of the anomaly (or point). For overall heat flow calculated details.
lations, the linear and point transmittances is simply added to
the clear value U-value for a given total assembly area (wall/  Q – Q0  A total
 = ---------------------
- =  U – U o   -------------- (1)
roof) to determine the overall thermal transmittance. Splitting T  L L
the total assembly area into areas affected by individual ther-
mal anomalies and clear field areas is not necessary.  Q – Q0 
 = ---------------------
- =  U – U o   A total (2)
Figure 6 provides a simple illustration of the procedures T
to calculate linear transmittances utilized in RP-1365, using a
parapet as an example. These procedures are loosely based on Where
European literature and standards (Anderson 2006, ISO  is the linear thermal transmittance (Btu/hft°F or W/m K)
10211: 2007 (E)).  is the point thermal transmittance (Btu/h°F or W/K)
Q is the assembly heat flow with the thermal anomaly for
1. Establish the thermal transmittance of the clear field
Atotal (Btu/h or W)
assemblies, Uo, wall, Uo, roof (Btu/h•ft2•oF or W/m2 K).
2. Calculate the individual clear field heat flow, Qo, wall, Qo, U is the assembly U-value with the thermal anomaly (Btu/
hft2°F or W/m2 K)
roof (Btu/h or W) for the assembly.
3. Calculate the overall clear field transmittance, Uo, = (Qo, Qo is the assembly heat flow without the thermal anomaly
wall + Qo, roof )/ T*Atotal (Btu/h•ft • F or W/m K). Atotal
2 o 2
for Atotal (Btu/h or W)
Figure 6 Clear field and assembly heat flows necessary for linear transmittance calculations.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
574
ASHRAE Transactions
T is the temperature difference (°F or °C) The clear field U-values calculated for RP-1365 are simi-
Uo is the assembly U-value without the thermal anomaly lar to the assembly U-values in Appendix A of ASHRAE 90.1
(Btu/hft2°F or W/m2 K) (2007) with one important exception. The values in the tables
of ASHRAE 90.1 include an allowance for heat flowing
Atotal is the total projected surface area (ft2 or m2)
through the framing element that attaches to the slabs (top and
THERMAL TRANSMISSION DATA bottom tracks). The clear field U-values calculated in RP-1365
do not include top and bottom tracks, which are accounted for
In the RP-1365 final report, the thermal performance data in the linear transmittance values at the slab as discussed previ-
of the 40 building envelope details are summarized in simula- ously and in the next section3.
tion result data sheets for each detail. An example result data
sheet can be found in Appendix A. Linear and Point Transmittances
Clear Field U-Value, Uo The results catalogue includes linear (value) and point
(value) transmittances for slab edges, parapets, glazing
Clear field U-values were calculated for two purposes in transitions, corners, interior wall intersections, concrete curbs
RP-1365; reference values to calculate the linear or point trans- on roofs, and panel joints for a variety of common assemblies.
mittance and to quantity the impact of thermal bridges uniformly Linear transmittances were calculated for a number of
distributed by a sufficient amount such that they can be assumed increments of insulation levels for many of the clear field
to modify the area based thermal transmittance (U-value) of the assemblies. It is worth noting that when the thermal resistance
assembly. For example brick ties, girts supporting cladding, and of the clear field assembles is at the higher end of the cases we
structural framing. These repeated and prevalent thermal bridges modelled, the linear transmittance is fairly constant4.
can have an important impact on assembly thermal transmit- However, when the thermal resistance of the clear field assem-
tance. As an example of this, clear field U-values were calculated blies was relatively low and the assemblies has strong lateral
for a number of exterior insulated steel stud walls with various
cladding attachments and several increments of insulation ther-
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
3.
The top and bottom tracks can contribute to around a 10% differ-
mal resistance (increments of R5 to R25 (Btu/h ft2 oF, 0.88 to 4.4
ence for a 8 foot (2.4 m) high steel stud assembly without exterior
W/m2 K) nominal thermal insulation). The results are presented insulation and the difference diminishes with increasing insula-
in Figure 7 in terms of nominal insulation and assembly effective tion outboard of the steel studs.
R-values to illustrate the impact of diminishing returns. Figure 7 4.
Less than 10% difference in linear transmittance for clear field
shows the diminishing returns of adding insulation to walls with assemblies with nominal insulation ranging from R-5 to R-25 Btu/
the same cladding attachment methods. h•ft2•oF (0.88 to 4.4 W/m2 K).
Figure 7 Diminishing returns of clear field u-values for exterior insulated steel stud assemblies with different cladding
attachments.
©2012 ASHRAE
575
heat flow paths (such as steel studs attached to the slab) the concrete block back-up wall, and R-15 (2.6 RSI) exterior insu-
calculated linear transmittance showed more variation and lation; assembly clear field U-value of 0.056 h•ft2•oF / Btu (0.32
even non-linear variation with increasing insulation. The vari- m2 K/W). There are 25% balconies across the elevation for each
ation can be traced to the presence of lateral heat flow paths floor level. There are a few possibilities for slab edge conditions
(often in three dimensions). Adding insulation to stop heat and values can be utilized to quickly evaluate the impact of
flow perpendicular to the plane of the assembly can force addi- slab edge detailing at the balconies and field area. Tables 1A and
tional heat laterally to the thermal bridge. 1B illustrate the potential impact of several possibilities of slab
The variability of the value is important for awareness edge detailing on the overall heat flow through an example
but does not appear to be a significant factor when using  opaque wall assembly for one floor level 5.
value in practice for evaluating major thermal anomalies at
intersecting components; either a maximum or average  Conventional Curtain Wall Spandrel Panel Thermal
factor can be assigned to details, with reasonable accuracy, Transmittance
independent of the level of insulation of the adjacent assem-
blies. The ability to assign factors to details that facilitate the Another sample detail modelled for RP-1365 was a
analysis of the impact of detailing on overall heat flow, rather conventional, stick-built curtain wall spandrel panel at a floor
than only assemblies, can be an important step towards more
thermally efficient building practice. 5.
Note that this evaluation can be easily extended to include other
For example, consider a wall elevation, 100 ft (30.5 m) intersecting components, for example at glazing, parapet, corner
length, 9 ft (2.7 m) floor to floor height, with brick veneer, intersections, for an entire elevation
Table 1A. Increase in Heat Flow Due to Slab Edge Detailing
Slab Edge Condition

Balcony Location Non-Balcony Location
Length (L) is 25 ft, 7.6 m Length (L) is 75 ft, 22.9 m
Balcony and Floor Slab Flush Shelf Angle Shelf Angle
Continuous Slab Floor Slab with Brick Attached Directly Detached from
Separated Veneer to Slab Floor Slab
Linear Transmittance (
0.36 (0.62) 0.13 (0.22) 0.34 (0.59) 0.26 (0.45) 0.18 (0.31)
Btu/ft•hr•oF (W/m K)
Increase in Heat Flow due to
Detail per location,
9.0 (4.7) 3.3 (1.7) 25.5 (13.5) 19.5 (10.3) 13.5 (7.1)
Qdetail /T
Btu/hr•oF (W/K)*
*
Qdetail/T = L*
Table1B. Increase in Overall Wall Heat Flow Due to Slab Edge Detailing
Total Heat Flow Clear Field Overall Heat Overall Ther-

due to Details, Heat Flow, Qo/ % Flow, U mal Resistance
Scenario
Qdetail/T T Btu/hr•oF Qdetail /Qtotal‡ Btu/ft2•hr•oF ft2•hr•oF/Btu (m2
Btu/hr•oF (W/K)* (W/K)† (W/m2 K) K/W)
Continuous Balcony & Flush
34.5 (18.2) 50.4 (26.8) 41 0.094 (0.54) 10.6 (1.9)
Floor Slab
Continuous Balcony Slab & Shelf
Angle Attached Directly to Floor 28.5 (15.0) 50.4 (26.8) 36 0.088 (0.50) 11.4 (2.0)
Slab
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Continuous Balcony Slab & Shelf

22.5 (11.8) 50.4 (26.8) 31 0.081 (0.46) 12.3 (2.2)
Angle Detached from Floor Slab
Separate Balcony Slab & Shelf
16.8 (8.8) 50.4 (26.8) 25 0.075 (0.42) 13.4 (2.4)
Angle Detached from Floor Slab
*
Total Heat Flow due to details is the addition of Qdetail from table 1A for the balcony and non-balcony locations for corresponding scenario
†
Qo/T = U*A = 0.056 h•ft2•oF / Btu (0.32 m2 K/W)* 900 ft2 (83.6 m2)
‡
Qtotal = Qo + Qdetail
576
ASHRAE Transactions
slab intersection with and without an additional 2 inches little available information that considers the complex three-
(50 mm) of closed cell spray foam insulation inboard of the dimensional heat flow paths past the insulation layers.
back-pan. Both systems were evaluated for varying levels of Figure 9 shows that adding insulation in the back-pan has
insulation in the back-pan R5 to R25 (Btu/h ft2 oF, 0.88 to 4.4 a similar diminishing rate of returns for both systems since the
W/m2 K) nominal thermal insulation. Figure 8 illustrates the aluminum structural members bypasses the back-pan insula-
modeled curtain wall system spandrel panel at a floor slab tion. Furthermore, adding R-11.5 (2 inches or 50 mm of foam)
intersection. insulation to the system only resulted in an increase in overall
(effective) thermal resistance of about R-4. This is disappoint-
This comparison is of particular interest because it is a ing but not surprising given the significant amount of heat flow
common element in mid- and high-rise construction and has that can flow through elements bypassing the thermal insula-
remarkably high thermal transmittance. Designers are increas- tion; slab, anchors, and perimeter stud framing. Table 2 and
ingly calling for the addition of insulation inboard of the back- Figure 9 summarize these results.
pans to improve thermal performance to meet energy stan- Spandrel panels are an integral part of many glazing
dards for opaque elements of the enclosure. However, there is systems, but are often considered an opaque assembly and are
treated as any wall assembly in load calculations. The trans-
mittance through spandrel sections of glazing systems are
highly dependent on the spacing of mullions, have compli-
cated three-dimensional heat flow paths when considering
slab edge effects, and are difficult to separate from the vision
sections. The thermal performance of spandrel panels was
identified in RP-1365 as an area where considerable more
work can be done to establish generic thermal performance
data for curtain and window wall spandrel sections.
USING THE PROCEDURES AND DATA IN

PRACTICE
Utilizing the procedures and data from RP-1365 is
straightforward. The separation of the clear field transmit-
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
tance for assemblies from the linear and point transmittance
values for details allows the contribution of heat flow for each
major thermal bridge to be added to the heat flow to the base
Figure 8 Modeled curtain wall system spandrel panel at assembly heat flow. The procedures and data contained in RP-
floor slab. 1365 do not require information about the area of influence
Figure 9 Diminishing returns of insulation in spandrel back-pan.
©2012 ASHRAE
577
Table 2. Overall Thermal Transmittance of Curtain Wall Spandrel Section [Btu/hr ft2 °F (W/m2 K)]
Back-Pan Insulation R-Value (RSI)
R-4 R-8 R-18 R-28
Un-insulated Stud Cavity 0.29 (1.66) 0.24 (1.35) 0.21 (1.19) 0.20 (1.14)
Spray Foam in Stud Cavity 0.14 (0.77) 0.12 (0.69) 0.11 (0.64) 0.11 (0.63)
and weighted averages to calculate the overall heat flow hr∙ft2∙oF or W/m2K
through the building envelope. This simplifies calculations for
Atotal = total projected surface area (ft2 or m2)
determining the heat loss through whole building elevations
because less information is needed and areas of influence do The procedures and data outlined in RP-1365 should be
not need to be calculated, which can be complicated for 3D utilized in further development of North American energy
intersections with overlapping effects. The following equation standards, such as ANSI/ASHRAE/IESNA 90.1. The proce-
illustrates how the overall heat flow can be calculated using dures outlined in this paper can provide more consistency for
clear field, linear, and point transmittances.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
accounting for thermal bridging in energy standards, reduce

uncertainty on the intent of requirements, and ultimately
Q = Q anomalies + Q o
(3) should lead designers to consider the significance of building
=     L   T +      T + Q o envelope details. Discussion of how the output of RP-1365 can
be incorporated into codes and standards and example calcu-
Where: lations using these procedures follow.
Q = overall heat flow through envelope including
anomalies, Btu/hr or W Incorporating Linear Transmittance into Codes
Qanomalies = additional heat flow for the anomalies (details), Btu/ and Standards
hr or W
Qo = clear field heat flow without anomalies (assembly), The difficulty in establishing a consistent approach for
Btu/hr or W accounting for and minimizing thermal inefficiencies has
likely restrained the full impact of thermal bridging from
T = temperature difference across assembly, oF or oC
being thoroughly addressed in North American energy stan-
 = linear transmittance of anomalies (details), Btu/hr∙ dards. For example, in ASHRAE 90.1 (2007), the compliance
ft∙oF or W/m K
paths are largely silent on major thermal bridges, like slab
 = point transmittance of anomalies (details), Btu/hr∙ oF edges, shelf angles, parapets, etc. and do not provide a consis-
or W/K tent approach to consider their effects. Unfortunately, this
L = the characteristic length of anomaly (ft or m) leads to confusion with designers on the intent of the require-
Note that the Qo for a unit temperature difference can be ments and often can lead to these features being ignored in
calculated by multiplying the tabulated clear field thermal design.
transmittances (Uo) total by the projected surface area of the
For example, for a mass concrete wall with insulation
envelope (Q/T=UA).
installed inboard the concrete can have continuous insulation
For many applications, such as whole building energy
in the clear field with a low U-value but typically has a large
modeling or specifications, the ultimate result maybe to
thermal bridge at the slab. Contrast this to an exterior insulated
determine the overall heat flow per unit area or U-value. An
steel stud assembly with z-girts penetrating exterior insulation
overall U-value can be calculated using the clear field U-
requiring more insulation to meet the same clear field U-value
value and linear and point transmittances using the following
formula. but easy to insulate the slab edge. The additional heat flow at
an un-insulated slab edge for a concrete wall can account for
  L +  a significant portion of the overall wall heat flow compared to
U = ---------------------------------------- + U o (4)
A Total a small amount of additional heat flow for a steel stud assem-
bly where the slab edge is insulated. This can result in the
Where: concrete wall having a much higher overall U-value than the
U = overall thermal transmittance including anomalies, steel stud assembly. Example calculations demonstrating the
Btu/hr∙ft2∙oF or W/m2K potential impact of details is presented in the following
Uo = clear field thermal transmittance (assembly), Btu/ section.
578
ASHRAE Transactions
With the procedures and data outlined in RP-1365, the
importance of the details can be easily quantified and can
encourage thermally efficient design at the detail level. By
integrating this methodology into codes and standards, the
codes themselves will be able to more comprehensively
address their own intent, which is to maintain a minimum level
of thermal efficiency in buildings. This can be achieved by a
few different approaches using the procedures outlined in RP-
1365:
a. Performance Approach: limit overall U-value for all
thermal bridging. Provide information in the form of clear
field transmittance (Uo), linear transmittance () and
point transmittance () and guidance to users to thor-
oughly consider the effects of thermal bridging.
b. Prescriptive Approach: limit the clear field assembly U-
value (Uo) and transmittances for details (- and -fac-
tors) for different types of construction. For example as
proposed by Janssens et al (2007).
c. Solution Approach: provide acceptable solutions,
including details and assemblies, considering all thermal
bridging for typical construction.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Ultimately a combination of all these approaches can be

incorporated in standards. Different building constructions
will then have a more even means of comparison when all their Figure 10 Brick veneer slab edge thermal anomaly.
details are taken into account and thermal efficiency can be
looked at objectively and consistently. This will allow stan-
dards to be more enforceable (and therefore likely more
widely adoptable) and greatly encourage minimizing thermal Ab = 0.6185 m •15.2 m = 9.4 m2 (101.2 ft2),
inefficiencies due to thermal bridging in building design.
Ao = 41.0 – 9.4 = 31.6 m2 (340.1 ft2)
Examples of Using the Procedures and Data Calculate the overall U-value
in Practice
 Ub  Ab + Uo  Ao 
Simple Example: compares the linear transmittance to U = -----------------------------------------------
area weighted method for a brick veneer shelf angle anomaly. A total
Calculate the overall U-value of the steel stud brick  1.162  9.4 + 0.287  31.6 
= ------------------------------------------------------------------
veneer assembly with a slab and shelf angle with exterior insu- 41
lation R-15 (2.6 RSI) and interior stud cavity insulation R-12 = 0.49 W  m 2 K  0.086 Btu  h  ft 2  F 
(2.1 RSI), using the following information and methods.
Figure 10 illustrates the assembly and detail. Linear Transmittance Method
• Gross Wall Height = 2.7 m (9 feet) For this brick veneer assembly, Uo = 0.287 W/m2K
• Gross Wall Length = 15.2 m (50 feet) (0.0562 Btu/hr•ft2•oF) and the linear transmittance (  ) of a
slab with a shelf angle is 0.5439 W/ m K (0.314 Btu/ hr•ft•oF).
• Gross Wall Area = 41.0 m2 (450 ft2)
Calculation
Area Weighted Method Calculate the overall U-value
The thermal transmittance for the area around the shelf

angle is 1.162 W/m2 K (0.205 Btu/hr•ft2•oF) for the effective L 15.2
U = -------------- + U o = 0.5439  ---------- + 0.287
lengths L1 = 205 mm (8.1 inch), L2 = 618.5 mm (24.4 inch) and A total 41
Uo = 0.287 W/m2K (0.0562 Btu/hr•ft2•oF).
= 0.49 W  m 2 K  0.086 Btu  h  ft 2  F 
Calculation
Calculate area for the thermal anomaly and area for the This example illustrates the simplicity of the linear trans-
clear field. mittance method compared to the area weighted method for
©2012 ASHRAE
579
Figure 11 Building elevation with thermal bridges at intersecting components.
Table 3. Example Building Components and Areas
Total Elevation
7200 ft2 (669 m2)
120 ft (36.6 m) height by 60 ft (18.3 m) width
Total Window Area

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
1200 ft2 (111.5 m2)

5ft (1.5 m) by 4 ft (1.2 m) windows, 5 per floor, 12 floors, 60 in total
Total Opaque Wall Area 6000 ft2 (557.5 m2)
thermal anomalies within opaque building envelope assem- tion installed inboard the poured-in-place concrete and a clear
blies. The amount of information that must be provided is field assembly U-value of 0.074 Btu/hr•ft2•oF (0.42 W/m2 K).
reduced and the calculation is simpler.6 A fictional building elevation with 20% window to wall
Entire Building Elevation Example: demonstrates how ratio is used to demonstrate how details can impact the overall
evaluating details using linear transmittances can highlight the heat flow differently depending on the construction and the
relative contributions of details on the overall heat loss through thermal efficiency of details at the intersection of building
the opaque portion of the building envelope. envelope components. The components and geometry of the
Two wall assemblies are compared: an exterior and inte- building elevation are summarized in Table 3.
rior insulated steel stud assembly with horizontal z-girts
Figure 11 illustrates the building elevation with thermal
supporting metal cladding and a poured-in-place concrete
bridges at the intersection of components identified. Table 4
mass wall. The steel-framed wall assembly does not have
summarizes the standard details and linear transmittances
continuous insulation and has a clear field assembly U-value
selected for this example and Table 5 summarizes the results
of 0.061 Btu/hr•ft2•oF (0.35 W/m2 K) with R-10 (1.8 RSI) exte-
of the calculations.
rior insulation and R-12 (2.1 RSI) insulation in the stud cavity.
The mass concrete wall has continuous R-11 (1.9RSI) insula- The overall heat flow for the concrete mass wall assembly
can be improved to the same level overall U-value as the steel
6. stud assembly, but not realistically by adding insulation to the
The weighted average method is further complicated for a whole
building elevation when accounting for the 3D intersections. clear field alone. For example, providing R-21 (3.7 RSI) of
Some have suggested that the overlapping effects be combined insulation inboard of the concrete wall with a clear field U-
using mitre corners as per ASHRAE 1145-RP (Kemp et al 2001). value of 0.04 Btu/hr•ft2•oF (0.23 W/m2 K) will only raise the
580
ASHRAE Transactions
Table 4. Description and Linear Transmittances for Standard Details
Intersection Detail Steel Framed Assembly Mass Concrete Assembly
The exterior insulation insulates the slab, The slab edge is exposed,
Slab Edge slab = 0.04 Btu/hr•ft•oF slab = 0.47 Btu/hr•ft•oF
(0.08 W/m K) (0.81 W/m K)
Extra framing at corner,

The insulation is continuous at corners as per the wall,
Corner corner = 0.09 Btu/hr•ft•oF
corner = 0
(0.16 W/m K)
Intersection to an R-20 (3.5 RSI) inverted roof with a

The parapet is un-insulated,
concrete deck with insulation on both sides of the parapet
Parapet slab = 0.45 Btu/hr•ft•oF
parapet = 0.28 Btu/hr•ft•oF
(0.78 W/m K)
(0.49 W/m K)
Extra framing and flashing around window parameter.

Window in concrete opening,
Window The flashing ends at the window thermal break,
window = 0.03 Btu/hr•ft•oF
Transition window = 0.05 Btu/hr•ft•oF
(0.05 W/m K)
(0.09 W/m K)
Table 5. Heat Flow Contributions of Overall Opaque Elevation
Steel Framed Assembly Mass Concrete Assembly
Heat Flow, Q/T Heat Flow, Q/T

Transmittance Type % Q/Qtotal % Q/Qtotal
Btu/hroF (W/K) Btu/hroF (W/K)
Clear Field 366 (193) 77 444 (234) 53
Slab Edge 29 (15) 6 337 (178) 40
Corner 11 (6) 2 0 0
Parapet 17 (9) 4 27 (14) 3
Window Transition 54 (29) 11 31 (16) 4
Total 477 (252) 100 841 100

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Overall Opaque U-value 0.08 (0.45) 0.14 (0.79)
overall U-value to 0.1 Btu/hr•ft2•oF (0.6 W/m2 K) or R-75 specific details that can be considered for future work.
(13.2 RSI) of insulation will raise the overall U-value to 0.08 Regardless of the assortment of assemblies and details that
Btu/hr•ft2•oF (0.45 W/m2 K). Ultimately a concrete wall can be modeled, all that is essentially needed is accepted
assembly can match the overall U-value of the example steel benchmarks of transmission values for common details to
assembly, with less insulation, but the heat flow at the floor establish a more consistent standard practice of accounting
slab needs to be addressed. Evaluating the impact of details for thermal bridging. The challenge of comprehensively
using linear transmittances makes this evaluation practicable. accounting for the effect of thermal bridges has restrained a
consistent approach for considering thermal bridging in
CONCLUSIONS building design, simply because of the complexities and
available information for North American construction. The
When considering the vast amounts of building assem- data and procedures contained in RP-1365 provide a good
blies and details used across North America, possible addi- start to overcoming the complexities and lack of information
tional research and work that can follow from RP-1365 can for accounting for the effect of thermal bridging.
appear considerable. The RP-1365 catalogue is based on RP-1365 provides thermal performance data for only
assemblies and details most commonly used across the conti- opaque assemblies; however, some details included transi-
nent and where information is most needed; however, there is tions to glazing assemblies. While further research in apply-
a much wider assortment of possible assemblies and region ing the linear transmittance method to glazing assemblies
©2012 ASHRAE
581
can be done, the current knowledge and techniques used in Enermodal Engineering Ltd, 1996. ASHRAE Research Proj-
analyzing glazing systems are extensive and thoroughly ect 785-RP – Building Insulation System Thermal
rooted in the industry. The exception is spandrel panels. Anomalies, Final Report.
Spandrel panels are an integral part of many glazing systems, ISO 10077-2:2003 (E) Thermal performance of windows,
but are often considered an opaque assembly and are treated doors and shutters – Calculation of thermal transmit-
as any wall assembly in load calculations. The transmittance tance – Part 2: Numerical method for frames. Geneva,
through spandrel sections of glazing systems are highly Switzerland.
dependent on the spacing of mullions, have complicated ISO 10211:2007 (E). Thermal bridges in building construc-
three-dimensional heat flow paths when considering slab tion – Heat flows and surface temperatures – Detailed
edge effects, and are difficult to separate from the vision calculations. Geneva, Switzerland.
sections. There is considerable work that can be done to Janssens, A, E. V. Landersele, B. Vandermarcke, S. Roels, P.
establish generic thermal performance data for curtain and Standaert, P. Wouters, 2007. Development of Limits for
window wall spandrel sections. the Linear Thermal Transmittance of Thermal Bridges
Finally, other further research identified in the RP-1365 is in Buildings. Proceedings of the IX International Con-
to extend the building envelope thermal performance design ference on the Performance of Whole Buildings, Clear-
catalogue with additional z-transfer and other time dependent water, FL.
temperature input functions and establish widely accepted Kemp S., et al. 2001. ASHRAE Research Project 1145-RP –
contact resistance values for accurate component modeling. Modeling Two and Three-Dimensional Heat Transfer
through Composite Wall and Roof Assemblies in
ACKNOWLEDGEMENTS Hourly Simulation Programs. ASHRAE Research Proj-
ect 1145-RP Final Report
Our team would like to thank ASHRAE Technical
Committee 4.4 (Building Materials and Building Envelope Kosny, J.P., J.E. Christian, E. Barbour, J. Goodrow, 1994.
Performance) for sponsoring this project and the generous Thermal Performance of Steel-Framed Walls. CRADA
support of the TC4.4 Project Monitoring Subcommittee Final Report, CRADA Number ORNL 92-0235.
(PMS) throughout this project. The PMS for ASHRAE RP- Kosny, J., A.O. Desjarlais, 1994. Influence of Architectural
1365 includes the following individuals: Marcus Bianchi Details on the Overall Thermal Performance of Residen-
(Chair), Alex McGowan, Rick Peters, and Anton TenWolde. tial Wall Systems. Journal of Building Physics, Vol. 18,
Finally we like to acknowledge ASHRAE, Inc. for the finan- July.
cial support that ultimately make projects like RP-1365 be Kosny, J., P. Childs, 2000. Validation of Heating 7.2 Simula-
realized for the benefit of everyone. tions Using Hot Box Test Data for RASTRA Wall Form
System with Expanded Polystyrene-Beads. Oak Ridge
REFERENCES National Laboratory Buildings Technology Center.
(Appendix E in ASHRAE 1145-RP final report).
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Anderson, B., 2006. Conventions for U-value calculations, Kosny, J., Childs, P., 2000. Dynamic Guarded Hot Box Mea-
BR 443, BRE. surements for RASTRA Wall Form System with
ASHRAE. 2009. ASHRAE Handbook—Fundamentals. Expanded Polystyrene-Beads. Oak Ridge National Lab-
Atlanta: American Society of Heating, Refrigerating oratory Buildings Technology Center, (Part of ASHRAE
and Air-Conditioning Engineers, Inc. Research Project 1145-RP final report prepared by
ASHRAE. 2007. ANSI/ASHRAE/IESNA Standard 90.1- Enermodal Engineering 2001).
2007, Energy Standard for Buildings Except Low-Rise Roppel, P., Lawton, M,. 2011. Thermal Performance of
Residential Buildings. Atlanta: American Society of Building Envelope Details for Mid- and High-Rise
Heating, Refrigerating and Air-conditioning Engineers, Buildings. ASHRAE Research Project RP-1365 Final
Inc. Report.
582
ASHRAE Transactions
APPENDIX A—EXAMPLE DATA AND RESULT SHEETS
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 12 Example Data Sheet.
©2012 ASHRAE
583
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 13 Example Result Sheet.
584
ASHRAE Transactions
SA-12-022
Parametric Investigation of
PCM Thermal Properties on
Temperature of Buildings in Toronto
M. Ebrahim Poulad Alan S. Fung, PhD, PE

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
ABSTRACT Generally, four different technologies are involved in
TES:
The Sustainable Urbanism Initiative Net-Zero Energy
(NZE) Townhouse in Toronto is simulated in TRNSYS 16. The 1. Water (sensible): this is mature and the most popular
incorporation of phase change material (PCM) with Type 204 technology in the market.
into the building envelope is modeled. Simulations are 2. Phase change material or PCM (latent): this technology is
conducted for the summer and winter months to determine the in demonstration phase and needs new materials devel-
potential reductions in the average hourly temperature fluc- opment.
tuations (AHTF). Sensitivity analysis is conducted for thermal 3. Sorption (physical): it is under development.
conductivity and melting range of the PCM. Ten millimeters 4. Thermo-chemical materials or TCMs (chemical): this
(0.4”) PCM is applied as one layer inside the building enve- technology is the most promising as they can be charged
lope. The best melting range and thermal conductivity for a and discharged so many times. They are under research to
typical PCM are sought. To do this, other physical and thermal develop new materials.
specifications of the PCM are fixed. All simulations consider
Toronto-716240 weather conditions. Although indoor air On the average, storage density of the sensible materials
is the lowest, PCMs are higher than sensible materials, sorp-
temperature swing is neither much sensitive to the melting
tion materials are higher than PCMs, and thermo-chemical
range nor to the thermal conductivity, it is lower when the PCM
materials have the highest storage density. To visualize the
is present. AHTF reduction of up to16% is found in mezzanine
importance of latent heat storage (PCM), in the case of water,
due to the PCM additions to the building envelope.
80 times more energy is required to melt 1 kg of ice than to
raise the temperature of 1 kg of water by 1°C. This means that
INTRODUCTION a much smaller weight and volume of material is needed for
storing a specific amount of energy.
Energy storage systems play a key role in improving While the sensible heat storage method is the most
energy utilization, as many energy sources including solar common, there is much activity in the research of thermal stor-
energy - are intermittent. Solar energy is available during the age for solar energy applications using PCM storage. PCM
day; while a demand for domestic hot water exists, and space storage relies on the materials latent heat as the mode of heat
heating or cooling during times of low solar radiation. This transfer as opposed to sensible heat as the transfer mode.
mismatch of availability and demand can be overcome by the Latent heat storage is a new and developing technology which
use of efficient thermal energy storage (TES) such that heat has attracted much attention due to its advantages over sensi-
collected during sunshine hours may be stored for later use ble heat storage, including smaller temperature fluctuations,
during the hours between sunset and sunrise (Farid et al., smaller size and lower weight to volume ratio (Canbazoglu et
2004). al., 2005).
M. Ebrahim Poulad is a PhD candidate and Alan S. Fung is an associate professor in the Department of Mechanical and Industrial Engi-
neering at Ryerson University in Toronto, Ontario.
©2012 ASHRAE
585
Sensible heat storage and latent heat storage are two basic Phase Change Material Storage
types of thermal energy storage (TES) techniques. In sensible Phase change is always either release or absorb some
heat storage, temperature of the storage material varies with heat. Latent heat storage is an efficient way of storing thermal
the amount of energy stored, for example, in solar heating energy due to two key factors. In one recent study, it was found
systems water is used for heat storage in liquid-based systems, that the storage time of hot water, the produced hot water mass
while a rock bed is used for air-based systems. Figure 1 illus- and total heat accumulated in the solar water-heating system
trates the temperature versus stored amount of energy (as heat) having a heat storage tank combined with PCMs were 2.59 -
absorbed by a typical ideal PCM material (latent heat source) 3.45 times that of the conventional solar water-heating system
compared with sensible heat source (e.g. water and concrete). without PCMs. This translates directly into a cost savings as
a smaller tank with less insulation could be used in conjunc-
Practically phase change (melting) temperature is not a hori-
tion with PCM storage (Canbazog˘lu et al., 2005).
zontal line because phase change temperature in heating
Phase change material storage relies on the PCM material
(melting) and cooling (crystallization) points are slightly to absorb, store, and release energy as they change state utiliz-
different. The length of the horizontal line represents the latent ing the latent heat of fusion. Ideally, PCMs must have a large
heat of the PCM. Storage of latent heat means storing heat in latent heat and a high thermal conductivity, have a melting
a material, which undergoes a phase transformation. The most temperature within the range of operation, be chemically
commonly used phase transformation is between the liquid stable, low in cost, non-toxic and non-corrosive (Farid et al.,
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
and solid states, but the phase change between two solid states 2004).
can also be used in principle. However, the latter usually has In solar thermal applications, it is necessary to provide the
a much lower storage density. When heat is fed into the storage required storage of solar energy collected during the day and
material, the material begins to melt once the phase change be able to utilize this stored energy for later use.
temperature has been reached. Although further heat is Several materials have been investigated to understand
their applicability to solar thermal applications. Issues such as
applied, the temperature of the material does not increase until
the degradation of thermal properties, phase segregation, and
it has melted completely. Only then the temperature rises again stability have been researched to ensure good systemic perfor-
(Canbazog˘lu et al., 2005). mance (Farid et al., 2004). Current research indicates that
PCM storage is an attractive option for thermal storage in PCMs such as salt hydrates, paraffins (e.g., RT58), and fatty
solar thermal applications because, given the proper PCM acids are potential candidates for thermal storage in solar
material, a large part of the solar energy available during the applications (Canbazog˘lu et al., 2005). More specifically,
fatty acids including capric, lauric, palmitic, and stearic acids
day could be stored. When the temperature in a hot water stor-
possess good thermal characteristics making them promising
age tank rises above the melting temperature of the PCM, the PCMs for solar applications (Farid et al., 2004). With a melt-
PCM begins to melt as it absorbs heat. When the energy stops ing range between 30°C (86°F) and 65°C (149°F) and latent
being supplied, the PCM will release its energy as it solidifies heat of transition varying from 153 to 180 kJ/kg (66 to 77 Btu/
(Whitman et al., 2011). lb), fatty acids could potentially be used in space and domestic
hot water applications as they have demonstrated good ther-
mal stability in terms of thermal cycling when used as latent
heat storage materials in solar thermal applications (Sari,
2003).
In terms of storage density and the avoidance of large
temperature fluctuations, latent heat energy storage is an effi-
cient way to store thermal energy. Other than increased effi-
ciency, advantages to this method of thermal storage
compared to sensible heat storage include a higher energy
storage density and smaller temperature fluctuations. Once the
PCM has reached its melting temperature, the energy added to
the storage system is used in the melting process, which
prevents large temperature fluctuations. There are however
difficulties when using latent heat thermal storage. Also,
PCMs may undergo density change, phase segregation, sub-
cooling, and lose stability under extended thermal cycling
(Farid et al., 2004).
Producing PCM consumes energy and may have some
Figure 1 Thermal behavior of a sensible material versus a environmental impact. Life cycle assessment (LCA) of PCM
PCM. was considered in Spanish building industry. Results show
586
ASHRAE Transactions
that the addition of PCM in the building envelope, although living house in Toronto. In SUI design, mezzanine is placed
decreasing the energy consumption during operation, does not between second and third floor to afford daylight for the
reduce significantly the global impact throughout the life time house.
of the building. For the hypothetical scenario considering Figure 2 shows a computer generated 3-D model of the
summer conditions all year around and a life time of the build- house. The building envelope of the SUI House is designed
ing of 100 years, the use of PCM reduces the overall impact by with the intention of minimizing the heat transfer to the
more than 10% (Gracia et al., 2010). outside, thereby saving energy and contributing to occupant
In Canada, most of the houses are made of wood. These thermal comfort.
buildings are light and inside temperature fluctuates with The external walls have been insulated with sprayed poly-
ambient temperature. Use of PCM in lightweight construction isocyanurate foam insulation, which provides an overall insu-
(e.g. a wood house) makes it possible to improve thermal lation value of R-60 (RSI-10.6). Roof assembly consists of
comfort and reduce energy consumption (Kuznik et al., 2010). drywall on 19 x 19 mm (0.75 x 0.75 inch) furring and 0.15 mm
Extensive review on dynamic characteristics and energy (5.91 mil) polyethylene vapor retarder attached to the bottom
performance of buildings using PCMs showed that (Zhu et al., of the 294 mm (11.57 inch) pre-engineered I-joists. Sprayed
2009): polyisocyanurate foam is applied between joists as roof insu-
lation. Table 1 shows the various layers used within the wall
• PCMs enhance building thermal and energy perfor- and floor of the building envelope respectively. Thermo-phys-
mance. ical properties of all layers are available in TRNSYS 16, but
• energy performance of buildings using PCMs are not the PCM. The PCM may be introduced as mass-less inside the
sufficiently studied. envelope or added as a massive layer mixed with other mate-
• the research on using optimal control strategies may rial (e.g. plaster).
maximize the potential of using PCMs. The roof has an insulation value of R-76 (RSI-13.4). The
windows used in the house have low emissivity and are argon
Modeling of Latent Heat Thermal Energy Storage
filled with a fiberglass frame and have an overall insulation
Exergy analysis is very important in developing a good value of R-4 (RSI-0.7). Walls below grade are of the insulating
understanding of the thermodynamic behavior of thermal concrete form and have 2.5 in. of rigid polystyrene board with
energy storage systems because it clearly takes into account a waterproof membrane. The overall insulation value of the
the loss of availability and temperature of heat in storage appli- below grade wall is R-35 (RSI-6.27).
cations, and hence it reflects the thermodynamic and
economic value of the storage operation. Most of the analyses TRNSYS16 Building Simulator
are based on first law of thermodynamics, which is inadequate
as a measure of the energy storage because the temperature of TRNSYS is an acronym for a “transient simulation
the surroundings and the effect of time duration through which program” and is a quasi-steady simulation model. This
heat is supplied are not considered. The energy analysis might program was developed at the University of Wisconsin by the
produce a workable design, but not necessarily one with the
highest possible thermodynamic efficiency. In contrast an
exergy analysis consideration leads to optimal design opera-
tion of thermal system (Verma et al., 2008).
Verma et al. (2008) showed that none of the availability
analyses cited have treated thermal energy storage systems
that utilize PCM. However, the study showed that elimination
of the time periods required to heat or to cool the storage mate-
rial above or below the melting temperature, respectively, can
improve the second law efficiency of the system.
Net Zero Energy Building (NZEB) Simulation

in Toronto
The Toronto Net Zero Energy house represents an award
winning design initiative that collaborates between the
Sustainable Urbanism Initiative (SUI) Toronto and a host of
architectural and engineering firms, with the objective of
increasing public awareness and adoption of energy efficient
homes in Canada. More information about the SUI building is Figure 2 Three dimensional computer representation of
given in its website (Siddiqui 2009). NZEB has not yet been the three identical Toronto NZEB (Siddiqui
built, but it has been the subject of much research as a typical 2009).
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
©2012 ASHRAE
587
Table 1. The Layers of the NZEH Wall (Left) and Floor (Right) (Poulad et al. 2011)
Wall Layers Floor Layers
Indoor air exposure Boundary with relevant zone

PCM Layer 10mm (0.4") PCM Layer 10mm (0.4")
Plaster, 13mm (1/2”)
Plaster, 13mm (1/2”)
Furring 19mm (3/4”)
Polyethylene Vapor Retarder, 0.15mm Timber floor 25mm (1")
(2x6) Wood Studs @600mm (24") O.C Common con 50mm (2")
Sprayed Polyisocyanurate closed cell foam 139mm RSI-6.5 I-Joist 50mm (2")
OSB Structural Sheathing with STO Gold Coat 13mm Rigid insulation-Extruded Polystyrene 200mm (8")(R-7)
Rigid insulation-Extruded Polystyrene 100mm RSI 3.48
Furring 19mm (3/4”)
Air space 25mm (1")
Face Brick 100mm (4") Plywood 10mm (0.4")
External ambient exposure Boundary with other zone
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
members of the Solar Energy Laboratory. TRNSYS is elaborate experimentation. Building simulation also provides
comprehensive and powerful building simulation software a valuable tool for validation of the experimental data.
that can provide whole building simulations with the added Prior to the development of the TYPE 204 PCM module
capability of easily integrating a variety of renewable energy in TRNSYS, the only manner in which the effects of PCMs in
and HVAC components. It is this software that would be buildings could be investigated was through the development
utilized for conducting detailed energy simulations on a vari- of an active layer within the building envelope. Ibanez et al.
ety of building models to assess the impact of integrating ther- (2005) presented a methodology in TRNSYS whereby,
mal mass and phase change materials (PCM) into the building through the definition of an active wall containing tubes
envelope. through which a fluid was circulated, the overall thermal effect
TRNSYS is a modular simulation program, based on the of phase change materials could be determined. Even though
FORTRAN programming language. It utilizes standalone this approach did not simulate the real heat transfer process
components and mathematical modules for a wide variety of through a PCM wall, the overall impact in terms of energy
applications such as heat pumps and PV panels etc, in a user- transfer was quite similar to what would be expected with a
friendly graphical interface. Each of these components can be PCM integrated wall (Ibanez et al., 2005).
connected together to represents the flow of information The TYPE 204 component was developed in FORTRAN
during the simulation. The TRNSYS engine calls the system and integrated into TRNSYS by a team based at the Helsinki
components based on the input file and iterates at each time- University of Technology (Lamberg et al., 2004). Utilizing the
step until the system of equations is solved. Weather data is finite difference method with a Crank-Nicholson scheme, the
needed to perform the simulations with TRNSYS. TRNSYS model simulates heat transfer through a 3-D PCM composite
runs through hourly values of various weather parameters wall component containing a total of 729 nodes (9 nodes each
included in a typical meteorological year (TMY) file. The in the x, y and z directions). At each node the conduction,
weather file included with TRNSYS contains detailed weather convection and radiation heat transfer along with the temper-
data for thousands of locations around the world (Klein et al., ature is calculated (Ahmad et al., 2006). The 3-D wall element
1998). can be defined precisely to specify the concentration and melt-
Each component in TRNSYS is defined as a TYPE and ing points of the PCM used. The properties of the composite
contains all the relevant mathematical parameters to integrate building materials used in conjunction with the PCM can also
it in to the overall TRNSYS model. Building models are be easily defined. To account for the changes in the specific
defined as TYPE. heat capacity of the PCM due to temperature variations during
phase change, the model uses the effective heat capacity (Cpe)
TRNSYS Type 204 PCM Components method to define the heat capacity at each phase, i.e., liquid or
Prior to the development of the TYPE 204 model in solid as follows:
TRNSYS, it was impossible to directly simulate the real effect
Latent Heat
of heat transfer through a wall containing PCM. While in the C pe = C p + ---------------------------------------------------------------------------------- (1)
Phase Change Temperature Range
past, most of the work was focused on the experimental anal-
ysis of building integrated PCM, more recently, with the The Type 204 PCM module in TRNSYS has the follow-
development of robust building simulation software, it is now ing input parameters that must be entered into the model to
possible to investigate in detail the thermal properties of a accurately represent a particular phase change material. These
wide variety of phase change materials without the need for properties are described in detail below:
588
ASHRAE Transactions
Number of Iterations: This parameter can be given any PCM density: The density of the pure PCM can be entered
value between one and infinity and is used primarily for the into the model using this parameter. It is 1000 kg/m3 (62.4 lb/
sake of accuracy. Utilizing any number more than one for iter- ft3) for the PCM.
ation would involve the solution of relevant heat transfer equa- PCM specific heat capacity (Cp): This parameter is
tions multiple times and generally provide more accurate concerned with the Cp of the PCM. It is an important charac-
solutions. The only drawback is increased computation time. teristic since it provides a measure of the energy storage/
To find a reasonable value for iteration, some preliminary release capacity of a particular PCM at a temperature outside
simulations were conducted on ASHRAE Standard 140-2001 the temperature range of phase transition.
(BESTEST) Case 600 in Toronto weather conditions. Total Density of other material in PCM node: The density of
energy demand to keep the indoor air temperature of the Case other material that has been integrated with the PCM can be
600 in range of 21°C (70°F) to 24°C (75°F) versus the number entered through this parameter. In this simulation, Plaster is
of iteration is plotted in Figure 3. Three iterations provide used as other material.
reasonably good accuracy in this work. The maximum differ- Cp of other material: The specific heat capacity of any
ence between the maximum number of iteration, 20, and 3 is other materials incorporated with the PCM can be entered
about 5%. This uncertainty is reasonable in this case because, through this parameter.
comparing with 20 iterations; it saves at least four days of Volume fraction of PCM in node: The overall concentra-
simulation time for each run. tion of PCM in a particular specimen can be entered through
Melting temperature: This characteristic is concerned this parameter. Since most studies characterize the overall
with the initial temperature during which the phase change concentration of PCM by weight, this value must be converted
material undergoes phase transition. into volume fraction to reflect the input requirements of the
parameter. For massless simulation, PCM is not mixed; there-
Crystallization temperature: The crystallization tempera-
fore, volume fraction is 1.
ture is determined by the point where the PCM changes phase
Set point in summer and winter is considered 21°C (70°F)
back to a solid. This temperature is always lower than the melt-
and 24°C (75°F) respectively. This is only applicable for
ing temperature. This is considered one degree lower than
investigating energy demand. Heating and cooling are off for
melting point.
indoor air temperature investigation.
Range in crystallization temperature: Unlike pure mate- To add the PCM Type 204 to TRNSYS, the following
rials such as water, which changes phase at a distinct temper- parameters were fixed in the text file named “ALKU”:
ature, most phase changes undergo the phase change process The number of nodes (i, j, k), (fixed) 9*9*9 = 729
within a temperature range. This parameter could be used to Dimensions of the wall component, [m]: 0.45, 0.45, 0.2
define the phase change range of a particular PCM. [meters], height (j), width (i) and depth (k)
Latent heat of PCM: This parameter measures the total Convective heat transfer coefficient of the surface of the
heat storage /release capacity of a particular PCM at the phase wall component, h = 2.0 W/m2K (0.35 Btu/h.ft2.°R)
change temperature range. (HILMA).
Figure 3 Effect of increasing iteration on SUI annual total energy demand.
©2012 ASHRAE 589

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---

Table 2. The NZEB Zones Descriptions and the PCM Sizes
Total PCM area in part, m2 (ft2)

Zone of the Window Area, Total Floor Total Wall Zone Vol., m3
House m2 (ft2) Area, m2 (ft2) Area, m2 (ft2) (ft3) Wall Ceiling Total
Garage 1.03 (9) 11.16 (102) 32.49 (298) 27.2 (757) 0 (0) 0 (0) 0 (0)
1st Floor 10.28 (94) 52.38 (481) 83.18 (764) 146.56 (4078) 80 (735) 50 (459) 130 (1194)
2nd Floor 7.39 (68) 58.46 (537) 86.22 (792) 163.57 (4552) 80 (735) 55 (505) 135 (1240)
3rd Floor 13.50 (124) 49.3 (453) 92.69 (851) 175.5 (4884) 85 (781) 45 (413) 130 (1194)
Mezzanine 33.08 (304) 22.85 (210) 86.23 (792) 171.64 (4776) 50 (459) 20 (184) 70 (643)
Total 65.28 (599) 194.15 (1783) 380.81 (3497) 684.47 (19046) 295 (2709) 170 (1561) 465 (4270)
Time step = 300 second (LASVALI) Table 3. Shading Schedule

Weighting factor of finite-difference method (The Crank-
Nicholson method: MENKERROI = 0.5) Time of the Year (Hour) % Shading
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Initial temperature of the nodes = 20°C (68°F)
0 to 2550 0
Indoor temperature (fixed) = 40°C (104°F)
Initial value of effective heat capacity Cpe = 2.5 kJ/kgK 2550 to 3300 50
(1.1 Btu/lbK)
3300 to 6560 100
SIMULATION 6560 to 8760 0
Simulation was conducted with TRNSYS using Type 204
PCM module developed in Helsinki, Finland. Interior temper-
defined as the temperature difference between start of melting
ature of the SUI Net Zero House with no PCM was taken as
point during heating the solid and start of crystallization point
baseline, and then the PCM added to simulate its effect on
during cooling the liquid. Crystallization temperature is
comfort of the house. The comfort is measured by temperature
always lower than melting point.
swing index (TSI). It, in turn, depends on the average hourly
Melting Range Sensitivity. Melting point of the PCM is
temperature fluctuations (AHTF), see Equations (2) and (3).
an important factor in absorbing/releasing heat from/to the air.
As shown in Table 2, 465 m2 of the total wall and ceiling area
To analyze the sensitivity of melting range, thermal conduc-
was covered with PCM of ten millimeters thickness (0.4 inch).
tivity of the PCM is fixed to a value of 0.3 W/mK (0.17 Btu/
This is equal to a latent energy of 409 MJ (114 kWh or 3.9E5
h.ft.°R). Melting range effects are investigated on indoor air
Btu). Therefore, theoretically, PCM can reduce the fluctuation
temperature (IAT). To quantify the temperature fluctuation or
of the inside air temperature by either absorbing or releasing
swing, a dimensionless temperature swing index (TSI) is
its latent heat. To investigate the effects of PCM on interior
defined and calculated by taking the average hourly tempera-
temperature, the heating and cooling systems were eliminated
ture fluctuations (AHTF) of each zone in 200 successive hours
from the building zones.
(1 to 200 in winter, the first week in January, and 4325 to 4524
Shading Schedule in summer, the first week in July) relative to AHTF of outdoor.
To control the radiation gain inside the building envelope  i200 = 1 ZT i – ZT i + 1
and reduce cooling load in summer and decrease heating load AHTF = ----------------------------------------------
- (2)
200
in winter, schedule is illustrated in Table 3. Each year is
composed of 8760 hours. where ZT is the zone temperature in °C and i is the specific
hour of year (could be any value from 1 to 8559). Then,
PCM Material Properties
TSI = Average of AHTF for the Zone-
For the sake of the simulation, a virtual phase change -------------------------------------------------------------------------- (3)
Average of AHTF for Outdoor
material is used. Table 4 shows the physical property of the
PCM that was simulated. The melting point and thermal For this investigation, heating and cooling are off. Mezza-
conductivity were varied from 22°C (72°F) to 26°C (79°F) and nine has the highest fluctuation in temperature out of all zones
from 0.1 W/mK (0.06 Btu/h.ft.°R) to 3 W/mK (1.73 Btu/ in NZEB (Table 2) because it has the maximum window area;
h.ft.°R), respectively. therefore, its average temperature is investigated.
Figure 4 illustrates the mezzanine temperature on a typi-
RESULTS cal summer (July 2nd) and winter (Jan. 2nd) day. Curves of “10-
Results are classified in two categories: effects of melting 12” and “15-17” overlap each other and the closest to “No
range and effects of thermal conductivity. Melting range is PCM” curve. On the other side, “20-21” and “21-22” curves
590
ASHRAE Transactions
Table 4. PCM Properties
PCM Properties Value
Density,  : kg/m3 (lb/ft3) 800 (50)
Thermal conductivity, k: W/mK (Btu/h.ft.°R) 0.3 (0.17)
Specific heat capacity, Cp: J/gK (Btu/lb) 1.6 (0.69)
Onset phase change temperature upon heating/cooling: Te °C (°F) 22/23 (72/73)
Latent heat of fusion on heating, L: kJ/kg (Btu/lb) 110 (0.43)

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 4 Temperature sensitivity in different PCM melting range in a typical winter, Jan. 2nd, (top) and summer, July 2nd,
(bottom) day (Both Figures have the same legends).
©2012 ASHRAE
591
provide more comfortable temperature (i.e., closer to set point, The analysis of all simulations is quite difficult, due to the
24°C) than the others in summer. amount of data generated. Therefore, only the main points are
highlighted here.
Daily fluctuation in indoor temperature changes from
7.6°C (13.7°F) and 6.5°C (11.7°F) to 5.3°C (9.5°F) and 4.9°C As Figure 5 shows, the minimum temperature fluctuation
(8.8°F) in the typical winter and summer day, respectively, by is seen for PCM with melting range of 21°C to 23°C. It should
using PCM (see Figure 5). Simulation results for other days of be mentioned that if there is enough PCM to store/release heat
the year show the same trend, i.e., No PCM has the maximum there should be no temperature fluctuation. In summer, during
swing and swing is lower in summer than winter days. Addi- the day ambient temperature is above set point range; there-
tionally, other zones have lower temperature fluctuations than fore, inside the house PCM absorb energy until it changes the
the mezzanine. phase completely. At night, ambient temperature is lower than
set point range; the heat that already stored inside the PCM is
Effects of Thermal Conductivity on IAT. For this inves- released to keep the IAT around its melting range. In winter,
tigation, heating and cooling are off. The thermal conductivity ambient temperature is almost always lower than set point
was changed from 0.1 to 3 W/mK. The average temperature of range; therefore the heater should work to keep IAT above
the mezzanine zone is investigated. Analysis of data (Figure 6) 21°C.
clarifies that the indoor average temperature is not sensitive to As shown in the last column of Table 5, the TSI of the
the PCM thermal conductivity, but the presence of the PCM mezzanine is higher in winter than summer (0.91 and 0.62,
keeps the zone temperatures comfortable (between 21°C and respectively). This is due to the direct radiation of sun through
24°C). Although “k=2” shows better performance in summer the windows during the day in winter and cold night in
and winter, all plots with different thermal conductivities are Toronto. In summer, shading (Table 3) reduces the radiation
almost overlapping each other in a narrow band. Other days of into the mezzanine; therefore, the indoor temperature fluctu-
the year show the same trend, i.e., No PCM has the maximum ation is lower with respect to the outdoor temperature fluctu-
distance from set points (in summer days, the temperature is ation (TSI). Analysis of the data in Table 5 proves that using
the furthest from 24°C, and in winter days, it is the furthest the PCM reduces the AHTF by 16% and 15% (comparing
from 21°C). AHTF in No PCM and mezzanine columns) in winter and
summer, respectively. As it is expected, by increasing the
SUMMARY AND DISCUSSION window areas, the difference between winter and summer
AHTF increases. This difference is about 0.03 for first and
It is worth mentioning that, for investigating thermal second floors and increases up to 0.16 for the mezzanine. The
conductivity effect, the melting range was fixed to the value maximum difference goes to No PCM condition, which
given in Table 4 (i.e., 22-23°C). To investigate the melting proves that by using PCM, the temperature fluctuation
range effects, thermal conductivity was fixed and taken from between winter and summer decreases. Regarding tempera-
Table 4. ture fluctuations in all zones, the PCM is more effective in
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 5 Daily temperature fluctuation (Mezzanine zone) on a typical winter (Jan. 2nd) and summer (July, 2nd) day in different
PCM melting range.
592
ASHRAE Transactions
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 6 Sensitivity of IAT to the melting range in a typical winter (top), Jan. 2nd, and summer (bottom), July 2nd, day (both
Figures share the same legend).
Table 5. AHTF of Different Zones and Outdoor Temperature

in the First Week of Jan. (Winter) and July (Summer)
Garage 1st Floor 2nd Floor 3rd Floor Mezzanine No PCM Outdoor TSI
AHTF (Winter) 0.19 0.34 0.21 0.35 0.69 0.82 0.76 0.91
AHTF (Summer) 0.15 0.31 0.19 0.2 0.53 0.62 0.86 0.62
Difference 0.04 0.03 0.02 0.15 0.16 0.2 0.1 0.29
©2012 ASHRAE
593
winter than summer. Due to the large window area of the TCM = Thermo-Chemical Material
mezzanine, when PCM is not used, the mezzanine has a higher TES = Thermal energy storage
temperature fluctuation than the outdoor in winter. In summer, TMY = Typical meteorological year
due to the shading and thermal inertia of the building envelope,
TSI = Temperature swing index
outdoor has a higher AHTF.
In the mezzanine, due to the thermal inertia, the peak ZT = Zone temperature
occurs 2 to 4 hours after the outdoor temperature is in its maxi-
mum point (Figure 6) in summer; and in winter, due to the Greek Notations
steady state of outside conditions, the peak occurs almost at  = Density (kg/m3)
the same time that ambient temperature is in its minimum.
Subscript
CONCLUSIONS p = Specific
The Net-Zero House was simulated using TRNSYS
incorporated with PCM module of the TYPE 204. The best REFERENCES
melting range and thermal conductivity of the PCM were Ahmad, Maha, Andre Bontemps, He´bert Sallee, and Daniel
sought to make the lowest IAT swing. Considering the set Quenard. “Thermal testing and numerical simulation of
points of 21°C (70°F) and 24°C (75°F), the best melting range a prototype cell using light wallboards coupling vacuum
for the PCM was found to be 21°C (68°F) to 23°C (71.6°F), isolation panels and phase change material.” Energy and
which is in lower section of the set points range. In terms of Buildings, 2006: 38 673–681.
thermal conductivity, the higher is the better. Although IAT Canbazog˘lu, Suat, Abdulmuttalip Sahinaslan, Ahmet
swing is neither sensitive to the melting range nor to the ther- Ekmekyapar, Go¨khan Y. Aksoya, and Akarsu, Fatih.
mal conductivity, the swing is lower when the PCM is added. “Enhancement of solar thermal energy storage perfor-
mance using sodium thiosulfate pentahydrate of a con-
FUTURE WORK(S) ventional solar water-heating system.” Energy and
As energy demand and peak power of buildings have Buildings, 2005: 37 235–242.
direct impact on economy, these effects must be studied. Castellón, C., A. Castell, M. Medrano, I. Martorell, and L. F.
Moreover, the effects of the thermal conductivity are better to Cabeza. “Experimental Study of PCM Inclusion in Dif-
be investigated by applying some small layers of PCM on the ferent Building Envelopes.” Journal of Solar Energy
envelope instead of one thick layer (e.g., 10 layers of one mm Engineering, Nov. 2009: 131 (4) 041006 (6 pages).
instead of one layer of 10mm). Farid, M.M., A.M. Khudhair, A.K. Razack, and Said Al-Hal-
laj. “A review on phase change energy storage: materials
AKNOWLEDGMENTS and applications.” Energy Conversion and Management,
This work was funded in part by the Solar Buildings 2004: 45 1597-1615.
Research Network under the Strategic Network Grants Gracia, Alvaro de, et al. “Life cycle assessment of the inclu-
Program and Discovery Grant of the Natural Sciences and sion of phase change materials (PCM) in experimental
Engineering Research Council (NSERC) of Canada. buildings.” Energy and Buildings, 2010.
Ibanez, Manuel, Ana Lazaro, Zalba Belen, and F. Cabeza
NOMENCLATURE Luisa. “An approach to the simulation of PCMs in build-
ing applications using TRNSYS.” Applied Thermal
AHTF = Average hourly temperature fluctuations (°C, °F)
Engineering, Vol, 2005: 25 1796-1807.
Cp = Heat capacity (J/g.K, Btu/lb.°R) Jokisalo, Juha, Piia Lamberg, and Kai Sirén. Suitability of
Cpe = Effective heat capacity (J/g.K, Btu/lb.°R) Building Construction Materials in Short-Term Energy
h = Convective Heat Transfer Coefficient (W/m2K, Storage - Office Room Simulations. Helsinki, Finland:
Btu/h.ft2.°R) Helsinki University of Technology, 1999.
H = Enthalpy Jokisalo, Juha, Piia Lamberg, and Kai Sirén. SUITABILITY
OF BUILDING CONSTRUCTION MATERIALS IN
HVAC = Heating, Ventilation, and Air Conditioning
SHORT-TERM ENERGY STORAGE - OFFICE ROOM
IAT = Indoor air temperature SIMULATIONS. Helsinki, Finland: Helsinki University
L = Latent heat of fusion on heating (kJ/kg, Btu/lb) of Technology, 1999.
k = Thermal conductivity (W/mK, Btu/h.ft.°R) Kuznik, Frederic, Joseph Virgone, and Kevyn Johannes.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
NZEB = Net Zero Energy Building “Development and validation of a new TRNSYS type
for the simulation of external building walls containing
PCM = Phase change material PCM.” Energy and Buildings, 2010: 42 1004–1009.
SUI = Sustainable Urbanism Initiative Lamberg, Piia, Reijo Lehtiniemi, and Anna-Maria Henell.
T = Temperature (°C, °F) “Numerical and experimental investigation of melting
594
ASHRAE Transactions
and freezing Numerical and experimental investigation Verma, Prashant, Varun, and S.K. Singal. “Review of mathe-
of melting and freezing.” International Journal of Ther- matical modeling on latent heat thermal energy storage
mal Sciences, 2004: 43 277–287. systems using phase-change material.” Renewable and
Poulad, M.E., A. Fung, and D. Naylor. “Effect of PCM Ther- Sustainable Energy Reviews, 2008: Vol. 12, pp. 999-
mal Conductivity on Energy Demand and Temperature 1031.
of the Net-Zero Building in Toronto.” CANCAM Pro- Whitman, C. A., M.B. Johnson, and M. A. White. “Charac-
ceedings. Vancouver, B.C., 2011. terization of Thermal Performance of a Solid-Solid
Schranzhofer, H., P. Puschnig, A. Heinz, and W. Streicher. Phase Change.” Thermochimica Acta, 2011.
Validation of a TRNSYS Simulation Model for PCM
Energy Storages and PCM Wall Construction Elements. Zhu, Na, Zhenjun Ma, and Shengwei Wang. “Dynamic char-
Graz, Austria: Institute of Thermal Engineering, Univer- acteristics and energy performance of buildings using
sity of Technology Graz, 2006. phase change materials: A review.” Energy Conversion
Sharma, Atul, V.V. Tyagi, C.R. Chen, and D. Buddhi. and Management, 2009: 3169-3181.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
“Review on thermal energy storage with phase change
materials and applications.” Renewable and Sustainable DISCUSSION
Energy Reviews, 2009: 13 318–345.
Siddiqui, Omar. “Thermo-physical characteristics of build- Mehdi Shahrestani, University of Reading, Reading,
ing integrated phase change materials (PCM) and their Berkshire, UK: 1) Is it sustainable to use some kinds of
applicability to energy efficient homes.” MASc. Thesis. PCM? 2) Is it possible to change the melting point of a specific
Ryerson University, 2009. PCM?
Siddiqui, Omar, Alan Fung, Humphrey Tse, and Dahai M. Ebrahim Poulad: 1) Yes, PCM is sustainable because it is
Zhang. “Modelling of the Net Zero town house in not exhaustive. So, many labs are producing different PCMs
Toronto using TRNSYS, and an analysis of the impact for different applications. They don’t lose their properties in
using thermal mass.” Energy Sustainability. Jackson- storing energy by the number of iterations they have been
ville, Florida USA, 2008. used. 2) It is not easy to change the melting range of a current
Sustainable Urbanism Initiative. Toronto: Sustainable PCM, but we can produce a PCM with a different melting
Design in Toronto’s Annex Net Zero Energy Healthy range that fits any specific demand. Please feel free to contact
Housing Proposal, 2007. me if you need more explanation.
©2012 ASHRAE
595
SA-12-023 (RP-1397)
Field Measurements of Thermal Conditions

During Surgical Procedures for the
Development of CFD Boundary Conditions
James S. McNeill Zhiqiang (John) Zhai, PhD Jean Hertzberg, PhD

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
ABSTRACT nosocomial (i.e. hospital-acquired) infections during the past

The hospital operating room (OR) thermal environment century. Although surgical site infection (SSI) rates are rela-
was investigated via field measurements that were conducted tively low, they continue to be a significant problem, leading
in four similar operating rooms at Denver Health Medical to increases in health care expense and morbidity and mortal-
Center to determine appropriate boundary conditions for ity associated with surgery. SSIs have been estimated at over
computational fluid dynamics (CFD) models. The primary 270,000 annually in the U.S or a rate of approximately 2.2%
measurements used infrared thermography to quantify the making SSIs responsible for $3.45 – 10 billion of additional
surface temperatures found inside the OR. Comparisons were costs to the health care system annually (Douglas & Scott
made to literature to determine realistic approximations for 2009; Gaynes et al. 2001). The increases in costs and risks
models of blockages and heat sources that can be used to accu- associated with SSIs are a burden on both the individual
rately model the indoor environment in CFD. Additionally, the patient and the patient’s family, as well as the entire health care
surface temperature of the surgical site of the patient was system. Infectious disease transmission in the hospital oper-
measured to evaluate the possibility of a thermal plume from ating room is a complex process and may be attributed to
the surgical site. A mean surgical site temperature of 25.6ºC multiple transient sources that may be either exogenous or
[78.1ºF] was measured, which was approximately 2.1ºC endogenous vectors of infection. Staphylococcus aureus has
[3.8ºF] cooler than the surrounding skin temperature. Addi- been determined to be a major cause of SSIs (Lidwell 1981;
tional relevant observations of the OR environment are Hambraeus 1988), and is carried by a significant portion of the
discussed in order to improve the engineering knowledge of population either on the skin or in the nasal cavity (R. E. O.
how these environments are used by medical personnel. Occu- Williams et al. 1959; Davies & Noble 1962; Noble 1975). The
pant heat gain was estimated at 54.3 W for an anesthetized fact that skin and nasal carriage of S. aureus is common among
patient to 102.9 W for a circulating nurse based on a combi- many people adds to the challenge of prevention of SSIs in the
nation of field measurements and surgical personnel thermal operating room, as there are typically many members of the
parameters from literature. The total cooling load in a general hospital staff that are involved in the patient’s care throughout
OR is approximately 2539 W with lighting providing 50% of the duration of the patient’s stay at the hospital. While, noso-
the heat gain. The supply air  T for a generic operating room comial infections may occur at any time during the patient
is approximately 3ºC based on ASHRAE Standard 170-2008 stay, the patient is at significant risk during surgery as the
air flow requirements. subcutaneous tissue is exposed to the environment.
S. aureus is typically made airborne through the natural
INTRODUCTION process of sloughing of skin cell or squames on which the
Advances in hospital cleanliness, sterility, and both surgi- bacteria has colonized. The skin cells have an aerodynamic
cal and aseptic techniques have led to significant reductions of diameter of 10-20 µm (Davies & Noble 1962; Noble et al.
James McNeill is a doctoral candidate and Zhiqiang (John) Zhai is an associate professor in the Department of Civil, Environmental, and
Architectural Engineering, University of Colorado, Boulder, CO. Jean Hertzberg is an associate professor in the Department of Mechanical
Engineering, University of Colorado, Boulder, CO.
596
ASHRAE Transactions
1963; Lees & Brighton 1972), putting the particles in the grav- valuable data on the thermal characteristics of the surgical
itational settling regime. Although it would be expected that staff, such as activity level and dress. A detailed and compre-
infectious particles would quickly settle from the air there is hensive listing of typical thermal-fluid conditions for hospital
anecdotal evidence that sources local to the patient and parti- operating rooms has not been identified in the literature.
cle resuspension from surfaces may play a role in the airborne Researchers have recently been interested in the influence
transmission process. The prevalence of S. aureus in hospitals of the surgical site on the local airflow environment. Memar-
requires that all members involved in the design, operation, zadeh and Manning (2002) hypothesized that the surgical site
and maintenance of operating rooms play a role in the infec- may be warm enough to produce a thermal plume that could
tion control plan. protect the surgical site from particle deposition. They
Aseptic surgical technique is currently the primary conducted extensive CFD modeling of the OR indoor air
defense against surgical site infections, but the airborne path- distribution, and found the lowest incidence of particle
way remains a critical and not well understood aspect in the impingement on the sterile objects of the room with a low
prevention of SSIs. Although there has been an increase in velocity unidirectional air flow from above the patient. The
minimally invasive surgical (MIS) techniques, which limit the boundary conditions used in their models included a surgical
area of surgical site that are exposed to the indoor environ- site temperature of 37.8 °C [100 °F] and were determined by
ment, there are still a large number of surgical cases where the a panel of experts consisting of engineers and physicians. This
surgical wound is significant in size and at greater risk to theory puts the surgical site plume and downward sterile air jet
airborne pathogens. Prevention of surgical site infections are as opposing forces acting against each other. Taking this into
particularly important in total joint replacement, where large consideration, they recommended that the supply air velocity
areas of exposure occur to the indoor environment and resis- at the diffuser face is low (< 0.127 m/s [25 fpm]) to allow the
tance to infection is inhibited by lower blood flow surrounding thermal plume to form.
the prosthesis. While prophylactic antibiotics are a useful There is currently insufficient information regarding the
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
preventative method, reliance upon this treatment has led to surface temperature for surgical sites in open wound surgical
the increased virulence of resistant-strains of infectious applications, although significant work was found in the liter-
diseases, such as methicillin-resistant Staphylococcus aureus ature regarding heat and moisture loss in burn patients (Fergu-
(MRSA) (Graffunder & Venezia 2002). A comprehensive son & Martin 1989; Ferguson et al. 1991; Martin et al. 1985).
aseptic surgical plan includes protocols for staff conduct, ster- The surface temperature differs from typical skin temperature
ilization of instruments and persons, and high indoor air qual- for several reasons: reduction in metabolic rate due to anes-
ity. The U.S. Centers for Disease Control and Prevention thesia, tourniquets and shunts applied to extremities, and evap-
(CDC) has produced guidelines specifically related to the orative cooling due to blood loss. Ferguson et al. (1991) found
prevention of surgical site infections (Mangram et al. 1999). that when the wounds were still wet the wound temperature
These protocols provide a set of procedures for minimizing the was typically 2-4 °C [3.6-7.2 °F] cooler than the healthy skin
risk of SSIs for patients by addressing patient care, indoor temperature, but converged to normal skin temperatures once
environment, and surgical technique during the pre-, intra-, the wounds dried out. Kurz et al. (1996) have shown that the
and post-operative periods. The design of ventilation systems core body temperature will also drop to 34°C [93.2 °F] from
remains a key method of protection against the spread of surgi- a normal temperature of 37°C [98.6 °F] due to perioperative m
cal site infections. during surgeries. Sessler (2000) finds that vasoconstriction is
CFD has become a popular tool for evaluating the oper- inhibited by the anesthesia causing a warming of extremities
ating room indoor environment by modeling the combined compared to normothermia, which provides higher extremity
characteristics of indoor air distribution, indoor air quality, temperatures than would be expected in a hypothermic condi-
thermal comfort, and infectious disease transmission. Most tion. There is also evidence that suggests the patient is more
CFD models have relied on engineering best practice for the likely to receive an infection while under anesthesia due to
development of boundary conditions due to the lack of perioperative hypothermia (Flores-Maldonado et al. 2001;
detailed data for the OR thermal environment. Accurate CFD Sessler & Akca 2002). This has led to the use of perioperative
modeling requires detailed knowledge of realistic boundary warming of the anesthetized patient, which causes the
conditions, without which erroneous or nonphysical results patient’s core body temperature to be maintained closer to the
may be produced. Hospital operating rooms are specialized normal level.
and unique environments that contain a significant amount of In addition to the role the surgical site temperature plays
equipment and activities that are not common in typical indoor for a thermal plume at the surgical site, the other thermal
environments. It is therefore necessary to develop a set of conditions for an operating room are also important in the
detailed boundary conditions specific to hospital operating determination of the overall room air distribution. Imaging
rooms that are based on measured data and observations of the and diagnostic equipment is becoming increasingly used
mechanisms by which the surgical procedure influences the during the surgical procedure with the advancement of
indoor environment. There have been a number of studies that modern hybrid operating rooms. The addition of this equip-
have focused on the measurement of thermal comfort of the ment brings further heat gains to the operating room. There is
surgical staff (Wyon et al. 1968; Mazzacane et al. 2007; currently insufficient data regarding the heat gains from medi-
Balaras et al. 2002; Mora et al. 2001), which have provided cal equipment. Hosni et al. (1999) provided some of the first
©2012 ASHRAE
597
data on the average and peak power consumption for several (ASHRAE Group E) and low-returns at opposing corners. The
common medical and laboratory equipment. In ASHRAE operating rooms were approximately 6.7 x 6.7 m (22 x 22 ft)
1343-RP Method of Testing and Reporting of Energy Use by in floor area with 3.0 m (10 ft) floor to ceiling heights. A pair
Medical Equipment, Koenigshofer et al. (2009) developed a of operating rooms shared a single sub-sterile room that
method of test to determine the heat gains from medical imag- contained the facilities for scrubbing-in and autoclaves for
ing equipment for the use on the heat loads for this equipment sterilization. The operating room contained a double door to
(e.g. CT scan, X-Ray, PET scan, etc.). the corridor to allow movement of the patient and large equip-
This paper documents measurements of thermal-fluid ment into and out of the operating room and a single door that
room characteristics to be used as boundary conditions for provided access to the sub-sterile room. The operating rooms
CFD simulations of hospital operating rooms. These include that were tested were at the far end from the entry to the oper-
relevant thermal measurements of surface temperatures that ating suite.
may induce buoyant plumes, typical objects that form block- The ventilation air for the entire suite was provided by a
ages to air flow patterns, and inlet and outlet conditions for air single central air handling unit (AHU) with two variable air
flow. The boundary conditions presented in this paper are volume (VAV) terminal units per operating room. Although a
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
reasonable for use in a general surgery environment, but would VAV system is installed, the operating rooms received
not accurately represent a specialty surgical environment that constant airflow rates at all times due to concern for particle
deviates significantly from ASHRAE Standard 170-2008 deposition in the sterile region. The main air handling unit
(ASHRAE 2008). Examples of specialty environments (AHU) provided high efficiency particulate arrestance
include cardiac, burns, and modern hybrid surgical rooms. (HEPA) filtration, humidity, and temperature control for the
These environments have significantly different ambient operating rooms, with fine temperature control being provided
temperature and/or heat gains from equipment. This paper at the terminal reheat units. Each operating room was commis-
aims to provide detailed boundary conditions for use in a sioned to provide positive pressurization of 2.5 Pa (0.01 in
general surgical environment and does not attempt to provide w.g.) to both the sub-sterile room and the corridor. Air flow
data that are useable for all types of surgical environments. measurements conducted by the hospital engineering staff are
Additionally, the boundary conditions that are presented in shown in Table 1 demonstrating that the air flow rates are
this paper are appropriate for steady-state numerical simula- within the requirements of ASHRAE Standard 170-2008.
tions of air flow and heat transfer in operating room indoor The majority of measurements and survey information
environments. were obtained during actual surgical procedures. This consti-
tuted the bulk of the work for this research. Observations and
METHODOLOGY measurements were made during 19 surgical procedures
during four separate visits to the hospital campus on different
Field measurements were conducted of thermal and
days. Types and quantities of objects, equipment, and person-
airflow characteristics of the indoor environment in order to
nel that are present during a typical surgical procedure were
obtain suitable boundary conditions for use in computational
identified during the survey process. This data represents
fluid dynamics simulations of hospital operating rooms. The
requirements of operating rooms are significantly different actual usage by the medical staff and provides valuable insight
from typical indoor environments, and specialized medical into the real use of operating rooms. The data is not intended
equipment and occupant thermal exchange parameters make it to be a statistically significant sample, but is rather provided to
necessary to further investigate the actual indoor environment give additional guidance and insight into the surgical environ-
characteristics. The room air flow is provided by downward, ment for use by the design engineer. The location of key medi-
non-aspirating, unidirectional diffusers that cover a large area cal equipment, particularly equipment that represented a
of the rooms ceiling. ASHRAE requires that the diffuser array significant heat gain, was noted during each surgical proce-
covers 0.30 m [12 in] beyond the edge of the surgical table in dure in order to be able to represent this equipment in subse-
all directions. The typical surgical table is 0.55 m x 1.9 m (22 quent laboratory or numerical experiments. Also, a detailed
in x 72 in), so the average diffuser array must cover an area of survey of equipment in the four separate operating rooms was
1.15 m x 1.14 m (46 in x 96 in). In-situ thermal measurements conducted during a separate site visit during off-hours to
were conducted on patients during surgical procedures using obtain further detail regarding equipment specifications.
non-contact infrared thermography. IR thermography In addition to general characteristics of the surgical proce-
provides high spatial resolution mapping of temperatures that dures and thermal conditions of the operating rooms, the surgi-
are useful in determining the temperature difference between cal site was identified as a critical boundary condition that
various objects in the operating room, including the surgical required further investigation. There is currently insufficient
site and the surrounding body. data in the literature regarding the thermal characteristics of
The measurements were conducted in the Department of the typical surgical site. The in-field investigation of the surgi-
Orthopaedics at Denver Health Medical Center in Denver, CO cal site focused on the characteristics of both the surgical site
during the month of March, 2009. The operating rooms are and surgical procedure that were deemed to have a possible
representative of typical modern operating rooms in the U.S. influence on the overall thermal characteristics, including
with downward, unidirectional, non-aspirating diffusers operating room number, patient age, patient height, and
598
ASHRAE Transactions
Figure 1 Example of the operating room in which observations and measurements were conducted.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Table 1. Air Flow Rates for Operating Rooms
OR# Dimensions Supply Air Flow Rate Exhaust Air Flow Rate ACH
6.71 x 7.93 x 3.04 m

1 1202 L/s (2547 CFM) 859 L/s (1821 CFM) 26.7
(22 x 26 x 10 ft)
6.71 x 7.93 x 3.04 m

2 1027 L/s (2177 CFM) 809 L/s (1716 CFM) 22.8
(22 x 26 x 10 ft)
6.71 x 6.71 x 3.048 m

3 1115 L/s (2363 CFM) 761 L/s (1613 CFM) 29.3
(22 x 22 x 10 ft)
6.71 x 6.71 x 3.048 m (22 x 22
4 1115 L/s (2363 CFM) 770 L/s (1631 CFM) 29.3
x 10 ft)
patient weight, application of surgical compression wrap, graphic camera was used to provide visual representation for
patient position, patient thermal treatment, draping, tourni- all thermal images that were obtained. The emissivity of skin
quet, estimated blood loss, surgical site location, and surgical has been determined by several researchers to be between 0.98
site size. Non-contact infrared thermography was used to and 0.99 (Hardy 1937; Cohen 1977), which is quite close to
safely measure the surface temperature of the surgical site the typical emissivity for indoor objects. A value of 0.98 is
from outside of the sterile region of the operating room. These used for all infrared thermal measurements in this study.
measurements provide data regarding the temperature differ- Particular objects of interest were evaluated by manually
ence between the surgical site and surrounding skin and room selecting the thermographic pixels and averaging the temper-
objects that may be responsible for generating thermal ature values. Therefore, objects of interest (surgical site, surgi-
plumes. cal staff, heat sources, etc.) were determined manually by
comparing the thermographic image with a photographic
Experimental Instrumentation image and selecting the relevant pixels. Once the individual
Non-contact infrared thermography was used in order to objects of interest were extracted from the thermal image a
determine the surface temperatures of heat sources within the statistical sample of the objects’ surface temperatures condi-
operating room, using a Fluke TI-30 thermal imager mounted tions was determined.
on a tripod parallel to a photographic camera. The thermal
imager provides a 120 x 160 pixel array of thermal measure- Occupant Heat Gain Parameters
ments, and the accuracy is ±2 ºC and NETD is 200 mK. The The skin temperature of a typical human is based on a
thermal imager measurements were validated by comparison dynamic response with the local environment, but for most
with a Type K thermocouple, and provided temperatures conditioned indoor environments the skin temperature ranges
within 0.33% of the thermocouple measurements. The photo- from 33 to 34°C (91 to 93 ºF) for normal sedentary activity
©2012 ASHRAE
599
Figure 2 Example of thermal image and photograph of equipment used in the operating room.
Table 2. Occupant Metabolic Rate and Clothing Insulation1
Task Dress Type Metabolic Rate (met) Clothing Insulation (clo)
Anesthesia Non-scrubbed 1.4 0.42
Nurse/Technician Non-scrubbed 1.6 0.42-0.78
Surgeon/Scrub Nurse Scrubbed 1.6 0.86
Patient Varies 0.7 0.6-1.0

Note: 1. Values for metabolic rate and clothing insulation were determined from Mora et al. (2001).
levels. The hypothalamus region of the brain controls the flow person’s clothing which alters their thermal comfort, while
of blood to the skin via the vasodilation and vasoconstriction also reducing the clothing surface temperature. Mora et al.
processes in order to maintain thermal comfort. A wide vari- (2001) used infrared thermography to evaluate the thermal
ation of occupant skin temperatures is observed, which comfort of the occupants during surgical procedures. They
depends on the occupant’s activity level, dress, local air concluded that the non-scrubbed surgical staff can be assumed
temperature, humidity and mean radiant temperature. The to have a metabolic rate of 1.4 met and clothing insulation of
range of conditions and tasks in the OR leads to a wide range 0.42 clo, while the scrubbed surgical staff have a metabolic
of heat gains and skin temperatures for the surgical staff and rate of 1.6 met and a clothing insulation of 0.86 clo. Details of
patient. the estimated metabolic rate and clothing insulation deter-
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
In the Guideline for the Prevention of SSI (Mangram et al. mined by Mora et al. are provided in Table 2.
1999), The CDC provides guidelines for the appropriate dress The values for clothing insulation and metabolic rate for
of surgical team members that are categorized as either the surgical staff and patient provided by Mora et al. are used
“scrubbed” or “non-scrubbed”. The non-scrubbed team in this research to evaluate the heat gain from the occupants.
members typically wear a “scrub suit” that consists of pants CFD simulations of the indoor environment often neglect the
and shirt, a surgical mask, surgical cap, and shoe covers. These calculation of radiation due to computational cost, giving the
are the minimum clothing requirements to reduce the spread of heat source boundary condition advantage over the surface
contaminants from clothing, shoes, and shedding from hair temperature boundary condition. The heat source boundary
and scalp. Other than the shoe covers and hair covers this dress condition is easily split into convective and radiant portions,
is consistent with typical warm weather dress for occupants. and is less sensitive to over and under predictions due to differ-
The scrubbed surgical team members are recommended to ences in the local thermal boundary layer from the experi-
wear sterile surgical gowns, sterile gloves, and eye protection ments. This allows the CFD user to correctly specify the total
in addition to the clothing for the non-scrubbed team room cooling load and accurately specify the correct cooling
members. This additional clothing forms an extra barrier of rate to obtain energy balance. The radiant portion of the heat
protection between the team member and the patient that gains are modeled by applying an additional heat flux to the
serves multiple duties by lowering the rate of dispersal of surfaces of the room. The skin temperature measurements
contaminants from the person, while also protecting the surgi- were converted into heat gains using the equations for human
cal team member from blood-borne pathogens that may result energy balance in order to develop occupant heat gain values.
from the surgical procedure. This additional clothing layer The ASHRAE equations for energy balance for sensible
created by the sterile gown increases the insulation value of the heat from a human were used to determine appropriate heat
600
ASHRAE Transactions
gains for the various occupants (ASHRAE 2009). The The surgical team also included a scrub nurse, anesthesiolo-
measured skin temperature was used as an input to this model, gist, and one or two circulating nurses. It was common for an
and expected values were compared to experimental results x-ray technician to be present for surgical cases involving
for agreement. The heat output of a surgical team member can orthopedic repair or hardware. The number of additional
vary significantly due to level of dress, activity, mean radiant specialist surgical staff may vary depending on the type of
temperature, and air flow. The appropriate heat gain from an surgical procedure that is being conducted. The maximum
occupant can be determined from the equations for energy number of surgical staff that were in the room during the
balance on a person. Considering only the sensible heat from observed surgical cases was nine and the minimum was five.
a clothed person, the following equation is used: The highest temperatures on the occupants in the room are
found on the head and hands. The occupants all wear hair
T sk – T op
q conv + q rad = ---------------------------------- (1) coverings and nitrile gloves, but these garments do not provide
R cl + 1   f cl h  significant insulating value and the surface temperature is near
the skin temperature in these locations. Figure 3 shows an
The operative temperature, Top is calculated by:
example of the thermal characteristics of several surgeons
hr Tr + hc Ta underneath the sterile air flow. It is apparent from this figure
T op = ---------------------------
- (2) that the primary heat sources are in the region of the head and
hr + hc
that the upper body of the surgeons is well covered. The ther-
The typical mean radiant temperature in the operating mal measurements show that temperature in the head region
room is equivalent to the average room air temperature. The that is 3-5°C [5-9°F] greater than that of the upper body
surgeons work with the surgical lights slightly above and surface temperature due to a lack of thermal insulation in this
behind their heads. They have an elevated mean radiant temper- area. Figures 3 & 4 demonstrates the large temperature varia-
ature exposure as compared to the other surgical members. The tions between the areas with little thermal insulation and the
mean radiant temperature of the surgical staff was measured by covered areas of the surgeon, with the face and hands of the
Mora et al. to be 22.5 – 26.0 ºC [72.5 – 78.8 ºF]. Although mean surgeon having the highest temperature.
radiant temperature was not directly measured in this study, the The sensible heat gains for each surgical team member
average measured background temperature for walls and large can be calculated using the measured values for skin and cloth-
equipment was approximately 21 ºC [69.8 ºF]. This is near the ing temperatures and estimates of mean radiant temperature,
ambient air temperature, which was in the range of 18 – 22 ºC air velocity, clothing insulation, and metabolic rate. The
[64.4 – 71.6 ºF]. It was therefore assumed for this analysis that results of this analysis are provided in Table 4. The two
radiant temperature was equal to the room air temperature. extremes for dress are the surgeon and nurse, and the corre-
The sensible heat gain from the surgical staff was deter- sponding heat gains are significantly different. The estimated
mined using the metabolic rate values for each occupant type sensible heat gain from a lightly dressed nurse is 134% higher
that were found in the literature in conjunction with the surface than that from a scrub nurse, and the scrub nurse has a sensible
temperature measured in this research. to metabolic heat ratio 45.9% while the ratio for the nurse is
70.4%. A nurse that is dressed in heavier clothing (i.e. an addi-
Cooling Load Calculations tional shirt underneath the scrubs) will lower the sensible heat
The cooling load is necessary for accurate modeling of gains by 22.5 W [76.7 Btu/hr].
indoor air distribution in non-isothermal air flows. Operating It was common to have between one to four surgeons
room air distribution is dominated by a unidirectional jet in the present during the surgeries. The average number of surgeons
sterile region of the room with a high jet width to room width that were present was 2.4, but as these observations were made
ratio, in the range of 0.3-0.5. The cooling load provided in this at a teaching hospital this value might be slightly higher than
paper is based on a field survey of a single suite of operating a non-teaching hospital. It was also common to have at least
rooms in order to provide some detail for CFD modeling. Field one of each of the following staff members: scrub nurse, circu-
surveys were conducted over the course of two months to lating nurse, and anesthesiologist. As the data was collected in
determine the typical characteristics of the general operating the department of orthopedics, it was quite common to have
room. one or two x-ray technicians that were present for the majority
of the surgical case. The use of a c-arm type x-ray machine was
RESULTS & DISCUSSION typical of the majority of the cases. For a typical orthopedic
surgical case it would be expectable to have 6 staff members
Staff Temperature present during the majority of the surgery.
The thermal conditions of several surgical staff members
Surgical Site Temperature
were measured using IR thermography. The thermal charac-
teristics have been segmented based upon temperature varia- The surgical site is the area of the body where a surgical
tions in the thermal images, as shown in Table 3. It was typical incision has been made. An open surgical incision exposes the
for two to four surgeons to be in the sterile field at a single time. subcutaneous tissue to the indoor environment, and deeper
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
©2012 ASHRAE
601
Figure 3 Thermal image of three surgeons surrounding patient.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 4 Thermal image of surgical site surrounded by surgeons.
Table 3. Staff Temperature surgical incisions may also expose the deep soft tissue and
organ space. There are a number of factors that may influence
Location on Body Temp (ºC) Temp (ºF) the surface temperature of the surgical site: the core body
Head - exposed 31.44 ± 1.09 88.59 ± 1.96 temperature, the wound size, the rate of blood loss, whether a
tourniquet is applied to the area, and the location on the body.
Head – covered 27.18 ± 1.69 80.92 ± 3.04 The core body temperature of the patient is also a function of
Surgical mask 26.20 ± 2.07 79.15 ± 3.73 anesthesia, body mass, age, and thermal treatment. It is very
common to apply perioperative warming to the patient to
Hands 28.76 ± 1.14 83.77 ± 2.05
reduce hypothermia associated with anesthesia. Patient warm-
Upper body – ing is typically achieved by using a warming blanket that is
24.91 ± 1.43 76.84 ± 2.57
exposed placed underneath the surgical draping across the patient’s
Upper body – cov- torso. The patient’s core body temperature is monitored by the
29.20 ± 0.99 84.56 ± 1.78 anesthesiologist and the warming blanket is activated and deac-
ered
tivated to maintain a desired core temperature. In the observed
Lower body 25.43 ± 0.83 77.77 ± 1.49
surgical procedures, the patient was warmed with air that was
602
ASHRAE Transactions
Figure 5 Thermal image of heat source caused by cauterizing the surgical site.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Table 4. Staff Heat Gains
Location on Body qsen, W/m2 qsen, Btu/hr-ft2 Qsen, W Qmet, W Qsen/ Qmet
Anesthesia 57.2 18.1 103 146 70.4%
Nurse – lightly dressed 57.2 18.1 103 146 70.4%
Nurse – medium dress 44.6 14.1 80.4 146 55.0%
Surgeon 47.6 15.1 85.6 167 51.3%
Scrub Nurse 42.6 13.5 76.6 167 45.9%
Patient – lightly covered 42.5 13.5 76.5 73.1 105%
Patient – well covered 30.2 9.6 54.3 73.1 74.3%
either 32°C, 38°C, or 43°C [89.6 °F, 100.4°F, 109.4 °F] the deposition of airborne pathogens. This hypothesis was
depending on the core body temperature. Seventeen of the 19 tested by measuring the temperature of the surgical site in the
monitored cases used warming blankets to maintain the 19 observed cases in order to ascertain quantitative evidence
patient’s body temperature, so this is a common practice in of the possibility for a plume to arise. The mean measured
many surgical procedures with anesthetized patients. surgical site temperature was 25.6 ± 2.1ºC [78.1 ± 3.7 ºF] with
IR thermal measurements were conducted on 19 separate a surrounding skin temperature of 29.5 ± 4.0ºC [85.1 ± 7.3ºF],
surgical cases, as described in Table 5, consisting of measure- producing a mean temperature difference of -2.1 ± 1.4ºC [-3.9
ments in a variety of different locations on the patient body ± 2.4ºF] between the surgical site and surrounding body. All
(body locations described per Table 6). The surgical cases took measurements of the surgical site were found to be cooler than
place in the department of orthopedic surgery, so the majority the surrounding skin due to evaporation of body fluids from
of procedures were on the lower torso, legs, and arms. the subcutaneous tissue, and the temperature was found to be
However, two cases were recorded of surgery on the pelvic about 13 ºC [8 ºF] above the room air temperature. The thermal
region, which had surgical sites in the lower torso region of the conditions of the surgical area of the room are largely domi-
body. The patients were typically in the supine position, but nated by the supply air, and so the supply air temperature is a
several cases used the lateral and lateral recumbent positions. better representation of the ambient temperature surrounding
Previous studies regarding airborne infectious disease the surgical site.
transmission to open wounds have theorized that a thermal Local treatments were observed to provide transient
plume formed by localized heat sources in the vicinity of the influences on the surgical site surface temperature, such as
wound may provide sufficient opposing momentum to prevent cooling during irrigation and debridement or temperatures in
©2012 ASHRAE
603
Table 5. Surgical Cases
Surgical Site
Case # Case Type Body Position Thermal Treatment
Locations
1 Pelvic fixator 4 Lateral recumbent None

2 Tibia hardware 6,8 Supine Warmed
3 Tibia repair 6 Supine Warmed
4 Ankle repair and hardware removal 8 Supine Warmed
5 Tibia & fibula repair 7,8 Supine Warmed
6 Calcaneous fixator repair 8 Lateral recumbent Warmed
7 Left fibula repair 7 Right lateral Warmed
8 Pelvic repair 4,4,4 Supine None
9 Irrigation and debridesment, Hardware left ankle 8,8,8 Supine Warmed
10 Pelvic fixator repair 4 Supine Warmed
11 Left ankle hardware recovery 7,7 Supine Warmed
12 Right tibia and fibula fixator and plate 8 Supine Warmed
13 Right and left femur fixator 5,5,6,5 Left lateral Warmed
14 Right ankle fixator and distal tibia and fibia repair 8 Supine Warmed
15 Right distal radius fixator 11 Supine Warmed
16 Right tibia hardware removal 7 Supine Warmed
17 Right hip dislocation, irrigation and debridesment 5 Left lateral Warmed
18 Right ulna fixator 11 Supine Warmed
19 Bilateral pubic fixator 4 Supine Warmed
excess of 35 ºC [95 ºF] following cauterization. These medical was warmer than the surrounding skin was a left femur
procedures cause abrupt and large changes to the surface fixator.This case also had a procedure on the right femur, in
temperature in the region of the surgical site, which may which the surgical site was cooled than the surrounding skin.
temporarily influence the local buoyancy driven airflow. The
The surgical site temperature of different body parts was
steady-state airflow patterns would not be significantly altered
by these transient phenomena, but the production of a tempo- found to vary based on the location on the body as shown in
rary thermal plume is likely. Table 6. The surgical site temperature on the extremities was
found to be slightly warmer than that of the surgical sites on
The formation of a thermal plume in the vicinity of the
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
surgical site would more likely be the result of the heat gain the torso region. This may be attributed to greater blood loss
from the surrounding skin temperature than from the surgical by evaporation in surgeries that involve the torso than the
site. Such a plume would involve air drawn across the surface surgeries on the extremities. Open chest surgeries were not
of the patient. This may be acceptable in the case of a well- available for measurement in this study, which may vary from
sterilized patient, but the patient body itself must be consid- this data. The histogram of surgical site to surrounding skin
ered as a possible source of endogenous pathogens. This temperature difference shown in Figure 8 shows that the surgi-
surface temperature can vary significantly due to differences cal site was generally cooler than the surrounding skin temper-
in metabolic rate, level of anesthesia, instances of infection, ature, but the data does not follow a normal distribution. There
and location on the body. The range that was measured was are likely several parameters that influence this temperature
24.3 – 29.8°C [75.8– 85.7°F] with an average surrounding difference that are discrete functions (e.g. patient has an infec-
skin temperature of 27.7°C [81.9°F]. In all but one case the tion). The dataset collected was insufficient to be able to ascer-
surgical site temperature was lower than the surrounding skin tain the types of parameters that might influence this
temperature. The case in which the surgical site temperature distribution.
604
ASHRAE Transactions
Table 6. Body Zone Location Measurements
Realiz- Surgical Site Surgical Site Normal Skin to Normal Skin to

Zone Region Cases
ations Temp, ºC Temp, ºF Surgical Site ?T, ºC Surgical Site ?T, ºF
1 Head 0 0 N/A N/A N/A N/A

2 Chest 0 0 N/A N/A N/A N/A
3 Abdomen 0 0 N/A N/A N/A N/A
4 Lower torso 6 20 25.2 ± 2.0 77.4 ± 3.6 -2.49 ± 1.9 -4.48 ± 3.4
5 Thigh 4 23 27.8 ± 1.5 82.0 ± 2.7 -0.26 ± 1.4 -0.47 ± 2.5
6 Knee 3 5 26.1 ± 3.0 79.0 ± 5.4 -0.74 ± 1.4 -1.33 ± 2.5
7 Leg 5 12 23.7 ± 1.0 74.7 ± 1.8 -3.33 ± 0.8 -5.99 ± 1.44
8 Foot 8 22 24.7 ± 1.3 76.5 ± 2.3 -3.28 ± 1.8 -5.90 ± 3.2
9 Upper arm 0 0 N/A N/A N/A N/A
10 Elbow 0 0 N/A N/A N/A N/A
11 Forearm 2 17 27.9 ± 0.2 75.0 ± 0.4 -1.59 ± 0.9 -2.86 ± 1.6
12 Hand 0 0 N/A N/A N/A N/A
Figure 6 Histogram of all measured surgical site Figure 7 Histogram of mean surgical site temperature
temperatures. averaged by surgical case.
heat gains in the operating room. Therefore, equipment

Equipment Temperature temperatures were measured in lieu of actual electrical loads.
Modern operating rooms contain a vast amount of The heat loads were estimated from manufacturer’s data and
specialized medical equipment that produce significant heat measured data in the literature where necessary. The surface
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
loads. Unfortunately, there is currently insufficient informa- temperatures are also provided in this paper in order to assess
tion regarding the heat produced by this equipment. ASHRAE the impact of the convective portion of the heat gain on the
1343RP developed a method of test for measuring the actual formation of thermal plumes.
loads from specialized medical equipment, in order to provide Typical equipment in the operating room includes the
more information for calculating the loads in health care facil- surgical light, surgical table, and anesthesia machine, but a
ities. The current research was conducted in a fully function- variety of other equipment also contributes to blockage of air
ing set of operating rooms, so non-contact thermal flow and internal heat gains. Table 7 describes the heat gains
measurements were the only allowable methods for measuring and dimensions from the typical medical equipment that was
©2012 ASHRAE
605
found in the ORs that were surveyed in this research. The having frequent usage is required to be turned on throughout
equipment described below constitutes equipment that is typi- the duration of the surgical procedure. Equipment with labeled
cally found in the majority of operating rooms, but additional as having infrequent usage may either be used for short dura-
equipment may be found in ORs designed for specific functions throughout the surgery or may not typically be required.
tions. A general description of how often the equipment was Infrequently used equipment is neglected from the heat gain
used is included in this table. Equipment that is denoted as total for a steady-state analysis, but may be considered if
detailed transient results are of interest. These data can serve
as a basis for objects to include in a CFD simulation of the
hospital indoor environment.
Equipment in the operating room was grouped into the
following categories for thermal measurements; surgical
lights, computer monitors, and misc. surgical equipment. The
measurements are limited but these temperatures do provide
some idea of the thermal conditions that are presented by vary-
ing types of equipment in the operating room. Several
measurements were conducted of surgical lights and other
equipment to determine the relative surface temperatures of
equipment that are found in the OR. Table 8 shows that the
surgical lights have the highest temperature of any of the
equipment in the operating room. As shown in Figure 9 the
surgical lights are typically located in the region of the surgical
staff and patient presenting a significant heat source in the ster-
ile region of the room.
The surgical lights in the ORs that were evaluated used
150W halogen bulbs and do not represent modern LED surgi-
Figure 8 Surgical site temperature difference with cal lights. LED lighting in surgical applications will provide
surrounding skin. far lower heat gains and surface temperatures, which will
Table 7. Survey of Medical Equipment
Nominal Heat Nominal Heat

Equipment Make/Model Dimensions (WxDxH), m Usage
Gain, W Gain, Btu/hr
GE Datex Omeda Aes- 0.83 x 0.93 x 1.36

Anesthesia machine 100 341 Frequent
tiva /5 (2.7 x 3.1 x 4.5 ft)
1.2 x 0.71 x 1.0

Medication station CareFusion Pyxis 104 355 Frequent
(3.9 x 2.3 x 3.3 ft)
0.21 x 0.18 x 0.24
Fluid warmer Hotline Fluid Warmer 180 614 Frequent
(0.69 x 0.59 x 0.79 ft)
Berchtold Chromophare 0.66 diameter
Surgical lights 300 1024 Frequent
D650 Plus (2.2 m)
0.11 x 0.36 x 0.44

Electrosurg. gen. Valleylab Force FX 150 512 Infrequent
(0.36 x 1.2 x 1.4 ft)
0.3 x 0.25 x 0.24
Blanket warmer Bair Hugger 750 800 2730 Intermittent
(0.98 x 0.82 x 0.79 ft)
Workstation N/A N/A 100a 341a Frequent
a a
LCD Monitor N/A N/A 100 341 Frequent
Mobile C-arm x-ray N/A N/A 480b 1638b Infrequent

Notes:
a. Values estimated from ASHRAE Handbook.
b. Source: Hosni et al. (1999)
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
606
ASHRAE Transactions
Table 8. Equipment Temperature
Equipment Temp, ºC Temp, ºF
Surgical Light 33.92 ± 4.14 93.06 ± 7.45
Computer Monitors 27.65 ± 6.01 81.77 ± 10.8
Misc. Equipment 26.00 ± N/A 78.80 ± N/A
Figure 9 Thermal image of surgical light above the heads of surgeons.
significantly reduce the convective gains to the indoor air. This rate of 20 ACH a supply air to room air  T of approximately
will both reduce the energy consumption of operating rooms, 3°C [5.4 °F] would be required to meet the steady-state cool-
as well as improve the behavior of the indoor air distribution ing load. This is an estimation of the minimum heat gain
by reducing thermal plumes and the impact of buoyancy- contributed by the staff and equipment that would be expected
driven flow. during a typical surgical procedure. In the sterile field, the
surgeons and surgical lights each provide equivalent heat gains
Cooling Load of approximately 300 W. For a trauma case where more than
Steady-state thermal-fluid analysis of indoor environ- 2.4 surgeons are attending the heat gain is biased towards the
ments requires that the cooling load is accurately estimated in surgeons.
order to determine the appropriate supply air temperature to
maintain thermal comfort. The primary constituents of the CONCLUSIONS
cooling load in the OR are the lights, equipment, and occu- Thermal conditions in the hospital operating room were
pants. The lighting accounts for approximately 50% of the determined for the development of realistic CFD boundary
load, while the equipment and occupants each constitute about conditions through a combination of analytical methods, data
25% of the load. The ambient lighting is a significant portion in the literature, and data from field measurements. The
of the overall heat gain to the space. temperature of the surgical site was determined to be approx-
The total room load was calculated using the values that imately 25.6 ± 2.1ºC [78.1 ± 3.8 ºF], which is well below the
were determined for the heat gains from equipment and occu- normal core body temperature of 37ºC [98.6 ºF]. The surgical
pants. The total room cooling load for staff plus frequently site was found to generally be cooler than the surrounding
used equipment, excluding the electrosurgical generator, body, which may be attributed to numerous factors, such as
mobile c-arm x-ray system, and blanket warmer is 2539 W. evaporative cooling from wetted tissue, reduced blood flow
The individual components and quantity for each component due to clamping or clotting, and perioperative hypothermia. In
of the cooling load are provided in Table 9. For a typical OR the limited experimental results presented in this paper, the
size (6.7 x 6.7 x 3.0 m [22 ft x 22 ft x 9.8 ft]) and an air change location of the surgical site on the extremities versus the core
©2012 ASHRAE 607

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---

Table 9. Operating Room Cooling Load project. The authors would also like to thank the staff of
Denver Health Medical Center, and particularly the Depart-
Total Total ment of Orthopaedics for allowing access to their operating
Equipment Qty Heat Gain, Heat Gain, rooms and for insight into the medical profession.
W Btu/hr
NOMENCLATURE
Anesthesia machine 1 100 341
Qmet = Metabolic heat gain, (W)
Medication station 1 104 355
Qsen = Sensible heat gain, (W)
Fluid warmer 1 180 614 qsen = Sensible heat gain (W/m2)
Surgical lights 1 300 1024 qconv = Convective heat gain (W/m2)
qrad = Radiative heat gain
Workstation 1 100 341
Tskin = Skin temperature averaged over the entire body (ºC)
LCD monitor - large 1 100 341 Top = Operative air temperature (ºC)
LCD monitor - small 2 200 682 h = combined convective and radiant heat transfer
coefficient (W/m2-K)
Ambient lights 8 1024 3494
Rcl = Clothing thermal resistance value (m2-K/W)
Anesthesiologist 1 103 351
fcl = Clothing area factor
Surgeons 3 85.6 292 Tr = Mean radiant temperature of occupant (ºC)
Scrub nurse 1 76.6 261 Ta = Room air temperature (ºC)
hc = Convective heat transfer coefficient (W/m2-K)
Circulating nurse 1 80.4 274
hr = Radiative heat transfer coefficient (W/m2-K)
Patient 1 54.3 185
REFERENCES
Total Heat Gain 1 2539 8663
ASHRAE. 2008. ANSI/ASHRAE Standard 170-2008. Venti-
lation of Healthcare Facilities. Atlanta: American Soci-
body did not appear to be major factor on the surgical site ety of Heating, Refrigerating, and Air-Conditioning
temperature. Local thermal conditions may also influence the Engineers, Inc.
conditions of the surgical site, such as cooling irrigation and ASHRAE. 2009. ASHRAE Handbook - Fundamentals.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
debridement or warming during electro-cauterization. These Atlanta: American Society of Heating, Refrigerating
processes may drastically change the local air patterns during and Air-Conditioning Engineers, Inc.
a shorter time duration that requires further investigation. Balaras, C.A., E. Dascalaki, A.A. Argiriou, & A. Gaglia.
The primary heat sources in the region of the surgical site 2002. HVAC systems and indoor conditions in Hellenic
are the surgical staff and the surgical lighting. The occupants hospital operating rooms. ASHRAE Transactions
have a wide range of heat gains due to large differences in 108(2):23–38.
activity and dress. The heat gains should be modeled appro- Cohen, M.L. 1977. Measurement of the Thermal Properties
priately for each type of occupant, because the influence on of Human Skin. A Review. The Journal of Investigative
thermal plumes may be significantly different for occupants in Dermatology 69(3):333-338.
the sterile field and those that are outside of the sterile field. Davies, R.R., & W.C. Noble. 1962. Dispersal of bacteria on
Progress is being made in surgical environments to use LED desquamated skin. The Lancet 280(7269):1295-1297.
surgical lighting, which would greatly aid in reducing the heat Ferguson, J.C., & C. J. Martin.1989. Burn wound evapora-
gains in the OR. This will also help reduce the possibility of tion - an evaluation of air diffusion resistances govern-
plumes forming around the surgical lights. ing heat transfer in a clean air unit. Clinical Physics and
The typical heat gains in the operating room represent a Physiological Measurement 10(4):319-330.
total cooling load of 2539 W. The majority of the contribution Ferguson, J.C., C.J. Martin& C. Rayner. 1991. Burn wound
comes from the lighting in the space. This load would require evaporation - measurement of body fluid loss by probe
a typical supply air T of 3°C [5.4 °F] with the air flow evaporimeter and weight change. Clinical Physics and
requirements found in ASHRAE Standard 170-2008. Physiological Measurement 12(2):143-155.
Flores-Maldonado, A., C.E. Medina-Escobedo, H.M.G.
ACKNOWLEDGMENTS Ríos-Rodríguez & R. Fernández-Domínguez. (2001).
This work was sponsored under ASHRAE 1397RP, Mild Perioperative Hypothermia and the Risk of Wound
“Experimental Investigation of Hospital Operating Room Air Infection. Archives of Medical Research 32(3):227-231.
Distribution”. The authors would like to thank the members of Gaynes, R.P., D.H. Culver, T.C. Horan, J.R. Edwards, C.
TC 9.6 – Healthcare Facilities for their support of this research Richards, & J.S. Tolson. 2001. Surgical site infection
608
ASHRAE Transactions
(SSI) rates in the United States, 1992-1998: the National the thermal environment and its effect on evaporation
Nosocomial Infections Surveillance System basic SSI and heat loss. Clinical Physics and Physiological Mea-
risk index. Clinical Infectious Diseases 33(Suppl surement 6(4):323-336.
2):S69-S77. Mazzacane, S., C. Giaconia, S. Costanzo, and A. Cusumano.
Graffunder, E.M., & R.A. Venezia. 2002. Risk factors asso- 2007. A Survey on the Thermal Conditions Experienced
ciated with nosocomial methicillin-resistant staphylo- by a Surgical Team. Indoor and Built Environment
coccus aureus (MRSA) infection including previous use 16(2):99-109.
of antimicrobials. Journal of Antimicrobial Chemother- Memarzadeh, F., & A.P. Manning. 2002. Comparison of
apy 49(6):999-1005. operating room ventilation systems in the protection of
Hambraeus, A. 1988. Aerobiology in the operating room--a the surgical site. ASHRAE Transactions 108(2): 3-15.
review. Journal of Hospital Infection 11(Suppl 1):68-76.
Hardy, J.D. 1937. The physical laws of heat loss from the Mora, R., M.J. English, and A.K. Athienitis. 2001. Assess-
human body. Proceedings of the National Academy of ment of thermal comfort during surgical operations.
Sciences of the United States of America, 23(12):631- ASHRAE Transactions 107(1):52–62.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
637. Noble, W.C. 1975. Dispersal of skin microorganisms. British

Hosni, M.H., B.W. Jones, & H. Xu. 1999. Experimental Journal of Dermatology 93(4):477-485.
results for heat gain and radiant/convective split from Noble, W.C., O.M. Lidwell, and D. Kingston. 1963. The
equipment in buildings. ASHRAE Transactions Size Distribution of Airborne Particles Carrying Micro-
105(2):527–539. Organisms. The Journal of Hygiene 61(4):385-391.
Koenigshofer, D., R. Guevara, & D. Nemecek. 2009. Method Scott, R.D. 2009. The direct medical costs of healthcare-
of Testing and Reporting Energy Use by Medical Equip- associated infections in U.S. hospitals and the benefits
ment. Final Report for ASHRAE 1343-RP. of prevention. Atlanta: Centers for Disease Control and
Kurz, A., D.I. Sessler and R. Lenhardt. 1996. Perioperative Prevention.
normothermia to reduce the incidence of surgical- Sessler, D.I. 2000. Perioperative heat balance. Anesthesiol-
wound infection and shorten hospitalization. New Eng- ogy 92(2):578-596.
land Journal of Medicine 334(19):1209-1215.
Lees, J., and W.D.Brighton. 1972. Simulated human skin Sessler, D.I., and O. Akca. 2002. Non-pharmacologic Pre-
scales. The Journal of Hygiene 70(3):557-564. vention of Surgical Wound Infection. Clinical Infectious
Lidwell, O.M. 1981). Airborne bacteria and surgical infec- Diseases, 35(11):1397-1404.
tion. The American Journal of Medicine 70(3):693-697. Williams, R E.O., M.P. Jevons, R.A. Shooter, C.J.W. Hunter,
Mangram, A.J., T.C. Horan, M.L. Pearson, L.C. Silver and J.A. Girling, J.D. Griffiths, and G.W. Taylor. 1959.
W.R. Jarvis. 1999. Guideline for prevention of surgical Nasal staphylococci and sepsis in hospital patients. Brit-
site infection, 1999. American Journal of Infection Con- ish Medical Journal 2(5153):658-662.
trol 27(2):97–134. Wyon, D.P., O.M. Lidwell, and R.E. Williams. 1968. Ther-
Martin, C.J., R. Wytch, and I.F. Muir. 1985. Exposure treat- mal comfort during surgical operations. The Journal of
ment of the burned patient - a computer simulation of Hygiene 66(2):229-248.
©2012 ASHRAE
609
SA-12-024
Modeling Phase Change Materials with a

Building Simulation Code
Developed in MATLAB
Sridhar Sadasivam Dahai Zhang Alan S. Fung, PhD, PEng Fabio Almeida
Student Member ASHRAE Member ASHRAE Student Member ASHRAE
ABSTRACT RADCOOL, BSim, DeST, IDA ICE, PowerDomus and

SUNREL have the capability to handle phase change prob-
In this paper, the existing numerical methods to simulate
lems (Crawley et al., 2005; Pedersen, 2007; Stetiu and Feustel,
phase change materials (PCMs) in building simulations are
1998; Rose et al., 2009).
reviewed and two forms of the modeling scheme are explained
in detail. The first formulation is a non-iterative temperature Since most PCMs undergo solid-liquid phase transition,
formulation which suffers from the requirement of extremely analytical solutions given to the phase change phenomenon
small time steps to produce accurate results. The second are described as solutions to the Stefan problem, or the so-
formulation is an iterative enthalpy formulation which through called moving boundary problem (Stefan, 1889). However,
the use of an iterative correction scheme at every time step, can the difficulties arise when the boundary interface needs to be
produce fairly accurate results for comparatively much larger continuously tracked between two phases. Moreover, the
time step values. A customized building simulation code was analytical solution is suitable for the isothermal phase change,
developed in MATLAB to incorporate the correction scheme. i.e. phase change occurs at a constant temperature, but not
The simulation results were compared with ESP-r for the stan- appropriate for a temperature range (Dalhuijsen and Segal,
dard BESTEST Case 600 model. It was found that there was 1986). It is impractical, therefore, to provide these analytical
significant discrepancy in temperature profiles at large time solutions in the building simulation when realistic non-
step values when PCMs were introduced. However, the differ- isothermal PCMs are applied, and other structures and
ence became smaller when the time steps in ESP-r was reduced systems are involved.
while maintaining the large time step in the MATLAB model.
Numerical models, as alternative approaches based on the
solutions to the differential equations, have been adopted
INTRODUCTION
widely in various building simulation tools (Jokisalo et al.,
The application of phase change materials (PCMs) in 2000; Pedersen, 2007; Heim, 2010). The effective heat capac-
building envelopes has gained growing interest in building ity method and the enthalpy method are the two preferred
industry due to their potentials in energy savings and peak load numerical methods. The former method describes the temper-
shifting associated with the HVAC system, as well as the ature change using the heat capacity Cp(T), which can be typi-
improved thermal comfort. Several pilot housing projects cally measured through differential scanning calorimetry
using integrated PCMs building envelopes have been devel- (DSC) in dynamic mode. The dynamic DSC operation,
oped in Europe (Schossig et al., 2005; Mehling and Cabeza, however, presents non-isothermal behaviour in PCMs during
2008). Numerical studies of the thermal performance of such phase change, which greatly depends on the DSC scanning
PCM integrated buildings often employ building simulation rate and the PCM sample size, leading to the deviations of the
tools. Crawley et al. (2005) comprehensively compared vari- PCMs heat capacity with respect to the temperature (Mehling
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
ous features of 20 building simulation software. Among these and Cabeza, 2008). The variation of heat capacity curves of the
tools, ESP-r, TRNSYS, EnergyPlus, DOE-2 with same PCM can be avoided by introducing the enthalpy
Sridhar Sadasivam is a PhD candidate in the School of Mechanical Engineering, Purdue University, West Lafayette, IN. Dahai Zhang and
Fabio Almeida are PhD candidates and Alan S. Fung is an associate professor in the Mechanical and Industrial Engineering Department, Ryer-
son University, Toronto, Canada.
610
ASHRAE Transactions
method, which presents heat capacity in terms of its integral tive heat capacity method with the explicit scheme and vali-
form, H(T), with respect to the temperature. Moreover, the dated with the experimental results (Kuznik et al., 2010). An
enthalpy method can correctly account for the subcooling implicit PCM module with the enthalpy method was imple-
effect (Gunther et al., 2007). The literature review on the defi- mented in EnergyPlus (Pedersen, 2007). A summary of
nition of both methods indicates the interpolation difference in published PCM modules in various building simulation soft-
the classification of the methods. Swaminathan and Voller ware and the corresponding methods/schemes is given in
(1993) classified the effective/apparent heat capacity method Table 1. It should once again be noted that the PCM property
into the enthalpy method as the term Cp,eff was eventually calculation scheme in this table specifically indicates the
calculated from the enthalpy H. However, recent development application of PCM module, not necessarily represents the
in building simulation with PCMs employed pre-defined heat methodology employed in the overall building simulation
capacity values as input data, such as in TRNSYS, ESP-r and package. An explicit scheme means that the heat capacity in
EnergyPlus etc. Hence, the two methods are treated as sepa- the energy conservation equation is calculated based on the
rate methods. nodal temperature at the previous time instant while an
An important factor in these numerical methods is the implicit scheme means that the heat capacity is corrected
calculation of the corresponding thermal properties (Cp or H) based on the nodal temperature calculated at the next time
at the new time step. For instance, in the effective heat capacity instant.
method the heat capacity value needs to be calculated in the
discretisation equations based on the current temperature of OBJECTIVES
the node. However since the dependence of heat capacity on
The objective of this study is to develop a customized
temperature is highly non-linear during phase change, an
code to solve the heat transfer through PCMs in buildings
explicit scheme is often used in many situations where the heat
using the implicit enthalpy scheme. The results are compared
capacity is calculated from the nodal temperature value at the
with one of the full building simulation software at the stan-
previous time instant. This can lead to errors, especially during
dard conditions. Upon choosing the paired software, consid-
phase change when the heat capacity changes rapidly with
eration was given to 1) the capability to handle PCM in
temperature. Here in the current paper, the explicit scheme
building structure, 2) authors’ experience in using the software
means that the program calculates the heat capacity of PCM
and 3) flexibility to vary time step in a wide range from the
based on the temperature at the previous time instant. This
magnitude of seconds to hours/days. Based on these criteria,
should be differentiated from the general usage of the term
ESP-r was selected as it applies the finite difference formula-
explicit in numerical schemes where it normally means that
tion on a control volume heat-balance to present all necessary
the current temperature at a node depends only on the old
energy flows in buildings (Clarke, 2001). Our customized
temperature of surrounding nodes. In this paper, all subse-
code was developed in the MATLAB environment, with the
quent references to the term explicit or implicit are only with
finite volume enthalpy method applied. Both explicit and
respect to calculating the thermal properties of PCM, such as
implicit schemes for calculation of PCM properties are
heat capacity or enthalpy.
presented and compared with ESP-r for various time step
The errors that occur when using an explicit scheme can
values. In order to explain the design process, the methodol-
be avoided by using small time stepping values. However, this
ogy of effective heat capacity and enthalpy methods with
may dramatically increase the simulation time in building
explicit and implicit schemes are further reviewed in depth.
simulation, especially when other mechanical systems are
involved. The other option is to use an implicit scheme, which
METHODOLOGY
often involves corrective iterations within a particular time
step (Swaminathan and Voller, 1993; Pederson, 2007). A The governing equations to describe PCM heat transfer
recent TRNSYS PCM module was developed using the effec- are the mass conservation equation, momentum equation and
Table 1. PCM Modules in Building Simulation Programs and the Corresponding Methodology
Reference Software PCM Numerical Method Discrete Form Heat Capacity Calculation
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Heim, 2010 ESP-r Effective heat capacity Finite volume Explicit

TRNSYS
Jokisalo et al., 2000 Effective heat capacity Finite element Crank-Nicholson
(Type 204)
TRNSYS
Kuznik et al., 2010 Effective heat capacity Finite difference Explicit
(Type 260)
Pedersen, 2007 EnergyPlus Enthalpy Finite difference Implicit
Rose et al., 2009 BSim Enthalpy Finite volume Explicit
©2012 ASHRAE
611
energy conservation equation. Although most PCMs undergo Explicit Scheme
solid-liquid phase transition, they are confined within the
Equation (7), when written in a discretized form becomes
building envelope through micro/macro-encapsulation as an
integrated structure. The governing equations in building H T old T
new
–T
old
simulation, hence, can be reduced to the energy conservation  ------- = C p eff  T  ------  C p eff ---------------------------- (8)
t t t
equation only. Further, the internal convective heat transfer
old
can be reasonably neglected, leading to the simplification of where Cp eff denotes the effective heat capacity at Told. The
the governing equation as: R.H.S of Equations (3) or (7) is typically approximated using
a finite difference or finite volume method and a discretized
---------------
 H  = .  k  T   expression of these equations is written for every node in the
(1)
t domain. The system of equations is then solved to obtain the
nodal temperatures at any given time instant.
where  is the density, k is the thermal conductivity, and H is
the enthalpy, which is the sum of sensible and latent heat. In The condition that causes inaccuracy in estimating the
the phase change process, the enthalpy can be defined as enthalpy change during one time step is described in Figure 1,
(Swaminathan and Voller, 1993): which shows the procedure of calculating enthalpy based on
the Cp,eff –T relationship. To and Te denote the onset and end of
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
T phase transition, respectively. The heat capacity remains more
H =  Cp mix dT + fL (2) or less constant outside the phase change temperature range
To and changes sharply during phase change. Consider any node
on the PCM where the temperature Told is just below To and
where the subscript mix refers to the thermal properties of the about to enter the phase transition region. Let Tnew be the
mixture of solid and liquid phases, o refers to the onset of temperature attained after time  t (time-step) such that To <
phase change process, L denotes the latent heat, and f is the Tnew < Te. The change in the enthalpy during this time period
fraction of liquid phase. Consider negligible variation of  and of  t is expressed as the area of the shaded region in
k with respect to time, Figure 1(a) when Equation (8) is used. However the actual
area under the Cp,eff curve is shown in Figure 1(b). Errors can
- k thus be introduced into the energy balance equation if  t is not
-----  0 and -----  0 ,
t t sufficiently small.
Equation (1) can be written as: The magnitude of this error is dependent on the time step
and the non-linearity of the Cp,eff –T curve. Larger the time
H 2 step, more the error and smaller the time step, lesser would be
 ------- = k  T (3)
t the error. Meanwhile, a sharp, highly non-linear Cp,eff -T curve
would result in a larger error compared to a slowly varying
Equation (3) is thus the simplified governing equation in curve. Errors of this kind not only occur at the beginning of
the enthalpy form. Imagine that the enthalpy expression in phase change, but also during phase change and at the exit of
Equation (2) can be substituted with an equivalent “sensible phase change. For building simulation, the errors can be
heat” formulation, the effective heat capacity, Cp,eff, is then further amplified when the PCM frequently experiences phase
introduced. During the phase change process, change. Since building simulations are typically yearly or long
T
term simulations, a small time step would be quite demanding
in terms of computational time.
H =  Cp eff dT (4)
To Implicit Scheme
In the entire temperature range, The implicit scheme can be applied to both the effective
heat capacity method as shown in the Cp,eff –T curve
H = C p eff T (5) (Figure 2(a)) and the enthalpy method as shown in the H –T
curve (Figure 2(b)). The H –T curve can be obtained from the
Therefore, Cp,eff –T curve by integrating Cp,eff over the temperature. The
application of implicit scheme often involves iterative solu-
H H T T
 ------- =  ------- ------ = C p eff  T  ------ (6) tions to correct the thermal properties at each time step. Peder-
t T t t
sen (2007) used the Gauss-Seidel iteration scheme for the
Substitute Equation (6) into (3), the governing equation in enthalpy corrections in EnergyPlus. Swaminathan and Voller
the effective heat capacity form can be written, (1993) proposed an optimal iterative correction scheme for the
enthalpy method to reduce the number of iterations.
T 2 The methodology proposed by Swaminathan and Voller
C p eff  T  ------ = k  T (7)
t (1993) is briefly reviewed herein. Similar to the Cp,eff –T curve,
612
ASHRAE Transactions
Figure 1 Enthalpy change over the Cp,eff –T curve, (a) erroneous calculation based on explicit scheme, (b) actual enthalpy
change.
(a) (b)
Figure 2 Figure 2. Non-linearity during phase change: (a) Cp,eff –T curve and (b) H –T curve
discontinuity in the slope of the enthalpy curve occurs at To Figures 3(a-c) show a point [a] on the H –T curve just
and Te. This is the primary reason that iterations are required below To and about to enter the phase transition region.
to compute the correct enthalpy and temperature when a node Assume that the new enthalpy and temperature (point [b]) are
is entering or exiting phase change. The L.H.S of Equation (3) within the phase change region. As shown in Figure 3(a), (H*,
would typically be approximated by a backward Euler finite T*) is the same as (Hold, Told) before correction and Equation
difference as follows,
(9) reduces to Equation (8).
* * new * old
H H
new
–H
old H + C p eff  T –T –H
 -------   ------------------------------ =  ------------------------------------------------------------------------- (9) H
old
H – C p eff  T
old new old
–T –H
old new
–T
*
t t t old T
 -------   ---------------------------------------------------------------------------------- = C p eff -----------------------
t t t
Here the new enthalpy Hnew is expressed using a 1st order (11)
Taylor series expansion around H*,
Thus the first iteration of this enthalpy scheme is similar
= H + dH
new * new * * * new *
H ------- T – T  = H + C p eff  T – T  (10) to the explicit effective heat capacity scheme explained in the
dT *
T previous section. However the iterative scheme doesn’t just
©2012 ASHRAE 613

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 3 Iterative correction scheme on the enthalpy method: (a) for the 1st iteration, (b) for the 2nd iteration, (c) illustration
of iterative correction.
stop after calculating the new nodal temperatures based on there is no discontinuity of slope between H* and Hnew. The
Equation (11). new enthalpy Hnew can now be expressed using a Taylor series
new expansion around H*. The iterations thus proceed till the
H calc —New enthalpy calculated from the 1st order
old enthalpy calculated from a Taylor series expansion around H*
Taylor series expansion around H ,
matches with the actual enthalpy at Tnew.
new
H curve —Actual enthalpy corresponding to the new Figure 3(c) shows the path followed by the iterative
new
temperature T . scheme during one particular time step. The errors are heavily
It is clear from Figure 3(a) that there is a marked differ- exaggerated for clarity of understanding and only two itera-
ence between the calculated and actual values of enthalpies, tions are shown here. Point [a] represents the old state of the
new new
i.e., H calc  H curve . The new enthalpy cannot be expressed node. Point [b] represents the new state that is calculated after
using a Taylor series expansion around Hold because of the the first iteration. As explained above, the algorithm recognises
discontinuity in slope of the enthalpy curve at To. The algo- that it has overshot from the material’s actual enthalpy curve
rithm recognises this difference between the two enthalpies and it chooses a new (H*, T*) which is given by point [c]. With
and chooses a new value of (H*, T*) to revaluate the new point [c] as (H*, T*), the second iteration is performed and the
enthalpy for the second iteration, as shown for point [c] in new state is calculated as point [d]. Since point [d] is found to
Figure 3(b). As can be seen, the H* for the second iteration is lie on the actual enthalpy curve, the algorithm converges.
new
equal to H calc at the end of first iteration, T* is the temperature The above method drastically reduces the potential error
new
corresponding to Hcalc from the H –T curve. Now T* > To and that could occur during phase change by using iterations at
614
ASHRAE Transactions
every time step. Thus the iterative enthalpy scheme does not finite volume method with a Crank Nicholson scheme thus
suffer from the requirement of small time steps unlike the making it similar to ESP-r. The initiative of developing the
explicit method explained in the previous section. MATLAB code is to study the PCM performance, and the
number of extra features similar to other whole building simu-
PCM Modeling in ESP-r lation tools is very minimal. The enthalpy method with the
The open source building simulation software ESP-r iterative correction as proposed by Swaminathan and Voller
includes an active materials module with which PCMs (1993) was implemented. The heat capacity was expressed
embedded in the building structures can be simulated. ESP-r using a step function, same as in ESP-r. One-dimensional
uses the finite volume method with options of implicit, explicit energy equation was applied to the building structure. Same as
or Crank Nicholson schemes. The apparent heat capacity in ESP-r, 9 nodes were set up for the finite volume analysis.
method with explicit scheme is applied (heat capacity calcu-
lated from previous temperature). The heat capacity during Model Description
phase change is expressed using a linear function as shown in
Equation (12). A low mass building, BESTEST Case 600 model,
described in ANSI/ASHRAE Standard 140-2007, Standard
C p eff  T  = aT + b ; T o  T  T e (12) Method of Test for the Evaluation of Building Energy Analysis
Computer Programs (ASHRAE, 2008) was used for the simu-
In current study, the effective heat capacity in ESP-r was lations to illustrate the numerical algorithms explained in the
expressed as a step function, i.e. a = 0 and b = L/  T, where L previous sections. This standard model was chosen to identify
and  T denote the latent heat capacity and phase transition the differences in predictions of ESP-r and MATLAB code.
temperature range, respectively. Although a step function like Figures 4(a) and 4(b) show the two house models used in this
heat capacity was applied, the general conclusions from this paper. Both models are identical in all aspects except for the
study are valid for any shape of heat capacity curve. In addi- south facing window in Case 2. The dimensions and the
tion, PCMs was modeled using a 4-layer structure in ESP-r construction materials of both models were chosen to match
instead of a single layer, i.e. a given thickness of PCM was with the BESTEST Case 600 model. The walls are made of a
divided into 4 layers and added to the structure with each wooden outer layer followed by layers of fiberglass quilt and
layer’s thickness as ¼ the original thickness. A total of 9 nodes plasterboard. A summary of the wall construction is shown in
were used in the 4-layer PCM structure. This was done to Table 2. In addition to the wall materials defined in the ANSI/
improve simulation accuracy during phase change (Almeida ASHRAE Standard 140-2007, an extra PCMs structure was
et al., 2010). added on the inside surface, with the thermophysical proper-
ties given in Table 3. For Case 2, the window was added to
MATLAB Building Simulation Code investigate the differences in the simulation results between
A customized building simulation code performing the two software when the direct solar gain was considered.
simple building simulations was developed using the commer- Different phase change temperatures were assigned to the
cial software package MATLAB. The MATLAB code uses PCMs in two models so that phase change occurs for a reason-
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 4 Schematics of two house models: (a) Case 1, no window, and (b) Case 2, south facing window.
©2012 ASHRAE
615
able amount of time during the simulation period. Since Case of 1 hour and with PCM embedded on the walls, ceiling and
2 (with window) permits more solar energy into the room, the floor. As can be seen, the temperature profiles match very
phase change temperature was chosen higher than that for closely when the temperatures are outside the phase change
Case 1. The thermal conductivity, density and heat capacity temperature range (19-20oC / 66.2-68oF). Since the nodal
remain the same before and after phase change. The heat surface temperature of inside walls is typically close to the
capacity of PCMs was expressed as a step function in both indoor ambient temperature, it also means that the results
ESP-r and MATLAB simulations. Toronto weather data
match very well when there is no phase change happening in
adopted from TRNSYS was employed in the simulation. All
the walls. However, when the ambient temperatures are
the simulation results are for the time period from May 1st to
July 31st. around the phase change temperature range, significant devi-
ation is observed between the two temperature profiles. More
distinguished differences between the results from ESP-r and
SIMULATION RESULTS
MATLAB can be seen in Figures 6(b) and 6(c), by zooming
onto certain regions in Figure 6. As shown in Figure 6(b), the
Case 1 (without Window)
ESP-r model exhibits much smaller amplitude of temperature
The results from MATLAB and ESP-r were first swing than the MATLAB model when it predicts the ambient
compared without any PCM incorporated. Figure 5 shows that temperature around the phase change temperature. This indi-
the two temperature profiles overlap on each other, with cates that the ESP-r model predicts a “better” thermal comfort
almost no discrepancy of results.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 6(a) shows the indoor ambient temperature Table 3. Thermophysical Properties of PCM used in
profiles obtained from ESP-r and MATLAB with a time step Current Study
Property Value
Table 2. House Construction Details
Thermal conductivity 0.3 W/m K (0.17 Btu/hr ft oF)
Wall layers Wood siding, fiberglass quilt and
(outer to inner) plasterboard, PCM Phase change Case 1: 19-20oC (66.2-68oF)
temperature range Case 2: 25-26oC (77-78.8oF)
Overall R value
1.944 m2K/W (11 h·ft²·°F/Btu) Heat capacity 1400 J/kg K (0.334 Btu/lb oF)
(wall without PCM)
Latent heat 80000 J/kg (34.4 Btu/lb)
Window U value
3 W/m2K (0.53 Btu/hr ft2 oF)
(Case 2) Thickness 8 mm (0.315 in)
Figure 5 Simulated indoor ambient temperature profiles from ESP-r and MATLAB with no PCM for time step of 1 hour,
May 1st to July 31st.
616
ASHRAE Transactions
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 6 Simulated indoor ambient temperature profiles from ESP-r and MATLAB (with iterative correction) with time step
of 1 hour, May 1st to July 31st.
©2012 ASHRAE
617
zone by maintaining the indoor temperature in a narrower can be even more exaggerated when HVAC and mechanical
range. In Figure 6(c), it can be seen that the two codes give systems are included. It should be noted that the file sizes and
different results when exiting the phase change and the results simulation run times reported above are specific to the 3 month
match more closely again when the temperatures are building simulations performed in this study.
completely out of phase change temperature range.
It was found that the results from the MATLAB model are Case 2 (with Window on South Wall)
almost independent of the value of time step used, i.e. reducing
The same room modeled in the previous section was
the time step did not change the MATLAB ambient tempera-
modeled with a window on the centre of the south facing wall
ture profile significantly. By reducing the time step in ESP-r to
(window size - 10% of south wall area). The phase change
2 minutes, the results were found to match more closely during
temperature range was changed to 25-26oC (77-78.8oF) in
phase change, as shown in Figure 7(a). The maximum error
order to increase the period for which the PCM experiences
between the two results was reduced to 0.2oC when the time
phase change. Since the PCM on the internal surfaces of the
step in ESP-r was taken as 2 minutes which is a significant
room is directly exposed to solar radiation, the difference in
reduction when compared to the maximum error of 3.6oC that
results between ESP-r and MATLAB is expected to be more.
occurred for an ESP-r time step of 1 hour. Figures 7(b) and 7(c)
were obtained for the same period as the regions shown in A double glazed window was used in the model, with a U
Figures 6(b) and 6(c), with a reduced time step in ESP-r. The value of 3 W/m2K (0.53 Btu/hr ft2 oF). The solar insolation that
observation in Figures 7(a-c) is consistent with our under- gets into the room through the window, as well as the distri-
standing of the two methods that were used in ESP-r and bution of this radiation on the internal surfaces are all taken
MATLAB. The explicit effective heat capacity method with a from ESP-r and given as inputs to the MATLAB model. This
temperature formulation used in ESP-r is expected to give is to ensure that there is no difference in the window modeling
accurate results only when the time step is reduced. The between the ESP-r and MATLAB codes. Any difference in
implicit enthalpy scheme used in MATLAB can however give results can thus be attributed to the difference in PCM model-
fairly accurate results even with large time steps of the order ing between the two codes.
of 1 or 2 hours. A comparison between ESP-r and MATLAB results was
In order to verify that the erroneous results in ESP-r with first done by setting the latent heat of PCM to zero, which is
large time steps were in fact due to the reasons explained equivalent to no PCM. As expected, the results were found to
above, the MATLAB model was run without any iteration match closely for the case of no PCM.
(explicit scheme). As explained in the Methodology section, Figure 9(a) shows ambient temperature profiles obtained
the first step in the iterative enthalpy method is exactly same from ESP-r and MATLAB with PCM from May 1st to July
as the effective heat capacity method. Since the iterations in 31st. Both results were obtained with a time step of 1 hour.
the enthalpy method were disabled, MATLAB gave results Similar to case 1, the temperature profiles match very closely
comparable to ESP-r with a time step of 1 hour for the same when the temperatures are outside the phase change temper-
time frame of Figures 6(a) and 7(a), as shown in Figure 8. It ature range (25-26oC / 77-78.8oF). More distinguished differ-
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
can be seen that both codes predict very close results, which ences between the results from ESP-r and MATLAB can be
further confirms our conjecture that the difference in results seen in Figures 9(b) and 9(c). As can be seen, ESP-r once again
between MATLAB and ESP-r was in fact due to the lack of predicts smaller amplitude of indoor temperature swing, simi-
iterative corrections in the effective heat capacity method used lar to results in Figure 6(b). The maximum difference in results
in ESP-r. was found as high as 5.7oC and hence the errors are even
An important parameter in computational studies is often higher for the model with window. Figure 10(a) shows the
the simulation runtime. On a state of the art personal computer, results comparison by reducing the time step in ESP-r to 2
it is found that the simulation run time in ESP-r is about 2 minutes and maintaining 1 hour time step for MATLAB. It can
seconds when the time step is held at 1 hour. This however be seen that the results match very closely (maximum error of
increases to almost 57 seconds when the time step is changed 0.4oC), even in the regions (b) and (c), as shown in Figures
to 2 minutes. While the actual run times may vary depending 10(b) and 10(c). It is clear from these plots that the ESP-r
on the actual computer used to run the simulation, it is however results approach MATLAB results when time step is reduced
clear that time step reductions in ESP-r produces a severe to as low as 2 minutes.
penalty on the simulation run time (almost a 30x increase). Figure 11 shows the number of iterative corrections
This once again emphasizes the need for an algorithm that can applied at each hourly time step in the MATLAB simulation
produce acceptable results at fairly large time step values. for case 1. As can be seen, the frequency when the iterative
Another disadvantage of using large time steps is the large size correction occurs is very high between hours 400 to 2000,
of the output files that are produced. The size of the ESP-r which corresponds to the regions in Figures 7(b) and 7(c).
output data for a time step of 2 minutes is about 130 MB which These high frequent iterations suggest that significant errors
is significantly larger than 4 MB output files that are produced may occur with the explicit scheme at large time steps. This
for a time step of 1 hour. The aforementioned disadvantages optimal iterative scheme can improve the accuracy of simula-
618
ASHRAE Transactions
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 7 Simulated indoor ambient temperature profiles from ESP-r and MATLAB (with iterative correction) with different
time steps, May 1st to July 31st.
©2012 ASHRAE
619
Figure 8 Simulated indoor ambient temperature profiles from ESP-r and MATLAB (without iterative correction) with time
steps of 1 hour.
tion for PCM heat transfer while remaining comparably large scheme used in current study can be applied to other building
time step values. simulation software to enhance simulation performance for
All the above results have been presented for 1 hour and PCMs.
2 minute time steps in ESP-r. The effect of intermediate time
steps in ESP-r is presented here by plotting the root mean ACKNOWLEDGEMENT
square error between the ESP-r and MATLAB room temper- This work was funded in part by the Solar Buildings
ature profiles for different time step values in ESP-r. This is Research Network under the Strategic Network Grants
shown in Figure 12 for both the models with and without Program and Discovery Grant of the Natural Sciences and
window. As can be seen, the root mean square error decays Engineering Research Council (NSERC) of Canada. The
rapidly and flattens out even at 16 time steps per hour (corre- authors would also like to acknowledge MITACS and OGS for
sponds to a time step of 3.75 minutes) for the model without their financial support.
window. However for the case with window, the root mean
square error decays more slowly and flattens out only at REFERENCES
around 30 time steps per hour (2 minute time step). Thus a 2-
minute time step has been chosen in all the comparisons Almeida F., Zhang D., Fung A.S. and W.H. Leong. 2010.
presented in this paper for maximum accuracy. Investigation of multilayered phase-change-material
modeling in ESP-R. Proceedings of 1st International
CONCLUSIONS High Performance Buildings Conference, West Lafay-
ette, USA.
Two methods to model phase change, namely the non-
iterative effective heat capacity method and the iterative ASHRAE. 2008. ANSI/ASHRAE Standard 140-2007, Stan-
enthalpy method have been explained in details in current dard method of test for the evaluation of building energy
paper. It is shown that the effective heat capacity method can analysis computer programs. Atlanta: American Society
lead to errors during phase change, whereas the enthalpy of Heating, Air-Conditioning and Refrigeration Engi-
method with iterative correction can give fairly accurate neers, Inc.
results even when the value of time step is quite large. This has Clarke, J. A. 2001. Energy simulation in building design.
been verified with case studies of two simple rooms subjected Butterworth-Heinemann.
to Toronto weather conditions. It was found that the building Crawley D.B., J.W. Hand, M. Kummert and B.T. Griffith.
simulation software ESP-r could give results that matched 2005. Contrasting the capabilities of building energy
with the MATLAB model using iterative enthalpy method performance simulation programs. A Joint Report, US
when the time step was reduced to as low as 2 minutes. DOE, University of Strathelyde UK, University of Wis-
MATLAB results were however found to be fairly indepen- consin-Madison USA and National Renewable Energy
dent of the time step value. The optimal iterative correction Laboratory USA.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
620
ASHRAE Transactions
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 9 Simulated ambient temperature profiles from ESP-r and MATLAB (with iterative correction) with time step of 1 hour,
with PCM case, May 1st to July 31st.
©2012 ASHRAE
621
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 10 Simulated indoor ambient temperature profiles from ESP-r and MATLAB (with iterative correction) with different
time steps, May 1st to July 31st.
622
ASHRAE Transactions
Figure 11 Number of iterative corrections applied at each hourly time step in the MATLAB code for case 2 with PCM.
Figure 12 Root mean square error between MATLAB and ESP-r temperature profiles.
Dalhuijsen A.J. and A. Segal. 1986. Comparison of finite Jokisalo J., Lamberg P. and K. Siren. 2000. Thermal simula-
element techniques for solidification problems. Int J tion of PCM structures with TRNSYS. Proceedings of
Numer Meth Engng, 23, 1807-29. 8th International Conference on Thermal Energy Stor-
Dutil, Y., Rousse, D. R., Salah, N. B., Lassue, S., & age, TERRASTOCK–2000. Stuttgart, Germany.
Zalewski, L. 2011. A review on phase change materials: Kuznik, F., Virgone, J. and J. Kevyn. 2010. Development and
Mathematical modeling and simulations. Renewable validation of a new TRNSYS type for the simulation of
and Sustainable Energy Reviews, 15, 112-130. external building walls containing PCM. Energy and
Buildings, 42, 1004-1009.
Gunther E., Mehling H. and S. Hiebler. 2007. Modeling of Mehling H. and L.F. Cabeza. 2008. Heat and cold storage
subcooling and solidification of phase change materials. with pcm: an up to date introduction into basics and
Modeling Simul. Mater. Sci. Eng. 15, 879-892. applications. Chapter 9, Applications for heating and
Heim D. 2010. Isothermal storage of solar energy in building cooling in buildings. D. Mewes and E.H.F. Mayinger ed.
construction. Renewable Energy, 35, 788-796. Berlin: Springer-Verlag Berlin Heidelberg.
©2012 ASHRAE
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
623
Nedjar B. 2002. An enthalpy-based finite element method Schossig, P., H.-M. Henning, S. Gschwander and T. Hauss-
for nonlinear heat problems involving phase change. mann. 2005. Micro-encapsulated phase-change materi-
Comput Struct, 80, 9-21. als integrated into construction materials. Solar Energy
Pedersen CO. 2007. Advanced zone simulation in EnergyP- Materials and Solar Cells, 89(2-3), 297-306.
lus: incorporation of variable properties and phase Stefan, J. 1889. Uber einige probleme der theorie der
change materials (PCM) capability. Proceedings of 10th warmeleitung. Sber Akad Wiss Wien, 98, 473–484.
International IBPSA Conference. Beijing. 1341-5. Stetiu, C. and H.E., Feustel. 1998. Phase-change wallboard
Rose, J., Lahme, A., Christensen, N.U., Heiselberg, P., Han- and mechanical night ventilation in commercial build-
sen, M. and K. Grau. 2009. Numerical method for calcu- ings. Lawrence Berkeley National Laboratory, Berkeley,
lating latent heat storage in constructions containing CA, USA.
phase change material. Proceedings of 11th Interna- Swaminathan C.R. and v.R. Voller. 1993. On the enthalpy
tional IBPSA Conference, Glasgow, Scotland, 400-407. method. Int. J. Num. Meth. Heat fluid flow, 3, 233-244.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
624
ASHRAE Transactions
SA-12-025
Occupancy Simulation in
Three Residential Research Houses
Philip R. Boudreaux Anthony C. Gehl Jeffrey E. Christian
Notice: This manuscript has been authored by UT-Battelle, Data comparing the energy use of each house will be presented
LLC, under Contract No. DE-AC05-00OR22725 with the U.S. and it will be shown that the second house used 37% less and
Department of Energy. The United States Government retains the third house used 67% less energy than the control house in
and the publisher, by accepting the article for publication, 2010. It will also be shown how an energy saving family can
acknowledges that the United States Government retains a further decrease energy use in the third house to 73% less over
non-exclusive, paid-up, irrevocable, world-wide license to the control house occupied by an average American family.
publish or reproduce the published form of this manuscript, or
allow others to do so, for United States Government purposes. INTRODUCTION
Three houses of similar floor plan are being compared for
ABSTRACT
energy consumption. The first house is a typical builder house
Three houses of similar floor plan are being compared for of 2400 ft2 (223 m2) in east Tennessee. The second house
energy consumption. The first house is a typical builder house contains typical retrofits available to a home owner such as
of 2400 ft2 (223 m2) in east Tennessee. The second house energy efficient appliances, windows and air-to-air heat pump,
contains retrofits available to a home owner such as energy as well as a sealed and insulated attic which contains the duct
efficient appliances, windows and air-to-air heat pump, as well work. The third house was built using optimum-value framing
as an insulated attic which contains the duct work. The third construction with photovoltaic modules and solar water heat-
house was built using optimum-value framing construction ing. Since occupancy puts an electrical load on the house, and
with photovoltaic modules and solar water heating. To simu- adds sensible and latent heat to the space, occupancy needs to
late typical occupant energy consumption researchers have set be simulated to evaluate residential building envelopes and
up appliances, lights, and plug loads to turn on and off auto- technology in these research houses. Since performance of the
matically according to a schedule based on the Building Amer- three houses is compared it is important that each house have
ica Research Benchmark Definition. As energy efficiency the same occupancy profile. If families were allowed to
continues to be a focus for protecting the environment, national occupy the houses then data would be skewed because of the
security, and conserving resources, experiments involving different living habits of each family. By simulating occu-
whole house energy consumption have shown to be extremely pancy this variable is removed and the data represents only the
valuable toward filling the need for real field performance data whole house performance.
without the enormous variability that comes with occupant The trio of three bedroom houses is located in the mixed-
behavior. This paper describes how to simulate the same aver- humid climate of east Tennessee. Performance of the second
age American family living in each of these research houses. and third houses is compared to the control house to evaluate
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
The process for achieving automated occupancy simulation the retrofit package and high performance residential home
will be discussed and the performance of the occupancy simu- construction. The control home, called the builder house, is a
lation system over the 2010 calendar year will be diagnosed. 2400 ft2 (223 m2) house with 100% incandescent lighting. A
Philip R. Boudreaux and Anthony C. Gehl are a R&D staff members and Jeffrey E. Christian is a retired R&D staff member in the Energy
and Transportation Science Division at Oak Ridge National Laboratory in Oak Ridge, TN.
©2012 ASHRAE
625
1.5 ton (5.3 kW) capacity heat pump with a seasonal energy HOW TO SIMULATE OCCUPANCY
efficiency ratio (SEER) of 13 supplies the first level, and a 2.5 The Building America Research Benchmark Definition
ton (8.8 kW) 13 SEER heat pump supplies the second level. (Hendron, R. 2008) was used as the source for designing the
The second house, called the retrofit house, has the same floor target control profile for the occupancy simulation system.
plan as the builder house but has significant retrofits. These The BA Benchmark attempts to capture the living habits of the
retrofits include high performance windows, a sealed and average American household, and how this affects energy
insulated attic, 3 ton (10.5 kW) 16 SEER single stage heat consumption. Through empirical formulas based on finished
pump with zone control, ducts in the sealed and insulated attic, floor area, number of bedrooms and number of occupants the
100% compact fluorescent (CFL) lighting, ENERGY STAR® BA Benchmark gives information for appliance cycles, hot
appliances, and a 50 gal (189.2 L) heat pump water heater. The water use, light loads, plug loads, etc. Described below are
third house, called the high performance house, has a similar how the occupancy simulation target profile was built using
floor plan as the first two houses and was built using optimum- the BA Benchmark, how the household shower, appliances,
value framing techniques for minimizing wood (which is a lights, sensible heat generators and latent heat generators are
thermal short in the envelope) and maximizing cavity insula- controlled, and finally how the whole house control is auto-
tion. Technologies used in the high performance house include mated and synchronized between all three houses. It should be
triple layer windows, 2 ton (7 kW) 16 SEER two stage heat noted that all energy values presented in this section are target
pump, ducts inside conditioned space, 100% CFL lighting, values. The occupancy controls and programming are
ENERGY STAR appliances, 2.5 kW peak photovoltaic (PV) designed to meet these targets. In the case of domestic hot
system, and a drain back solar water heating system with elec- water the water volume and temperature are the targets and in
tric back-up. For more detailed information on these houses the case of the heat pumps the air temperature is the target. In
see the DOE report (Christian J., et. al. 2010). the discussion section a diagnosis of the occupancy simulation
system performance is presented and actual measured energy
The occupancy simulation in each house consists of two use for the 2010 calendar year is shown.
separate systems. One system uses a custom appliance control
program that interfaces to a digital output module and relays. Domestic Hot Water
The program controls the master shower to simulate domestic
According to the BA Benchmark (Hendron, R. 2008), the
hot water draws, range, dishwasher, refrigerator and freezer
domestic hot water usage of showers, baths and sinks can be
doors, clothes washer, clothes dryer, and latent heat generator. computed with the equations in Table 1. Note that Nbr is
The other system controls the lights and sensible heat gener- number of bedrooms. The water use for the clothes washer and
ators using wireless mesh network technology for controlling dishwasher are not included because these appliances are
switches and plugs. The Building America Research Bench- automatically operated daily.
mark Definition (Hendron, R. 2008) was used as a basis for With three bedrooms in each of the research houses, the
building the control profile for occupancy simulation. Using total daily water use is 60 gal (227.1 L) at 105°F (40.6°C). The
the finished floor area and number of bedrooms, a target target hourly domestic hot water consumption can be
profile based on the average American family was built. In this computed for each end use from the BA Benchmark (Hendron,
report, the process used to construct the occupancy simulation R. 2008). In Figure 1, the collated target hourly consumption
target profile, and the design of the control hardware and soft- curve for sinks, showers and baths is presented. This consump-
ware will be discussed. Measured data from the houses is tion pattern is an average of many homes in America and does
presented to demonstrate how the occupancy simulation not capture the use in one household. Because of this concern,
performed. Energy data shows that the high performance discrete shower discharges were programmed which total to
house uses 67% less, and the retrofit house uses 37% less 60 gal (227.1 L). These five discharges are shown as square
energy than the builder house. The researchers will also markers in Figure 1.
discuss how behavior can further reduce energy use in the To simulate domestic hot water use of showers, sinks, and
high-performance house by 73% over an average American baths, the master shower is used. A solenoid valve, flow meter,
family living in the control house. and thermistor immersion probe were installed before the
Table 1. Domestic Hot Water Usage Per Day for End Use (Hendron, R. 2008)
Water Usage for Three

End Use Water Usage, gal Average Temperature
Bedroom House, gal
Shower 14 + 4.67 x Nbr 28.01 (106 L) 105°F (40.6°C)
Bath 3.5 + 1.17 x Nbr 7.01 (26.5 L) 105°F
Sinks 12.5+ 4.16 x Nbr 24.98 (94.6 L) 105°F
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
626
ASHRAE Transactions
Figure 1 Water (105° F) consumption of showers, sinks and baths for a 3 bedroom house.
showerhead. The shower valve was adjusted until the water is operated five times a week then the total cycles for the year
coming out of the shower was approximately 107°F (41.7°C) will be 260.
and then left at that position to yield a daily average shower
392   0.5 + N br  6 
temperature of about 105°F (40.6°C). Because the valve is Clothes washer cycles per year = (1)
always on, a check valve was installed on the hot water side so  12.5 lbs / W test
that the hot water would not drain through to the cold water
side. Occasionally the shower valve needs to be adjusted (this 0.84
Dryer cycles per year = (2)
was done twice during 2010, once in February and once in  Clothes washer cycles per year
September) because the cold water temperature coming into
Table 2 presents the profile for the washer and dryer
the house changes. In this mixed humid climate the cold water
control. To correctly simulate a wet load entering the dryer,
comes into the water heater at 45°F (7.2°C) in February and
a mister was installed in the dryer to moisten the load before
75°F (23.9°C) in August. The flow rate is measured at the
the drying cycle starts. Notice the mister time varies for each
showerhead in each house. The inverse flow rate is then multi-
house (see shaded area in Table 2). In the builder house the
plied by the volume needed to be discharged (see Figure 1) to
mister adds 0.97 gal (3.7 L) of water to the load and in the other
derive a “time-on” for the solenoid valve. To maintain consis-
two houses the mister adds 0.64 gal (2.4 L). This is because the
tent flow rates the shower head is periodically cleaned (this
washer in the retrofit and high performance houses are high
was done once in May of 2010).
efficiency front loading units and the unit in the builder house
The solenoid valve is controlled by a relay connected to is a traditional top loading washer. The high efficiency wash-
a digital output module by an Ethernet cable. This module is ers spin more water out of a load and this is taken into account
controlled by the custom appliance control program. This soft- when the dryer load is misted. The different misting times in
ware has independent loops for each device it controls, when the retrofit and high performance houses were needed because
the current time matches the scheduled “on-time” for the the mister flow rate is slightly different at each house. The
shower the program toggles a digital bit on the digital output washer is set to wash with hot water and both the washer and
module corresponding to the shower solenoid valve. Once the dryer contain approximately nine pound loads, which is the
solenoid has been on for the appropriate duration, the program average load of a washer of this capacity according to the Code
toggles the bit off, which closes the relay and solenoid valve, of Federal Regulations (CFR. 2011).
stopping the shower water flow. To control the washer and dryer the units were disassem-
bled to gain access to the control boards. The buttons for power
Clothes Washer and Dryer
and start were wired at both poles and then connected to relays
Equations 1 and 2 describe the clothes washer and dryer so the contact closures could be controlled remotely and auto-
cycles per year (Hendron, R. 2008). Wtest is the maximum matically. These relays are connected to a digital output
clothes washer test load weight, which is 15.4 lbs (6.99 kg) for module which is connected to the computer running the appli-
the washers in each of the three houses. Based on Equation 1 ance control program. When it is time to start a cycle the
and 2, the clothes washer cycles per year is 318 and the dryer program toggles the appropriate bit for turning power on to the
cycles per year is 267. If the washers are operated six times a unit, waits a few seconds, then toggles another bit correspond-
week, then 312 cycles a year will be attained and if the dryer ing to the “start cycle” button. The dryer mister is controlled
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
©2012 ASHRAE
627
Table 2. Washer and Dryer Profile
Dryer Mist
Washer Start Wash Days* Dryer Start Dry Days*
Duration, Seconds
7:00:00 4 19:00:00 4 Builder 240
17:00:00 4 9:45:00 1,7 Retrofit 192
8:00:00 1,7 11:35:00 1,7 High Perf. 157
10:00:00 1,7
*Sunday =1, Wednesday=4, Saturday=7
Table 3. Range Profile
Oven Start Oven Duration, Minutes Oven Days*
12:00:00 40 1,2,3,4,5,6,7
16:00:00 160 1,2,3,4,5,6,7

*1-7 represents day of the week
by a relay and solenoid valve. All user inputs for the type of rather a standard for testing refrigerator doors. Since no better
cycle are set manually at the appliance front panel. Washers citable reference was found these values were used.
are set to a large load with heavy cottons, hot wash and cold Both the freezer and refrigerator doors are opened and
rinse, a normal soil level, and extra high spin (for the retrofit closed by a pneumatic system consisting of an air compressor,
and high performance houses). All dryers are set to a normal an electrically controlled valve and pneumatic piston with an
sensor dry cycle with a high dry temperature. arm and hinge for opening the doors. The electrically
controlled valve is wired to a relay that is controlled with the
Dishwasher
appliance control program. To add thermal mass 16.2 gal (61.4
Equation 3 describes the dishwasher cycles per year L) of water was placed in the refrigerator and 54.7 lbs (24.8 kg)
(Hendron, R. 2008). For a three bedroom house Equation 3 of ice was placed in the freezer in each house. Each refriger-
yields 215 cycles. However, to meet the average water use of ator has a combined refrigerator and freezer volume of 23 ft3
the dishwasher in a three bedroom house (see Table 10 in the (0.65 m3).
BA Benchmark) of 1825 gal (6908 L) per year, it should be
cycled 326 times a year since the dishwasher in the builder Range
house was measured to use 5.6 gal (22.7 L) per cycle. If the
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
dishwasher is cycled 6 times a week then a total of 312 cycles Equation 4 describes electricity consumption per year for
for the year will result. Note that the dishwashers in the retrofit an electric range (Hendron, R. 2008). For three bedrooms
and high performance house use less water than in the builder Equation 4 yields 605 kWh per year, which is 1.658 kWh per
house. The dishwashers are started at 7:30 PM on Sunday day.
through Friday. Dishwashers in each house are loaded with
eight place settings, and are set for normal wash and heated Range energy consumption (kWH/yr) = 302 + 101  N br (4)
dry. The dishwasher is controlled in the same fashion as the
clothes washer and dryer, with the buttons on the control board Since using the range top in an unoccupied house presents
wired to relays and controlled remotely through Ethernet cable a fire hazard the oven is used to consume the whole range load.
by the appliance control program. For the builder house it took 200 minutes of baking at 350°F
(176.7°C) to consume 1.658 kWh. This time was used for the
Dishwasher cycles per year = 215   0.5 + N br  6  (3) ranges in the other two houses as well. Table 3 presents the
schedule for range control.
Refrigerator
The range is controlled with a 240V relay installed in-line
The Code of Federal Regulations (CFR, 2009) states that with the oven element. The oven is set to bake at 350°F
the refrigerator should be opened 24 times a day and the (176.7°C). When the relay is open there is no current flowing
freezer six times with the doors staying open for 12 seconds. to the oven element, but when the relay is closed the element
This citation does not convey human occupancy habits but is heats the oven.
628
ASHRAE Transactions
Hard Wired Lighting Load hour and after the light has been on for the prescribed time, the
To determine the profile for lighting, the researchers first controller turns the light off.
computed a profile for the builder house (100% incandescent)
and applied this profile to the lights in the retrofit and high Sensible Heat Loads
performance houses (100% CFL). Equations 5 – 7 describe the
total lighting energy (Hendron, R. 2008), where FFA is Two ceramic heaters, one downstairs in the living room
finished floor area and is 2400 ft2 for each house. Note that the and one upstairs in the master bath, plugged into wireless mesh
1.12 factor in each equation is needed to correct for a house network enabled outlets, are used to add sensible heat to the
with 100% incandescent lighting as is described in the BA space. The outlets are controlled in the same way as the light
Benchmark. The total lighting energy in the builder house is switches discussed above. The heaters add heat contributed to
computed to be 2520 kWh/yr or 6.904 kWh per day. the space by miscellaneous electric loads (MELs), plug-in
lighting, and people. Since the occupancy patterns change
1.12  0.8 between weekend, weekday, upstairs and downstairs, different
Interior hard-wired lighting (kWh/yr) = (5)
  FFA  0.8 + 455  profiles will be used for adding sensible heat to the space
contributed by people. Table 5 shows sensible heat depending
Garage lighting (kWh/yr) = 1.12  100 (6) on zone for MELs, plug-in lighting and people (Hendron, R.
Exterior lighting (kWh/yr) = 1.12  250 (7) 2008). The MELs per day was computed by summing the vari-
able miscellaneous electric loads, fixed miscellaneous loads
In Figure 2, the lighting energy target per hour for one day (gas/electric) and fixed miscellaneous loads (all electric)
is presented for the builder house. This plot was created using described in the BA Benchmark. The daily total is split
the hourly lighting curve given in the BA Benchmark and between the living area and bedroom area, with 75% of the
multiplying the curve by 6.904 kWh. The hard wired lights in energy being consumed in the living area and 25% being
the builder house that were equipped to be controlled were consumed in the bedroom area. The plug-in lighting load was
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
measured with a power meter. These values along with also computed using the BA Benchmark and split up the same
Figure 2 were used to create a profile to describe the runtime way as the MELs. Note that the plug-in lighting in the retrofit
for each light load. Table 4 shows an abbreviated version of the and high performance homes is 75% less than the energy in the
lighting profile. This profile was also used in the retrofit and builder house. This is because the builder house is simulating
high performance house for controlling lights, where the light- incandescent plug-in lights and the retrofit and high perfor-
ing energy is expected to be lower because of CFL lighting. mance house is simulating CFL plug-in lights. The sensible
The hard wired lights have wireless mesh network heat per hour from the three occupants was computed by
enabled switches that can be remotely controlled. The control- multiplying the sensible load per person per hour for each zone
ler is a computer without monitor, keyboard or mouse that from the BA Benchmark (0.0674 kWh for the living area and
wirelessly communicates with each light switch. The control- 0.0615 kWh for the bedroom area) by the number of occu-
ler reads a database that contains the target profile in Table 4. pants. Equation 8 describes the number of occupants in a
The “on-time” for the lights is always at the beginning of the house based on the number of bedrooms (Hendron, R. 2008).
Figure 2 Daily lighting profile for the builder house of 2400 ft2 (223 m2) with 100% incandescent lighting.
©2012 ASHRAE
629
Table 4. Abbreviated Lighting Profile, Minutes
Hour of the Great Master Outside Front

Bed2 Breakfast Kitchen Master Bed Office
Day Room Bath Front Porch
1 23.71
2 23.71
3 23.71
4 23.71
5 50.49
6 25.22 60.00
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
7 60.00 12.40
8 20.00 60.00 20.00 4.66
9 13.72 5.00
16 10.00 10.00 10.00 10.00 21.64
17 30.00 60.00 15.00
18 35.00 48.62
19 31.36 60.00
20 31.35 60.00 60.00
21 60.00 60.00 17.35 10.00
22 60.00 60.00
23 43.78 30.00
24 57.78
Table 5. Sensible Heat by Source Per Day or Hour
Living Room Bedroom
MELs energy per day, kWh 4.845 1.615
0.976* 0.325*
Plug-in lighting energy per day, kWh
0.244** 0.081**
Total occupant energy per hour, kWh 0.202 0.185

*builder house
**retrofit and high performance houses
Number of occupants = 0.5  N br + 1.5 (8) Figure 4 shows the sensible heat energy target of the
combined MELs and plug-in lighting for the retrofit and high
Occupancy curves, from the BA Benchmark, showing performance house. These curves were calculated using
fraction of total occupants in the space were multiplied with Table 5 and the BA Benchmark (Hendron, R. 2008). The daily
values from the row titled “Total occupant energy per hour” in target for the plug load (combined MELs and plug-in lighting)
Table 5 to calculate the total sensible heat energy from people is 7.761 kWh for the builder house and 6.785 kWh for the other
that needs to be added to the space (Hendron, R. 2008). two houses.
Figure 3 shows these targets for zones (living and bedroom)
and time of week (weekday and weekend). The daily occu- The curves from Figures 3 and 4 can be added and will
pant sensible heat target for the weekday is 3.012 kWh and for result in sensible heat curves for each zone and time of week.
the weekend is 3.587 kWh. After the heaters are measured for power consumption, the
630
ASHRAE Transactions
Figure 3 Total sensible heat contributed by occupants per hour for two zones and time of week for a three bedroom house
of 2400 ft2 (223 m2).
Figure 4 Sensible heat target made up of MELs and 100% CFL plug-in lighting per hour for two zones in the retrofit and
high performance house (same for weekday and weekend).
“time on” for each heater for the weekday and weekend peri- “time-on” values. The actual power of each heater varies
ods can be calculated using Equation 9 below. In Equation 9, slightly, so the profile is customized to match individual heat-
En is the sensible heat energy in kilowatt-hours needed for a ers. To illustrate this, the three 0.5 kW heaters varied from
particular hour, Pheat is the measured power of the heater in 0.494 kW to 0.542 kW.
kilowatts, and tn is the time on in minutes.
t n = E n   60  P heat  (9) Latent Heat
Table 6 shows the final target schedule for the sensible The latent heat generators were commissioned in Febru-
heat generators, the left side of the table shows the energy that ary of 2011. These systems consist of a point-of-use 2.5 gal
needs to be added to the space per hour for each zone and time (9.5 L) water heater insulated in a large trash can. The duty
of week, the right side of the table shows the “time-on” needed cycle on the water heater elements are changed according to
for the heater to achieve the desired load. This “time-on” how much latent heat needs to be added to the space. The units
depends on the measured power of the heater. Typically a have water level sensors and are automatically refilled as
larger 1.5 kW heater is used in the living zone and a smaller 0.5 needed. Since the units add sensible heat to the space as well
kW heater is used in the bedroom zone, so this will affect this value is subtracted from the sensible heat curves for each
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
©2012 ASHRAE
631
Table 6. Example Sensible Heat Generator Target Profile
Total Energy, kWh Time, Minutes
Weekday Weekday Weekend Weekend Weekday Weekday Weekend Weekend

Hour of Day
Living Bedroom Living Bedroom Living Bedroom Living Bedroom
1 0.181 0.245 0.181 0.245 7.2 29.4 7.2 29.4
2 0.172 0.242 0.172 0.242 6.9 29.0 6.9 29.0
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
3 0.167 0.240 0.167 0.240 6.7 28.8 6.7 28.8
4 0.167 0.240 0.167 0.240 6.7 28.8 6.7 28.8
5 0.161 0.238 0.161 0.238 6.4 28.6 6.4 28.6
6 0.187 0.247 0.187 0.247 7.5 29.6 7.5 29.6
7 0.318 0.165 0.217 0.257 12.7 19.8 8.7 30.8
8 0.326 0.136 0.326 0.167 13.1 16.3 13.1 20.1
9 0.243 0.062 0.320 0.062 9.7 7.4 12.8 7.4
10 0.184 0.053 0.260 0.053 7.4 6.3 10.4 6.3
11 0.189 0.055 0.265 0.055 7.6 6.5 10.6 6.5
12 0.189 0.055 0.265 0.055 7.6 6.5 10.6 6.5
13 0.184 0.053 0.260 0.053 7.4 6.3 10.4 6.3
14 0.189 0.055 0.265 0.055 7.6 6.5 10.6 6.5
15 0.198 0.058 0.274 0.058 7.9 6.9 11.0 6.9
16 0.211 0.062 0.287 0.062 8.4 7.4 11.5 7.4
17 0.252 0.076 0.362 0.076 10.1 9.1 14.5 9.1
18 0.377 0.092 0.411 0.092 15.1 11.0 16.5 11.0
19 0.509 0.102 0.442 0.102 20.4 12.3 17.7 12.3
20 0.524 0.107 0.457 0.107 20.9 12.9 18.3 12.9
21 0.534 0.110 0.467 0.110 21.4 13.3 18.7 13.3
22 0.514 0.104 0.514 0.104 20.6 12.5 20.6 12.5
23 0.365 0.180 0.365 0.180 14.6 21.6 14.6 21.6
24 0.223 0.259 0.223 0.259 8.9 31.1 8.9 31.1
house and the sensible heat generator targets were corrected cooking as described in the BA Benchmark (Hendron, R.
appropriately (after February 2011). 2008).
To compute the target profile for latent heat generation,
the occupancy patterns in the BA Benchmark were used. 3412.14  0.30
Electric range latent load (BTU/yr) = (10)
According to the BA Benchmark 570 BTU (0.167 kWh) and   302 + 101  N br 
420 BTU (0.123 kWh) of latent heat are emitted from three
occupants in the living area and bedroom area respectively. According to Equation 10, 1697 BTU (0.497 kWh) of
This results in a daily latent heat load target from occupants of latent heat is added to the space each day from cooking. One
2.243 kWh for the weekday and 2.698 kWh for the weekend. latent heat generator in the kitchen is used for adding latent
Latent heat from cooking also needs to be added to the space heat in the living area due to living space occupancy and cook-
because no food is actually being cooked when the oven is ing. Another latent heat generator in the master bathroom is
cycled each day. Equation 10 describes the latent heat from used to add heat based on bedroom zone occupancy. Figure 5
632
ASHRAE Transactions
Figure 5 Latent heat generator target profile including people and cooking for a three bedroom house.
shows the latent heat hourly target for the latent heat genera- counted using energy data recorded at one minute intervals for
tors depending on time of week and zone. 2010. A similar technique was used for the shower in counting
gallons of water flowing through the shower head. The shower
Software and Synchronization temperature was averaged for the year and compared to the
As has been stated, a custom appliance control program target temperature. For heat pump control performance the
running on a laptop connected to a digital output module and inside air temperature for the year was averaged and compared
another computer controlling wireless mesh enabled switches to the thermostat set point. The transition months of April,
and plugs are the two systems used to control the occupancy May and October were not included in this average. The plug
simulation hardware. The appliance control runs 24 hours a load control diagnostics was straight forward and only
day and the code contains many simultaneous loops, one for involved comparing the measured energy for the year to the
each appliance being controlled. The target profiles discussed energy target (found in Table 5 and is the sum of the MELS and
above are read into the appliance control program by a text file. plug-in lighting yearly energy). To determine the performance
This file resides on a file transfer protocol (FTP) server. A of the lighting control the total measured energy for interior,
scheduled task on the laptop downloads the profile every garage, and exterior lighting in 2010 was compared to the
night. This file is then read into the program for instructions on target lighting energy. The target lighting energy was
simulating occupancy. Although changes are not typically computed by multiplying the lighting profile, which describes
made to the profile, this method allows one file to be changed. how long and when each light in the home should turn on and
Since all three houses download this profile from the FTP off, by the measured wattage of each fixture in each house. For
server it assures that the houses are controlling synchronously. each house the duration that each light is on is the same, but the
The database file for the wireless mesh system is also read target is different because each fixture has different wattages
from an FTP server every night. To change the control of the (the retrofit and high performance house lighting target is
lights or sensible heat generators, the file on the FTP server is lower than the builder house, because they use CFL bulbs).
replaced. Table 7 presents the targets and actual measured cycles,
energy, water volume, temperature or time for individual
DISCUSSION systems for the 2010 calendar year. Also presented in the table
are the percent error between the actual performance and the
The occupancy simulation scheme presented above was
target.
implemented in all three research houses in 2009 except the
latent heat generator which was not commissioned until Note that the highest percent errors in Table 7 are from the
February of 2011, therefore the following data does not reflect range in the retrofit and high performance houses. The error in
adding latent heat to the space. To determine how well the the range control occurs during power outages. The electron-
occupancy simulation performed, successful appliance cycles ics in the range do not reset to allow control after the power is
(dishwasher, clothes washer, and dryer) were counted by restored, therefore human intervention is required. This is not
analyzing 2010 energy data and compared, in Table 7, to what the case for the builder range. The 2010 total range energy will
is specified in the target profile described above. To diagnose be corrected, as if each range operated the full 73000 minutes,
the range control, the total minutes that the range was on was before comparing the energy consumption of the homes. For
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
©2012 ASHRAE
633
Table 7. 2010 Measured System Energy Compared to Target Energy
BA Yearly Target Measured 2010 Data Percent Error
High High High

Builder Retrofit Builder Retrofit Builder Retrofit
Performance Performance Performance
Range,
73000 71891 65814 62595 -2% -10% -14%
minutes
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Shower 21109 20295 21460
21900 (82900 L) -4% -7% -2%
volume, gal (79906 L) (76825 L) (81234 L)
Shower
Temperature, 105 (41°C) 104 (40°C) 102 (39°C) 100 (38°C) 1% 3% 5%
°F
Dishwasher,
312 313 312 308 0% 0% -1%
cycles
Clothes
312 296 309 308 -5% -1% -1%
Washer, cycles
Clothes Dryer,
260 253 253 241 -3% -3% -7%
cycles
Conditioned
Space Summer: 76 (24°C), 74 (23°C), 75 (24°C), 75 (24°C), 3%, 1%, 1%,
Temperature, Winter: 71 (22°C) 71 (22°C) 72 (22°C) 71 (22°C) 0% 1% 0%
°F
Plugs,
2833 2476 2476 2870 2464 2401 1% 0% -3%
kWh
Lights,
2520 921 637 2615 920 661 4% 0% 4%
kWh
this correction the ratio of total measured energy of the range 3993 and 1929 (base 65°F or 18.3°C) respectively. The 2010
to measured minutes that the range was on for 2010 was calcu- calendar year contained 4260 heating degree days and 2037
lated which resulted in an “energy per minute” value. This cooling degree days.
ratio was then multiplied by the target time that the range Table 8 also shows individual sub-metered energy totals.
should have been on for the year. The refrigerator, clothes washer and dishwasher in the retrofit
Table 8 shows the energy use of each sub-metered system and high performance houses are ENERGY STAR appliances.
and the whole house energy use for 2010 as well as percent This explains the savings for the retrofit and high performance
savings for the retrofit and high performance house over the refrigerators and washer (high efficiency front loader) and
builder house. The measured data for the washer and dryer in dryer combination when compared to the builder house.
the retrofit house for June and July and the water heater in the However, notice that the dishwasher energy in the builder
retrofit and high performance homes in January was corrected house is the lowest. It is not clear why the dishwashers in the
(Christian J., et. al. 2010). The retrofit house used 37% less retrofit and high performance homes used more energy than
energy and the high performance house used 67% less energy the builder dishwasher.
than the builder house. Without the PV the high performance Energy consumption for water heating in the high perfor-
house used 52% less energy than the builder house. The energy mance house was 1093 kWh, for the retrofit the energy
savings in the high performance house is due mostly to water consumption was 1586 kWh and for the builder house 3830
heating, lighting, and space conditioning energy savings over kWh. The high performance house has a drain back solar water
the builder house. An EnergyGauge® model (EnergyGauge. heater with an 85 gal (321.8 L) storage tank with electric back-
2011) of the three homes was completed and predicted a up. The retrofit house has a heat pump hybrid water heater and
yearly energy use of 21543 kWh, 14828 kWh, and 11519 kWh the builder house has an electric water heater. The high perfor-
(without PV) for the builder, retrofit, and high performance mance house with the solar water heater used 72% less energy
homes respectively (Christian J., et. al. 2010). This model for water heating than the electric water heater in the builder
assumes long term normal heating and cooling degree days of house in 2010. Note that this water heating energy savings in
634
ASHRAE Transactions
Table 8. 2010 Total and Sub-system Measured Energy.
Percent Difference Over Builder House

Builder Retrofit High Performance
Retrofit High Performance
Refrigerator Energy,
479 411 408 -14% -15%
kWh
Dishwasher Energy,
248 320 312 29% 26%
kWh
Range Energy, kWh 576 553 569 -4% -1%
Washer and Dryer

1127 835 847 -26% -25%
Energy, kWh
Water Heater Energy,

3830 1586 1093 -59% -72%
kWh
HP1 + HP2 Energy,

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
11824 7825 5029 -34% -58%

kWh
Plug Energy, kWh 2870 2464 2401 -14% -16%
Lighting Energy, kWh 2615 920 661 -65% -75%
PV Energy, kWh 0 0 -3487
Total Energy, kWh 23569 14914 7833 -37% -67%
the retrofit and high performance homes is also due to the performance house saved 58% in energy for space condition-
energy efficient clothes washer and dishwasher which both ing over the builder house.
use less hot water than the corresponding units in the builder The MELs plus plug-in lighting target value is 2833 kWh/
house. yr for the builder house and 2476 kWh/yr for the retrofit and
high performance house for the entire year. Measured energy
The builder house has two heat pumps both 13 SEER, one
values for plugs compare favorably for 2010. Hard-wired
2.5 ton (8.8 kW) in the unconditioned attic serving the upstairs
lighting in the builder house is 100% incandescent, lighting in
and one 1.5 ton (5.3 kW) in the garage serving the downstairs.
the retrofit house is 100% CFL, and lighting in the high perfor-
Ducts are not in the conditioned space and have a leakage rate
mance house is 100% CFL with pin-style lamp bases which
to the outside (as compared to the blower air flow) of 9%
save energy over E85 base CFLs because the number of trans-
downstairs and 16% upstairs. In the retrofit house a single 3 formers is minimized. The lighting consumption for the
ton (10.5 kW) 16 SEER heat pump is used. This single unit has builder house was 2615 kWh, for the retrofit 920 kWh and for
dual zones for upstairs and downstairs. All ducts are in the the high performance house 661 kWh. The high performance
sealed and insulated attic with a leakage rate to the outside of house saved 75% in lighting energy. This energy savings was
6%. The high performance house has a single 2 ton (7 kW) 16 also due to a retrofit to the high performance house of LED
SEER dual speed compressor heat pump with zone dampers lighting in October 2010. This retrofit saved about 7-8 kWh a
for conditioning upstairs and downstairs. All ducts are in the month in lighting energy over the pin style CFL bulbs.
conditioned space and have a leakage rate to the outside of 0%. The comparisons thus far have focused on the equipment
The envelope tightness has a large effect on space conditioning and envelope performance rather than behavior because the
efficiency. The builder house had a blower door result of 5.7 occupancy control is based on an average American family
air changes per hour (ACH) at 1 lb ft-2 (50 Pa), the retrofit and is the same in each house. Energy consumption differ-
house was measured at 3.4 ACH at 1 lb ft-2 (50 Pa) and the high ences between the houses are dominated by lighting, water
performance house measured at 2.4 ACH (Christian J., et. al. heating and energy efficient space conditioning equipment in
2010). The high performance house has an energy recovery order of highest to lower percent savings in the high perfor-
ventilator for bringing in fresh air from the outside; the energy mance house over the builder house. Space condition savings
from this device is added to the heat pump energy. For space are due to efficient equipment, energy recovery ventilator,
conditioning, Table 8 shows that the builder house used 11824 tighter envelopes, and lower duct leakage. Another large vari-
kWh, the retrofit house used 7825 kWh, and the high perfor- able in energy efficiency is the habits of the occupants. The
mance house used 5029 kWh of energy in 2010. The high research houses described above are a powerful household
©2012 ASHRAE
635
laboratory where the researchers have full control over equip- since ducts are in the conditioned space, air sealing of ducts is
ment and behavior because of the occupancy simulation not as important. Table 9 shows a comparison of the high
systems. A brief discussion of how behavior can affect energy performance house yearly energy use broken down into sub-
consumption follows. system with an average family and the energy conscious
family described above. It should also be noted that the follow-
How Does Behavior Affect Energy Consumption ing is not an extensive list of all energy saving habits that could
The Department of Energy (DOE) has many energy be practiced, but simply how some select habits might
saving tips that require changed behaviors in order to save decrease energy usage.
energy (ENERGY STAR. 2011, EERE. 2011). The authors Energy data for the dishwasher in the high performance
chose behaviors described by the DOE that could easily be house was analyzed, and it was determined that an energy
estimated or programmed into the occupancy simulation so savings of 21% is possible if the heated dry were not used.
that yearly energy savings due to changed behavior could be Using a clothes line in place of using the clothes dryer for half
estimated. What if energy conscious families moved into the of the year’s dryer cycles would save 378 kWh per year (half
high performance house, how would energy consumption of the dryer’s measured energy). It was stated earlier that LED
compare to an average American family? The opportunities lighting was used in the house for three months and this saved
for changed living habits and estimated energy savings will be 7-8 kWh per month. If the LED lights were used year around
discussed. The opportunities discussed are as follows: not an additional 68 kWh per year could be saved.
using heated dry on the dishwasher, using a clothes line instead About 13% of MEL energy is contributed by standby
of clothes dryer for half of the annual dryer cycles, setting back loads (Roth, K. 2008). A total of 2401 kWh for 2010 was used
the thermostat, using LED bulbs in hard-wired and plug-in for plug loads with 4.79% of the energy contributed from CFL
lighting, using power strips with auto-switches to eliminate plug-in lighting. If standby loads were eliminated, by using an
standby loads, washing clothes with cold water and minimiz- auto-switching power strip, then the MEL energy could be
ing bath and shower times. It should be noted that many of the reduced to 1989 kWh. If the energy conscious home owners
tips at the aforementioned websites are already practiced in the also replaced the CFL plug-in lighting with LED lighting then
high performance house such as setting the water heater to 16% of plug-in lighting energy could be saved, resulting in a
120°F, using ENERGY STAR appliances and lighting, and yearly total of 97 kWh. So a total of 2086 kWh would be used
Table 9. Yearly Energy Use for an Average Family and Energy Conscious Family in the High Performance House
Average American Energy Conscious Percent

Energy Conscious Habit
Family Family Difference
Refrigerator Energy,
408 408 0%
kWh
Dishwasher Energy,
312 246 -21% Eliminated heated dry
kWh
Range Energy,
569 569 0%
kWh
Washer and Dryer Energy,

847 469 -45% Used clothes line to dry half of the loads
kWh
Water Heater Energy, Wash clothes with cold water, take half
1093 732 -33%
kWh the time for showers and baths
HP1 + HP2 Energy, Winter set point: 69°F (20.6°C),

5029 4727 -6%
kWh Summer set point: 78°F (25.6°C)
Plug Energy, Eliminate standby loads, Use all LED

2401 2086 -13%
kWh plug in lighting.
Lighting Energy,
661 593 -10% Use all LED hard wired lighting
kWh
PV Energy, kWh -3487 -3487 ----
Total Energy, kWh 7833 6343 -19%
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
636
ASHRAE Transactions
for plug loads with the energy conscious family living in the in the retrofit house and the water heater in the retrofit and high
high performance house. performance house were also corrected. This correction was
There are two ways, among many available, that this needed due to appliance failure.
energy conscious family might save energy in water heating. For 2010 the retrofit house had an energy savings of 37%
First, they could wash clothes in cold water. Second they could over the builder house and the high performance house had an
reduce the water use for showers and baths to half of the energy savings of 67% over the builder house. An energy
current volume. To estimate the savings of these habits a savings of 19% in the high performance house was estimated
detailed look at water use in the high performance house in if the home were occupied by an energy prudent family over
2010 was made. The solar water heater consumed 1093 kWh the average American family. This research shows the impor-
to heat 19768 gal (74830 L) of water. Approximately 7% of tance of using occupancy simulation controls in uninhabited
this water use was from the dishwasher, 8% from the clothes research homes. Without the occupancy controls the yearly
washer, 49% from showers and baths, and 36% from sinks. If measured loads could not be compared with a model. It was
half of the shower and bath water volume were eliminated and also learned that controlling a house in this manner requires
all of the hot water for the clothes washer, then a total of 13306 constant attention to both detect and correct control errors,
gal (50368.7 L) of hot water would have been used in 2010; appliance failures, and re-lamping needs. The energy compar-
this is a savings of 33%. If this savings is applied to the energy isons of the three homes show that energy efficient design and
used for water heating then a total 732 kWh would be used for technology, as well as energy conscious living habits can
2010 with these energy conscious habits being practiced. decrease the load of a home by 73%. This research was funded
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Currently in the summer the thermostat is set to 76°F by the Tennessee Valley Authority and the United States
(24.4°C) and in the winter it is set to 71°F (21.7°C), if these set Department of Energy Building America Program.
points were changed to 78°F (25.6°C) in the cooling season
and 69°F (20.6°C) in the heating season then a total of 6% REFERENCES
could be saved in the space conditioning load. This was esti-
Christian, J., T. Gehl., P. Boudreaux, J. New, and R. Dockery.
mated by running a calibrated EnergyGauge model of the high
2010. Tennessee Valley Authority’s Campbell Creek
performance house. Table 9 shows a comparison of the high
Energy Efficient Homes Project: 2010 First Year Perfor-
performance house yearly energy use broken down into cate-
mance Report July 1, 2009- August 31, 2010. Report
gory with an average family and the energy conscious family
ORNL/TM-2010/206. Oak Ridge National Laboratory,
described above. The total energy use for the average family
USA. http://info.ornl.gov/sites/publications/files/
in the high performance house minus PV energy is 7833 kWh.
pub26374.pdf
The total energy use of the energy saving family minus the PV
Code of Federal Regulations. 2009. Uniform Test Method
energy is 6343 kWh for the year; this is a 19% savings over an
for Measuring the Energy Consumption of Electric
average family in the high performance house. A 73% total
Refrigerators and Electric Refrigerator-Freezers. Title
energy savings results with an energy efficient family in the
10. Part 430, Subpart B, Appendix A1.
high performance house over the average American family in
the builder house. Code of Federal Regulations. 2011. Uniform Test Method
for Measuring the Energy Consumption of Automatic
SUMMARY and Semi-Automatic Clothes Washers. Title 10. Part
430. Subpart B. Appendix J1.
Details describing hardware and software used to auto-
mate occupancy simulation have been given above. The logic EERE. 2011. Energy Savers Tips. http://www.energysav-
and sources used to build the target profiles for the occupancy ers.gov/tips/
simulation were also presented. When the performance of the ENERGY STAR. 2011. ENERGY STAR @ home tips. http:/
occupancy simulation was diagnosed for 2010, errors at 7% /www.energystar.gov/index.cfm?c=prod-
and below were shown except in the range control in the retro- ucts.es_at_home_tips
fit and high performance house. This is because the control of EnergyGauge. 2011. EnergyGauge, Energy and Economic
these two appliances failed when the power went out at the Analysis Software. http://www.energygauge.com/
homes which happened occasionally when severe storms Hendron, R. 2008. Building America Research Benchmark
came through the area. Sometimes control errors were not Definition. Report NREL/TP-550-44816. National
noticed until a few days after it failed, resulting in missed Renewable Energy Laboratory, USA.
appliance cycles. This measured range data was corrected Roth, K., K. McKenney, and J. Brodrick. 2008. Small
before comparing the energy consumption of the research Devices, Big Loads. http://www.tiaxllc.com/publica-
houses. The measured energy data for the washer and dryer tions/small_devices_big_loads.pdf
©2012 ASHRAE
637
SA-12-026 (RP-1509)
Study of the Degradation of Typical HVAC

Materials, Filters and Components Irradiated
by UVC Energy—Part I: Literature Search
Robert E. Kauffman
ABSTRACT systems is a summary of the project final report (Kauffman

2011). Two books published by J. Rabek (1995 and 1996)
Ultraviolet light with a primary wavelength of 254 nano-
were the basis for the conducted literature search and contain
meters (0.01 mil), referred to as UVC, is being used to kill
thousands of references on research performed to study poly-
microbes in the air and on the surfaces of coils/drain pans in
mer photodegradation mechanisms. Although the majority of
HVAC systems and on the surfaces of food processing equip-
the reported UV studies were performed with UVA [320-400
ment. UVC also is being combined with catalytic surfaces to
nm (0.013 – 0.016 mil)] and UVB (280 – 320 nm (0.011 –
remove volatile organic compounds from the air. Since many
0.013 mil)], a significant number of the studies explored the
organic compounds present in HVAC systems and processing
effects of UVC on polymer degradation. The results of the
equipment may be exposed to the UVC light due to poor shield-
conducted literature search were broken down into the follow-
ing or design, the potential of material degradation leading to
ing subcategories for discussion.
component failure, and possibly, total system malfunction is
significant. Standardized methods for exposing, then evaluat- 1. UVC absorption by the fresh polymeric surface
ing, nonmetallic materials have been developed for UVA and 2. Photodegradation reaction mechanism
UVB but not UVC. Consequently, a research program consist- 3. Effects of UV wavelength on degradation products
ing of a literature search and the UVC exposure of numerous 4. Effects of contaminants and additives on degradation rate
polymers and nonmetallic compounds representative of 5. Effects of UV irradiance on degradation rate (Reciprocity
common materials present in HVAC systems was performed. Law)
The following paper describes the results of the literature 6. Correlation between laboratory and HVAC UVC results
search performed to identify the photodegradation mecha- 7. Identification of laboratory photoreactors
nisms of various polymers exposed to UVC, the photoreactors 8. Identification of photodegradation monitoring tech-
capable of performing long-term, accelerated UVC exposure niques
tests of multiple samples and the analytical techniques best-
suited for evaluating the photodegradation suffered by the UVC ABSORPTION BY THE FRESH
UVC exposed samples. POLYMERIC SURFACE
In order for UV to degrade a polymeric surface, chromo-
INTRODUCTION
phoric groups (chemical groups capable of absorbing UV
The literature search presented herein describing the light) must be contained in the chemical structure of the poly-
different degradation mechanisms of the materials and mer or in compounds dispersed in the polymer matrix. Exam-
nonmetallic components typically subjected to long-term ples of chromophoric groups present in common polymers are
UVC [254 nanometers (nm), 0.01 mil] exposure in HVAC carbonyls (polyesters, nylon, acrylates, polyimide, etc.) and
Robert E. Kauffman is a distinguished research chemist at the University of Dayton Research Institute, Dayton, OH.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
638
ASHRAE Transactions
phenyl rings [polystyrene, epoxy, poly (ethylene terephthal- the UV absorption of the degraded polymer products. Alter-
ate), etc]. nately, if the scission step occurs at the end or in a short side
Since they do not contain chromophoric groups, polyeth- chain of the original polymer, the UV exposure produces vola-
ylene and polypropylene do not absorb UV light above 220 nm tile carbonyl products (ketones/aldehydes, carbon dioxide,
(0.0087 mil) (Rabek 1996), and therefore, should be totally carbon monoxide) and a shortened polymer chain with a free
resistant to sunlight (UVA and UVB) as well as UVC. Poly radical end. In contrast to scission, the cross-linking steps
(vinyl chloride) does not absorb light above 290 nm (0.011 cause the molecular weight of the polymer to increase and the
mil) (Rabek 1996), and therefore, should only be slightly flexibility of the polymer to decrease.
susceptible to UVA and UVB degradation. In addition to the general reaction scheme in Figure 1,
However, in practice, polyethylene, polypropylene and several polymers have reaction schemes specific to their
poly (vinyl chloride) are all highly susceptible to UVA - UVC chemical structure. For example, poly (vinyl chloride) also
degradation (Denizligil and Schnabel 2003, Hamid 2000, undergoes dehydrochlorination when exposed to UV (Deni-
Rabek 1995 and 1996, Schnabel 1981, Scott 1965 and Hamid zligil and Schnabel 2003, Kaczmarek 2009 and Rabek 1996)
2000). The UV reactivities of the polymers are attributed to the as shown in Figure 2.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
presence of UV absorbing additives and impurities formed The resulting double bonds (C = C) are chromophoric
during polymerization, processing and/or storage. The impu- (increase the UVC reactivity of the reaction product) and are
rities can be both internal (polymerization) and external responsible for the yellow-red coloring of the exposed poly-
(processing/storage) and include hydroperoxides, carbonyl mer. The fact that the poly (vinyl chloride) and any other chlo-
and unsaturated bonds, catalyst residues, additives, metal rinated polymers would outgas hydrochloric acid gas indicates
traces, etc. (Rabek 1995 and 1996). Consequently, the capa- UVC exposure would have the potential to promote corrosion
bility of a polymer to absorb UV energy is affected by both the of surrounding metallic surfaces during polymer degradation.
chemical structure of the polymer as well as the presence and Polycarbonates, polyurethanes, poly (phenyl acrylate)
location of the chromophoric impurities. and epoxy resins undergo the photo-Fries rearrangement
(responsible for yellowing of aged polymers) shown in Figure
PHOTODEGRADATION REACTION MECHANISM 3 when exposed to UV (Rabek 1995 and 1996). The rear-
Once the organic compound absorbs the UV light, the ranged polymers produced in Figure 3 are more susceptible to
produced excited state must be capable of producing radicals UV degradation than the original polycarbonate polymer.
and other reactive species for the UV degradation mechanism
to continue. The reaction scheme in Figure 1 is basically the EFFECT OF UV WAVELENGTH
same whether the polymer free radicals (P•) are generated by ON DEGRADATION PRODUCTS
thermal or UV energy (photolysis). Although the focus of the literature search was on the
In the presence of oxygen, the scission steps are generally specific UVC wavelength of 254 nm (0.01 mil) [253.7 nm
dominant with regard to the cross-linking steps (Rabek 1995 (0.00999 mil) to be precise] photodegradation of polymers,
and 1996). If the scission step occurs in the polymer back the majority of the UV experiments in the identified references
bone, the UV exposure reduces the tensile strength of the were performed in the UVA and UVB region or with an
exposed polymer and the carbonyl groups (ketone/aldehyde)
attached to the end of the severed polymer molecules increase
Figure 2 Dehydrochlorination reaction of poly (vinyl

chloride).
Figure 1 Photodegradation reaction scheme of

polymeric materials. Figure 3 Photo-Fries rearrangement of polycarbonate.
©2012 ASHRAE
639
unspecified wavelength. However, several authors (Rabek In addition to the type of degradation products produced,
1995) have noted that the reaction products obtained with the wavelength of the UV light affects the depth of the polymer
UVA/UVB light are lower in quantity (produced radicals have surface undergoing photodegradation. For low absorbing
less energy) but similar in composition to those produced with polyethylene and polypropylene, UV light in the 290 – 360 nm
254 nm (0.01 mil). For instance, both poly (vinyl chloride) and (0.011 – 0.014 mil) range caused significant scission reactions
polycarbonate produce yellow compounds (Figure 2 and 3) to depths of 1.5 – 0.4 mm (59 – 16 mil) (depth decreased with
when exposed to either UVA/UVB or UVC light (Rabek increasing wavelength) (Shyichuk 2005). For high absorbing
1996). Under 254 and ~300 (UVB) nm (0.01 and 0.012 mil) polystyrene, polycarbonate, acrylonitrile–styrene and poly
light, polypropylene decomposition produced similar prod- (methyl methacrylate), (Nagai 2004) reported that the chem-
ucts such as ketones, peroxy acids, peroxy esters, etc. (Aslan- ical changes due to exposure to UVA/UVB was less than 5
zadeh and Kish 2005). microns (0.20 mil) deep. Comparing the depth penetration of
Any differences in the UVA/UVB and 254nm (0.01 mil) an acrylic resin, 50% of the incident light penetrated to a depth
(UVC) photodegradation products arise from the higher of 10 mm (390 mil) for 364 nm(0.014 mil) light, to 1 mm (39
energy of the UVC light (able to break bonds stable to lower mil) for 313 nm (0.012 mil) light and to only 0.01 mm [10
energy of UVA/UVB). Polyacetal undergoes depolymeriza- microns (0.39 mil)] for 254 nm (0.1 mil) light (Feller 1994).
tion to only produce formaldehyde when exposed to 360 nm
EFFECTS OF CONTAMINANTS AND ADDITIVES
(0.014 mil) UV light but produces ethyl alcohol and ethylene
ON DEGRADATION RATE
glycol as well as formaldehyde when exposed to 254 nm (0.01
mil) (Chiang and Huang 1999). In addition to the Photo-Fries As previously discussed, the photodegradation of many
rearrangement causing the polycarbonate to yellow, the higher non-absorbing polymers are attributed to the presence of UV
energy of UVC is able to directly break (C-O) bonds in the absorbing impurities formed during polymerization, process-
original polycarbonate structure to form reactive free radicals ing and/or storage. The impurities can be both internal
and volatile products (Hamid 2000). (synthesis) and external (processing/storage) and include
Regardless of the UV energy, once the photodegradation hydroperoxides, carbonyl and unsaturated bonds, catalyst
is initiated, the scission degradation products increase the UV residues, additives, metal traces, etc. (Rabek 1995 and 1996).
absorption of the polymer surface due to the presence of Consequently, the capability of a polymer to absorb UV
carbonyl groups (C=O) and double bonds (C=C). Based on the energy is affected by both the chemical structure of the poly-
experimental results reported in the identified references mer as well as the presence and location of chromophoric
(Rabek 1995 and 1996, Rabney and Rabek 1975 and Hamid impurities. For instance, the susceptibility of polystyrene to
2000), Table 1 was compiled to compare the susceptibilities of photodegradation is highly dependent on how it was synthe-
different materials to photo-initiated scission and gas produc- sized (Rabek 1995). Radically prepared polystyrenes are more
tion [dependent on the efficiencies of both the photolysis (UV susceptible to UV degradation than anionically prepared poly-
absorption) and scission steps in Figure 1]. The materials are styrenes since they contain double bonds, in-chain peroxide
listed in order of decreasing yield, i.e., decreasing scission linkages, and other oxygen containing groups, all of which
(gas production) with same level of UV exposure indicates increase the UV absorption/reactivity of the internal and exter-
increasing resistance to UV degradation. nal polystyrene matrix.
In contrast to polystyrene, the UV susceptibilities of poly-
The results in Table 1 indicate that the susceptibility of the ethylene and other extruded polymers are dependent on their
different polymers to scission are more dependent on polymer thermal history. The hydroperoxides formed at low tempera-
composition than on the UV wavelength used in the experi- tures [50 - 90°C (122 – 194°F)] are isolated and have minimal
ment. Many of the researchers reported that the quantum photo-initiating effects (radical produced by impurity is not
yields of scission and other reactions (cross-linking, gaseous transferred to polymer). Whereas, the hydroperoxides formed
products, etc.) were independent of the wavelength of UV irra- at high [135 - 160°C (275 – 320°F)], prolonged temperatures
diance. on the surfaces of the extruded polymers are associated and
Regardless of the wavelength used in the studies, the have significant photo-initiating capabilities (Rabek 1995).
primary volatile product of the UV irradiation was water orig- Consequently, low density polyethylene would be expected to
inating from the hydroxyl radical abstraction of a hydrogen be more resistant to UV photodegradation than high density
atom from the polymer (Figure 1). Carbon dioxide and carbon polyethylene due to its lower extrusion temperatures.
monoxide were also common degradation products resulting In addition to unwanted impurities, additives such as
from the scission reaction in Figure 1 when the carbonyl group plasticizers are added intentionally to polymers to aid
was at the end of the polymer radical. Other reported volatile processing and improve the flexibility of the final product.
products such as hydrochloric acid (Figure 2), formaldehyde Commercial plasticizers such as di-n-octyl adipate/phthalate
(polyacetal), formic acid [poly (ethylene terephthalate)], etc. esters are capable of acting as photo-initiators for poly (vinyl
were more dependent on the composition of the polymer being chloride). Consequently, the degradation products of
irradiated than the wavelength of the UV light. commercial poly (vinyl chloride) products can originate from
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
640
ASHRAE Transactions
Table 1. Quantum Yieldsa of Chain Scission and Gas Evolution
for UV Irradiation of Different Polymers
UV Wavelength Quantum Yield

Polymer
(nm) of Scission (Gas)
Poly(phenyl isopropyl ketone) 254 0.17 – 0.22
Poly(methyl vinyl ketone) 254 0.025
Poly(methyl methacrylate) 254 0.02 – 0.04
300 0.016 - 0.005 (0.0002)
Poly(methyl phenyl silane) 313 0.017
Poly(vinyl chloride) 254 - 400 (0.015) (loss of HCl)
Poly(methyl acrylate) 254 0.013
Poly(  -methyl styrene) 254 0.007

Poly(vinyl acetate) 254 0.005, 0.05 (0.0065)
EPDM Rubber ----b 0.003
Poly(ethylene terephthalate) 254 0.0016 (0.0002)
280 - 360 0.0005 (0.0002)
Polystyrene 254 0.0015 -0.0005 (0.000045)
Cellulose 254 0.001 – 0.0007
Poly(vinyl pyrrolidone) 254 0.00043
Natural Rubber (cis-1,4-polyisoprene) 254 0.0004 (0.001)
Polysulphones 254 0.00084
Nylon 6 254 0.0007
Polycarbonate 260 -300 0.0007 – 0.0001
Polyacrylonitrile 254 0.0002 – 0.00077

Polyurethanes 254 (0.00014)
Mixed phenyl – methyl polysiloxanes 254 (0.000026)

a Quantum Yield = Number of molecules decomposed/Number of photons absorbed
b Unspecified
both the plasticizer as well as the polymeric matrix (Deni- EFFECTS OF UV IRRADIANCE ON THE
zligil and Schnabel 2003). DEGRADATION RATE (RECIPROCITY LAW)
In contrast to plasticizers, carbon black has been reported
to inhibit the UV photodegradation of different polymers such One of the primary purposes of the literature search was to
as polyacetal (Chiang and Huang 1999) and polyethylene identify research that indicated that short-term, high irradiance
(Scott 1965). All of the references identified during the liter- UVC experiments could be used to accurately predict the
ature search were concerned with the protective effects of performance of polymers exposed to long-term, low irradiance
carbon black with regard to weathering (UVA/UVB), not UVC, i.e., UVC exposure obeys the reciprocity law. The reci-
UVC exposure. The presence of other inorganic fillers capa- procity law is obeyed if the photodegradation of the polymer is
ble of forming a protective coating (chalking) on the surface of dependent only on the total energy of UV exposure (irradiance
the polymer or acting as an energy sink for the UV energy x time) and is independent of the time or irradiance level taken
absorbed by the polymer would be expected to have an inhib- separately. In other words, the degree of photodegradation
itory effect on the 254 nm (0.01 mil) photodegradation of would be the same whether produced by 10,000 microwatts per
polymers. square centimeter (µW/cm2) [64,500 microwatts per square
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
©2012 ASHRAE
641
inch (µW/in2)] for 100 hours or 500µW/cm2 (3220 µW/in2) for orders by the accelerated UVA/UVB laboratory and standard-
2000 hours. ized outside weathering tests (Hamid 2000).
When dealing with materials without fillers such as poly- The surface photodegradation results produced by UVC
ethylene, researchers have reported that the rate at which UV lighting in the accelerated laboratory tests and HVAC systems
(wavelength unspecified) irradiated polyethylene absorbs are expected to have better correlation than the results of the
oxygen (oxidation) is proportional to the square root of the UVA/UVB laboratory tests and weathering tests since the
light irradiance (Scott 1965). Also, the researchers determin- exposure environments of the laboratory photoreactors and
ing the quantum yields for different polymers in Table 1 HVAC systems are controlled, surface erosion is minimal and
reported that the quantum yields of scission and other reac- sun light is not involved. The sample vibration (rotating plat-
tions (cross-linking, gaseous products, etc.) were independent form) and the frequent air flow (maintain constant tempera-
of the irradiance. ture) during the accelerated laboratory tests are expected to
More to the point of this project, research with UV light help simulate the effects of HVAC operation on the photodeg-
in the wavelength and irradiance ranges of 290 to 400 nm radation rates (enlarge microcracks, remove particles, etc.) of
(0.011 to 0.015 mil) and 3,600 to 32,200 µW/cm2 (23,200 to exposed materials. The expected UVC laboratory test corre-
207,700 µW/in2), respectively, demonstrated that the photo- lation with the HVAC systems is further improved when the
degradation of acrylic-melamine coatings obeyed the reci- exposed materials are at a controlled temperature in the HVAC
procity law (Chin 2005). Fourier Transform Infrared (FTIR) system.
analyses of the irradiated coatings determined that the rates of To further increase the probability of reciprocity between
scission, photo-oxidation and mass loss were directly propor- the photodegradation rates produced by the laboratory and
tional to dosage regardless the time of irradiance. The refer- HVAC systems, the UVC irradiances of the accelerated labo-
ences identified by Chin (2005) also stated that the ratory tests were selected to be with in an order of magnitude
photodegradation of acrylic coatings, poly (vinyl chloride), of those used in HVAC systems (Kauffman 2011). For
polycarbonate, poly-?-methylstyrene, acrylonitrile butadiene- instance, the UVC irradiances for air and cooling coil treat-
styrene and poly (butylene terephthalate) obeyed the law of ments range from 1 to 20,000 µW/cm2 (6.4 to 129,000 µW/in2)
reciprocity (wavelength unspecified). (IUVA 2005) and 10 to 500 µW/cm2 (64.5 to 3220 µW/in2)
(Kowalski 2009), respectively. Consequently, the irradiance
CORRELATION BETWEEN LABORATORY range of the developed accelerated laboratory tests was
AND HVAC UVC RESULTS between 900 and 14,000 µW/cm2 (5800 and 90,300 µW/in2).
Even for the deeper penetrating UVA/UVB, in which the rate
The main goal of this project is to allow HVAC designers
of photodegradation is limited by oxygen diffusion, the labo-
to select the best material for use in an UVC application and
ratory tests obeyed reciprocity (Chin 2004) when the lower
to allow maintenance personnel to estimate the useful life of
UV irradiance level was within 10x of the upper irradiance
an in-service HVAC component based on its UVC dosage. For
level used. The range of UVC irradiance levels that obey reci-
the laboratory test results to be the most useful for the designer
procity should be even wider since UVC mainly produces scis-
or maintenance personnel, the law of reciprocity must be
sion reactions (Rabek 1995 and 1996) and its depth of
obeyed by both the laboratory test and HVAC results. The reci-
penetration is reported to be only 10 microns (0.39 mil) (Feller
procity law is obeyed if the photodegradation of the material
1994 and Shyichuk 2005). Consequently, the UVC scission
is dependent only on the total energy of UV exposure (irradi-
step (Figure 1) will always occur in the presence of air, regard-
ance x time) and is independent of the irradiance time, geom-
less of the UVC irradiance level, so that the rate of photodeg-
etry, chamber, environment, etc.
radation will only depend on the level of UVC irradiance
Various factors such as: (criteria for reciprocity) independent of the oxygen diffusion
rate.
• wind, particle erosion, rain,
One factor that could lead to large differences in the labo-
• large temperature variances between day and night, ratory and HVAC results is the wavelength spectrum of the
• air polluants (ozone, nitrogen oxides, etc.), UVC light used by the HVAC system. According to the IUVA
• wavelengths/angle of sunlight versus UVA/UVB used in Draft Guideline IUVA-G01A-2005, the two most common
lab test, types of UV germicidal lamps are the medium and low pres-
• light and dark cycles, etc sure mercury lamps. The emission spectra of the low and
medium pressure lamps are quite different as shown in
are listed in ASTM G151 and ISO 4892-1 to explain the poor Figure 4. The emission spectrum of the UVC lamps to be used
correlation between accelerated laboratory (ASTM G154) in the accelerated laboratory tests matched the spectrum of the
and standardized outside (D1435) weathering tests (Scott low pressure lamp in Figure 4.
1965 and Rabek 1995 and 1996). Even with their lack of Consequently, the photodegradation rates of the UVC
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,
correlation, studies have shown that the susceptibilities of laboratory tests and HVAC systems are expected to have good
different materials to photodegradation are ranked in similar correlation and reciprocity when the HVAC systems employ
642
ASHRAE Transactions
Figure 5 Identified photoreactor (reproduced with
permission from Luzchem).
Figure 4 Comparison of medium and low pressure

mercury lamp spectra with a typical germicidal
instrumented to monitor the irradiance reaching the surface of
effectiveness curve (reproduced from IUVA
the test solutions and the test chambers are temperature regu-
2005).
lated to ensure repeatable experimental conditions. The LZC-
ICH2 photoreactor (Luzchem Research, Inc. Ottawa, Ontario)
has an inside chamber that is 30 cm wide, 30 cm deep and 22
low pressure mercury lamps. Poor correlation is expected cm high) (12 in. x 12 in. x 8.5 in.) that is lined with an alumi-
between the photodegradation rates of UVC laboratory tests num alloy (Al 5052-H32) to maximize UVC reflections and is
and HVAC systems employing medium pressure lamps [depth equipped with 16 UVC lamps (Figure 5) to ensure the UVC
of penetration and accelerated reactions dependent on wave- irradiance is consistent throughout the reaction zone. A rotat-
length (Feller 1994, Nagai 2004 and Shyichuk 2005)]. ing octagon turntable 21.6 cm wide with 8.9 cm sides (8.5 in.,
3.5 in. sides) can be added to the photoreactor to ensure all of
One other significant wavelength that is produced by a
the samples receive equal levels of UVC exposure to allow the
low pressure lamp is the 185 nm (0.0073 mil) line. The quartz
photodegradation rates of successive tests to be compared.
envelope of the UVC lamp in Figure 4 absorbs the 185 nm
The spectral output of the UVC lamps employed by the ICH2
(0.0073 mil) so it is not transmitted to the environment. If the
photoreactor is well characterized with ~ 96% of the UV
lamp envelope is produced from synthetic quartz (Heraeus
energy (92.5% UV, 7.2% visible and 0.5% infrared) produced
2009) or has a defect, then both the 185 and 254 nm (0.0073
by the 254 nm (0.01 mil) line [similar to 254 nm (0.01 mil)
and 0.01 mil) lines are transmitted resulting in ozone produc-
lamps in Figure 4].
tion [185 nm (0.0073 mil) line]. A large portion of the gener-
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
ated ozone is converted into diatomic oxygen (O2) and a Even though the ICH2 lamps are designed to produce
singlet oxygen (O·) (Jones and Wayne 1969 and Qu 2005) by only 254 nm (0.01 mil), ozone measurements were made
the 254 nm (0.01) line. The quantum yield of ozone photolysis inside the photoreactor to ensure ozone was not being
by 254 nm (0.01 mil) is nearly unity, i.e., ratio (number of produced by a lamp defect [185 nm (0.0073 mil) transmitted
molecules converted/number of photons absorbed) is almost 1 through the defect]. Measurements determined that the ozone
for ozone compared to <0.02 for most polymer photodegra- level inside the photoreactor was below the instrument detec-
dation (Rabek 1995 and Table 1). In the presence of water tion limit of 0.03 parts per million by volume confirming the
[water treatment (Spartan water treatment 2009)] or water proper functioning of the photoreactors. Since varying ozone
vapor [surface cleaning systems (Kim 1996)], the ozone concentration is one of the factors contributing to poor corre-
photolysis by 254 nm (0.1 mil) results in the production of lation between accelerated laboratory tests and weathering
highly reactive hydroxide radicals (Figure 1) and oxygen. tests (Scott 1965), ozone destruction by the UVC lamps in
both the accelerated laboratory tests and HVAC systems
IDENTIFICATION OF LABORATORY should further aid in their mutual reciprocity.
PHOTODEGRADATION CHAMBERS
IDENTIFICATION OF PHOTODEGRADATION
In order to compare the photodegradation rates of differ- MONITORING TECHNIQUES
ent type materials in the irradiation range of 900 and 14,000
µW/cm2 (5800 and 90,300 µW/in2) photoreactors designed to The final topic of the literature search was to identify the
perform chemical and biochemistry experiments involving analytical technique(s) best suited to monitor the photodegra-
UVA – UVC exposure were identified. The photoreactors are dation of the non-metallic samples exposed to 254 nm (0.01
©2012 ASHRAE
643
mil) UV light. Although 254 nm (0.01 mil) exposure causes i.e., UVC volatilization of polymer surface left behind protec-
the irradiated molecules to undergo free radical induced tive layer of inorganic filler.
oxidation (Figure 1) followed by scission reactions, and to a
The Energy Dispersive Spectrometric (EDS) elemental
lesser extent cross-linking reactions (Rabek 1995 and 1996),
analyses of the polymer samples containing fillers also indi-
the affected molecules are concentrated close to the surface
cated that the UVC exposed surface contained a much higher
(Feller 1994, Rivaton 2002). Also, the oxidation of the surface
inorganic content than the interior material (Kauffman 2011).
molecules (C=O and C=C bonds increase) increases the 254
nm (0.01 mil) absorption by the surface further decreasing the As opposed to Figure 6, the FTIR and EDS spectra of
depth of the UVC penetration into the material (Rabek 1995 polymers not containing fillers were virtually identical for the
and 1996). Consequently, even though photodegradation interior material and UVC exposed surface (Kauffman 2011).
strongly affects the mechanical properties and chemical Surface analyses indicated that the UVC exposed organic
composition of the exposed surface, the bulk properties and surface had a distinctive pattern, i.e., mass loss of clean surface
composition of the polymer remain unaffected so that struc- concentrated at susceptible locations (grain structure, crystal-
tural integrity tests such as tensile strength, impact and flexi- linity, etc.) as previously discussed (Rabek 1995 and 1996).
bility are considered impractical for monitoring UVC
Consequently, all of the samples underwent organic mass
photodegradation during the accelerated laboratory tests.
loss during UVC exposure in agreement with Figure 1 indi-
Since the identified literature was not focused on HVAC cating that photolysis (direct bond breaking) and photooxida-
applications, several scrap samples (Witham 2009) from tion scission reactions producing volatile products (mass loss)
HVAC material UVC testing were obtained for initial study occur at a much higher rate than cross-linking reactions
(dosage levels unknown). The scrap samples were analyzed (harden surface). The polymers without fillers would be
with different analytical techniques to identify the type of expected to undergo constant mass loss regardless of dosage
surface photo-oxidation mechanisms occurring and to identify while the polymers containing fillers would lose mass until an
the analytical test(s) with the best potentials for monitoring the exterior layer of inorganic fillers formed to protect the interior
accelerated 254 nm (0.01 mil) photo-oxidation tests. Since the molecules from further UVC exposure and photodegradation
original, unexposed material was not available for compari- reactions (Figure 6). Several references were identified which
son, a portion of each sample’s surface was scraped away to stated that surfaces exposed short-term to 254 nm (0.01 mil)
reveal unexposed material for analysis. had mass loss and other changes in morphology (Kaczmarek
For polymers with fillers, the FTIR spectra of the unex- 2005 and 2009, Rabek 1996 and Soto-Oviedo 2002). Longer-
posed materials (inside) had distinctive organic peaks (C-H, term 254 nm (0.01 mil) tests with linear polyethylene with out
C=O, C-O, etc.) not present in the spectra of the UVC exposed fillers (Ranby and Rabek 1975) demonstrated that mass loss
surfaces as illustrated by the representative spectra in Figure 6, occurred for the entire 800 hour test as indicated by the
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 6 Representative FTIR spectra of exposed surface and inside (revealed by scraping) of sample containing fillers after
UVC exposure.
644
ASHRAE Transactions
constant rate of oxygen adsorption by the exposed surface depths of the UVC formed craters by sequentially scanning
(fresh material exposed by mass loss reacts with oxygen). across the protected then exposed surface areas. The primary
In addition to the HVAC scrap material analyses, addi- drawback of the surface profilometers is that the sample
tional polymers without fillers were exposed to UVC for surface has to be polished to a level finish prior to masking and
different lengths of time and then visually inspected for color UVC exposure.
changes as performed in ASTM methods such as D6290 or
International Standard ISO 4582. The surface of a polyure- SUMMARY
thane sealer turned light orange after only 10 hours of UVC The results of the literature search indicate that even
exposure [~4,000 µW/cm2 (25,800 µW/in2) with UVC pen though the rates of UVC photodegradation are affected by a
light]. After 50 and 200 hours of UVC exposure, the surface wide range of factors (processing parameters, contaminants,
remained light orange in color, i.e., extended UVC exposure fillers, additives, etc.) and vary by several orders of magnitude
did not further darken the exposed areas since the oxidized for different polymers (Table 1), the primary photodegrada-
species were not accumulating due to mass loss. In contrast to tion mechanisms are photolysis (direct bond breakage) and
the polyurethane sealer, the polished surface of a polyacetal photooxidative scission (Figure 1). The results of the literature
rod did not change color (exposed area remained white but search and initial UVC irradiated samples further indicate that
became dull) even after 200 hours of UVC exposure. the primary change in the surface of the UVC irradiated
Based on these initial results and the literature search, samples is mass loss as opposed to the compositional changes
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
surface compositional changes were not pursued for monitor- observed during UVA and UVB tests. For polymers without
ing UVC photodegradation damage. Although the FTIR spec- fillers, the mass loss due to UVC exposure is expected to obey
trum peak heights are well-suited for quantification, the the law of reciprocity, i.e., photodegradation of the polymer is
produced spectra did not change with dosage due to surface dependent only on the total energy of UV exposure (irradiance
mass loss (no fillers) and became independent of the polymer x time) and is independent of the time or irradiance level taken
(filler accumulation due to polymer loss). Color changes, separately. For polymers with fillers, the mass loss is expected
although simple to perform with a colorimeter, are strongly to obey the law of reciprocity until the layer of protective inor-
affected by surface finish and do not change with exposure ganic particles/fibers forms on the exposed surface inhibiting
time (surface craters observed), and consequently, do not further mass loss. Analytical balances for measuring sample
correlate well with degree of photodegradation. Finally even weight loss or surface profilometers for measuring the depths
though monitoring the rates of oxygen absorption (Ranby and of craters formed in masked sample surfaces were identified as
Rabek 1975) and/or volatile compounds produced during the best techniques for monitoring the degree of surface UVC
mass loss could be used to monitor the rate of UVC degrada- photodegradation.
tion of a polymer surface, the large numbers of samples to be The literature search also identified photoreactors with a
irradiated in the accelerated UVC laboratory tests made the rotating turntable and intermittent air flow (simulate HVAC
monitoring of oxygen consumption and/or produced gases environment) that were capable of irradiating multiple small
from individual samples impractical. samples under controlled temperature conditions as required
Consequently, an analytical technique capable of moni- by the accelerated UVC laboratory test to be developed.
toring surface mass loss was selected as the best technique for According to the literature search, selecting irradiance levels
monitoring surface UVC photodegradation. The most direct within an order of magnitude of those used in HVAC systems
technique for monitoring mass loss from a surface is measur- would further increase the probability of reciprocity between
ing the sample weight loss using an analytical balance [sensi- the photodegradation rates produced by the developed labora-
tive to 0.1 milligram (3.5e-06 ounce) changes]. The primary tory tests in Parts 2 and 3 of this project and the rates experi-
drawback of measuring weight loss is that small changes in the enced in HVAC systems.
water content of the sample (due to heating, room air humidity
REFERENCES
changes, etc.) will affect the sample weight, and consequently,
the accuracy of the UVC mass loss calculations. The second ASHRAE Handbook. 2008. 2008 HVAC Systems and Equip-
technique of monitoring surface mass loss is masking the ment. Chapter 16. Ultraviolet Lamp Systems. Atlanta:
surface of the sample so that the surface exposed to the UVC American Society of Heating, Refrigerating and Air-
suffers mass loss/height reduction compared to the original Conditioning Engineers, Inc.
surface protected from the UVC (crater formed in exposed Aslanzadeh, S. and M. Kish. 2005. Photodegradation of
area compared to protected area). Scanning electron micro- polypropylene thermal bonded non-woven fabric. Poly-
scope microphotographs, which are informative as to the mer Degradation and Stability 90 (3): 461 – 470.
morphology changes of the exposed surface, are limited by ASTM D1435. 2005. Standard Practice for Outdoor Weath-
expense and inability to measure crater depth. Optical micro- ering of Plastics. Philadelphia: American Society for
scopes using multiple focusing depths are able to monitor Testing and Materials.
crater depths but are laborious and time-consuming. Conse- ASTM D6290. 2005. Standard Test Method for Color Deter-
quently, surface profilometers were used to quantitate the mination of Plastic Pellets.
©2012 ASHRAE
645
Philadelphia: American Society for Testing and Materials. Kauffman, R. 2011. Study the degradation of typical HVAC
ASTM G151. 2006. Standard Practice for Exposing Nonme- materials, filters and components irradiated by UVC
tallic Materials in Accelerated Test Devices that Use energy (1509-RP). Final Report. American Society of
Laboratory Light Sources. Philadelphia: American Soci- Heating, Refrigerating and Air-Conditioning Engineers,
ety for Testing and Materials. Inc.
ASTM G154. 2006. Standard Practice for Operating Light Kim, C., Chung, C. and S. Moon. 1996. Removal of organic
Apparatus for UV Exposure of Nonmetallic Materials. impurities from the silicon surface by oxygen and UV
Philadelphia: American Society for Testing and Materi- cleaning. Korean J. of Chemical Engineering 13(3): 328
als. – 330.IL
Chiang, W. and C. Huang. 1999. Polymer Blends and Alloys. Kowalski, W. 2009. ULGI for Cooling Coil Disinfection, Air
1st Ed. Chapter 2. Compounding and compatibilization Treatment, and Hospital Infection Control. Report pre-
of high-performance polymer alloys and blends. G. pared for American Air & Water, Hilton Head, SC http:/
Shonaike and G. Simon, ed. New York: Marcel Dekker. /www.americanairandwater.com/UV-pdf/UVGI-
Chin, J., Nguyen, T., Byrd, E. and J. Martin. 2005. Valida- Report.pdf
tion of the reciprocity law for coating photodegradation. Kowalski, W.J. and W.P. Bahnfleth. 2002. UVGI design
JCT Research 2 (7): 499 - 508. basics for air and surface disinfection. HPAC 72
Denizligil, S. and W. Schnabel. 2003. Photooxidation of (1):100-110. http://www.engr.psu.edu/ae/wjk/uvh-
polyvinyl chloride at 254 nm in the absence and pres- pac.html
ence of phthalate plasticizers, Die Angewandte Makro- Kowalski, W.J., and Dunn, C.E. 2002. Current trends in
molekulare Chemie 229(1): 73 – 92. UVGI air and surface disinfection. INvironment Profes-
Feller, R. 1994. Accelerated aging: photochemical and ther- sional 8(6):4-6.
mal aspects, The J. Paul Getty Trust. http:// Kuzina, S. and A. Mikhailov. 2001. Photo-oxidation of poly-
www.getty.edu/conservation/publicatins/ mers 4. The dual mechanism of polystyrene photo-oxi-
pdf_publications/aging.pdf dation: a hydroperoxide and a photochain one.
Grossetête, T., et al. 2000. Photochemical degradation of European Polymer Journal 37 (11):2319-2325
poly(ethylene terephthalate)-modified copolymer, Poly- Levetin, E., et al. 2001. Effectiveness of germicidal UV radi-
mer 41(10):3541-3554. ation for fungal contamination within air-handling units.
Hamid, S. 2000. Handbook of Polymer Degradation, 2nd Ed. Applied and Environmental Microbiology 67 (8): 3712-
Chapter 15: Wavelength sensitivity of photodegradation 3715.
of polymers and Chapter 18: Fundamental and technical Nagai, N., Matsunoe, T., and T. lmai. 2005. Infrared analysis
aspects of the photooxidation of polymers. New York: of depth profiles in UV-photochemical degradation of
Marcel Dekker. polymers. Polymer Degradation and Stability 88(2):
Heraeus. 2009. http://www.hi-techlamps.com/pdfs/uv/ 224-233.
hLow%20Pressure%20UVC.pdf Qu, Z. et al. 2005. The photodissociation of ozone in the
ISO 4582. 2007. Plastics – Determination of changes in Hartley band: A theoretical analysis. J. Chem. Phys.,
color and variations in properties after exposure to day- Vol. 123 (7):074305.
light under glass, natural weathering or laboratory light Rabek, J. 1995. Polymer Photodegradation: Mechanisms
sources. Geneva: International Organization for Stan- and Experimental Methods. 1st Ed. London: Chapman
dardization. Hall.
ISO 4892-1. 1999. Plastics – Methods of exposure to labora- Rabek, J. 1996. Photodegradation of Polymers: Physical
tory light sources, Part 1: general guidance. Geneva: Characteristics and Applications. 1st Ed. Berlin:
International Organization for Standardization Springer-Verlag.
IUVA. 2005. IUVA Draft Guideline IUVA-G01A-2005. Ranby, B. and Rabek, J. 1975. Photodegradation, Photo-oxi-
Scottsdale: International Ultraviolet Association. dation and Photostabilization of Polymers: Principles
Jones, I. and R. Wayne. 1969. Photolysis of ozone by 254-, and Applications. 1st Ed. New York: John Wiley and
313- and 334-nm radiation. J. Chemical Physics, Sons.
51:3617 – 3618. Rivaton, A. et al. 2002. Comparison of the photochemical
Kaczmarek, H., Felczak, A. and D. Bajer. 2009. Photoxida- and thermal degradation of bisphenol-A polycarbonate
tive degradation of carboxylated poly(vinyl chloride). and trimethylcyclohexane–polycarbonate. Polymer Deg-
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Polymer Bulletin 62: 503 – 510. radation and Stability 75: 17–33.
Kaczmarek, H. et al. 2005. Effect of short wavelength UV- Schnabel, W. 1981. Polymer Degradation: Principles and
irradiation on ageing of polypropylene/cellulose compo- Practical Applications. Chapter 4: Photodegrada-
sitions. Polymer Degradation and Stability 88(2):189- tion,.Hanser International. Macmillan Publishing Co.,
198. Inc., New York.
646
ASHRAE Transactions
Scott, G. 1965 Atmospheric Oxidation and Antioxidants, 1st Soto-Oviedo, M. and M. De Paoli. 2002. Photo-oxidative
Ed. Chapter 7, Oxidative deterioration of saturated oil degradation of poly(epichlorohydrin-co-ethylene oxide)
and polymers. New York: Elsevier Publishing Company. elastomer at 254 nm. Polymer Degradation and Stability
76 (2): 219 – 225.
Shyichuk, A. et al. 2005. Comparison of UV-degradation Spartan Water Treatment. 2009. http://www.spartanwater-
depth-profiles in polyethylene, polypropylene and an treatment.com/index.html
ethylene-propylene copolymer. Polymer Degradation Witham, D. 2009. Telephone conversation with author and
and Stability, 88 (3): 415 – 419. sample submission April 14, 2009.
©2012 ASHRAE 647

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---

SA-12-027 (RP-1509)
Study of the Degradation of

Typical HVAC Materials, Filters and
Components Irradiated by UVC Energy—
Part II: Polymers
Robert E. Kauffman J. Douglas Wolf
ABSTRACT ers) exposed to an unknown dosage of UVC, analytical tech-

niques capable of determining mass loss were selected to
This paper summarizes the results of the research
monitor the photodegradation of samples during long-term
performed to evaluate the UVC photodegradation of different
UVC exposure tests.
polymers used in HVAC systems. The photodegradation of the
Therefore, the long-term UVC resistances of 14 pure
polymers was monitored using mass loss as discussed in Part
polymers were evaluated during Part II of this research project.
I of the project (Kauffman 2012). The polymers evaluated in
The UVC photodegradation evaluations were based on the rate
this project did not contain fillers so that the long-term rela-
of mass loss and were performed at irradiance levels of 1000,
tionship between rate of mass loss (rate of photodegradation)
4000, 7000 and 11,000 microwatts per square centimeter (µW/
and UVC irradiance could be studied in the range of 1000 to
cm2). A linear rate of mass loss versus irradiance plot indicates
11,000 microwatts per square centimeter without the influ-
that the photodegradation mechanism is independent of irra-
ence of inorganic particle accumulation. The UVC rates of
diance or exposure time. A linear plot also indicates that the
mass loss for each studied polymer were determined and then
law of reciprocity (Chin 2005) is obeyed, i.e., photodegrada-
plotted versus their respective irradiances. For the majority of
tion of the polymer is dependent only on the dosage of UV
the studied polymers, linear plots were produced demonstrat-
exposure (irradiance x time) and is independent of the time or
ing that the law of reciprocity was obeyed by UVC photodeg-
irradiance level taken separately. The photodegradation eval-
radation, i.e., the photodegradation mechanism was
uations and rates of mass loss versus irradiance plots are
independent of irradiance or exposure time and the polymer
presented herein for 14 of the numerous polymers used in
mass loss caused by UVC could be predicted based solely on
HVAC applications.
dosage.
EXPERIMENTAL
INTRODUCTION
Since the photodegradation mechanisms of UVC were Polymers
not understood, a research project was conducted to study the The polymers used in this study were obtained in the
effects of UVC on different materials, filters and components forms of rods or solutions (solvent evaporated to obtain poly-
used in HVAC systems. In Phase I of this project (Kauffman mer film) and contained no fillers, plasticizers, colorants or
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
2012), a thorough literature search was performed on the UVC other modifying compounds. The polymers obtained as rods
photodegradation of pure polymers (no fillers or plasticizers), were acrylonitrile butadiene styrene (ABS), acrylic, high
on photoreactors and on analytical techniques best-suited for density polyethylene (HDPE), low density polyethylene
monitoring UVC photodegradation. Based on the literature (LDPE), nylon, phenolic, polyacetal, polybutylene terephthal-
search and initial analyses of polymers (with and without fill- ate (PBT), polycarbonate, polyethylene terephthalate (PET),
Robert E. Kauffman is a distinguished research chemist and J. Douglas Wolf is a senior research physicist at the University of Dayton
Research Institute, Dayton, OH.
648
ASHRAE Transactions
perfluoroethylene and polypropylene. The epoxy and formvar circle in the center (minimize heating of sample surface by
polymers were obtained as solutions and their respective films UVC/increase heat transfer from sample). Approximately 15
were obtained by solvent evaporation/chemical reaction. samples with aluminum collars are shown placed in a circle on
the elevated platform in Figure 1.
Photoreactor
As previously described in Part I, the LZC-ICH2 photo- Stylus Profilometer
reactor (Luzchem Research, Inc. Ottawa, Ontario) equipped A stylus profilometer (KLA Tencor Instruments, Milipi-
with a rotating platform (Figure 1) was used to perform the tas, CA) was used to monitor the mass loss (crater depth) from
temperature controlled, irradiation monitored UVC experi- the exposed surfaces of the polished polymer samples. To
ments at 254 nanometers (nm). To maximize the surface UV negate any effects on the crater depth measurements from vari-
exposure for the higher irradiance experiments, the platform ations in the polished polymer surface, marks were made in the
was modified with support legs so that the samples could be unexposed polymer surface to ensure the 2-D surface profiles
rotated 7.6 centimeters (3 inches) below the 8 upper UVC were made across the same unexposed and exposed surfaces
lamps as shown in Figure 1. The maximum irradiance level during each weekly surface measurement. The surface varia-
measured on the platform was ~ 12,000 µW/cm2. To obtain the tions in the original polished surface were measured prior to
reduced (1000, 4000 and 7000 µW/cm2) irradiance levels, placing the sample on the platform for UVC exposure to estab-
screens with varying hole sizes were placed between the upper lish the baseline (± 1 micron) for the crater depth measure-
UVC lamps and the samples on the platform as shown in ments (important for samples that experience less than 2
Figure 1. Screens with 10, 40 and 50% open surface were used micron depth loss during entire UVC exposure).
to produce the 1000, 4000 and 7000 µW/cm2 irradiance levels,
respectively. Accelerated UVC Exposure Procedure
Sample Preparation Based on the literature search and preliminary test results
(Kauffman 2011, 2012), the accelerated UVC exposures of
For this project, circular samples [12.5 millimeter (mm) the listed polymers were performed as follows:
(0.5 inch) diameter)] were used so that the turntable could hold
over 100 samples allowing duplicate UVC exposures of the 1. LZC-ICH2 photoreactor with 8 lamps along top and
selected materials to be run simultaneously. The polymer appropriate screen between the rotating platform and
samples (no filler) were cut into disks with a thickness of 6 mm lamps (Figure 1) to obtain desired irradiance on platform
(0.25 inch) and polished to a smooth, flat finish (required for surface.
surface profilometer measurements). Aluminum metal collars 2. Polymers cut to fit into aluminum sample collar, polished
were then placed over the samples so that all of the polished smooth, surface marked for consistent positioning of
surface was masked except for a 3.2 mm (0.125 inch) diameter sample collar and to ensure repeatable starting point for
surface profiling. Profile initial polished surface to estab-
lish surface baseline for crater profiles.
3. Record UVC irradiance on platform surface prior to
samples being placed along circumference of platform.
4. Turn ON photoreactor with temperature set to 298°K
(25°C, 77°F). Platform rotates at 6 RPM.
5. Run UVC exposure test for 168 hours.
6. Record UVC irradiance on platform surface.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
7. Remove samples from platform one at a time for surface
profiling.
8. Check crater/surrounding surface for residue (regardless
of polymer, time or irradiance, no residues were ever
observed).
9. Place sample back into holder and repeat UVC exposure
(Step 3).
The determination to end the UVC exposure test for each

sample was based on the analytical results obtained for that
Figure 1 Photoreactor with screen between sample sample. In some cases, the exposure test was stopped due to a
turntable and UVC lamps to obtain reduced lack of detectable photodegradation. For instance, the UVC
irradiance levels. Samples with aluminum collars exposure test of aluminum foil tape was ended at 1000 hours
on turntable. (~ 6 weeks) because the surface profile measurements
©2012 ASHRAE
649
detected less than 0.5 micron surface loss versus exposure weekly intervals were calculated and then plotted versus expo-
time (Kauffman 2011). sure time for each irradiance level tested. As seen in Figure 3,
the average crater depth versus exposure time plots for PBT
Surface Profile Measurements are linear regardless of the irradiance level or length of expo-
To illustrate the UVC crater formation process in a poly- sure and are dependent only on the level of irradiance (indi-
mer surface, representative 2-dimensional surface profiles are cating law of reciprocity obeyed). The calculated slopes of the
shown in Figure 2 for a polished polybutylene terephthalate best-fit lines listed in Figure 3 represent the rates of crater
(PBT) polyester sample exposed to 11,000 µW/cm2 for 1171 formation for the 11,000 [0.013 micron/hour (µm/hr)], 7000
hours. The surface profiles determined that the UVC exposure (0.010 µm/hr), 4000 (0.0066 µm/hr) and 1000 (0.0046 µm/hr)
created a crater (no residue observed, mass loss occurred) that µW/cm2 irradiance tests. The coefficients of correlation (R2)
was ~ 3500 microns in width as dictated by the width of the are all above 0.99 indicating the excellent fit of the calculated
3200 microns (3.2 mm, 0.125 inch) hole in the center of the lines.
aluminum collar sample holder.
The increasing depths of the UVC created crater were Bunsen-Roscoe Law of Reciprocity
measured every 44 - 168 hours (1 week) and the depth of the To determine if the Bunsen-Roscoe law of reciprocity was
crater increased at a constant rate to a depth (with respect to the obeyed for each UVC photodegraded polymer, the calculated
positive profile of the unexposed surface) of approximately 19 rates of crater formation were then plotted versus irradiance.
microns (0.019 mm, 0.0007 inch) below the original polished If the crater formation rate versus irradiance level plot was
surface in Figure 2 after 1171 hours of exposure to 11,000 µW/ linear, the best-fit line has an excellent R2 (> 0.9) and the y-
cm2. To convert each measured surface profile into a single intercept was zero (no crater formation at 0 µW/cm2), the law
data point, the average crater depth was calculated over the of reciprocity was obeyed, i.e., the rate of formation was
middle 3 mm (0.12 inch) of each measured crater. The crater dependent on the dosage (irradiation level x time) and was
depth averaging also negated any effects of roughness, uneven independent of the irradiation level or exposure time sepa-
mass loss, etc. on the repeatability of determining the bottom rately. The slope of the crater formation rate versus irradiance
of the produced crater. level plot has the terms µm/(hr x µW) of UVC and could be
used by a HVAC designer or maintenance personnel to predict
Rate of Mass Loss Calculations
the crater depth/mass loss of a room temperature film/sheet
To determine the rates of crater formation for the different (no filler) at any exposure time based solely on the irradiance
polished polymers, the average crater depths measured at level.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 2 Stylus profile of the polybutylene terephthalate surface exposed to 11,000 µW/cm2 for up to 1171 hours. (Positive
profile of unexposed surface due to uneven polishing).
650
ASHRAE Transactions
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 3 Plots of average crater depths versus hours of UVC irradiance for polybutylene terephthalate (PBT) polished
samples exposed at 1000, 4000, 7000 and 11,000 µW/cm2. Calculated slopes listed in descending order of
irradiance.
REPRESENTATIVE PHOTODEGRADATION increasing with exposure time as in Figure 4. The observed

MECHANISMS OF POLYMERS crater roughness is in agreement with the literature (Rabek
1995 and 1996) and is attributed to crystallinity effects.
Introduction To determine the rates of crater formation in the exposed
Fourteen different types of polymers were polished flat HDPE surfaces for the 1000, 4000, 7000 and 11,000 µW/cm2
and exposed to four different levels (1000, 4000, 7000 and tests, the average depths along the bottom surfaces of the
11,000 µW/cm2) of UVC irradiation for up to 2700 hours. roughened craters (1000 – 4000 microns in Figure 4) were
Similar to the results for the PBT polyester illustrated in calculated and plotted on the same graph versus their respec-
Figures 2 and 3, the majority of the polymers produced smooth tive exposure time as shown in Figure 5.
craters at a constant rate regardless of exposure time or irra- The rates of crater formation (plots in Figure 5) in the
diance level. However, some of the polymers produced rough- HDPE surface are linear and the best-fit lines have excellent
ened craters at formation rates that changed with exposure R2 (> 0.98) regardless of the irradiance level, length of expo-
time/crater depth. The results for high density polyethylene sure or crater roughness and are totally dependent on level of
(HDPE), low density polyethylene (LDPE) and polyacetal irradiance, i.e., rate proportional to irradiance level, indepen-
polymers (no fillers) are presented as illustrations of the proce- dent of exposure time (indicating law of reciprocity obeyed).
dure/calculations used to determine the long-term crater The calculated slopes of the best-fit lines listed in Figure 6
formation rates regardless of the irradiance level, crater represent the rates of crater formation for the 11,000 (0.041
morphology or formation rate changes. µm/hr), 7,000 (0.031 µm/hr), 4,000 (0.017 µm/hr) and 1,000
(0.011 µm/hr) µW/cm2 irradiance tests. The calculated slopes
Roughened Crater Formation, Constant Rate in Figure 5 are approximately 3 times higher those in Figure 3
of Degradation indicating the mass loss of the HDPE is 3 times higher than
that of PBT, i.e., PBT is 3 times more resistant to UVC than
The numerous surface profiles obtained for a polished
HDPE.
disk of HDPE exposed to 11,000 µW/cm2 for up to 1171 hours
are shown in Figure 4. As opposed to the PBT craters in
Roughened Crater Formation, Rate of Degradation
Figure 2, the bottom of the crater produced by the UVC expo-
Increases with Exposure Time
sure of the HDPE surface became increasingly roughened
after 331 hours of exposure making the profiles of the The numerous surface profiles obtained for a polished
produced crater increasingly difficult to discern (roughness disk of polyacetal exposed at 4000 µW/cm2 for up to 1512
larger than depth increases). Similar crater bottom roughening hours are shown in Figure 6. Similar to the HDPE craters in
was also observed for the HDPE samples exposed to 1000, Figure 4, the crater produced by the UVC in the polyacetal
4000 and 7000 µW/cm2 (Kauffman 2011) with the roughness surface became roughened with increased exposure time.
©2012 ASHRAE
651
Figure 4 Surface profiles of high density polyethylene (HDPE) polished samples exposed at 11,000 µW/cm2 for up to 1171
hours.
Figure 5 Plots of average crater depths versus hours of UVC irradiance for high density polyethylene polished samples
exposed at 1000, 4000, 7000 and 11,000 µW/cm2 for up to 1700 hours. Calculated slopes listed in descending order
of irradiance.
652
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
ASHRAE Transactions
Figure 6 Surface profiles of polyacetal polished samples exposed at 4000 µW/cm2 for up to 1512 hours.
However, the high rate of crater formation in Figure 6 is Smooth Crater Formation, Rate of Degradation
greater than the increase in roughness allowing the crater Decreases with Exposure Time
profile to remain discernable even at longer exposure times.
The numerous surface profiles obtained for a polished
Regardless of the irradiance used in the polyacetal test, the rate
disk of low density polyethylene (LDPE) exposed at 4000 µW/
of crater formation was slow for the first 500 hours then
cm2 for up to 1680 hours are shown in Figure 8. Similar to the
became rapid for the rest of the exposure time [photodegrada-
PBT craters in Figure 2, the craters produced by the UVC in
tion increases UVC absorption/degradation rate (Chiang and
the LDPE samples are smooth for the entire test period regard-
Huang 1999)].
less of irradiance level. However for Figure 8 and all of the
To quantitate and compare the rates of photodegradation other irradiance levels (Kauffman 2011), the rate of crater
at the different irradiance levels, the average depths along the formation slows considerably with extended exposure. The
bottom surfaces of the craters (1000 – 4000 microns in crater depth at which the rate changes was inversely propor-
Figure 6) were calculated (negate roughness of the crater tional to the irradiance level, e.g., rate change occurred at a
bottoms) and plotted on the same graph versus their respective depth of 8 microns for 11,000 µW/cm2 but not until 15 microns
exposure time as shown in Figure 7. for the 1000 µW/cm2 test (Kauffman 2011). The change in rate
As seen in Figure 7, the rate of crater formation (rate of is attributed to the depletion of reactive species (oxidation
photodegradation) plots all have two linear sections (each products such as hydroperoxides produced on surface during
section has an excellent fit with R2 > 0.96) with the break- extrusion process) as the exposure time/crater depth increases
points occurring after 300 – 500 hours of UVC exposure. [HDPE contains reactive species throughout bulk (linear plots
Regardless of the irradiance level, the rates (slopes) calculated in Figure 5) due to its higher processing temperature (Rabek
for the tests before 500 hours (top slope for each irradiance 1995)].
level in Figure 7) are approximately 10% of the rates (slopes) To quantitate and compare the rates of photodegradation
for the remainder of each test. The 7000 µW/cm2 (R in at the different irradiance levels, the average depths along the
Figure 7) was repeated with a new sample to confirm the rates bottom surfaces of the craters (1000 – 4000 microns in
for the 11,000 and 7000 µW/cm2 are similar, i.e., law of reci- Figure 8) were calculated and plotted versus exposure time on
procity not followed. Regardless of the initial rates of degra- the same graph. As seen in Figure 9, the rate of crater forma-
dation, the final calculated slopes listed in Figure 7 [11,000 tion (rate of photodegradation) plots all have two linear
(0.10 µm/hr), 7000 (0.11, 0.095 µm/hr), 4000 (0.069 µm/hr) sections (R2 > 0.97): rapid rate of crater formation for the first
and 1000 (0.051 µm/hr) µW/cm2 irradiance tests] are the rates 8 -15 microns then a much slower rate for the remainder of the
of crater formation to be used when predicting the long term test. For each irradiance level, the initial rates (slopes) are
photodegradation rate of polyacetal. similar to those of HDPE in Figure 5 (reactive species present)
©2012 ASHRAE
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
653
Figure 7 Plots of average crater depths versus hours of UVC irradiance for polyacetal polished samples exposed at 1000,
4000, 7000 and 11,000 µW/cm2 for up to 1700 hours. Calculated slopes listed in descending order of irradiance
(initial slope listed first). Last set of slopes marked R for the repeat 7000 µW/cm2 test.
Figure 8 Surface profiles of low density polyethylene polished samples exposed at 4000 µW/cm2 for up to 1680
hours.(Positive profile of unexposed surface due to uneven polishing).
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
654
ASHRAE Transactions
Figure 9 Plots of average crater depths versus hours of UVC irradiance for low density polyethylene polished samples
exposed at 1000, 4000, 7000 and 11,000 µW/cm2 for up to 2200 hours. Calculated slopes listed in descending order
of irradiance (initial slope listed first).
while the final slopes are slower than those calculated for both Law of Reciprocity for Representative
the PBT and HDPE tests (reactive species depleted) in Figures Photodegradation Mechanisms
3 and 5, respectively. Regardless of the initial rates of degra-
To determine if the PBT, HDPE, polyacetal and LDPE
dation, the final calculated slopes listed in Figure 9 [11,000
photodegradation rates obey the law of reciprocity, the four
(0.0045 µm/hr), 7000 (0.0040 µm/hr), 4000 (0.0017 µm/hr) rates of crater formation (slopes in equations from Figures 3,
and 1000 (0.0016 µm/hr) µW/cm2 irradiance tests] are the 5, 7 and 9) determined for each polymer were plotted versus
rates of crater formation to be used when predicting the long their respective irradiance levels as shown in Figure 11. The
term photodegradation rate of LDPE. final slopes for polyacetal and LDPE were used for Figure 11
since extended photodegradation would be expected to occur
Photographs of Craters Formed by Representative at the final rate determined in the UV studies not the initial
Photodegradation Mechanisms rate. The initial rate was also not considered for Figure 11
since it would be affected by the degree of polishing and outer
To support the surface profile results, the changes in the layer contaminants produced by extrusion parameters
polished surfaces resulting from exposure to the highest irra- (LDPE), handling, etc. not present in the bulk of the material.
diance level (11,000 µW/cm2 for 1700 – 2700 hours) were The dimension of mm per 1000 hours in Figure 11 is equal to
photographed at high magnification. The photographs in micron per hour. The dimension of most interest to the HVAC
Figure 10 show that the produced craters are circular and designer or maintainer would depend on the UVC exposed
follow the outline of the hole [3.2 mm (0.125 inch) diameter] component: mm for drip pan versus micron for wire insula-
in the middle of the aluminum sample collars. The photo- tion. Since the irradiance levels produced by the UVC lamps
graphs show that rough craters were created in the HDPE and slowly decreased with time (less than 3% over 2500 hour test),
polyacetal surfaces while smooth craters were created in the the average irradiance level was calculated (Kauffman 2011)
LDPE and PBT surfaces. Consequently, the photographs are and used in the plots in Figure 11 (the error bars on the HDPE
in agreement with the crater roughness measured by the data points represent the variance in the irradiance level during
surface profilometer (photographs have no depth perception). each test).
In fact, the smooth crater in the LDPE surface (depth of 11 As can be seen in Figure 11, the rate of crater formation
microns) is barely discernable from the unexposed surface versus irradiance plots for PBT, HDPE and LDPE all obey the
(only difference is that the polishing scratches that are still law of reciprocity (R2 > 0.93 and plots linear) in the irradiance
visible on the unexposed surface were removed by UV expo- range of 1000 to 11,000. The fact that the PBT, HDPE and
sure in the formed crater). LDPE plots all have a y-intercept slightly greater than 0 (crater
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
©2012 ASHRAE
655
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 10 High magnification photographs of polished polybutylene terephthalate (PBT), high density polyethylene (HDPE),
polyacetal and low density polyethylene (LDPE) samples exposed to 11,000 µW/cm2 for up to 2700 hours. Circular
exposed areas: sample holder hole.
formation at 0 µW/cm2) either indicates that the relationship rial UVC (254 nm) resistance at different irradiance levels
between the rate of crater formation and irradiance changes (obey law of reciprocity). For materials whose photodegrada-
(photodegradation mechanism changes) below 1000 µW/cm2 tion does not obey the law of reciprocity, the data will at least
(law of reciprocity not obeyed) or the best-fit line needs to be allow the materials to be rated/ranked with respect to UVC
adjusted. Consequently, additional research below 1000 µW/ resistance [similar to material rankings based on UVA and
cm2 is needed for the polymers in Figure 11 to allow designers UVB test results (Rabek 1995 and 1996, Hamid 2000)]. Based
and maintenance personnel to correctly predict the reduced on the UVC tests at the four different irradiance levels, the 14
rates of mass loss for the polymer surfaces (no filler) exposed polymers polished and tested during this program were cate-
to levels below 1000 µW/cm2 in HVAC systems. gorized into the following UVC resistance ratings (based on
Although the plot of the polyacetal crater formation rate plastic chemical ratings):
versus irradiance is linear in Figure 11 with a R2 = 0.86, the
rates for 7000 and 11,000 µW/cm2 are very similar and the plot • A – No effect (aluminum tape, all organic materials
indicates there would be significant crater formation at 0 µW/ experience mass loss)
cm2 (y-intercept = 0.0358). If a polynomial was used to fit the • B – Minor Effect (mainly cosmetic changes)
data (dashed line), a curve was produced that had a y-intercept • C – Moderate Effect (some cracking/pitting, LDPE in
 0 but the correlation coefficient decreased. Additional Figure 10)
research above 11,000 µW/cm2 is needed to determine if the
• D – Severe Effect (structural damage, not recom-
rate of crater formation levels off above 7000 µW/cm2 as
mended, HDPE in Figure 10)
suggested by Figure 11 or more data points are required to
more accurately fit the rate of crater formation versus irradi-
The wide range of UVC resistances of the different poly-
ance plot for polyacetal.
mers shown in Figure 11 represents the two worst ratings, C
and D. Even though polyacetal and HDPE degrade 10 and 4
POLYMER UVC RESISTANCE EVALUATIONS
times faster than PBT, respectively, all three polymers are
The primary goal of this project is to produce data that categorized as D. So even within a particular rating, one poly-
will enable HVAC designers and maintainers to predict mate- mer will be better suited for UVC exposure than another
656
ASHRAE Transactions
Figure 11 Plots of crater formation rate versus irradiance level for polished samples of PBT, HDPE, polyacetal and LDPE.
For y-axis:mm/1000 hours same as microns/hour.
(could designate polyacetal, HDPE and PBT as D-, D and D+, for 1200 hours is barely perceivable (unexposed area has
respectively). LDPE in Figure 11 is representative of a C rating polishing marks) in Figure 10. For perspective on the photo-
for UVC resistance degrading 3 times slower than PBT. degradation of polymers with a C rating, a slope of 30 x 10-08
The polymers evaluated during this project are listed in microns/(hour x irradiance) (polyimide in Table 1) would
Table 1 by rating and in descending order of resistance to predict losses of 0.30 and 3.6 microns (0.0003 and 0.0036 mm,
photodegradation. The slopes of the best fit lines and y-inter- 0.012 and 0.14 mil) after 1000 hours of exposure to 1000 and
cepts are included in Table 1 for quantitative comparison. 12,000 µW/cm2, respectively. Consequently, electrical wiring
The rate of crater formation versus irradiance level slopes with an unprotected polyimide insulation of 25 microns (2.5
listed in Table 1 for all of the polymers given a D rating with mm, 1 mil) would lose 50% of its protection leading to short-
regard to UVC resistance had best fit lines with R2 > 0.95 but ing/arcing in less than 5 years (8760 hours/year) of exposure
y-intercept = 0.001 – 0.0358 (linear plot of polyacetal). For to 1000 µW/cm2.
perspective on the photodegradation rates of polymers with a The rate of crater formation versus irradiance level slopes
D rating, a slope of 200 x 10-08 microns/(hours x irradiance) listed in Table 1 for all of the polymers given a B rating with
(polypropylene in Table 1) would predict surface losses of 2 regard to UV resistance (Kauffman 2011) had best fit lines
and 24 microns (0.002 and 0.024 mm, 0.1 and 1 mil) after 1000 with R2 down to 0.5 due to the much smaller craters (only
hours of exposure to 1000 and 12,000 µW/cm2, respectively. nylon had a best fit line with R2 > 0.9). However, the y-inter-
The severe surface damage caused by 11,000 µW/cm2 on PBT cepts were lower for the B rated polymers than for the C and
(2500 hours), polyacetal (1200 hours) and HDPE (1200 hours) D polymers. Formvar had apparent breakpoints similar to
are shown in Figure 10. those in Figure 9 [perfluoroethylene may have had a break-
The rate of crater formation versus irradiance level slopes point at 1500 hours for 11,000 µW/cm2 but longer testing is
listed in Table 1 for all of the polymers given a C rating with needed to confirm (Kauffman 2011)]. For perspective with
regard to UV resistance (Kauffman 2011) had best fit lines regard to the slopes, a slope of 10 x 10-08 (natural nylon) would
with R2 > 0.9 and y-intercept = 0.001 – 0.002. The plots of predict losses of 0.24 and 3.6 microns (0.00024 and 0.0036
average crater depths versus hours of UVC irradiance for the mm, 0.01 and 0.14 mil) after 3000 hours of exposure to 800
acrylic rod and cast epoxy had breakpoints similar to LDPE in and 12,000 µW/cm2, respectively.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 9 (rate of photodegradation slowed with increasing As a final comparison of the polymers tested during this
exposure time) and so the final rates are listed in Table 1. The project, the photodegradation rate versus irradiance level plots
moderate LDPE surface damage caused by 11,000 µW/cm2 of representative polymers with B, C and D ratings were
©2012 ASHRAE
657
Table 1. Tested Polymers Categorized into Ratings B, C and D
Rate of Crater Formation Y-Intercept

Polymer
[x10-8 micron/(irradiance x hour)] (micron/hour at 0 µW/cm2)
Rating B
Aluminum Tape* < 0.1 0
Formvar 4 < 0.0001
Natural Nylon 10 0.0007
Perfluoroethylene 20 < 0.0001
Rating C
Polyimide 30 0.0010
LDPE 30 0.0011
ABS 30 0.0014
Cast Epoxy 40 0.0016
Polycarbonate 60 0.0019
Phenolic Resin ** **
Rating D
Acrylic 80 0.0022
PET 90 0.0049
PBT 90 0.0031
Polypropylene 200 0.001
HDPE 400 0.0033
Polyacetal 700 0.0358

* Rating A
** Phenolic resin rating based only on 11,000 µW/cm2 results
combined in Figure 12. The plots in Figure 12 show that there mass loss was minimal with the depth of the created crater
is a much larger difference between the D and C ratings than being measured in microns using a surface profilometer. Due
between the C and B ratings. The photographs in Figure 10 to the small surface changes measured each week, the initial
also indicate that the damage experienced by D rated polymers polymer surfaces had to be polished to remove any roughness
(PBT, HDPE and polyacetal) is much greater than that expe- and the UVC exposures had to be performed for several
rienced by C rated polymers (LDPE). High magnification months to obtain reliable rates of mass loss that were not
scanning electron microscope microphotographs (Kauffman affected by species concentrated on the material’s surface,
2011) were needed to characterize/compare the shallow e.g., oxidized species resulting from extrusion process.
craters produced in B rated polymers. The polymers tested in Part II of this research did not
contain any fillers or plasticizers so that the rates of degrada-
SUMMARY
tion would only be dependent on the composition of the poly-
This project has demonstrated that all polymers are mer. Consequently, the information presented herein would be
degraded by UVC radiation of 254 nm. In agreement with the useful to HVAC component manufacturers in selecting the
literature search (Kauffman 2011 and 2012), the photodegra- base polymer best suited for UVC exposure or for developing
dation by the UVC caused primarily chain scission at the poly- component designs to protect UVC susceptible polymers from
mer’s surface. The main effect of UVC exposure was crater UVC exposure. Although not reported in this paper, the crater
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
formation/mass loss as opposed to the cross-linking/bulk formation rates of nylon containing carbon black and different
degradation reported for materials exposed to UVA and UVB. type poly (vinyl chloride) materials containing different plas-
Since the photodegradation was localized at the surface, the ticizers and fillers exposed to 1000 – 11,000 µW/cm2 UVC
658
ASHRAE Transactions
Figure 12 Comparison of crater formation rate versus irradiance level plots for representative polymers with B, C and D
ratings.
have been determined (Kauffman 2011) and will be reported Hamid, S. 2000. Handbook of Polymer Degradation, 2nd Ed.
with the other filler/additive containing materials in Part III of Chapter 15: Wavelength sensitivity of photodegradation
this research. of polymers and Chapter 18: Fundamental and technical
aspects of the photooxidation of polymers. New York:
Although all of the tested polymers obeyed the law of
Marcel Dekker.
reciprocity for the tested irradiance range of 1000 – 11,000
Kauffman, R. 2011. Study the degradation of typical HVAC
µW/cm2, the fact that the linear rate of crater formation versus
materials, filters and components irradiated by UVC
irradiance plots had y-intercepts > 0 (crater formation will
energy (1509-RP). Final Report. American Society of
occur at 0 µW/cm2) indicates that the photodegradation mech-
Heating, Refrigerating and Air-Conditioning Engineers,
anism changes below 1000 µW/cm2 (law of reciprocity not
Inc.
obeyed). Consequently, additional research is needed to deter-
Kauffman, R. 2012. Study the degradation of typical HVAC
mine the reduced rates of mass loss between 0 and 1000 µW/
materials, filters and components irradiated by UVC
cm2 so that polymer lifetimes can be predicted correctly for the
energy. Part I. Literature search (1509-RP). ASHRAE
10 – 500 µW/cm2 levels used in many locations of HVAC
Transactions (In Press)
systems (Kowalski 2009).
Kowalski, W. 2009. ULGI for Cooling Coil Disinfection, Air
Treatment, and Hospital Infection Control. Report pre-
REFERENCES
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
pared for American Air & Water, Hilton Head, SC http:/

/www.americanairandwater.com/UV-pdf/UVGI-
Chiang, W. and C. Huang. 1999. Polymer Blends and Alloys.
Report.pdf
1st Ed. Chapter 2. Compounding and compatibilization
Rabek, J. 1995. Polymer Photodegradation: Mechanisms
of high-performance polymer alloys and blends. G.
and Experimental Methods. 1st Ed. London: Chapman
Shonaike and G. Simon, ed. New York: Marcel Dekker.
Hall.
Chin, J., Nguyen, T., Byrd, E. and J. Martin. 2005. Valida- Rabek, J. 1996. Photodegradation of Polymers: Physical
tion of the reciprocity law for coating photodegradation. Characteristics and Applications. 1st Ed. Berlin:
JCT Research 2 (7): 499 - 508. Springer-Verlag.
©2012 ASHRAE
659
SA-12-028
Design of Roofs for Increased Solar Potential

BIPV/T Systems and their Applications
to Housing Units
Caroline Hachem Andreas Athienitis, PhD, PE Paul Fazio, PhD, PEng

Student Member ASHRAE Member ASHRAE Member ASHRAE
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
This paper presents a design methodology of housing The design of net zero energy solar buildings involves a
twofold approach of enhancing energy efficiency while opti-
units’ roofs for increased solar potential, and the application
mizing active solar energy production using photovoltaics and
of some of these designs to energy efficient housing units. The thermal collectors. A net zero energy house (NZEH) generates
basic principle is the utilization of complete near-south facing as much energy as its overall consumption, over a typical year
roof surfaces for solar collectors that generate both heat and (Torcellini, 2006).
electricity. The design of roof surface morphology incorpo- Reduction of energy consumption can be achieved
rates the application of building-integrated photovoltaic/ther- through several measures, such as airtight, well insulated
mal (BIPV/T) system that generates simultaneously electrical building envelope, implementation of HVAC efficiency
and thermal energy. The methodology employs a correlation measures, including the use of heat pumps, combined with
between thermal and electrical output of the BIPV/T system, geothermal energy or solar collectors, and finally the use of
and analysis of the effect of combinations of tilt and orientation energy efficient appliances. Window properties and size, espe-
cially on the equatorial façade, can maximize passive heating.
angles of the BIPV/T system. Optimal configurations are
Purchased heating energy can be significantly reduced by
implemented in the roof of a rectangular housing unit. Gable
means of solar heat gains. A well designed passive-solar build-
roof with 45° tilt angle is used as reference. The basic rectan- ing may provide 45% to 100% of daily heating requirements
gular shape is redesigned to follow the contour of optimal roof (ASHRAE, 2007).
geometries. Energy performance, in terms of heating and cool- Near-equatorial facing roof surfaces are considered opti-
ing demands, is analyzed to evaluate the effect of roof shape. mal for capture of solar energy for electricity and heat gener-
Some roof shape designs enable increasing the potential of ation, and therefore for the integration of photovoltaic/thermal
BIPV/T systems by up to 30%, as compared to the reference systems. Since the study presented in this paper is conducted
roof. The effect of roof design on heating and cooling loads is for the northern hemisphere, the term “south facing” is
not highly significant (<5%). While redesigning unit shapes employed hereunder to refer to “equatorial facing” roofs,
facades and units.
does not significantly affect their energy performance, it can
The basic principle applied in this study is the utilization
provide some design flexibility. In addition to optimizing
of complete near-south facing roof surfaces as extended
energy production, the integration of photovoltaic/thermal
surface solar collectors that generate both heat and electricity.
systems in surfaces with different orientations enables a spread The technological approach applied is to design the roof outer
of the timing of peak electricity generation over up to 3 hours. layer as a building-integrated photovoltaic/thermal (BIPV/T)
This feature can be beneficial in reducing mismatch between system - a technology that combines PV modules and heat
demand and supply of the grid. extraction devices to produce simultaneously power and heat
Caroline Hachem is a PhD student, Andreas Anthienitis is a professor, and Paul Fazio is a professor in the Department of Building, Civil,
and Environmental Engineering at Concordia University in Montreal, Quebec.
660
ASHRAE Transactions
(Tripanagnostopoulos, 2001). Heat extraction from the PV residential construction as compared to add - on PV/T systems
rear surface is usually achieved using circulation of a fluid (air which are usually attached to the outer layer of the construc-
or water) with low inlet temperature. The extraction of thermal tion, requiring thus additional mounting systems.
energy serves two main functions, first it is used for space The current study is part of an ongoing research that aims
heating and solar hot water applications, and secondly it helps at investigating the effect of design parameters of housing
in cooling the PV modules, thus increasing the total energy units of different shapes, isolated and in neighborhood
output of the system (Charron, 2006). patterns, on their solar potential and energy performance
The performance of a PV system depends mainly on the (Hachem et al., 2011a, 2011b). The study presented in this
tilt angle and azimuth of the collectors, local climatic condi- paper focuses on the design of roof shapes of two-storey hous-
tions, the collector efficiency, and the operating temperature of ing units to increase the thermal and electrical potential of
the cells. During the winter months, the insolation can be BIPV/T systems that cover complete near-south facing
maximized by using a surface tilt angle that exceeds the lati- surfaces. The intent is to explore some deviations from the
tude angle by 10-15º. In summer an inclination of 10–15º less basic hip or gable roof design used commonly for rectangular
than the site latitude maximizes the insolation (Duffie and floor plans, while allowing both architectural flexibility and
Beckman, 1991). The PV system is usually mounted at an energy production advantages. The effect of different roof
angle equal to the latitude of the location, to reach a balance designs on energy performance and the balance between
between winter and summer production (Kemp, 2006). The energy generation and energy use is studied as well.
orientation of the PV panels affects both the amount of elec- Section 2 presents the methodology of investigating
tricity generation and the time of peak generation. PV system energy production. The roof design approach presented in
orientation can be selected to better match the grid peak load section 3 includes three main parts: Housing units’ character-
(Holbert, 2009). This can affect the annual value of the istics, roof design, and redesign of the geometry of the housing
produced electricity, especially in locations where electricity units to follow the contour of the roof. Section 4 presents the
value changes with time of use. analysis of the results; it includes the effect of different roof
Pelland and Poissant (2006) report that BIPV systems shapes on electricity and heat generation, the effect of roof
have the potential to supply about 46% of Canada’s residential design, as well as the effect of house shape redesign, on overall
electricity. An open loop air-based BIPV/T system is particu- energy performance. The concluding section highlights the
larly suited to cold sunny climates such as much of Canada. main results and their significance.
This air-based BIPV/T concept utilizes circulating outdoor air
behind the BIPV panel with the aid of a variable speed fan. The METHODOLOGY
concept of the BIPV/T system is illustrated in Figure 1.The The main focus of this study is the investigation of the
circulated air assists in cooling the panel and recovering heat potential of BIPV/T electrical and heat energy production and
that can be used for space or water heating. Covering a overall energy performance of housing units, associated with
complete roof surface with a BIPV/T system has the advantage different roof designs. The ultimate goal is to maximize the
of forming an outer layer which acts as the weather barrier in combined potential of annual electricity generation and heat
addition to producing heat and electricity. production to achieve net zero energy or energy surplus
BIPV/T, with specially developed PV products, has the balance. With regard to heat generation potential, it should be
potential to meet all the building envelope requirements (such borne in mind that the annual thermal energy produced by the
as mechanical resistance, thermal insulation, etc….) in addi- BIPV/T system does not account for a “usefulness” factor, and
tion to producing heat and electricity. This multiple function- therefore not all the heat produced by the system can be
ality of BIPV/T system may improve the cost effectiveness of considered as useful heat.
Heat generation is computed in this study over an
assumed heating period, from mid-October to mid-April.
However, heat generation can be employed for water heating
and for various appliances, resulting in an increase of the over-
all potential use of the output of the BIPV/T system year
round.
Roof designs that range from the simplest to increasingly
complex multi-surface shapes are first developed for rectan-
gular shape housing unit, and then the units are redesigned to
fit the shape of these roofs. The sequence of the methodology
is therefore: unit shape design, roof design, unit shape re-
design.
A rectangular unit with gable roof is used as reference for
Figure 1 Cross-section illustrating an open loop BIPV/T comparative evaluation of BIPV/T potential of roof shapes
system. and energy performance.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
©2012 ASHRAE
661
Modelling and Simulations properties, the slab program produces average monthly
temperature of the slab, which is input in EnergyPlus to carry
EnergyPlus building simulation software (EnergyPlus,
out the simulations for the computation of energy demand for
2010) is employed in the simulations using annual weather
heating and cooling.
data for Montreal, Canada (latitude 45°N). The heating degree
days (HDD) for Montreal are ca.4519 HDD (the Weather
Network). BIPV/T Modelling
SketchUp/OpenStudio (EnergyPlus, 2010) is employed Electricity Generation. The Equivalent One-Diode
to generate geometric data for EnergyPlus. Each housing unit Model (or “TRNSYS PV” model, Eckstein, 1990) employed
is modelled as a single conditioned zone. The Conduction in EnergyPlus is selected to perform electricity generation
Finite Difference algorithm is selected as the heat balance simulations of the BIPV/T systems. The TRNSYS model
algorithm. This solution technique employs a one-dimension employs a four-parameter empirical model to predict the elec-
finite difference method to represent the construction trical performance of PV modules (see Duffie and Beckman,
elements. A short time step of 10 minutes is used in the simu- 2006).
lations. The current-voltage characteristics of the diode depend
on the PV cell’s temperature. The model automatically calcu-
Weather Data lates parameter values from input data, including short-circuit
The weather files of the building simulation program current, open-circuit voltage, current at maximum power, etc.
EnergyPlus are used for the simulations (EnergyPlus, Weather (Griffith and Ellis, 2004). For this study, the PV array is
Data Sources). The weather data file, which is based on selected from EnergyPlus database to provide approximately
CWEC – Canadian Weather for Energy Calculations, provides 12.5% efficiency, under standard conditions (Athienitis et al.,
hourly weather observations. These observations represent an 2011). The electrical conversion efficiency decreases by some
artificial one-year period, specifically intended for building 0.45% for each °C increase of cell temperature from the
energy calculations (EnergyPlus, Weather Data Sources). The temperature under standard conditions. For Montreal, the
data collected for this year includes hourly values for solar annual potential of PV electricity generation of south facing
radiation, ambient temperature, wet bulb temperature, wind surfaces at latitude tilt angle is about 1200 kWh per kWpeak
speed, wind direction and cloud cover. of installed PV (NRCan, 2007).
Two design days, a sunny cold winter day (in January) - Thermal Generation. A correlation between thermal
WDD, and a sunny hot summer design day (in June) - SDD, and electrical output is developed for several representative
are used to represent two sunny days with extreme tempera- clear days of the year, as well as for several roof slopes. A
ture, to study the incident solar radiation on roofs. Addition- constant air flow of 2m/s (6.56ft/s) in the BIPV/T system is
ally, a whole year weather data set is used to estimate the selected to ensure high efficiency. This air velocity represents
annual electricity generation and winter heat production of the a commonly used value in BIPV/T applications (Athienitis et
BIPV/T system installed on near south-facing roof surfaces, as al., 2011). The correlation can be used to determine the ther-
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
well as the energy use for heating and cooling of the housing mal energy as well as the outlet air temperature of BIPV/T
units. systems of different complex roof designs used in this study.
Since the present paper explores BIPV/T potential, such an
Solar Radiation and Shading Calculations approach is appropriate. In practice, for design purpose a more
Hourly direct solar radiation is computed using the Ener- detailed model would be advisable.
gyPlus program. The computation is based on the ASHRAE The correlation is obtained using a transient quasi-two-
model of clear sky (ASHRAE, 2003) applied to Montreal dimensional finite difference model. The BIPV/T system is
(45ºN). This model is the default model used by EnergyPlus to divided into several control volumes represented each by a set
estimate the hourly clear-day solar radiation for any month of of simultaneous equations. Figure 2 depicts the thermal
the year. The solar radiation accounts for the direct beam and network of one control volume of the BIPV/T system. The
diffuse radiation, as well as for radiation reflected from the various thermal conductances including that associated with
ground and adjacent surfaces. The shading algorithm handles the air flow (MCair) are presented in Figure 2. The PV panels
self-shading geometries such as L shape. This algorithm is are assumed to have negligible resistance and thermal capacity
based on coordinate transformation methods (Groth and (Liao et al., 2007). The bottom surface of the air cavity is
Lokmanhekim, 1969) and the shadow overlap method assumed to be well insulated.
(Walton, 1983). The governing equation used to describe the explicit finite
difference method for a thermal network, corresponding to a
Slab on Grade Modeling node i and time interval p, is expressed by Athienitis (1998):
The slab program (EnergyPlus, 2010), is used to compute
q i +  j  U  i,j  T  j,p  
the temperature of the underside surface of the slab (in contact T  i,p + 1  = ----------------------------------------------
- (1)
with the ground). Taking into account the slab and ground  j  U  i,j  
662
ASHRAE Transactions
The outlet air temperature at each control volume is deter-
mined as:
Qu  i 
T outlet  i  = T in  i  + ------------------- (6)
M  C air
where Tin (i) is the inlet air temperature at each control

volume, M is the mass flow rate of air, Cair is the specific heat
of air, and Qu (i) is the heat carried by the air flow and is deter-
mined using the finite explicit model at each control volume
as:
Q u  i  = A s  [h c   T pv  i  – T air  i   + h c   T b  i  – T air  i  ] (7)
where As is the area of each BIPV/T section, hc is the convec-

tive heat transfer coefficient in the cavity, Tpv is the tempera-
ture of the PV panel, Tb is the temperature at the bottom side
of the cavity and (Tair) is the air temperature, computed in each
Figure 2 Schematic illustrating the thermal network in one control volume using equation 8 (Charron et al 2006).
control volume of the BIPV/T system (Node i)
(refer to eq. 1 for i and j, and to eqs. 1-8 for T pv  i  + T b  i   T pv  i  + T b  i  x  2  a
T air  i  = ---------------------------------
- +  T in  i  – ---------------------------------
-  e (8)
meanings of symbols). 2 2
where x is the length of the studied section, and

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
where T is the temperature of a node at a specific time, qi is the M  C air

a = ------------------- (9)
heat source at the node in question, j describes the adjacent w  hc
nodes, and R(i, j) is the resistance between nodes i and j.
where W is the cavity width.Based on Equations 2 and 4, the
The total electrical energy is determined as function of the thermal efficiency can be determined as:
efficiency and the solar radiation as follows:
Qu
Q e =  pv  A  G (2)  thermal = ------   pv (10)
Qe
where (  pv ) is the PV efficiency (Eq.3), A is the surface area The ratio (Qu/Qe) determines the correlation between the
of the roof and G is the solar radiation. The electrical effi- heat and electricity generation of the BIPV/T systems. Qu/Qe
ciency of the BIPV (  pv ) system is computed using the for selected sunny days, of each month of the year for a 45o tilt
following linear equation (Skoplaki and Palyvos, 2009; angle is presented in Figure 3a. Figure 3b presents the results
Whitaker et al., 1991): of the simulations of the BIPV/T systems with various tilt
angles, for the (WDD). The ratio of solar thermal production
 pv =  STC  1 –   T pv – T STC   (3) to the electricity production (Qu/Qe) varies between 3 and 3.5
(mean value of 3.1 and standard deviation of 0.2). Therefore,
where  STC is the efficiency of the PV cells under standard on average, a value of Qu=3Qe is adopted in this study for an
test conditions (STC),  is the PV module temperature coef- air speed of 2m/s in the BIPV/T system.
ficient, and TSTC is the standard test condition temperature
(25ºC). A value of 12.5% is assumed for  STC . DESIGN APPROACH
The total thermal energy generated by the BIPV/T (Qu)
can be expressed as follows: Rectangular Units
The basic shape for housing units is a rectangle with
Q u =  thermal  A  G (4) constant floor area of ca 60m2 (645.6ft2). A two-storey hous-
ing option is adopted in this study since it represents one of the
Qu is determined as function of the inlet air temperature most common types of single family homes in Canada (Athi-
entering the BIPV/T system, assumed as the outside air enitis, 2007). The characteristics of the housing units are
temperature (To), and the outlet air temperature of the system detailed in Table 1. The basic design of the units relies on
(Toutlet-BIPV/T) (outlet air of the last control volume), (Athien- passive solar design principles (Chiras, 2002) and rules of
itis et al., 2011): thumb (CMHC, 1998). The aspect ratio of the rectangle ratio
of the south facing façade to the perpendicular façade is 1.3.
Q u = M  C air  T Outlet – BIPV  T – T o  (5) This ratio is within the optimal range for passive solar design
©2012 ASHRAE
663
Figure 3 a) Qu/Qe for 45o tilt angle roof, for one sunny day of each month, over a year, b) Qu/Qe for WDD of roofs with
different tilt angles.
Table 1. Main Characteristics and Electric Loads of Housing Units
Exterior wall: 6 RSI (R-34)

Roof: 10 RSI (R-57)
Thermal resistance values
Slab on grade: 1.2 RSI (R-7)
Slab perimeter: 6 RSI (R-34)
Thermal mass 20cm (7.8 in) concrete slab
Window type Triple glazed, low-e, argon filled (SHGC=0.57), 1.08 RSI (R-6.1)
Area of south glazing as percentage of south facing facades 35%
Shading Strategy Interior blinds
Shading control Blinds shut at indoor air temperature of 22 °C
Occupants 2 adults and 2 children, occupied from 17:00 - 8:00
Setpoint temperatures Heating set point 21ºC (70°F), cooling set point 25ºC (77°F)
Infiltration rate 0.8ACH @50Pa (1.0lb/ft2)
Ventilation rate 0.35ACH
Assumptions for electrical loads
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Lighting 3kWh/m2 /yr (9.63E+02 Btu/ft2/yr) (Sartori el al., 2010)
DHW 2.75kWh/day/person (9.48E+03 Btu/day/person) (Sartori et al., 2010)
Major appliances 1600kWh/yr (5.52E+06 Btu/yr) (Pohgarian et al., 2008)

Minor appliances 1100kWh/yr (3.79E+06Btu/yr) (Charon, 2007)
664
ASHRAE Transactions
in northern climate (Athienitis and Santamouris, 2002). A as compared to the system with 45° tilt angle, while the maxi-
geothermal heat pump with a seasonal coefficient of perfor- mum monthly increase of generation is about 6% (occurring in
mance (COP) of 4 is assumed to supplement the passive and December). In general, high slope roofs (>40°) are favored in
active solar heating. Ground source heat pumps (GSHPs) can climates such as those in most of Canada due to snow accu-
supply heat of up to quadruple the energy of the electricity they mulation; however, this factor is not considered directly in this
consume, by using ground extracted heat (NRCan, 2008). An study.
intelligent shading control is considered. The interior blinds Orientation Angle. The orientation angle of a surface,
are assumed to be shut when the indoor air temperature relative to south, can be defined as the angle between south and
exceeds 22°C, throughout the cooling season. A ventilation the projection on a horizontal plane of the normal to this
rate of 0.35 air change per hour (ACH) is assumed. This value surface. Orientation affects the solar potential in two ways: the
conforms with ASHRAE standard 62.2 requirements for air amount of generation and the time of peak generation. Annu-
change rate associated with a determined house size and occu- ally, the highest energy yield is associated with a south facing
pancy (ANSI/ASHRAE Standard 62.2; 2010). system. Deviation of the orientation of the system from the
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
south by up to 40° west or east leads to an approximate reduc-

Lighting and Appliance Loads tion of up to 5% of the annual generation of electricity
Electrical loads for major and minor appliances, for light- (Figure 4). The heat generation for the assumed heating period
ing and for domestic hot water (DHW) are assumed based on is reduced by up to 9%, for the orientation angle of 40° west
a variety of sources dealing with the electrical load in energy or east. A rotation of the system by 60°, west or east of south,
efficient and net zero energy houses (NZEH) (e.g. Armstrong results in a reduction of some 12% of the total annual electric-
et al., 2009, Sartori et al., 2010, Pohgarian et al., 2008). Major ity generation and of 20% of the heat generation during the
appliances include refrigeration equipment (freezer and heating period. The effect of the orientation on electricity
refrigerator), dishwasher, washing machine, clothes dryer and generation, for the design days and the total year generation
cooking appliances. Minor appliances include a wide range of associated with a 45° tilt angle BIPV/T system is presented in
equipments used in the kitchen and for entertainment Figure 4. The effect is measured as the ratio of the electricity
purposes. These loads are summarized in Table 1. Lighting generated to the generation of a south facing BIPV/T system
consumption can be limited to 3kWh/m2/yr (9.63E+02 Btu/ft2/ (orientation = 0).
yr) for a NZEH in mid-latitude locations, based on the assump- Monthly generation shows that in the summer months, the
tion that a NZEH is expected to optimize daylight utilization orientation of the BIPV/T system towards the west or east
(Sartori et al., 2010). results in electricity generation that is close or slightly higher
Hot water energy consumption can be limited to a daily than south orientation. The effect of orientation on monthly
average of 2.75kWh (9.48E+03 Btu/day) per occupant electricity generation is presented in Figure 5.
(Sartori et al., 2010), based on the assumption of hot water The orientation of the BIPV/T system affects not only the
usage of 50L/day/person (13.2Gal/day/person). This value is value of the electricity generation, but also the time of peak
derived from information provided in the literature (e.g. IEA generation. For a south facing system, the peak generation is
task 26, EN 15316 (66.6 L/person (17.59 Gal/person)) and the at noon. Rotation of the BIPV/T system towards the west
Canadian EQuilibrium Initiative (56.25L/person (14.86Gal/ results in shifting the peak radiation to the afternoon and vice
person))), with the assumption that it is possible to reduce versa for east rotation. A 30° orientation (east or west), enables
significantly the daily hot water (DHW) consumption, using a shift of peak generation time by to up to one and half hours
different methods (e.g., installing low flow fixtures, such as relative to solar noon. A 60° orientation enables a three hours
low-flow showerheads, faucets, etc.). shift of peak. Roofs that combine both east and west orienta-
tions can lead to a spread of peak generation time, reaching six
Roofs Design: Design Considerations hours.
This section summarizes the main considerations for the In some cases, return on annual energy produced may be
design of roofs for increased capture of solar radiation a more important object than the total energy produced, partic-
(Hachem et al., 2011c). ularly in locations where prices of electricity vary with time of
Tilt Angle. Tilt angle of a surface can be defined as the day. For a net-zero energy house or community the annual net
angle between the normal to this surface and a vertical axis. income becomes an important variable to optimize as the value
The annual electricity generation of the BIPV/T system is not of the electricity produced may be higher than that consumed.
significantly affected by a tilt angle that ranges between 30° This involves consideration of orienting the BIPV/T systems
and 50°. For a 60° tilt angle, the annual electricity production to match peak generation with high electricity demands,
is reduced by some 7% as compared to the 45° tilt angle. This enabling thus larger annual income from selling excess elec-
is mainly due to the reduction in production in the months tricity to the grid.
ranging from April to September, as compared to roofs with Combination of Tilt and Orientation Angles. The
lower tilt angle. For instance the electricity generation of the yearly study shows that for the winter months between
BIPV/T system with 60° tilt angle is reduced by 16% in June, September and March, the most effective BIPV/T systems are
©2012 ASHRAE
665
Figure 4 Effect of the angle of orientation on the electricity generation.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 5 Effect of the angle of orientation on the monthly electricity generation of the BIPV/T systems.
those that combine a tilt angle of 45° to 60° with a south facing nologies are not addressed. The photovoltaic system is
orientation. These configurations allow the highest yield of assumed to cover the total area of all south and near-south
electricity generation as well as heating energy. In the summer facing roof surfaces. In practice, a small percentage of the roof
months, lower tilt angles and rotation, particularly west, are area is used for the mounting structure, for framing and for
advantageous. other technical considerations. It is further assumed that, given
the dynamic nature of current PV technology development,
Roof Design Applications any future technology will be capable of accommodating any
specific requirements raised by proposed roof designs, such as
This part of the study, explores various design possibili- PV modules of varying shape and size as well as inverters for
ties of roof shapes for a rectangular house, to maximize solar different BIPV/T orientations.
energy potential. The design proposed in this study is concep- Three basic geometries of roofs are studied. The first
tual, aimed at exploring the potential of different roof design geometry is a commonly used hip roof with a 45o tilt angle and
to increase the potential of BIPV/T systems for electricity and varying side angles (Figure 6). The second and third types of
heat generation. Technical considerations relating to PV tech- roofs are designed relatively independently of the shape of the
666
ASHRAE Transactions
house, employing a multi-faceted roof surface combining a Two main configurations are considered, as well as some
range of tilt and orientation angles. variations of these options. The mid plate is south oriented
Hip Roof. The hip roof is designed with the ridge running while the side plates are rotated by equal angles, the east plate
east-west (E-W) at the center of the plan area (Fig. 6). The towards the east and the west plate towards the west. In the first
design parameter is the side angle (the angle the hip surface option, the orientation angle of the side plates is 15°, while in
forms with the horizontal plane (Figure 6)). This angle affects the second option this angle is 30°. The orientation angle is, as
mainly the surface area. Three values of the side angles are defined above, the angle between south and the projection on
studied: 45°, 60°, and 90°. The tilt angle is kept constant at 45o. a horizontal plane of the normal to the surface (e.g. the surface
BIPV/T system is assumed to cover the total south facing denoted ABCD in Figure7b).
surface of the roof. For each of the two configurations, a few variations of tilt
angles are used to explore the effect on the overall solar poten-
Split-Roof. The south-facing portion of the roof is tial of the roof. The two configurations of the plate roofs are
divided into three plates of differing orientations and tilt presented in Figure 7 and details of the combinations are
angles. A BIPV/T system is assumed to cover the total area of included in Table 2.
each of these plates.
Although the study assumes availability of suitable tech-
nology to cover the full surface, investigation of employing
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
PV strips of currently available systems, results in a maximum

loss of 4% of the total area.
Folded Plate. Folded plate roof design refers in this paper
to the shape of the roof, not necessarily to the structural
system. The folded plate roof design is composed of triangular
plates with various orientations. Two basic shapes are
designed. The first configuration is composed of four plates,
with the two side plates facing south (see Figure 8a). The
second basic shape consists of three plates with the central
plate facing south. Figure 8c shows a variation of this shape
composed of two basic units.
Variations of the basic shapes are composed of two, and
possibly more, units of the basic shape. Figure 8 presents the
three configurations analyzed in this study. The first configu-
ration (Figure 8a) is the 4-plate basic shape, with the central
Figure 6 Hip roof design. plates rotated 15o east and west and the side plates having 45o
Figure 7 Split- roof designs: (a) configuration 1, side plates with 15° orientation from south; (b) configuration 2, side plates
with 30° orientation from south.
©2012 ASHRAE
667
Table 2. Design Consideration Split- Roofs and Folded Plates’ Roofs
Complex Side Number Combinations of orientation and tilt of the plates*

Roof angle of surfaces *The first number refers to the orientation and the second number refers to the tilt angle;
Center plate 0°, 45°; 0°, 50°; 0°, 45°;

Configuration 1
Side plates 15°(E,W), 40° 15° (E,W), 40° 15° (E,W), 30°
Split Roofs 90° 3
Center plate 0°, 45; 0°, 50°; 0°, 45°;
Configuration 2
Side plates 30°(E,W), 40° 30° (E,W), 40° 30° (E,W), 30°
4 - (Basic folded Center plate 0°, 45°;

Configuration 1
plates) Side plates 15° (E, W)
Folded 7(2 basic folded Center plate 0°, 45°;

90° Configuration 2
plates plates) Side plates 30° (E, W)
6 (two 3-plate Center plate 0°, 45°;

Configuration 2
basic shapes) Side plates 30° (E, W)
Figure 8 Folded plate roof designs, (a) configuration 1-- basic 4 folded plates with 15° orientation;, (b) Configuration 2 -- two
basic 2folded plates units with 30° orientation, (c) Configuration 3 – two basic 3 folded plates roof with 30 °
orientation.
tilt angle. The second configuration (Figure 8b) is composed associated with the split-roof option, and two shape designs
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
of two basic 4-plate units, with central plates rotated 30° east correspond to the folded roof options.
and west. The third configuration (Figure 8c) is composed of For the split-roof option, the same re-design principle is
two three-plate basic shapes with side-plates rotated 30o east applied in roof configurations 1 and 2 (Table 2). This principle
and west. Details of the folded plates’ configurations are consists of splitting the south façade into surfaces that are
presented in Table 2. oriented according to the associated roof portions. Figure 9
shows the redesign of the rectangular shape associated with
Unit Shapes: Redesign the split- roof option of configuration 1.
In this section the rectangular shape of the housing units For the folded plate roof option, a similar concept as for
is re-designed to fit the outline of the roof. This mainly the split-roof is applied in configuration 1 (basic 4 folded
involves redesigning the south facing facades (south and north plates, Table 2). The south façade is split in the middle and
in the folded plates option, see below). The floor area is main- then the east part is oriented east (15°) while the west side is
tained fixed. Four new shapes are obtained, two designs are oriented 15° west (Figure 10a). For the second and third
668
ASHRAE Transactions
Figure 9 Split- roof option with: (a) rectangular shape, (b) redesigned south facing facade.
Figure 10 Folded plates roof option with: (a) rectangular shape, (b) redesigned south facing facade.
configurations (Figure 8b and c, Table 2), a saw tooth concept Simulations using EnergyPlus were performed to deter-
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
is applied to the south facade, in order to conform to the roof mine the energy generation potential of all roof shapes
concept (Figure 10b). Due to the symmetry of the folded plate presented above. The thermal generation of the BIPV/T
roof design, the north façade undergoes the same redesign system, for a constant air flow of 2m/s (6.56ft/s), is assumed
procedure as the south façade. to constitute 3 times the amount of electricity generation for a
Configurations that involve surfaces of varying orienta- given period (based on an approximate model, see above).
tions (such as split-roof and folded plates) may result in large Results are presented for the electricity generation annually
overhangs. Modification of the south façade to follow the roof and for the winter and summer design days. Heat generation
outline can be applied to avoid such overhangs. Concepts like results are over the assumed heating period (mid-October to
saw tooth, however, may be more applicable to projects with mid-April) and for the winter design day. The combined
larger size, such as large residential development or commer- annual electricity and winter heating energy production are
cial buildings, which can support increasing architectural also presented. The main observations drawn from the analy-
complexity. sis of all roof designs are summarized below.
RESULTS • The side angle in the hip roof design, affects mainly the
south facing roof area. The largest south roof area is
Roof Morphology Effect obtained with the larger side angle (90°) corresponding
BIPV/T Potential. This is assessed by comparing the to a gable roof (Figure 6). The electricity and heat gen-
electricity and heat generation of each roof design to those of eration of a hip roof with a 45° side angle is reduced by
the gable roof. approximately 40% as compared to the gable roof. The
©2012 ASHRAE
669
comparison between different roofs and the gable roof in Table 4.  T represents the air temperature rise inside
are presented in Table 3. the BIPV/T system.
• The split-roof design has two main characteristics: it
• Folded plate roof design enables obtaining various orien-
enables larger south facing roof area (about 48m2
tations for the same rectangular plan roof. Furthermore,
(516.66ft2) as compared to 40m2 (430.55ft2) for the
this roof shape has significantly higher south facing sur-
gable roof) and it facilitates combinations of orientation
and tilt angles, which allows obtaining spread of peak face area than the gable roof (see Table 3). The different
generation time of up to 3 hours, (Figure11). The results configurations analyzed in this paper do not show a sig-
(Table 4) indicate that there is no significant change of nificant difference in the potential of their BIPV/T sys-
the energy potential between the different combinations, tems (maximum difference of 4% is observed). The roof
of configuration 1 and 2 (3% or less). A significant options with 15° orientation allows a spread of time of
increase in the annual energy production is however peak generation of approximately 2 hours while the 30°
obtained using this roof design, as compared to the enables 3 hours difference. The six plate folded roof
gable roof. For instance, configuration 1 (with 15° ori- (Figure 8c) electricity generation exceeds the generation
entation) exceeds annual electricity generation of the of the gable roof by approximately 30% (Table 4). The
gable roof by 17% and heat generation for the assumed BIPV/T potential of these roof options are presented in
heating period by 15%. The annual electricity genera- Table 4.
tion, the heat generation for the heating period and the
combined energy potential, are presented in Table 4 Effect of Roof on Energy Performance of Units
(Table 4a is in metric units and 4b in imperial units).
The potential heat generation on the WDD and the aver- The effect of different designs of hip roof (changing the
age air temperature difference (  T) are also presented angle of side angle) on the heating and cooling loads is not
Table 3. Comparison of All Roof Design Options to the Gable Roof
Split-Roofs
Hip Roofs Folded Plates
(Orientation,
(Side Angle) (Orientation)
Tilt of Side Plates)
15°(E, 30°(E,W), Conf. 2 Conf. 3

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
45° 90° 60° Conf. 1 (15)

W),40° 40° (30) (30)
South facing (m2) 26 32 40 48 48 50 53 53

roof Area (ft2) 279.2 343.7 429.6 515.5 515.5 537 569.2 569.2
Ratio of annual electricity gener-
0.65 0.81 1 1.17 1.15 1.25 1.27 1.29
ation to Gable roof
Figure 11 Electricity generation on design days for the west and east oriented plates in the 30°(E,W), 40° roof option, (a)
SDD, (b) WDD.
670
ASHRAE Transactions
significant (5% or less). For instance, the heating load required and less). Results of these comparisons are presented in
for the housing unit with gable roof (reference case) is approx- Table 5.
imately 5% larger than for the unit with hip roof of 45° side
The increase in heating demand for rectangular housing
angle.
units is correlated with the increase of the roof area. The larger
The heating load of split-roof design is increased by 6%
and 7% for configuration 1 and 2, respectively, relative to a hip the roof the higher the heating load. However this increase in
roof with 45° side angle. The comparison to the reference case heating load is more than counterbalanced by the increased
(rectangular unit with a gable roof), shows that the increase in potential of such roofs, to generate both electricity and heat.
heating load is not significant (2%). Cooling load is decreased The decrease in cooling load can be explained by the shade
by 6-7% as compared to the hip roof (45° side angle) and by cast on the south façade, due to the overhang resulting from the
approximately 10% as compared to the gable roof. Table 5 use of split-roof and folded plates roof options (e.g. see
summarizes the results of the comparison of heating and cool- Figure 9a).
ing loads to rectangular units with a hip roof and with gable
roof. Effect of Redesign on Energy Performance
Heating load of rectangular units with folded plates roof
is increased by 9% for configuration 1 and 10% for configu- The rectangle shape with split and folded plate roofs is
rations 2 and 3, as compared to the rectangular unit with hip redesigned to conform to the shape of the roofs. Simulations
roof (Table 2). Comparing these configurations to the refer- are conduced to verify the effects of the redesigned shape on
ence unit, the increase in heating load is less significant (5% heating and cooling loads.
Table 4a. Results of Energy Potential of the Multi-Faceted Roof Design Options in International System of Units (SI)
Heat Average Air

Annual Electricity Electricity Heat
Generation Combined Change
Electricity Generation Generation Generation
Roof Options (Heating Generation Temperature
Generation WDD SDD (WDD)
Period) (kWh/yr) T
(kWh/yr) (kWh/day) (kWh/day) (kWh/day)
(kWh/yr) (WDD)(°C)
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Gable roof 7561 10068 17629 29.6 18.8 88.9 10
15°(E,W),40° 8815 11544 20358 33.6 22.3 101 9.5

Split-roofs
30°(E,W), 40° 8636 11169 19806 32 22.4 100 9
Conf. 1 (15) 9460 12708 22168 37.6 23 113 10.2
Folded Plates Conf. 2 (30) 9636 12974 22717 38 23.8 114.3 9.7
Conf. 3 (30) 9743 12851 22486 38.2 24.3 115 9.7
Table 4b. Results of Energy Potential of the Multi-Faceted Roof Design Options in Empirical Units (I-P)
Heat Average Air

Annual Electricity Electricity
Generation Combined Heat Change
Electricity Generation Generation
Roof Options (Heating Generation Generation Temperature
Generation WDD SDD
Period) (Btu/yr) (WDD) (Btu) T
(Btu/yr) (Btu/day) (Btu/day)
(Btu/yr) (WDD)(°F)
Gable roof 2.61E+07 3.47E+07 6.08E+07 1.02E+05 6.48E+04 3.07E+05 50
Split-roofs 15°(E,W),40° 3.04E+07 3.98E+07 7.02E+07 1.16E+05 7.69E+04 3.48E+05 49.1
30°(E,W),
2.98E+07 3.85E+07 6.83E+07 1.10E+05 7.72E+04 3.45E+05 48.2
40°
Folded Plates Conf. 1 (15) 3.26E+07 4.38E+07 7.64E+07 1.30E+05 7.93E+04 3.90E+05 50.36
Conf. 2 (30) 3.32E+07 4.47E+07 7.83E+07 1.31E+05 8.21E+04 3.94E+05 49.46
Conf. 3 (30) 3.36E+07 4.43E+07 7.75E+07 1.32E+05 8.38E+04 3.97E+05 49.46
©2012 ASHRAE
671
Table 5. Comparison of the Effect of Roof Shapes on Heating and Cooling,
for the Rectangular Shape and the Redesigned Shapes
Comparison to the 45° Hip Roof Comparison to Gable Roof (Reference Case)
Rectangular Unit Redesigned Units Rectangular Unit Redesigned Units
Heating Cooling Heating Cooling Heating Cooling Heating Cooling
Split-roof
Conf.1 1.06 0.94 1.06 0.97 1.02 0.89 1.01 0.92
Conf.2 1.07 0.93 1.02 1.17 1.02 0.89 0.97 1.11
Folded Plates
Conf.1 1.09 1.01 1.04 1.03 1.04 0.96 0.99 0.98
Conf.2, 3 1.10 0.90 1.09 1.45 1.05 0.86 1.04 1.38
The results indicate that heating load for the redesigned rectangle with gable roof produces some 2% more than it
units decreases as compared to the rectangular shapes with the consumes. By comparison, electricity production of housing
same roof configurations. This decrease of heating load is not unit with hip roof of 45° side angle is some 35% less than
significant, however (a maximum of 5%). Comparing the consumption.
redesigned units to the reference case (rectangle with gable Housing units with split-roof options, generate as much as
roof) shows that heating load is not significantly affected. A 17% more electricity than they consume, while the surplus of
maximum increase of 5% is observed for the folded plate electricity generation in units with folded plates roofs reaches
option (configurations 2 and 3). some 29%. The results in terms of percentage of energy
Cooling load is increased for the redesigned units that production to energy consumption of all rectangular shapes as
involve orientation of the south facing façade by 30° east and well as redesigned shapes, with different roof configurations
west. This is due to the large window areas on these facades are presented in Table 6. The energy generation to energy
(35% of the facade), which was originally intended for the true consumption ratio of redesigned units is identical to that of the
south facing facade (before redesigning the units). Results of corresponding rectangular units (Table 6). This allows the
the comparison of the redesigned units associated with each architect/engineer some flexibility of design, knowing that
roof configuration, to the rectangular unit with 45° hip and this design is not affecting significantly the energy perfor-
with gable roof (reference), are presented in Table 5. The mance.
comparison is presented in term of the ratio of the heating and
cooling loads of each configuration to this of the rectangular CONCLUSION
unit with 45° hip, and to the reference case. This paper presents a design methodology to increase the
potential of BIPV/T systems for roofs while improving the
Evaluation of Energy Balance energy performance of housing units. The methodology
consists of first considering a rectangular shape, with passive
In this section energy demand and supply are compared design, then designing solar-optimized roofs followed by
for the different roof configurations studied. Heat generation redesign of the shape, based on performance. The paper also
by air circulation is not considered in this analysis, and there- presents a comparative evaluation of net energy supply/
fore energy supply/demand balance is expected to be further demand of all design alternatives. Major conclusions from the
enhanced once the heat generation for space and water heating study are as follows:
is taken into consideration. Since the heat generation is
directly proportional to electricity generation, increase of • Solar optimization of roof design requires optimal
electricity generation has the twofold effect of increasing choice of orientation and tilt angles of roof surfaces.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
supply and reducing demand of energy for space heating and Both orientation and tilt angle of the BIPV/T system
hot water. affect its overall energy generation. The annual electric-
The total consumption of electricity for lighting, DHW ity generation of the BIPV/T system is not significantly
and appliances, in addition to the computed heating and cool- affected by a tilt angle that ranges between 30° and 50°
ing energy consumptions, for rectangular shape units with for the latitude considered (45° N). Deviation of the sur-
different roof configurations is presented in Figure 12, along- face azimuth angle of the BIPV/T system from due
side the energy production of the corresponding units. The south by up to 40° west or east leads to an approximate
672
ASHRAE Transactions
Figure 12 Energy demand and production for all rectangular units, with different roof configurations.
Table 6. Ratio of Energy Production to Consumption
Hip Roof Split-Roof Folded Plates

Gable Roof
Side Angle Config.1 Config.1 Config.2 Config.1 Config. 2
= 45°
Annual energy (kWh) 4848 7677 8814 8636 9460 9743

generation (Btu) 1.57E+07 2.61E+07 3.04E+07 2.98E+07 3.26E+07 3.36E+07
(kWh) 7496 7527 7530 7535 7551 7548
Annual energy use
Rectangular (Btu) 2.58E+07 2.60E+07 2.60E+07 2.60E+07 2.60E+07 2.60E+07
shape
Ratio of energy production
0.65 1.02 1.17 1.15 1.25 1.29
to energy use
(kWh) 7527 7517 7522 75742
Redesigned Annual energy use
(Btu) 2.60E+07 2.59E+07 2.59E+07 2.61E+07
units
Ratio of energy production to energy use 1.17 1.15 1.26 1.29
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
©2012 ASHRAE
673
reduction of 5% of the annual generation of electricity more than its energy use. Manipulation of roof design
while a rotation of 60°, west or east of south, results in a can help in improving production/consumption ratio.
reduction of about 12% in the total annual electricity Multi-faceted roofs, in addition to increasing production
generation. The orientation of a roof affects as well the associated with increased surface area, produce several
time of peak generation. This can be of particular advan- peaks of generated electricity, due to the different orien-
tage in cases where the value of electricity varies with tations of surfaces.
the time of day.
• Multi-faceted roofs, such as folded plate and split-roof This simulation study is conducted under a northern mid-
configurations, can significantly increase electricity pro- latitude climatic zone, similar to the climate of Montreal,
duction and heat generation, primarily through Canada (45°N). However, the methodology implemented in
increased effective surface area. Dividing the reference the study can be applied to different locations and climates,
gable shaped roof surface into three plates with varying with few modifications. The main design modification of the
tilt/orientation angles can increase electricity generation roof would be the range of the tilt angles of various roof
by up to 17%. Replacing the gable roof with a folded configurations. A tilt angle that approximates the latitude of a
plate surface increases electricity generation potential specific location is optimal for BIPV/T systems, as demon-
by up to 30%. Varying surface orientations in such roof strated in this study. Other considerations will include reduc-
designs enables spread of peak electricity generation ing snow accumulation, which favors higher roof slopes.
over up to 3 hours, potentially reducing the impact on Complex roof systems, such as folded plate, are associ-
the electric grid. ated with increased cost due to larger BIPV/T system and
• The effect of roof design on heating and cooling demand increase of complexity in the manufacturing. However, the
is not significant because there is a ventilated attic over cost rise can be counterbalanced by an increased potential of
the insulated ceiling. For rectangular units, increase of BIPV/T system which results in two potential benefits: 1)
heating load reaches a maximum of 5%, as compared to increasing the revenue of the system by selling the excess elec-
the rectangle with gable roof. Heating load increases tricity to the grid, 2) shifting the timing of the PV electricity
slightly with increased roof area; however this load is generation which can increase the system value by up to 20%
more than counterbalanced by the increase in electricity (Borenstein, 2008). On the other hand roof systems such as
and heat generation. folded plates can be designed as structural roofs, saving there-
• Redesigning the shape of units to match roof geometry fore on the cost of structural elements required in the tradi-
results in a minor reduction of heating load as compared tional roof systems, and providing extra functional space
to the rectangular units with the same roof design. This (when spatial structural action is utilized). A cost/benefit
approach allows for flexibility of design, whereby the assessment of these roof systems can be the subject of future
designer can opt for a rectangular shape with a sophisti- investigations.
cated roof design or to modified rectangular-based The significance of this study lies in highlighting the role
shapes, where the facades follows the contour of the of roof morphology in increasing the overall solar energy
roofs. generated, as compared to a regular hip/gable roof, as well as
enabling spread of peak electricity generation time. The paper
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
• In attempting to achieve a balance between energy
demand and energy production it should be noted that demonstrates an approach to design net zero or surplus energy
heating and cooling demand constitutes no more than houses, while maintaining maximum design flexibility that
10%-15% of total energy demand, when an energy effi- can accommodate functional and other considerations not
cient heat pump is used. The rest of energy consumption related to energy efficiency. This approach consists of design-
is attributed to appliances, water heating and other items ing first the approximate shape of the housing unit (rectangu-
that are not affected by parameters considered in this lar for example), designing the roof for the desired size of
study. The main objective, therefore, is to maximize BIPV/T systems and then redesigning the unit shape, without
electricity production, even at the expense of some significantly affecting the energy performance. Finally, it
increase in heating and cooling load. should be stressed that the methodology presented in this
paper needs to be applied at the conceptual design stage.
• The general comparison between energy consumption,
assuming energy efficient measures, and the energy pro-
REFERENCES
duction, shows that the use of multi-faceted roofs allows
reaching energy surplus status. Unit shapes with split- A440.2-09/A440.3-09. Fenestration energy performance/
roof design produce up to 17% more energy than their User guide to CSA A440.2-09, Fenestration energy per-
total energy consumption. With a folded plates roof formance.
option the surplus of energy generation can reach up to ANSI/ASHRAE Standard 62.2-2010, 2010. Ventilation and
30%. Rectangular shape with hip roof (45°side angle) Acceptable Indoor Air Quality in Low-Rise Residential
produces some 65% of the energy use. The rectangle Buildings. Atlanta, GA: American Society of Heating,
with a gable roof, on the other hand, produces about 2% Refrigerating and Air-Conditioning Engineers.
674
ASHRAE Transactions
Armstrong, M., Swintona, M.C., Ribberink, H., Beausoleil- Hachem, C., Athienitis, A., Fazio, P., 2011a. Parametric
Morrison, I., Millette, J., 2009. Synthetically Derived investigation of geometric form effects on solar poten-
Profiles for Representing Occupant-Driven Electric tial of housing units, Solar Energy, 85 (9), 1864-1877.
Loads in Canadian Housing, Journal of Building Perfor- Hachem, C., Athienitis, A., Fazio, P., 2011b. Investigation of
mance Simulation. Solar Potential of Housing Units in Different Neighbor-
ASHRAE. 2003. 2003 ASHRAE Handbook—HVAC Appli- hood Designs, Energy and Buildings, 43 (9), 2262-
cations. Atlanta, GA: American Society of Heating, 2273.
Refrigerating and Air-Conditioning Engineers, Inc. Hachem, C., Athienitis, A., Fazio, P., 2011c. Design of Roof
ASHRAE. 2007. Chapter 33 Solar Energy Use. ASHRAE Morphology for Increased Solar Potential of BIPV/T
Handbook HVAC Applications. Atlanta, GA, American Systems, ISES Solar World Congress, 28 Aug.- 2 Sep.,
Society of Heating, Refrigerating and Air-Conditioning Kassel, Germany.
Engineers, Inc. Holbert, K.E., 2007. An Analysis of utility incentives for res-
Athienitis, A.K., 1998. Building Thermal Analysis, second idential photovoltaic installations in Phoenix, Arizona,
ed. Mathcad Electronic Book, Mathsoft Inc., Boston. 2007 39th North American Power Symposium (NAPS)
Athienitis, A. K., and Santamouris, M., 2002. Thermal anal- Washington State University. Downloaded on January
ysis and design of passive solar buildings, 10, 2010 at 13:37 from IEEE Xplore.
James & James Science Publishers. Kemp, W.H., 2006. The Renewable Energy Handbook: A
Athienitis, A., 2007. Design of a solar home with BIPV-ther- Guide to Rural Energy Independence, Off-Grid and Sus-
mal system and ground source heat pump, 2nd Canadian tainable Living. Tamworth, ON: Aztech Press, 2006.
Solar Buildings Conference, June 10 – 14. Calgary, Liao, L., Athienitis, A. K., Candanedo, L. and Park, K.W.,
Canada. 2007. Numerical and experimental study of heat transfer
Athienitis, A.K., Bambara, J., O’Neil, B., Faille, J., 2011. A in a BIPV–thermal system, ASME Journal of Solar
prototype photovoltaic/thermal system integrates with Engineering 129 (4) pp. 423–430.
transpired collector, Solar Energy, 85:139-153. MathCAD 14, 2007. Parametric Technology Corporation
Borenstein, S., 2008. The Market Value and Cost of Solar (PTC), 140 kendrick Street, MA 02494 USA.
Photovoltaic Electricity Production. Working paper 176 Natural Resources Canada (NRCan), 2007. Photovoltaic
(Berkeley, CA: University of California at Berkeley, potential and solar resources maps of Canada. Retrieved
Center for the Study of Energy Markets). February 1, 2011, from https://glfc.cfsnet.nfis.org/
Charron, R and Athienitis, A., 2006. Design and Optimiza- mapserver/pv/rank.php?NEK=e
tion of Net Zero Energy Solar Homes. ASHRAE Trans- Natural Resources Canada (NRCan), 2008. Heating and
actions. 112 Pg. 285. Cooling with a Heat Pump. Office of Energy Efficiency
(OEE), http://oee.nrcan.gc.ca/publications/infosource/
Charron, R., 2007. Development of a Genetic Algorithm
pub/home/heating-heat-pump/booklet.pdf, visited on 17
Optimisation Tool for the Early Stage Design of Low
April, 2011.
and Net-Zero Energy Solar Homes, Unpublished Ph.D.
Thesis, Concordia University. Pelland S. and Poissant Y., 2006. An evaluation of the poten-
tial of building integrated photovoltaics in Canada, 31st
Chiras, D. 2002. The Solar House: Passive Heating and
Annual Conference of the Solar Energy Society of Can-
Cooling. White River Junction, VT: Chelsea Green Pub-
ada (SESCI). Aug. 20-24th, Montréal, Canada.
lishing, 2002.
Pogharian, S., Ayoub, J., Candanedo, J., and Athienitis,
CMHC, 1998. Tap the Sun: Passive Solar Techniques and
A.K., 2008. Getting to a net zero energy lifestyle in Can-
Home Designs. Ottawa: Canada Mortgage and Housing
ada: the Alstonvale net zero energy house. In 3rd Euro-
Corporation.
pean PV Solar Energy Conference. Valencia, Spain.
Duffie, J.A. and Beckman, W.A., 2006. Solar Engineering of Skoplaki, E., Palyvos, J. A., 2009. On temperature depen-
Thermal Processes. Wiley. dence of photovoltaic module electrical performances:
Eckstein, J. H., 1990. Detailed Modeling of Photovoltaic A review of efficiency/power correlations, Solar energy,
Components. M. S. Thesis 83, 614- 624.
– Solar Energy Laboratory, University of Wisconsin, Madi- Sartori, I., Candanedo J., Geier, S., Lollini R., Garde F.,
son: 1990. Athienitis, A., Pagliano, L., 2010. Comfort and Energy
EnergyPlus. 2010. Version 5. 0. Lawrence Berkeley National Efficiency Recommendations for Net Zero Energy
Laboratory, Berkely, CA. Buildings, EuroSun Conference, Graz, Austria, 28 Sep-
EnergyPlus, Weather Data Sources, 2011. http:// 1 Oct.
apps1.eere.energy.gov/buildings/energyplus/ The weather network, http://www.theweathernetwork.com/
weatherdata_sources.cfm, visited on September 3, 2011. statistics/degreedays/cl7025250. Visited on Nov. 2010.
Google SketchUp Plugins, 2011. http:// Torcellini, P.A., Crawley, D.B., 2006. Understanding zero-
sketchup.google.com/intl/en/download/plugins.html energy buildings. ASHRAE Journal 48 (9), 62–64, 66–
Griffith, B.T. and Ellis P.G., 2004. Photovoltaic and Solar 69.
Thermal Modeling with the EnergyPlus Calculation Tripanagnostopoulos Y., Nousia Th., Souliotis M.and
Engine. World Renewable Energy Congress VIII and Yianoulis P., 2002. Hybrid Photovoltaic/Thermal solar
Expo Denver, Colorado. August 29–September 3, 2004. systems. Solar Energy 72, 217-234.
©2012 ASHRAE --`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---

675
Groth, C.C., and Lokmanhekim, M., 1969. Shadow – A Whitaker, C.M., Townsend, T. U., Wenger, H. J., Illiceto, A.,
New Technique for the Calculation of Shadow Shapes Chimento, G. & Paletta, F. 1991. Effects of Irradiance
and Areas by Digital Computer. Second Hawaii Interna- and Other Factors on PV Temperature Coefficients.
tional Conference on System Sciences. Honolulu, HI, Conference Record of the 22nd IEEE Photovoltaic Spe-
Jan. 22-24, 1969. cialists Conference, (Las Vegas, NV 7-10.10.1991), pp.
Walton, G.N. 1983. The Thermal Analysis Research Pro- 608-613. ISBN 0-87942-635-7.
gram Reference Manual. National Bureau of Standards.
(now NIST) February 1983.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
676
ASHRAE Transactions
SA-12-029
Achieving Net-Zero Energy Conditions

in Retrofit—A Case Study for a Performing
Arts Building Using an Hourly
Simulation Approach
D.J. O’Connor C.C. Yavuzturk, PhD J.E. Fuller, AIA
ABSTRACT 19.9% by 2025. Heating ventilation and air conditioning

(HVAC) systems contributed to 33.9% of the total primary
A series of building energy analyses are presented using
energy consumption of commercial buildings (DOE, 2010).
building mathematical models and an hourly simulation
This percentage increases to 56.4% when lighting and domes-
approach for an existing performing arts building. The build-
ing is located on the grounds of the University of Hartford in tics hot water systems are included.
West Hartford, CT. The analyses include the incorporation of A significant impact can be made in the reduction of oper-
a series of renewable energy technologies, specifically ground ating expenditures and the burden placed on fossil fuel
source heat pumps and solar photovoltaic arrays along with resources by addressing the energy consumption in existing
additional energy savings measures with the ultimate goal of buildings. Landsberg et al. (2009) report that an energy effi-
achieving net zero or near net zero energy conditions in retrofit. cient retrofit of lighting in commercial buildings could
The results allowed for the detailed assessment of potentially produce energy savings up to 29%. Energy reductions
most effective measures in relatively older structures. Sensi- between 7% and 23% could be achieved by calibrating and
tivity analyses are presented for the cumulative impact of vari- monitoring building and system controls. Landsberg et al.
ous individual sub-systems and energy savings measures on (2009) also report that equipment upgrades such as the addi-
the overall thermal behavior of the building and its energy use tion of variable frequency drives and variable air volume fans
intensity. A final assessment for the achievement of a net zero could also decrease energy consumption up to 16%. Accord-
energy building is made, along with a discussion of lessons ing to DOE (2010) buildings are responsible for 40% carbon
learned from the system design process. A series of recom- dioxide emissions in the U.S. with 59.7% of carbon dioxide
mendations and design guidelines are developed. emissions produced by space heating, cooling, ventilation,
water heating and lighting systems. Considering the magni-
INTRODUCTION AND BACKGROUND tude and associated sources of energy consumption, the mini-
The dwindling natural resources for energy production mization of energy use in any building will reduce
and the ever rising cost of energy necessitate a judicious environmental pollutants from combustion of fossil fuels.
approach to the design and retrofit of energy-consuming Even 1% reduction in fossil fuel produced electricity in the
devices and systems in buildings, as significant amounts of United States will provide a significant reduction in carbon
energy is consumed in the operation of built environments. dioxide. If coal were the sole energy source for electricity
According to the U.S. Department of Energy (DOE, 2010), the production, 1% reduction in commercial building energy
United States consumed 97.8 Quad (103.2 EJ) of the world’s consumption would result in the elimination of 3.00 × 1010 lbs
primary energy in 2010 and by 2025 consumption is projected (1.36 × 1010 kg) of carbon dioxide (CO2) from the atmosphere.
to surpass 107.9 Quad (113.8 EJ). In 2010, commercial build- Approximately 2.27 × 1010 lbs (1.03 × 1010 kg) of CO2 would
ings were responsible for 18.8% of the primary energy be eliminated for distillate fuels as the source of energy and
consumed in the United States; this is projected to rise to about 1.65 × 1010 lbs (7.48 × 1010 kg) of CO2 for natural gas. It is
D.J. O’Connor is a project manager at NORESCO in Westborough, MA. C.C. Yavuzturk is an associate professor and J.E. Fuller is a profes-
sor in the College of Engineering, Technology and Architecture at the University of Hartford. West Hartford, CT.
--`,`,``,,`,`,```,`,,`
©2012 ASHRAE
677
obvious that with proper incorporation of energy efficient a comprehensive review of all publications is beyond the
technologies in commercial buildings, significant economic intentions of this paper a significantly condensed discussion is
and environmental benefits can be realized. provided. However, a number of studies are directly notewor-
An energy-conscious approach to the design of primary thy, specifically studies relevant to retrofit applications.
energy-consuming devices in buildings, utilizing renewable A series of general strategic guidelines in the design of
energy technologies, sustainable materials, and state-of-the- energy-efficient buildings are provided by Doty (2006) and
art methods is of significant importance for new constructions. Landsberg et al. (2009) while Glaeser (1989) focuses on the
Of equal or if not more importance are energy savings in exist- thermal behavior of insulating glass surfaces, Emmerich et al.
ing buildings that also present considerable opportunities via (2007) on energy savings potential based on air-tightness of
retrofit and redesign. The list of potential measures is well commercial buildings, and Schell et al. (1998) via demand
documented (Landsberg et al. 2009), and may of course be controlled ventilation. A study on the benefits of building
extended by considering permutations on the combination of retrofits using solar energy applications is given by Dalenback
various measures for energy efficiency. Nevertheless, what- (1995) and Nagpal (2006) with specific focus on photovoltaic
ever measures may be considered a whole-building approach applications. Christian et al. (2006) report that a net-zero home
is crucial to properly evaluate the potential for reductions in could be realized with the installation of additional PV panels
energy consumption for the proper assessment of cumulative combined with a tightened envelope, ground source heat pump
effects of various energy efficiency applications. As there are systems, and efficient HVAC equipment. Energy savings
multiple degrees of freedom in the operation of each system potential using energy-efficient HVAC systems are reported
that impacts energy consumption in a building, development by McCracken (2004), Chiasson et al. (2005) while Gonzalez
of detailed and coupled simulations of all building systems (2007) emphasizes the importance and significance of provid-
and sub-systems is necessary to accurately assess the whole- ing energy management in a building based on building auto-
building dynamics and the interaction of various systems with mation systems.
each other in hourly or shorter time-steps. A detailed energy
simulation approach yields multitude of information on the METHODOLOGY
thermal and energy performance for spaces, zones, systems, In order to determine the benefits of proposed energy effi-
buildings, and plants. Performing detailed building simula- cient retrofits to the Lincoln Theater building, a mathematical
tions before the measures are physically implemented could model of the performing arts center was created in a building
prevent costly repair and retrofit efforts that may provide little simulation environment eQUEST (2009). The theater’s
to no benefit in reducing the energy consumption of the build- current construction, space use characteristics, and overall
ing. Indeed, energy retrofits without detailed building simula- energy performance were used to establish a baseline building
tions may be counter-productive and result in greater energy model. Additional models for the building were subsequently
consumption for a building. constructed with consideration of a number of potential
This study examines various retrofit solutions for energy energy efficiency retrofits and modifications. Models that
efficiency of an existing performing arts center as a case study achieved a reduction in energy consumption with retrofits and
with special focus on incorporating ground source heat pumps modifications were assessed for their energy consumption per
systems and photovoltaic arrays. A base model of the building unit area of usable space, energy use intensity (EUI).
was constructed in an hourly simulation environment to repre-
sent the existing energy consumption profile and its energy Baseline Building Model
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
systems performance. Detailed building energy simulations The thermal zoning of the building yielded a total of 15
were performed for each of the proposed retrofits, and zones. The building upper level was modeled with nine zones
comparative analyses were conducted between the base and while the lower level was modeled with five, and the orchestra
retrofit models to assess the effectiveness of energy efficiency pit was modeled as a single thermal zone. The building foot-
applications. This study also develops possible systems and prints and thermal zoning along with three-dimensional
component inter-connect permutations and configurations of ‘exploded’ view of the model are shown in Figures 1 and 2.
technologies that best yield a net-zero, or near net-zero, build- The baseline model for the air-side HVAC systems is
ing energy performance as well as a series of design recom- shown in Figure 3. The building thermal load demands are
mendations for achieving net-zero conditions in similar presently satisfied by four packaged rooftop units (PRTUs).
buildings. Two 25-ton (87.92 kW) PRTUs serve the audience seating
The Literature Review for the study centered on relatively area, the stage, the orchestra pit, and the surrounding perime-
recent major studies focusing on the means to achieve net-zero ter areas. The area behind the stage is served by a 10-ton
or near net-zero energy consuming buildings. It should be (35.17 kW) PRTU and the lobby area is served by a 7.5-ton
noted that there is a significant number of recent publications (26.38 kW) PRTU.
dealing with the issues related to achieving net-zero condi- Each of the thermal zones was coupled to one of the four
tions in building, as well as dealing with individual renewable previously defined HVAC systems. The interior PRTU, the
energy technologies that would lead to net-zero conditions. As first HVAC system, included the interior zones of the theater:
678
ASHRAE Transactions
Figure 1 Building footprints and thermal zoning.
Figure 2 Three-dimensional view of Lincoln Theater.
the orchestra pit, the stage, and the audience seating area. The envelope, ventilation requirements, air distribution system
interior PRTU’s thermal zones are identified as A thru D. The along with step-by-step model input sequences are provided
perimeter zones surrounding the interior zones are identified by O’Connor (2010).
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
as E thru J, and are served by the exterior PRTU. The third

HVAC system, the west wall PRTU, serves the four zones Building Models Incorporating Renewable
behind the stage and are identified as K thru N. The fourth Energy Technologies
PRTU serves the lobby and is identified as O. A detailed Energy models of the theater building were developed to
description of the building along with details of the building provide means with which to evaluate the energy savings
©2012 ASHRAE
679
Figure 3 Packaged rooftop unit—typical for all four HVAC systems as modeled.
potential by replacing the current PRTUs with ground source established using peak building load conditions and consider-
heat pump systems (GHSPs), using an hourly simulation ing 20% propylene glycol solution and stipulating average
approach. Photovoltaic panels were also added to the baseline total ground temperature change in the range of 5°F (2.8ºC) to
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
design model of the theater to assess their effectiveness on 7°F (3.9ºC). High efficiency motors for the pumps were
reducing energy consumption. defined with variable frequency drives (VFD). A vertical bore-
The GSHPs consisted of a water loop heat pump (WLHP) hole field was chosen and sized for the greatest expected heat-
and ground loop heat exchangers (GLHXs) as shown in Figure ing or cooling load for the GLHX. The hypothetical borehole
4. Figure 5 represents the four major components of the GSHP field consisted of 288 boreholes in approximately square
system, water source heat pump, ground loop heat exchangers,
configuration with a depth of 250 ft (76.2 m) for each bore-
primary loop, and secondary loop.
hole.
The fifteen previously defined thermal zones were The PV panels are multi-crystalline Silicon 3.43 ft (1.05
coupled to each of the twelve GSHP systems. The remaining
m) in height and 5.12 ft (1.56 m) in width with a maximum
thermal zones were defined as plenums rather than condi-
power voltage of 54.70 V and a maximum current of 5.58 A.
tioned zones. The plenums provided a non-ducted path for the
Each panel is rated at a capacity of 305.23 W per 1 sun (1000
flow of air from the thermal zones to their respective HVAC
W/m2) of solar irradiation. All of the PV panels were located
equipment. These plenum zones started 9 feet (2.74 m) above
on the roof of the theater building. Models were developed
the occupied spaces and ended at the bottom of the theater’s
roof. Two of the plenum zones were located above the two with PV panel roof coverage of both 40% and 80% of the avail-
spaces adjacent to the west wall (identified as K and L in able total roof surface area. Electric energy production from
Figure 1) of the theater. The third plenum was located above the PV modules is stipulated to occur during the day. However,
the stage of the theater. due to the building’s expected high occupancy levels at night,
the majority of the theater’s energy consumption would take
Primary and secondary loops, their associated pumps, and
place in the evening. Since the production and consumption of
a ground loop heat exchanger (GLHX) were defined and sized
electrical energy were not coincident, electrical energy
in the model to satisfy the building load peaks. Previously
produced by PV panels was stipulated to be sold to the utility,
defined thermal zones and their associated plenum zones were assuming the utility grid to act as a perfect electrical energy
assigned to the secondary loop. The secondary loop provided storage medium.
the sink and source for the heat pumps to reject and extract heat
for their respective thermal zones. The primary loop provided Annual Simulations Using
the thermal conduit needed for the transfer of heat between the the Baseline Design Model
secondary loop and the GLHX. The GLHX allowed for the Building thermal loads and system energy consumption
transfer of heat between the ground and the primary loop as simulations were performed for the baseline design model of
needed to meet the heating and cooling loads of the building. the theater. The building occupancy levels and schedules were
The total flow rate of heat transfer fluid in the ground was calibrated so that the overall cooling and heating requirements
680
ASHRAE Transactions
Figure 4 Air-side HVAC—ground source heat pump system.
Figure 5 Four major components of a ground source heat pump system.
matched the capabilities of currently installed PRTUs. The • Increased thermal resistance of the building envelope
resulting output established a validated baseline, and was used with respect to the baseline model by 25, 50, 75, and
to evaluate the potential energy savings in comparison with 100%,
subsequent model simulations with retrofit modifications. • Reduction to the rate of infiltration by 50%,
Energy Simulations for the Baseline Model with • Addition of economizers with outside air dry-bulb con-
Retrofits and Modifications. A series of parametric runs trol to the PRTUs,
were conducted to assess the effects on the energy consump- • Addition of economizer with outside air dry-bulb and
tion of the individual retrofits and modifications to the base- dual dry-bulb control to the PRTUs,
line design model. Multiple hypothetical system • Addition of economizer with outside air dry-bulb and
reconfigurations were developed: dual enthalpy controls to the PRTUs,
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
©2012 ASHRAE
681
• Addition of demand-controlled ventilation (DCV) and BUILDING SIMULATION RESULTS
economizers with outside air dry-bulb and dual enthalpy
controls to the baseline design model PRTUs, Annual Energy Consumption Using
• Addition of energy recovery ventilation (ERV) and the Baseline Model
economizers with outside air dry-bulb and dual enthalpy The results for the energy simulation using the baseline
controls to the baseline design model PRTUs, design model are shown in Figures 6 and 7. Figure 6 provides
• Increase in the baseline design model HVAC equipment a graphical and numerical representation of the Annual
efficiency by a modest 1% Energy Consumption by end-use while Figure 7 illustrates the
same results in percentages of the total building energy
consumption.
A minimum energy consumption model for the theater
was developed considering the cumulative impact to the base- Building energy simulation results using the baseline
line design model with PRTUs. This additional model model of the Lincoln Theater yield an annual electrical
included individual system reconfigurations from the above consumption of 224,880 kWh (767.3 MBtu). Yearly energy
list that provided the greatest energy savings. consumption from natural gas was 3,026 MBtu (886,883
kWh). Lighting contributed to 73% of the total electrical
Energy Simulations for the Baseline Model of the consumption and almost all of the remaining electrical utili-
Theater with GSHPs. A minimum energy consumption zation was from ventilation fans and space cooling. Space
model was also defined for the baseline model with GSHPs
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
heating contributed to almost all of the gas consumption. The
only. The model with GSHPs included selected individual energy simulation for the baseline design model provided
components from the above list that cumulatively provided the annual energy consumptions of 37,710 kWh (128.7 MBtu)
greatest energy savings from their individual simulation runs. and 3,019.9 MBtu (885,045 kWh) respectively for cooling and
Energy Simulations—Annual Energy Use Intensity heating thermal loads. The maximum cooling load occurred
for the Models of the Theater. The site energy use intensity on Aug 2nd and was 100.033 kW (341 kBtu/hr). The maximum
(EUI) was analyzed for the baseline design model with and heating load was 2.3 MBtu/hr (674 kW) and occurred on the
without cumulative energy retrofits and modifications. The 16th of January. The simulation results display significant
EUI sensitivity analyses included: concurrence with the actual energy consumption profile of the
building based on comparisons to the annual building utility
• Baseline model with PRTUs with and without cumula- bills with expected space use. The baseline model of the
tive retrofits and modifications to the envelope and theater as such provided the energy consumption reference for
HVAC equipment all subsequent potential energy saving retrofits and modifica-
tions. Nevertheless, it should be noted that the accuracy of the
• Baseline model with GSHPs with and without cumula-
prediction between the baseline simulation model and the
tive retrofits and modifications to the envelope and
actual energy consumption of the performing arts building is
HVAC equipment
of minor consequence within the scope of this study as the
focus of interest is on the assessment of the relative effects of
Energy Simulations—Annual Energy Consumption renewable energy technologies and energy savings measures
Comparisons with the Inclusion of Photovoltaic Modules. on the baseline energy consumption profile of the building as
The possible benefits from the inclusion of photovoltaic expressed in energy use intensities (EUI). The evaluation of
modules on the roof of the Lincoln Theater were included in the relative impact is anticipated to allow for the subsequent
the site EUIs using two different comparison points. development of non-dimensional metrics and extrapolations
The first analysis compared the EUI calculated using the to other buildings with similar baseline energy consumption
baseline model to the baseline model with PV panels, which profiles.
hypothetically covered 40% and 80% of the available roof
area. The models with PV panels included the minimum Energy Use Intensity for the Theater
energy consumption that could be obtained from incorporat- Two energy simulation runs were performed to allow for
ing the cumulative energy retrofits and modifications to the site energy use intensity (EUI) comparisons. These runs
models of the theater and its existing HVAC equipment. included the baseline model with and without the below listed
The second analysis compared the EUI calculated using most beneficial cumulative energy retrofits and modifications.
the baseline model to the baseline model with GSHPs and with Figure 8 shows the comparison of the site EUIs with the
PV panels that covered both 40% and then 80% of the avail- theater’s existing PRTUs. Figure 9 shows for the comparison
able roof area. The simulation models with PV panels of the EUIs with GSHPs in place of the theater’s existing
included the minimum energy consumption that could be PRTUs. The site EUIs provided in Figure 8 were for the
obtained from incorporating the cumulative energy retrofits following energy models of the theater:
and modifications to the models of the theater and its existing
HVAC equipment. • Configuration A: Baseline model
682
ASHRAE Transactions
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 6 The Lincoln Theater’s annual energy consumption by end-use using the baseline model.
Figure 7 The Lincoln Theater’s annual energy consumption by end-use using the baseline model—percentages of the total
building energy consumption.
©2012 ASHRAE
683
• Configuration B: Baseline model + envelope thermal • Configuration I: Baseline model with GSHP - envelope
resistance increased by 100% thermal resistance increased by 100%, reduction in out-
• Configuration C: Baseline model + envelope thermal side air infiltration by 50%, demand control ventilation
resistance increased by 100% and reduction in outside (DCV), economizers with outside air dry-bulb and dif-
air infiltration by 50% ferential enthalpy controls.
• Configuration D: Baseline model + envelope thermal
resistance increased by 100%, reduction in outside air As shown in Figure 8, all of the models for the theater
infiltration by 50%, PRTUs with economizers with out- provided energy savings compared to the baseline model.
side air dry-bulb and differential enthalpy controls However, there were three modifications to the baseline model
• Configuration E: Baseline model + envelope thermal with existing PRTUs that provided the greatest savings in
resistance increased by 100%, reduction in outside air energy consumption. The first of these modifications was the
infiltration by 50%, PRTUs with demand control venti- additional increase in the thermal resistance of the envelope
lation (DCV) and economizers with outside air dry-bulb insulation, which provided a 6.8% reduction in energy
and differential enthalpy controls consumption. The other two significant energy savings came
• Configuration F: Baseline model + envelope thermal from the implementation of demand control ventilation
resistance increased by 100%, reduction in outside air (DCV) and energy recovery ventilation (ERV) that provided
infiltration by 50%, PRTUs with energy recovery venti- reductions in energy consumption of 34.3% and 34.2%
lation (ERV), DCV, economizers with outside air dry- respectively. The total EUI reduction for the baseline model of
bulb and differential enthalpy controls. the theater with cumulative energy retrofits and modifications
• Configuration G: Baseline model + envelope thermal was 60.2%. The EUI for the baseline model was reduced by
resistance increased by 100%, reduction in outside air 81.70 kBtu/ft2 (257.6 kW-hr/m2) per year from 135.7 kBtu/ft2
infiltration by 50%, PRTUs with energy recovery venti- (427.9 kW-hr/m2) to 54.0 kBtu/ft2 (170.3 kW-hr/m2) per year.
lation (ERV), DCV, economizers with outside air dry- Substantial energy consumption reductions were also realized
bulb and differential enthalpy controls, and an increase when the PRTUs in the baseline model were replaced with the
in HVAC equipment’s efficiency by 1%. GSHPs.
When compared to the baseline design model’s EUI of
The site EUIs in Figure 9 are given for the following 135.7 Btu/ft2 (427.9 kW-hr/m2) per year (Figure 9), the GSHP
energy models of the theater: model reduced the Theater’s EUI by 64.3% to 48.4 kBtu/ft2
(152.6 kW-hr/m2) per year. When the GSHP model, with addi-
• Configuration A: Baseline model tional cumulative energy consumption retrofits and modifica-
• Configuration H: Baseline model with Ground Source tions, was compared to the GSHP model, the EUI for the
Heat Pump (GSHP) GSHP model was reduced by 28.0% from 48.4 kBtu/ft2 (152.6
Figure 8 Site EUIs for the baseline model with PRTUs with cumulative energy consumption retrofits and modifications.

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---

Figure 9 Site EUIs for the baseline model, and baseline models with GSHPs with cumulative energy consumption retrofits
and modifications.
kW-hr/m2) to 34.8 kBtu/ft2 (109.7 kW-hr/m2) per year. There For the models with PRTUs, PV modules with 40% and
was a 74.3% total decrease in the baseline model’s EUI 80% roof coverage (Figure 10) reduced the site EUI from
provided by the GSHP with cumulative energy consumption 135.7 kBtu/ft2 (427.9 kW-hr/m2) to 36.5 kBtu/ft2 (115.1 kW-
retrofits and modifications. hr/m2) and 20.9 kBtu/ft2 (65.9 kW-hr/m2) respectively. The
40% and 80% PV roof coverage models provided 73.1% and
Energy Use Intensity Considering the Inclusion of 84.5% respective decreases in site energy. The GSHP models
Photovoltaic Modules with PV also provided significant energy savings.
With 40% and 80% PV roof coverage, Configurations L
Energy consumption benefits from the inclusion of photo- and M respectively, the GSHP models (Figure 11) reduced the
voltaic modules hypothetically installed on the roof of the site EUI from 135.7 kBtu/ft2 (427.9 kW-hr/m2) to 18.1 kBtu/
Lincoln Theater were evaluated via hourly system simulations ft2 (57.1 kW-hr/m2) and 1.4 kBtu/ft2 (4.4 kW-hr/m2), which
to obtain the site EUIs for the baseline model with PRTUs were respective decreases of 86.6% and 98.9%.
(Figure 10) and with GSHPs (Figure 11). These simulations Figure 12 illustrates the end-use energy consumption
were performed with the below listed minimum energy profile in the performing arts building considering the cumu-
consumptions that could be obtained from retrofits and modi- lative impact of the energy savings measures and the imple-
fications to the theater’s envelope and HVAC equipment. The mentation of renewable energy technologies (ground source
models stipulated 40% and 80% PV roof coverage. heat pumps and solar PV with 80% roof coverage).
As a result of the simulated replacement of the current
HVAC system in the building, the total energy consumption
• Configuration J and K - Baseline model + envelope ther-
for space heating (natural gas fuel consumption) that is shown
mal resistance increased by 100%, reduction in outside
in the baseline is shifted in its entirety to consumption of elec-
air infiltration by 50%, PRTUs with energy recovery
tricity in the models with energy savings measures and renew-
ventilation (ERV), DCV, economizers with outside air
able energy technologies. Consequently, the total electricity
dry-bulb and differential enthalpy controls, and an
use in the building when compared to the baseline model is
increase in HVAC equipment’s efficiency by 1%.
shown to increase somewhat. However, since the source of
• Configuration L and M - Baseline model with GSHP + electricity is primarily from solar PV for the simulation model
envelope thermal resistance increased by 100%, reduc- that includes 80% roof coverage, the total energy consumption
tion in outside air infiltration by 50%, demand control from sources not produced on site per unit area decreases
ventilation (DCV), economizers with outside air dry- significantly. Furthermore, the significant impact on energy
bulb and differential enthalpy controls. savings is noticeable with incorporation of demand control
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
©2012 ASHRAE
685
Figure 10 Site EUIs for the baseline design model, and baseline models with PRTUs with cumulative minimum energy
consumption retrofits and modifications—40% and 80% PV roof coverage.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 11 Site EUIs for the baseline design model, and baseline models with GSHPs with cumulative minimum energy
consumption retrofits and modifications—40% and 80% PV roof coverage.
686
ASHRAE Transactions
Figure 12 The Lincoln Theater’s annual energy consumption by end-use for the baseline model and models with GSHPs with
cumulative minimum energy consumption retrofits, modifications and 80% PV roof coverage.
ventilation coupled to the ground source heat pump systems. Although, only modest energy savings were realized for
It should also be noted that the decrease in the space heating, the simulation runs considering the following measures they
cooling and fan energy consumption is due to reduced building should not be overlooked: Orientation of the building (obvi-
loads after the installation of solar PV arrays on the roof of the ously not a retrofit measure, unless significant changes in the
performing arts building as with the presence of the PV arrays building thermal zoning can be made due to modified space
a significant portion of the roof surface is not exposed to direct use requirements); economizer and optimized control of the
solar radiation. In addition, the PV arrays also yield additional economizer; and further efficiency increases in existing
increase in the overall thermal resistance of the roof, reducing HVAC equipment.
building heating loads.
Demand control ventilation was one of the most effective
measures in reducing the energy consumption for the theater
CONCLUSIONS AND RECOMMENDATIONS
building due to the large number of people in the theater during
Analyses on the Lincoln Theater case study indicate that performances, which occurred mostly during evening hours,
net-zero or near net-zero energy conditions appear to be and the significantly varying occupancy levels during the day.
achievable in a performing arts building in retrofit, consider- Although DCV provided a considerable decrease in energy
ing the incorporation of ‘off-the-shelf’ renewable energy tech- consumption of the theater, it may not provide the same degree
nologies and energy savings measures that utilize hourly of savings for other buildings depending on the space use char-
system control and operating strategies. The most significant acteristics. It should be noted that DCV will be ineffective for
reductions in building energy consumption expressed in EUI buildings that experience no changes in occupancy levels,
have been achieved through the addition of demand control because the amount of ventilation required will remain
and energy recovery ventilation, the replacement of PRTUs constant. For buildings that will benefit, DCV may be imple-
with a ground source heat pump system, and the addition of mented with CO2 space sensors and fresh air intake and
photovoltaic arrays. exhaust dampers on the HVAC equipment.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
©2012 ASHRAE
687
Energy recovery ventilation also provided a notable function that minimizes total building energy consumption
decrease in energy consumption for the packaged roof top and total cost of energy or both, using an hourly system model-
units. The savings were due to energy transferred to/from ing and simulation approach. A minimization algorithm that is
exhaust air to pre-condition the large amount of outside air specific for a building and its ambient climate conditions
introduced for ventilation. Consideration should be given to would also allow for the optimal sizing of building mechanical
retrofitting ERV to existing HVAC equipment. When replac- systems and components, reducing typically high system first-
ing HVAC equipment consideration should be given to install- costs, making the adoption of renewable energy technologies
ing new equipment with integrated ERV. Although, it should associated sub-systems in retrofit economically more compet-
be noted that buildings with low ventilation rates due to low itive in significant ways.
occupancy levels might only marginally benefit from ERV.
The replacement of PRTUs with ground source heat REFERENCES
pumps accounted for a large reduction in energy consumption ASHRAE Standard 62. 2010. (American Society of Heating,
for the Lincoln Theater building. Multiple zoning with GSHPs Refrigerating and Air Conditioning Engineers). Stan-
will provide better thermal control for the respective thermal dards for Natural and Mechanical Ventilation.
zones, while decreasing the energy use intensity significantly. Chiasson, A.D, C. Yavuzturk, and W.J. Talbert. 2005.
However, dividing larger thermal zones into smaller ones “Design of school building HVAC retrofit with hybrid
would most likely require additional ductwork. geothermal heat-pump system.” Journal of Architectural
The production of electric energy from roof mounted Engineering. 10.3: 103-111.
photovoltaic panels can greatly offset the energy consumption Christian, Jeffrey E, L. Richards, P. Childs, J. Atchley, and
of a building. However, building electrical loads occurring H. Moon. 2006. “Energy efficiency, SIPS, geothermal,
and solar PV used in near zero-energy house.” ASHRAE
during night hours will only benefit from the electrical output
Transactions. 112.2: 275-285.
of PV, if utility electrical grid can be assumed to be analyzed
Dalenback, Jan-Olof 1996. “Solar energy in building renova-
as an off-site electrical storage medium, which may be approx- tion.” Energy and Buildings. 24.1, 39-50.
imated reasonably via net metering. DOE. 2010. “U.S. Department of Energy Efficiency and
It should be noted that the energy consumption profile of Renewable Energy, 2010. Buildings Energy Databook
the theater building may not be representative of similar 2010” http://buildingsdatabook.eren.doe.gov/docs/Data-
performing arts buildings. The theater’s lack of fenestration Books/2010_BEDB.pdf
and its highly concentrated and varying occupancy levels are Doty, Steve. 2006. “Strategic facility guidelines for
not typical of most commercial buildings. In general, the accu- improved energy efficiency in new buildings.” Strategic
racy of energy simulations for any type of building is depen- Planning for Energy and the Environment. 26.3: 7-33.
dent upon a properly constructed energy model that best Emmerich, Steven J., T. P McDowell, and W. Anis. 2007.
approximates the thermal conditions at the site. In order to “Simulation of the impact of commercial building enve-
ensure the results are accurate, the building’s orientation, lope airtightness on building energy utilization.”
ASHRAE Transactions. 113: 379-399.
physical characteristics, internal loads, and its air-side and
eQUEST. (1998-2009). Quick Energy Simulation Tool. Ver-
water-side heating, ventilation, and air-conditioning equip-
sion 3.63b, build 6500. http://www.doe2.com/equest
ment should be carefully evaluated and accurately represented
Glaeser, Hans J. 1989. “Coated Heat-Insulating Glasses.”
in the mathematical model. Multiple simulation runs for Glastechnische Berichte. 62.3: 93-99.
multiple system configuration scenarios leading to parametric Gonzalez, Roberto. 2007. “Energy management with build-
studies should be performed to ensure the identification and ing automation.” ASHRAE Journal. 49.1: 26-32.
possible implementation of the most energy efficient design. Landsberg, D. R, M. Lord, S. Carlson, and F. Goldner. 2009.
An energy model of an existing building that is accurately Energy Efficiency Guide for Existing Commercial
represented and thoroughly analyzed for all possible reduc- Buildings: The Business Case for Building Owners and
tions in energy consumption, can provide the information Managers. ASHRAE. ISBN: 978-1-933742-63-2.
needed to produce a highly energy efficient building through Nagpal, Shreshth. 2006. “Integration of photovoltaics in
retrofits and straightforward modifications. The energy use building envelope: Effect on whole building energy per-
intensities for the models of the Lincoln Theater are examples formance.” International Solar Energy Conference.
of the potential savings that can be realized. With combined 2006.
retrofits and modifications net-zero and near net-zero EUIs O’Connor, D. J. 2010. “Simulation of the Thermal Perfor-
were shown to be possible for existing performing arts build- mance of a Net Zero Energy Performance Arts Build-
ing”. Master of Engineering (MEng) Thesis. Mechanical
ings using the example of the Lincoln Theater building.
Engineering. College of Engineering, Technology and
Finally, it should also be noted that the problem of the best Architecture. University of Hartford.
combination of renewable energy technologies and energy Schell, M.B., S.C. Turner, and R.O. Shim. 1998. “Applica-
savings measures along with system control and operation
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`--
tion of CO2-based demand-controlled ventilation using

strategies for a building has multiple degrees of freedom. It is ASHRAE Standard 62: optimizing energy use and ven-
therefore one that requires optimization based on an objective tilation.” ASHRAE Transactions. 104.2: 1213-1225.
688
ASHRAE Transactions
SA-12-030
HVAC Systems Commissioning

in a Manufacturing Plant
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Yunhua Li Bei Zhang Mingsheng Liu, PhD, PE
Student Member ASHRAE Student Member ASHRAE Member ASHRAE
Lixia Wu Jinrong Wang, PE Tom Lewis

Member ASHRAE Member ASHRAE Member ASHRAE
ABSTRACT ing commissioning is driven more strongly by non-energy

objectives such as thermal comfort, indoor air quality, and
As one of the most cost-effective ways to improve energy
overall building performance with an emphasis on ensuring
efficiency, commissioning plays an important role in achieving
that the building functions according with the design intent,
building energy savings goals. In this paper, a specific case
study of energy saving-aimed commissioning was imple- whereas the commissioning of existing building is more
mented in an existing building by using the methods of retro- strongly driven by energy efficiency objectives by identifying,
fitting and optimizing system control and operation. The case resolving the operating problems, optimizing system energy
study site is a manufacturing plant with a gross area of use [Mills et al, 2004; Liu et al. 2002]. As one of the most cost-
1,212,940 sq. ft. (112,685 sq. meters). The work scope of the effective ways to improve energy efficiency, commissioning
case study covered the chilled water system with four 1200 ton plays an important role in achieving building energy saving
(4,219 kW) chillers, heating system with two 70,000 lb/hr goals. In this article, a specific case study with energy saving-
(31,751kg/hr) steam boilers, and thirty-four air handling oriented commissioning was conducted in a manufacturing
units. After a detailed walk-through of the plant, a series of plant to optimize the performance with the main methods of
system energy saving opportunities were identified for the retrofitting and optimizing the system operation and controls.
existing system operation such as chilled water bypassing the The case study plant, originally built in 1959, had a
building, blending of supply and return chilled water, high comprehensive and self-sufficient system including chilled
oxygen level in flue gas of the boiler, and extra fan power water system, processing water system, boiler system,
consumption for constant air volume systems and the like. After compressed air system, and air conditioning system. Later,
system retrofits and implementation of advanced control strat- system retrofits were made gradually such as reconstruction of
egy, system overall performance was greatly improved. The the original coal fired boiler into dual fuel (natural gas and oil)
compared utility data before and after the system optimization fired boiler with an automatic control system, installation of
indicate a total annual saving of $508,558 with the total elec- variable frequency drives (VFDs) for the chilled water pumps
tricity consumption reduced by 15% and total gas consumption and so on so as to improve the system performance. However,
reduced by 7%, considering the economic impact. In the paper, for the whole plant, great energy saving potential still exists
the system information, major retrofits, and advanced control allowing for further increases in system efficiency. In 2008, a
sequences and operational data were introduced and detailed system inspection was conducted and potential
discussed. energy saving opportunities were identified and proposed. The
retrofitting and most system optimizations were implemented
INTRODUCTION
in 2009. After that, a one-year follow-up was also provided for
Building commissioning is a process to improve the operator training and troubleshooting. Based on the whole
building energy performance from its design phase, construc- inspection and implementation process, this article briefly
tion phase, and operation phase [ASHRAE 2007]. New build- introduces the existing system information, points out the
Yunhua Li and Bei Zhang are doctoral students and research assistants in the Department of Architectural Engineering, University of
Nebraska–Lincoln, Omaha, NE. Mingsheng Liu is president and CTO of Bes-Tech Inc., Omaha, NE. Lixia Wu is a director of engineering
at Bes-Tech Inc., Philadelphia, PA. Jinrong Wang is the manager and Tom Lewis is a senior technical analysis engineer of commercial and
industrial solutions at Omaha Public Power District, Omaha, NE.
©2012 ASHRAE
689
energy saving opportunities, and presents the system retrofits consumption. The blending of chilled water supply and return
and improved control strategies. The operational data before not only reduces chiller’s COP, but may also degrade the
and after the system optimization were also analyzed and AHUs’ performances and cause building comfort issues.
compared with a total annual energy savings presented. Therefore, a single loop system with bypass line fully closed
was proposed as well as other related system retrofits and
SYSTEM INFORMATION controls to eliminate these issues and improve the system
The case study plant is located in Omaha, Nebraska. The performance.
central plant with a gross area of 71,803 sq. ft. (6,671sq.
meters), consists of chiller system and boiler system. The air Boiler System
handling units serve office building and main plant with a total
area of 1,212,940 sq. ft (112,685 sq. meters). The whole Due to the reduction of production, the boiler operated
under partial load for most of the time. It ran with low effi-
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
system works with a 24/7 schedule.
The chiller system supplies chilled water mainly for the ciency under lower boiler load with the flue gas oxygen level
air handling units (AHUs) for several buildings. The chilled as high as 16% which indicated too much excess air. However,
water system, shown in Figure 1, was designed as a primary- each boiler had two burners operating together in parallel
secondary system consisting of five chillers in parallel: four which can actually work separately as two individual burners.
1200 ton (4,219 kW) electric centrifugal ones and a steam- In addition, when the constant speed FD fans ran, the related
driven one which was seldom used recently. Each chiller was damper was closed most of the time which resulted in unnec-
equipped with a constant speed primary pump of 50 HP (36.8 essary fan power consumption. Similarly, the feed water
kW). Five 200 HP (147.1 kW) variable speed secondary pumps ran with constant speed all the time, whereas the boiler
pumps in parallel delivered the chilled water to the AHUs. In feed water valves were found to operate in low position range
recent years, four electric chillers were still in operation,
which were covered in the work scope.
The boiler system included three 70,000 lb/hr (31,751kg/
hr) dual-fuel fired boilers with two of them still in operation.
Each boiler was built with two burners operated in parallel,
one constant speed forced draft (FD) fan equipped for each
burner, and one induced draft (ID) fan with fluid drive at the
boiler stack outlet. The feed water system consisted of two 50
HP (36.8 kW) constant speed pumps delivering feed water to
each boiler drum.
The air conditioning systems were comprised of thirty-
one single-zone constant air volume (CAV) AHUs serving the
manufacturing area, three single duct variable air volume
(VAV) AHUs, and one DX rooftop unit (RTU) serving the
office area. For the thirty-one CAV units, there were three
types of cooling coil configurations: type 1, cooling coil with
cooling valve only; type 2, cooling coil with cooling valve and
Figure 1 Chilled water system diagram.
constant speed booster pump; type 3, cooling coil with cooling
valve and variable speed booster pump, as seen in Figure 2.
The single duct VAV units were equipped with type 1 cooling
coils. The DX RTU, equipped with two reciprocating
compressors, supplied constant air flow to the office room.
ENERGY SAVING OPPORTUNITIES

After a detailed inspection, the following energy saving
opportunities were identified.
Chilled Water System

The existing chilled water system, as seen in Figure 1, was
a constant primary flow and variable secondary flow system,
which caused excessive chilled water to bypass the chillers or
the building through the bypass line under off-design condi-
tion when the chiller design flow did not match with the build-
ing side flow, which resulted in significant pump power Figure 2 Three types of configurations for cooling coils.
690
ASHRAE Transactions
most of the time, which caused extra pump power consump- Table 1 summarizes the detailed retrofits for all the
tion and degraded the valve lifespan. So, the boiler system effi- systems.
ciency could be improved by optimizing the burner operation
sequence, reducing the excessive combustion air, and optimiz- SYSTEM CONTROL OPTIMIZATION
ing FD fan control and feed water pump control.
Chilled Water System
Air Conditioning System
Existing Control: The chiller start/stop was determined
The thirty-one CAV AHUs with constant speed motors by the operators based on their experiences.
consumed excessive fan power under partial load conditions, Operation of the primary pump was based on the opera-
of which twenty-one were designed with two-position inlet tion of related chiller. When a chiller was to start, the associ-
guide vanes for the supply fans. VFD was proposed to be ated primary pump was started first. It ran at constant speed
installed for each fan to improve the fan efficiency to over- and would not stop until the chiller was stopped.
come the drawbacks of inlet guide vanes as well. The three The secondary pump speed was modulated to maintain
VAV systems, designed with constant static pressure set point, the differential pressure of remote coil at set point of 1.5 psi
caused excessive fan head and fan power. Automatic static (13.1kPa). All pumps online ran at the same speed. The start/
pressure reset was proposed for the improved control. For the stop of a lag pump was based on 90% and 40% principle.
CAV DX rooftop unit, the existing compressor staging control When the online pump speed was higher than 90% of design
caused unstable space air temperature which can be further speed, a lag pump was started; when the online pump speed
optimized. Meanwhile a VFD was proposed to be installed for was lower than 40%, a lag pump was stopped. When a lag
the supply fan. pump was started, it ran at the same speed as the lead pump of
Based on these inspection results, retrofits and new 90%, which could cause pump cavitation at the moment it was
control strategies were recommended and implemented on the started.
site in 2009. The system chilled water supply temperature was fixed at
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
44°F (6.7°C) all the time.

SYSTEM RETROFITS Improved Control: In the single loop system without
bypass pipeline and primary pumps, the chilled water flow
For the chilled water system, a series of tests were through chillers was variable and always matched with the
conducted before converting the system to a single loop building-side flow. With the chiller control panel installed, the
system to verify its feasibility. The test results demonstrated chiller can communicate with the system controller. The
that when fully closing the isolation valve on the bypass line sequence of chiller staging was developed. When the load
and shutting down primary pumps, the secondary pumps had ratios of all chillers were higher than 80%, a lag chiller was
the capability to provide enough differential pressure for the started; when the load ratios of all chillers were lower than
remote AHU, since the cooling valve on the remote AHU still 40%, one chiller was shut down, which was based on the prin-
worked within functional range. ciple that chiller has a higher efficiency when running between
For air handling units, all the CAV systems were 40% and 80% load ratios [Liu et al. 2002].
converted to VAV systems. Two Fan Airflow Stations (FASs) The chilled water supply temperature set point (CHWST
[Liu 2006] were installed for two VAV units to optimize the StPt) was reset based on the system chilled water flow rate, as
static pressure. seen in Figure 3. If more than one chiller were online, or only
Table 1. Summary of System Retrofits
System Retrofits
Install new control panels for chiller 1-3

Chilled water system Remove the flow control valves and primary pumps
Shut off the isolation valve on the bypass pipeline
Install two VFDs for FD fans of the two boilers in operation
Boiler system
Install two VFDs for the feed water pumps
Lock the inlet guide vane fully open, shut off the reheat valve and fully open the bypass damper permanently.
CAV AHUs
Install VFDs for all supply fans and return fans.
VAV units Install Fan Airflow Stations on two VAV units
Install VFD on supply fan

DX RTU
Replace the existing two-position actuator with a new modulating actuator for the outside air damper.
©2012 ASHRAE
691
one chiller was online and the system chilled water flow was hr (3,516 kW), the operation of two burners could cause the
higher than 60% of chiller design flow, the set point was set at system steam pressure higher than the set point during summer
44°F (6.7°C). If only one chiller was online and the chilled which indicated energy waste. Therefore the single/double
water flow was lower than 60%, this set point was reset from burner mode was built. When the system heat demand was
44°F (6.7°C) to 48°F (8.8°C) to maintain the system chilled lower than 120 therms/hr (3,516 kW), one burner was shut
water flow rate above the chiller allowable low limit flow down and the system ran in single burner mode. When the
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
recommended by the chiller manufacturer. system heat demand was higher than 120 therms/hr (3,516
For the secondary pump control, the best pump efficiency kW), the offline burner was fired. With single burner mode
method was applied to find the critical speed when calling for built, the boiler can operate with the minimum firing rate of 60
or shutting down a pump. For this site, it was found that the therms/hr (1,758 kW).
critical speed had less influence on the pump power consump- With VFD installed for FD fans, the burner combustion
tion. Therefore, 90% and 40% speed limit was reserved. The airflow was controlled by modulating the FD fan speed with
only distinction was that when a lag pump was called, it would the FD fan damper full open all the time, which delivered the
speed up gradually until reaching the same speed as the lead burner with required airflow but consumed less fan power.
pump, which solved the pump cavitations issue. Among the factors that influenced the boiler efficiency,
excess air was an important one. Practically, for complete
Boiler System combustion, more air must be supplied than theoretically
required [David 2002]. However, too much excess air
Existing Control: The steam pressure set point of the
degraded the boiler efficiency by heating excess cold air to
boiler system was set at 110 psi.
stack temperature [Energy efficiency handbook]. Using
When a boiler was running, two burners ran together at oxygen level in flue gas as an indicator for excess air, usually,
the same firing rate with a total minimum rate of 120therms/ 3% oxygen level in flue gas means 3% efficiency drops [Boiler
hr (12,661MW). The combustion airflow of each burner was Efficiency Improvement guide]. Considering the control and
maintained at its set point by modulating the FD damper with system changes, the old boiler air-fuel curve when the system
the FD fan running at the constant speed. The Air/fuel ratio was built need to be calibrated. A test was conducted to adjust
data were the same as that when the system was built years the curves for both single burner and double burner modes.
ago. After the adjustment of the air/fuel ratio, the new oxygen level
The feed water system line pressure was maintained at a under various firing rates before and after the test were given
fixed set point of 120 psi (827.4 kPa) with feed water pump in Figure 4, which demonstrates the reduction of excess air and
running at constant speed all the time. the improvement of the boiler efficiency.
Optimized Control: The system steam pressure set point However, because the boiler code had a strict requirement
was reduced to 50 psi based on the requirements of production on the amount of the lowest combustion air of a boiler,
and equipment. although a burner was shut down in single burner mode, the
With the existing control and new steam pressure set associated FD fan still needed to run to supply excessive
point, when the boiler heat demand was less than 120 therms/ combustion airflow so as to maintain the total airflow in
Figure 3 Chilled water supply temperature reset.
692
ASHRAE Transactions
Figure 4 Oxygen level before and after optimization.
compliance with the boiler code requirement, under which When the economizer was off, the outside air damper was set
condition the efficiency of the boiler was the best that can be at the minimum position.
achieved. The space air temperature and supply air temperature
With VFD installed in the feed water system, the system were controlled differently for different systems, as described
line pressure set point was maintained by modulating the feed in Table 3.
water pump speed with the line pressure set point reset from Figure 6 and Figure 7 compares the results of two control
70 psi (482.6 kPa) to 120 psi (827.4 kPa) based on the system methods used for DX unit. Better supply air temperature
heat demand and feed water valve position. Since the boiler control was achieved.
drum level was controlled by the feed water valve, as long as For the three single duct VAV systems, the supply air
the feed water valve operated in the functional range, the boiler temperature was controlled by the cooling valve and its set
drum level was in good control. The trending data before and point can be reset based on the outside air temperature. The
after the optimization in Figure 5 show that the feed water existing control of space temperature was unchanged. The
valve operated at an average opening of 78% for optimized static pressure was maintained by modulating the supply fan
control compared with 20% for the existing control, which speed with the set point reset based on the supply airflow
indicates better valve performance and less pump power calculated from the Fan Airflow Station, as shown in Figure 8.
consumption.
RESULTS
Air Conditioning Systems
After implementing the proposed control strategy on the
Existing Control: For the thirty-one AHUs, the supply chilled water system, boiler system, and air conditioning
fans ran with constant speed. The air temperature was systems, the space thermal comfort was well maintained,
controlled using the methods listed in Table 2. For all the overall energy efficiency and reliability was greatly improved,
systems, the economizer was not used. For the three VAV and fan power consumption was significantly reduced as seen
units, the system static pressure was maintained at a constant in Figure 9 and Figure 10 which compare the electricity and
set point by modulating the VFD speed. gas consumption under different average daily temperatures
Optimized Control: The Temperature-based econo- before and after commissioning. Table 4 lists the consumption
mizer was used for all the single zone AHUs and single duct and savings for electricity and gas. The total annual electricity
VAV systems, with the following sequences: 1) The econo- and gas cost were reduced by $1,069,000 and $201,859
mizer was activated when outside air temperature was bellow respectively. However, the depression of economy had some
65°F (18.3°C), and deactivated when the outside air tempera- impacts on the energy usage in this plant by reduced occu-
ture was over 68°F (20°C). 2) In economizer mode, the outside pancy and machines usage, and so on. Since there was no
air damper and return air damper were modulated together to available data log to calculate this portion, 60% of energy
maintain the mixed air temperature set point which was two savings was assumed due to the economy impact based on the
degrees lower than the supply air temperature set point. 3) comparison of the numbers of employees and operating
©2012 ASHRAE
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
693
a) Existing Control
b) Optimized Control
Figure 5 Feed water system pressure control comparison.
Table 2. The Space and Supply Air Temperature Control with the Existing Control
Systems Space air Temperature (Tspa) Supply Air Temperature (Tsa)
Single zone units (with type 1

Controlled by modulating the cooling valve Not controlled
and type 2 cooling coil)
Controlled by modulating the booster pump speed Not controlled
cooling coil)
The compressors were staged to maintain Tspa set point at 74°F

(23.3°C). When Tspa > 75.5°F (24.2°C), the first compressor was
DX rooftop unit Not controlled
turned on. If Tspa could not be maintained within 5 minutes, the sec-
ond compressor was turned on.
Controlled by modulating the damper and reheat coil Controlled by modulating the
Single duct VAV units
at the terminal box cooling valve without reset
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
694
ASHRAE Transactions
Table 3. The Space and Supply Air Temperature Control with the Improved Control
Space Air
Systems Supply Air Temperature (Tsa)
Temperature (Tspa)
Controlled at 55°F (12.8°C) by modulating the cooling valve. When fan

speed was at the minimum, Tsa set point would be reset from 55°F
and type 2 cooling coil) Controlled by modulating the (12.8°C) to 60°F (15.6°C) based on the Tspa.
fan speed with the minimum
speed of 20% Controlled at 55°F (12.8°C) by modulating the booster pump speed.
When fan speed was at the minimum, Tsa set point would be reset from
cooling coil)
55°F (12.8°C) to 60°F (15.6°C) based on the Tspa.
When Tsa > 65°F (18.3°C) or supply fan speed > 80%, the first com-
pressor was turned on; when Tsa < 53°F (11.7°C), the first compressor
Controlled by modulating the
was turned off; when the first compressor was on, if Tsa > 60°F
DX rooftop unit fan speed with the minimum
(15.6°C) for 2 minutes, the second compressor was turned on; when
speed of 50%
both two compressors were on, if the Tsa < 53°F (11.7°C), the second
compressor was turned off.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 6 DX unit supply air temperature with existing control.
Figure 6 DX RTU staging control with improved method.
©2012 ASHRAE
695
Figure 8 Static pressure control.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Figure 9 Electricity consumption comparison.
Figure 10 Gas consumption comparison.
696
ASHRAE Transactions
Table 4. Energy Consumption and Savings Before and After Retrofitting
Pre Post Sub-Savings
Consumption (kWh) 47,677,486 29,477,281 18,200,205

Electricity
Cost ($) 2,030,499 1,735,721 1,069,537
34,789
consumption (DTH) (GJ) 185,939 (196,165) 151,150 (159,463)
Gas (36,702)
Gas cost ($) 1,597,872 877,032 201,859
Total savings ($) - - 1,271,396
Adjusted total savings ($) - - 508,558
machines. With these taken into consideration, the energy Boiler Efficiency Improvement Guide. http://www.clean-
savings due to the commissioning were 15% and 7% for elec- boiler.org/
tricity and gas consumption respectively, with a total amount
of cost about $508,558. The total cost including retrofits and E.Mills, H.Friedman, T.Powell, N.Bourassa, D.Claridge,
services are about $501,120. Therefore, the payback is about T.Haasl, M.A.Piette. 2004. The Cost-Effectiveness of
one year. Commercial-Buildings Commissioning. A Meta-Analy-
sis of Energy and Non-Energy Impacts in Existing
SUMMARY Building and New Construction in the United States.
This paper presents a detailed commissioning process for Report LBNL-56637. Lawrence Berkeley National Lab-
the HVAC system of an existing manufacturing plant. oratory.
Commissioning covers the retrofits and the optimization of the
David C. Farthing. Combustion control strategies for single
operation and system control for the chilled water system,
central heating system, and all air handling units based on the and dual element power burners. Steam Digest. 2002
building owner requirements. Significant annual energy compiled for the industrial technologies program. U.S.
savings were achieved. However, considering that the plant Department of Energy. Energy efficiency and renewable
experienced an economic decline in recent years, not all the energy. pp29-34
energy savings were attributed to the result of commissioning.
Energy Efficiency Handbook. Council of Industrial Boiler
60% of total savings was attributed to the economic decline
Owners (CIBO).
based on the reduction of employment and machine usage, so
that a total amount of $508,558 savings was finally achieved, Liu G., 2006. Development and applications of fan airflow
about 15% reduction for electricity use and 7% reduction for station and pump water flow station in heating, ventilat-
gas consumption. The payback is one year, which demon- ing and air-conditioning (HVAC) systems. PhD thesis.
strates the cost-effectiveness of the commissioning services. University of Nebraska-Lincoln.
REFERENCES Liu M., D.E.Claridge, and D. Turner, 2002. Continuous
ASHRAE. 2007. ASHRAE Handbook- HVAC Applications. Commissioning Guidebook: Maximize Building Energy
Atlanta: American Society of Heating Refrigeration and Efficiency and Comfort, Federal Energy Management
Air Conditioning Engineers, Inc. Program.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
©2012 ASHRAE
697
SA-12-031
Investigation of CO2 Tracer Gas-Based

Calibration of Multi-Zone Airflow Models
Steven C. Snyder T. Agami Reddy, PhD, PE William P. Bahnfleth, PhD, PE

Associate Member ASHRAE Fellow ASHRAE Fellow ASHRAE
ABSTRACT the behavior and performance of existing buildings. Building-

specific guidance from a calibrated multi-zone model
This paper proposes a methodology to calibrate multi-
provides greater confidence (as compared to a non-calibrated
zone airflow models in existing buildings. The methodology
model) in analyzing IAQ issues during normal diurnal and
has been demonstrated with the widely used multi-zone model
seasonal operation, or in formulating response plans and
CONTAM and applied to two buildings, one synthetic and one
analyzing the effectiveness of possible mitigation measures
actual. The methodology builds on previous work with the cali-
when the building is subject to either accidental or intentional
bration process enhanced by the use of carbon dioxide tracer
air-borne contaminant releases. In such cases, the programs
gas test data. A key concept highlighted in this paper is that
have to be calibrated, i.e., the numerous model parameters
calibration of airflow models is a highly over-parameterized
need to be tuned so that simulated output closely matches
problem and that some form of model reduction is imperative.
observed system performance under some baseline condi-
This is achieved by performing an airflow-based sensitivity
tions. This paper proposes a calibration methodology intended
analysis to identify influential system drivers, a tracer gas-
to improve the prediction accuracy of multi-zone software
based sensitivity analysis to identify macro-zones (i.e. groups
programs.
of rooms that can be combined into one zone for the purposes
of predicting or studying dynamic airflow behavior under
different types of stimuli), and finally, a process of tuning the Objectives and Scope
model parameters. It is found that for mechanically ventilated
CONTAM (Walton and Dols 2008) is a multi-zone model
commercial buildings, calibrating multi-zone models by
software program for indoor air quality and ventilation anal-
reconciling differences between measured and predicted
ysis. The main applications of CONTAM include assessing
tracer gas behavior is more strongly impacted by those factors
adequacy of ventilation rates, design and analysis of smoke
which significantly influence room air change rates than by
systems, assessing indoor air quality performance, predicting
airflow path leakage parameter tuning.
contaminant dispersion, and estimating personal exposure.
The objective of this paper is to propose a new calibration
INTRODUCTION
methodology and to demonstrate it using CONTAM. Previous
The modeling and simulation of airflow dynamics in research efforts by Bahnfleth and colleagues (Firrantello et al.
buildings has many applications including analyzing indoor 2005, Firrantello 2007a, Firrantello et al. 2007b, Sae Kow
air quality (IAQ) and ventilation flows, predicting contami- 2010, Bahnfleth et al. 2012) have led to the development of a
nant dispersion dynamics, and calculating personal occupant calibration methodology which is based on measured heating,
exposure under extreme events. Traditionally, multi-zone ventilation, and air-conditioning (HVAC) airflow rates and
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
airflow software programs have been used to study such inter-zonal airflow directions; this has been partially validated
effects during the design phase. Similar to detailed building using collected field data. This paper builds upon, and subse-
energy use simulation programs, they can also be used to study quently improves, the previous work by developing a refined
Steven Snyder is an energy services engineer for Johnson Controls Inc., Philadelphia, PA. T. Agami Reddy is a professor with faculty appoint-
ments in both the Design School and the School of Sustainable Engineering in the Built Environment at Arizona State University, Tempe, AZ.
William Bahnfleth is a professor and Director of the Indoor Environment Center in the Department of Architectural Engineering of The Penn-
sylvania State University, University Park, PA.
698
ASHRAE Transactions
calibration methodology that explicitly utilizes measured tuning a model for the same building used by Firrantello et al.
tracer gas data to improve model prediction accuracy. (2007b) and comparing results.
In these previous research efforts, CO2 testing was only
Previous Work on Multi-Zone Model Calibration used to determine model adequacy as suggested by the ASTM
D5157 metrics. Also, the results of these studies showed that
A specially developed version of CONTAM called the even after implementing the tuning algorithm, the models
Project Creation Wizard (PCW) has been developed by were still incorrect in predicting certain airflow directions and
Vandemusser (2007) which runs on the same simulation certain ASTM D5157 metrics remained unsatisfactory. There-
engine but differs from CONTAM in that it has a more “user fore, further improvements in model calibration methodology
friendly” graphical user interface and more built-in default were deemed necessary. Tracer gas data was not incorporated
model parameters that help reduce the time and effort needed into the calibration effort in these studies; hence, one of the
to develop an airflow model. PCW was also designed to help improvements suggested in this paper is to integrate the
with model calibration by integrating a “tuning” process collected tracer gas data into the calibration methodology.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
which accepts actual measured data and updates the model
based on those measurements to improve model accuracy. The Important Research Assumptions
calibration methodology integrated into PCW is based on the
tuning algorithm developed by Firrantello et al. (2007b) which The calibration methodologies developed in this paper
combines analytic and heuristic methods while utilizing build- and in the previous work are not universally applicable and are
ing site measurements (e.g., HVAC airflow magnitudes and subject to some significant research assumptions. Extrapola-
inter-zonal airflow directions) to improve model prediction tion of the developed methodology beyond these boundaries is
accuracy. Two metrics were used to evaluate model quality not recommended.
throughout the tuning process: the percentage of correctly
predicted airflow directions and the percentage of satisfactory 1. Inherent in the use of multi-zone models is the assump-
statistical metrics presented in ASTM Standard D5157 tion that all zones are under “well-mixed” conditions
(ASTM 2003). (i.e., a discrete set of state variables).
2. PCW assumes a simple AHU model for which the duct
The tuning algorithm (Firrantello et al., 2007b) was
distribution systems are represented as a single volume of
developed using two synthetic buildings and refined through
air. The details of the simple AHU model are discussed in
testing on two real buildings. It includes the following
the PCW user manual (Vandemusser 2007).
recommendations: (1) measure all possible interior airflow
directions and create a multi-zone model, (2) measure all 3. All zones use the PCW default air leakage relationships
bulk air handling unit (AHU) airflows, (3) identify all flow (i.e., all airflow path flow elements are one-way power
paths for which the simulation has incorrectly predicted flow law relationships with fixed coefficients and exponents).
direction, (4) identify the lowest unmeasured branch levels In a multi-zone model, airflow paths are building features
of HVAC flows associated with the incorrect airflow direc- which allow air travel between zones and the airflow
tions, (5) measure the associated HVAC airflows in those element is the mathematical relationship between the
areas, (6) revise the model based on these measurements (this magnitude of flow through an airflow path and the pres-
becomes the new “best guess” model), (7) evaluate model sure difference across the path. One-way flow using
quality and repeat starting at step (2) if unsatisfactory, and (8) power law models take the following form:
perform a regression analysis to estimate the difficult or
Q = k  P  n (1)
impossible to measure factors (e.g., average interior and
exterior leakage, shaft leakage, and terrain constant/coeffi- where Q is the volumetric airflow rate,  P is the pressure
cient). Firrantello et al. (2007b) found that the measurement difference across the flow path, k is the airflow coefficient
of bulk HVAC airflows resulted in the greatest improvements and n is the airflow exponent, which are referred to as
in model quality metrics. “leakage parameters”. In PCW, leakage severity can be
Sae Kow (2010) utilized the PCW program to implement, assigned automatically by the program as either “leaky”,
and further validate, the tuning algorithm developed by Firr- “normal”, or “tight”. These leakage parameter assign-
antello et al. (2005) and Firrantello et al. (2007b) using addi- ments are based on data surveyed from real buildings
tional measured data. The field tests were conducted on two (Persily 1998).
buildings on a university campus. Sae Kow (2010) evaluated 4. All simulations are performed as transient contaminant
the improvement in model development efficiency with PCW dispersion and steady-state airflow simulations. This
as compared to CONTAM, the performance of the PCW means that the building’s pressure distribution and the
model tuning process under different operating conditions, the resulting volumetric airflow rates through each flow
suitability for a model calibrated under one condition in path are calculated and remain fixed for one set of
predicting different conditions (e.g., open vs. closed doors), conditions (i.e., constant weather condition and
and the repeatability of PCW calibration by developing and constant air volume HVAC system operation) while
©2012 ASHRAE
699
contaminant concentrations are allowed to vary over model can be based on several metrics such as (a) the
time under these steady-state airflow conditions. percentage of correctly predicted airflow directions
5. A constant air volume (CAV) condition is assumed for the (which is arguable as a metric since the magnitude of the
HVAC system in all simulations. flow and not simply the flow direction influences indoor
6. All tracer gas releases are assumed to occur in an air air contaminant dispersion), or (b) ASTM Standard
handling unit. D5157 statistical metrics or some modified form thereof
which evaluates predicted vs. measured CO2 concentra-
GENERAL METHODOLOGY tions.
The specific intent of the proposed methodology is to
We note that the previously developed tuning algorithm
improve the previously developed tuning algorithm (Firran-
(Firrantello et al. 2007b) considered only one set of operating
tello et al. 2005, Firrantello et al. 2007b) by explicitly using
conditions (i.e., it did not account for varying ambient temper-
collected CO2 tracer gas data during the process of calibrating
atures, wind speeds, wind directions, and assumed leakage
multi-zone airflow models. The various steps of the proposed
parameters). Thus, a detailed calibrated model for only one set
methodology are summarized below:
of typical conditions may provide a false sense of precision,
1. Preliminary Model Tuning: Develop a “somewhat” real- and may not be accurate when extrapolated to different condi-
istic multi-zone model of the building by calibrating tions. Our first intent, therefore, was to determine the sensi-
based on the previously developed tuning algorithm tivity of the multi-zone model to varying ambient conditions
(Firrantello et al. 2007b). This step is necessary for and leakage assumptions. Next, the hypothesis was that model
enhancing the robustness of the subsequent calibration reduction, i.e. reducing the complexity of the model by group-
steps. ing rooms into “macro-zones”, would yield a more aggregated
2. Sensitivity Analysis: Evaluate whether the building model that may provide greater accuracy under varying condi-
airflow dynamics are climate or HVAC system domi- tions. With macro-zones identified, tracer gas (CO2) concen-
nated; if so, by how much. This sensitivity analysis is tration data could then be used to find aggregate (and realistic)
meant to identify the significant or important drivers of leakage parameters for each of the airflow paths of each
the system, and is needed to verify whether calibration macro-zone. Such an updated model would then provide more
performed under one set of operating and climatic factors accurate prediction of airflow behavior under the varying
still applies for other conditions. conditions.
3. Identify Macro-Zones for Model Reduction: Calibrating
DESCRIPTION OF SYNTHETIC AND REAL
multi-zone airflow models is a highly over-parameterized
BUILDINGS ANALYZED
problem. Model reduction is warranted and is achieved
by grouping rooms into “macro-zones” or clusters of Initially, the methodology steps were developed and
rooms with similar airflow and tracer gas dynamics under refined using a synthetic building (i.e., one whose dynamic
varying conditions. Macro-zone identification can be behavior can be entirely simulated on a computer and does not
achieved by performing a tracer gas-based sensitivity need any real building data). This is similar to the procedure
analysis. A preliminary hypothesis is that the great effort used by Firrantello et al. (2005) for the development of the
of performing factorial sensitivity tests with tracer gas tuning algorithm. The synthetic building used was a single
releases could be avoided by simply selecting the macro- story building that has 29 modeled zones (25 rooms, 4 corridor
zones based on room air change rates. spaces) and three air handling units serving three HVAC zones
4. Tracer Gas Release Tests: Perform tracer gas release tests (north perimeter, south perimeter, and core). Details of this
in the building being analyzed and place sensors in at building can be found in Snyder (2011).
least one room of each identified macro-zone. Some An actual building, namely the MBNA building, was also --`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
amount of replication is strongly advised. selected for which measured CO2 data, HVAC airflow rates,
5. Model Tuning and Calibration: Tune the flow parameters and airflow directions have been collected by Sae Kow (2010).
of the multi-zone models to improve the match between The PCW model of this building that was tuned by Sae Kow
measured and predicted tracer gas concentration dynam- (2010) using the previously developed tuning algorithm (Firr-
ics in each macro-zone. If flow parameter tuning is antello et al. 2007b) served as the starting point for our anal-
unsuccessful, investigate factors that significantly influ- ysis. The MBNA building is a three story, 44,000 ft2 [4090 m2]
ence the room air change rates (i.e., HVAC airflow rates, office and administration building which is served by three
room and system volumes, outside air percentage, etc.). variable air volume (VAV) air handling units. Two of the units
6. Evaluate Model Adequacy: Evaluate the adequacy of the (AHU1 and AHU2) are located in an unconditioned basement
updated model based on some metric. Since a robust and the third (AHU3) is located in a mechanical room on the
metric for model adequacy has yet to be determined, and third floor of the building. The evaluation of our methodology
since previous research has questioned the confidence in on the MBNA building focused exclusively on just the third
conclusions from any one metric, the evaluation of the floor since both Firrantello (2007a) and Sae Kow (2010) found
700
ASHRAE Transactions
that releases in AHU3 resulted in no cross contamination to Table 1. Synthetic Building Sensitivity
the rest of the MBNA building. Similarly, releases in either Analysis Summary
AHU1 or AHU2 resulted in no cross contamination on the
third floor. Therefore, for releases in AHU3, the third floor can Wind Speed Ambient Temperature (°F) [°C]
be considered to be an isolated building. Also, the proposed (mph) [m/s] 0 [-18] 70 [21] 100 [38]
methodology requires room-level CO2 measurements and the
data made available by Sae Kow (2010) only included room Leaky Leaky Leaky
level measurements on the third floor of the MBNA building.
0 [0] Average Average Average
It is important to note that there are a number of difficulties in
measuring and interpreting the use of tracer gas tests in build- Tight Tight Tight
ings as described in several publications (for example, Turk et
Leaky Leaky Leaky
al., 1989).
5 [2.2] Average Average Average
AIRFLOW-BASED SENSITIVITY ANALYSIS Tight Tight Tight
The objectives of an airflow-based sensitivity analysis Leaky Leaky Leaky
are: (1) to evaluate whether the airflow dynamics of the build-
10 [4.5] Average Average Average
ing are climate or HVAC dominated, (2) identify the signifi-
cant or important drivers of the system, and (3) assist in Tight Tight Tight
reducing the model complexity by identifying macro-zones.
Specifically, this analysis determines the sensitivity of the
building’s airflow dynamics to changing climate conditions Table 2. Subsequent Simulations for Wind Effect
and building leakage assumptions (i.e., ambient air tempera-
Wind Direction Leakage Severity
ture, wind direction, wind velocity, and building leakage
severity). Since the other main influential driver is HVAC North West Leaky Tight
system airflow rates, low sensitivity to these climate and leak-
age factors would indicate HVAC system dominance of West Leaky Tight
airflow dynamics. Ambient temperature = 70°F [21°C]
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
The airflow-based sensitivity analysis involves several Wind Speed = 10 mph [4.5 m/s]
steps. First, possible influential drivers (wind speed, wind
direction, ambient temperature, envelop leakage coefficients) Indoor Temperature = 72°F [22°C]
of the airflow system are selected and representative nominal
ranges are identified for each. Next, a full factorial experimen-
summarized in Table 2, were performed to analyze the impact
tal design is used to set up the simulation runs. A 2k factorial
of wind direction.
analysis is used assuming that the relationships are linear.
Then, a semi-quantitative airflow-based sensitivity analysis is After analyzing several methods of plotting the simula-
performed based on the magnitude and direction of airflow tion data, it was discovered that scatter plots provided the best
through each airflow path. In the absence of a robust response visual representations of the dynamic airflow behavior. Each
metric and due to the deterministic nature of multi-zone scatter plot compares the airflow magnitudes and directions of
models (which limits the ability to perform statistical signifi- two different scenarios (one scenario per axis) with only one
cance tests), a formal sensitivity analysis where main and different condition. The unit on each axis is volumetric flow
interaction effects are calculated has not been undertaken. rate in cubic feet per minute (CFM). The direction of airflow
Rather, graphical representations, especially scatter plots, are is accounted for in the sign of the flow rate. Each point in the
used to distinguish whether the building is HVAC dominated scatter plot represents flow through a single airflow path.
or climate dominated. Further, one can also determine whether These scatter plots allow for a simple visualization of how
the airflow-based sensitivity analysis is able to provide varying one condition impacts the magnitude and direction of
insights that would help to identify macro-zones. the airflows. It was at this stage that it was determined a 2k
factorial analysis would suffice since the relationships
The initial sensitivity analysis, as summarized in Table 1,
between the factor levels and the effects appear to be linear.
performed on the synthetic building, included a 33 full-factorial
experimental design (27 PCW simulations) to determine the Within these scatter plots, points which fall in the posi-
effects of varying ambient temperature, wind speed, and build- tive/positive and negative/negative quadrants imply that the
ing leakage severity on the direction and magnitude of airflow altered condition did not result in a change in the airflow
through all airflow paths. Common temperature and wind direction in that airflow path. On the other hand, points that
conditions were used along with the default building leakage fall within the positive/negative and negative/positive quad-
severity conditions in PCW. Four subsequent simulations, rants show that the altered condition did change the airflow
©2012 ASHRAE
701
direction in that airflow path. Points that exhibit a “y = x” rela- involved a 24 full factorial experimental design (16 PCW
tionship indicate that the altered condition did not change the simulations) to analyze the impacts of wind speed, wind direc-
magnitude of airflow through that airflow path. Points that tion, ambient air temperature, and leakage severity. A typical
fall off of this line indicate that the altered condition did meteorological year (TMY) weather file was used to deter-
change the airflow magnitude. Therefore, scatter plots show- mine typical weather conditions near the building. Similar to
ing a “y = x” relationship signify HVAC dominance of airflow the results of the synthetic building, it was found that these
dynamics. external drivers had negligible influence under “normal” and
Figure 1(a) provides an example scatter plot of two simu- “tight” conditions (see Snyder 2011 for details on this real
lations from Table 1 showing the effect of varying ambient air building analysis). Therefore, these conclusions reiterate the
temperature. Figure 1(b) provides an example of the effect of fact that buildings of this size which are mechanically venti-
varying wind direction. The effect of air temperature appears lated are likely to have airflow dynamics dominated by the
to be small with little airflow magnitude and direction HVAC system.
changes. The change in wind direction, however, was found to
alter the direction of flow through many airflow paths; this is IDENTIFY MACRO-ZONES FOR MODEL
concluded from the points in the negative/positive and posi- REDUCTION
tive/negative quadrants of Figure 1(b). After analyzing all A key concept of this paper was the recognition that cali-
scatter plots, it was found that wind speed and wind direction brating a multi-zone airflow model such as CONTAM or PCW
both had significant influences on airflow dynamics. This with numerous parameters specified by the user is a highly
significance, however, is a function of building leakage sever- over-parameterized problem. This is a result of having only a
ity (i.e., the effect is larger when the leakage severity is “leak- few data points and many parameters to tune. When faced with
ier”). Most of the scatter plots for this analysis indicated such problems, the traditional approach is to reduce the order
HVAC dominance. of the model and then perform system identification. In this
Since the synthetic building analysis revealed that a 2k case, the macro-zone approach offers such a reduction. We
factorial is more appropriate and that wind direction is also an identify macro-zones, i.e. groups of rooms which have similar
important factor, the analysis on the real building (MBNA) airflow dynamics under varying conditions, and then identify
(a)
(b)
Figure 1 Airflow-based sensitivity analysis scatter plots for synthetic building: (a) effect of ambient air temperature and
(b) effect of wind direction under “leaky” conditions.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
702
ASHRAE Transactions
aggregate flow parameters of these macro-zone flow elements released corresponds to the same procedure used by Sae Kow
which can be used to update the model. The approach adopted (2010). This was specifically modeled in PCW by specifying
was to identify macro-zones based on the results of a tracer an initial concentration of 2000 PPM in each room served by
gas-based sensitivity analysis (i.e., analyzing CO2 decay the AHU where the release occurred. Thus, the HVAC system
curves, decay coefficients, and peak concentrations). volume is initially ignored, and the CO2 curves show only the
Macro-zone formation based on tracer-gas simulations decay dynamics and not that during the uptake of CO2.
can be conducted in several ways. First, CO2 decay curves can Several methods of plotting the resulting tracer gas
be plotted for each room. Then, macro-zones can be identified release data were explored. It was found that CO2 concentra-
by visually observing the decay rates and peak concentrations tion curves for individual rooms were the most promising for
in each room under varying conditions. Visually identifying analyzing tracer gas dynamics and for identifying macro-
macro-zones, however, can be daunting if many rooms are to zones. Since there were several AHU’s where releases could
be considered at the same time. An alternative method is to use occur, it became apparent that the release location has a signif-
natural log transformation of the CO2 concentration data icant impact on which rooms are affected by the contaminant.
thereby linearizing the exponential decay curves for each The tracer gas-based sensitivity analysis, summarized in
room and then perform linear regression to identify the decay Table 3, were consistent among themselves, i.e., similar room
coefficient (which is the slope of the transformed decay line) groupings were obtained for all the combinatorial experimen-
for each room. Macro-zones can then be found by identifying tal design runs under all conditions, indicating HVAC domi-
rooms with similar coefficients under varying conditions. The nance of tracer gas behavior. This agrees with the results from
peak concentrations in each room should also be compared. the airflow-based sensitivity analysis. As discussed previ-
Unfortunately, this approach was not found to be suitable ously, wind speed and wind direction could have significant
because of higher order exponential decay dynamics (most influence, especially under “leaky” conditions. However, only
likely due to return air reintroducing CO2 into the supply air slight variations in decay rates and small amounts of cross
stream). Therefore, a natural log transformation did not result contamination between zones were observed due to these
in linear transformation of the data from which a decay coef- changing conditions. Whether these slight variations are
ficient could be identified. significant or not would depend on the circumstances under
which calibration is being performed, the type of contaminant
Macro-Zone Identification for Synthetic Building being analyzed, and the resulting difference in occupant expo-
sure. Such factors are to be specified by the analyst at the onset
Using PCW, simulations were performed where CO2 was of the calibration process depending on the circumstance and
released in each AHU. The synthetic building has three air the criticality of the consequence.
handling units, one serving the north perimeter zones, one By analyzing the tracer gas curves, the synthetic building
serving the south perimeter zones, and one serving the core was reduced from 29 model zones to 9 macro-zones consisting
zones. For these simulations, CO2 concentration decay data of rooms with similar behavior under varying conditions. The
was obtained for each room. Table 3 provides a summary of all macro-zones identified did not change under the various
of the tracer gas releases simulated. The amount of tracer gas conditions of both the airflow-based and tracer gas-based
sensitivity analyses. This confirms that the airflow dynamics
Table 3. Synthetic Building Tracer Gas Simulation are HVAC dominated. Wind had the most impact resulting in
Summary slight variations in tracer gas behavior as well as cross contam-
ination between zones. However, the significance of these
Release Location Leakage Severity Wind Direction slight variations in decay curves needs to be evaluated by the
analyst under the specific circumstances. The particular
North AHU Average North assumptions used for simulating the release in the synthetic
building did not account for system volume, and thus all rooms
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
South AHU Average North

have the same peak concentrations. Therefore, only the decay
Core AHU Average North rates are analyzed. Consequently, it was concluded that peak
129 Exit Average North concentrations should also be considered for macro-zone
identification by accounting for system volumes.
North AHU Leaky West
South AHU Leaky West Macro-Zone Identification for Real Building
North AHU Leaky South Using CO2 tracer gas simulations to identify macro-
zones was undertaken by simulating a release in AHU3 and
South AHU Leaky South plotting CO2 concentration curves for each room on the third
Ambient Temperature = 70°F [21°C] floor of the MBNA building. With three variables (wind
speed, wind direction, and leakage severity) at two factor
Wind Speed = 5 mph [2.2 m/s] levels each, a 23 factorial analysis was used to set up these
©2012 ASHRAE
703
Table 4. Real Building (MBNA) Tracer Gas CO2 simulations as summarized in Table 4. Ambient temper-
Simulation Summary ature was ignored since it was determined to be insignificant
from the airflow-based sensitivity analysis. Again, instead of
Wind Speed
Wind Direction Leakage Severity
performing a formal quantitative sensitivity analysis and
(mph) [m/s] calculating effects, the impact of changing conditions on
apparent macro-zones were visually determined.
10 [4.5] N Leaky
Figure 2 shows the individual CO2 concentration curves
10 [4.5] NW Tight for each room on the third floor of the MBNA building for the
10 [4.5] N Leaky first set of conditions described in Table 4 (i.e., 10 mph [4.5 m/
s] north wind and “Leaky”). Clearly, there is a wide range of
10 [4.5] NW Tight responses to one release in the AHU. For the synthetic build-
ing, an initial CO2 concentration was specified in each room
0 [0] N Leaky
ignoring the supply duct volume. Therefore, all rooms are
0 [0] NW Tight initially at the same concentration and all that was observed
was the concentration decay with time. Here, however, with a
0 [0] N Leaky release more realistically simulated in the supply duct volume,
0 [0] NW Tight the uptake and peak are seen along with the decay of CO2.
The overall goal of these tracer gas simulations was to be
Indoor Air Temperature = 72°F [22°C]
able to identify macro-zones that do not change under varying
Ambient Air Temperature = 67°F [19°C] conditions; however, the complexity of Figure 2 does not
allow for such classification. With so many concentration
Figure 2 Simulated CO2 concentration curves for individual rooms of the MBNA building for one tracer gas-based
combinatorial run done during sensitivity analysis.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
704
ASHRAE Transactions
curves on one plot it is too difficult to identify individual ing and the synthetic building indicated that the building’s
rooms or overlapping curves. Properly identifying macro- airflow dynamics are largely HVAC dominated, these newly
zones requires closer inspection of these graphs. Therefore, formed macro-zones were compared to the room air change
the tracer gas-based sensitivity analysis was performed by rates calculated from the HVAC airflows measured by Sae Kow
plotting concentration curves for each condition in Table 4 for (2010). Table 5 shows that the macro-zones identified are
individual rooms. Figure 3 shows an example of these tracer directly related to the air changes per hour in each space. Conse-
gas sensitivity plots for Room 1 of the MBNA 3rd floor. The quently, from the definition of macro-zones, the tracer-gas
slight variations observed lead us to conclude that wind direc- behavior in these rooms is directly influenced by room air
tion and wind speed have little impacts on the concentration change rates. Despite the apparent HVAC dominance, climate
curves. However, leakage severity does have a large impact conditions and leakage severity also have impacts. For each
when there is wind. This is indicated by the faster decay in the condition, the slight variations in peak concentrations and decay
rooms that are “leaky” when there is high wind. The concen- rates, as discussed previously, from wind changes and leakage
tration curves for “tight” leakage severity and high wind are severity changes, are not large enough to shift around macro-
almost identical to when there is no wind and for “leaky” or zones as identified by the air change rates. However, the signif-
“tight” rooms. Even closer inspection of this graph revealed icance of such variations may depend on the release scenario
that wind in any condition increases the rate of decay of the and the type of contaminant.
tracer gas. This can be attributed to higher air change rates
resulting from wind induced infiltration. Similar conclusions MODEL CALIBRATION USING CO2 DATA
were drawn from the sensitivity plots of the other rooms.
With macro-zones identified, the next step is to use actual
This tracer gas-based sensitivity analysis provided some measured CO2 data to calibrate the model. To collect tracer gas
useful insights into the building’s airflow dynamics under data, CO2 releases need to be performed in the building and
various conditions. Wind speed and wind direction changes CO2 sensors need to be placed throughout the building to
resulted in slight variations in the decay rate in rooms with record CO2 concentrations. The identification of macro-zones
exterior walls impacted by the wind and with “leaky” condi- assists in determining where it would be advantageous to place
tions. However, as noted with the synthetic building, changing CO2 sensors. Ideally, there should be a CO2 sensor in at least
climatic conditions mostly yielded insignificant differences in one room of each macro-zone. The procedures and equipment
tracer gas behavior. The results also showed that the decay
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
needed to perform such tracer gas tests are described by Firr-
rates and peak concentrations for individual rooms were antello (2007a) and Sae Kow (2010). Next, new “aggregate”
almost identical under all conditions. Therefore, macro-zones flow parameters (flow coefficients and exponents), one set for
could be formed by visually comparing peak concentrations exterior flow paths and one set for interior flow paths, are esti-
and decay rates for various graphs under one set of conditions. mated for each of the identified macro-zones. Since all rooms
The 33 rooms of the third floor of the MBNA building were in each macro-zone have similar airflow dynamics, one can
reduced to 8 macro-zones by identifying rooms with similar tune the flow parameters for each flow path in a macro-zone
tracer gas behavior. Figure 4 is an example of CO2 concentra- by the same amount. These aggregated flow parameters
tion curves for each room in two of the identified macro-zones. simplify the calibration process, reduce the number of param-
Note that the tracer gas behavior is essentially identical for each eters that need to be calibrated, and hopefully provide aggre-
room within each macro-zone. Since results from the real build- gate leakage parameter values that are more representative of
Figure 3 Room 1 CO2 tracer gas-based sensitivity analysis meant to illustrate the fact that the dynamics are close under
different combinatorial runs.
©2012 ASHRAE
705
Figure 4 CO2 concentration curves for macro-zones 1 and 2.
Table 5. Observed Macro-Zones for MBNA Building match between predicted and measured CO2 concentration
and Measured Room Air Changes curves for each macro-zone. The quality of the match between
the two curves can be evaluated visually or by some statistical
Identified Macro-Zones measure, i.e. root mean square error. The factorial analysis
will evaluate the effect of increasing and decreasing flow
1 2 3 4 5 6 7 8
parameters in each macro-zone on how well the predicted CO2
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Air Changes (1/hr) concentrations match the measured data. If significant differ-
ences between predicted and measured concentrations remain
5.9 5.3 8.2 7.8 15.4 10.7 3.6 0.0 after tuning the flow parameters, then possible measurement
6.1 4.7 8.9 7.6 0.0 errors and uncertainties for factors impacting room air
changes (i.e., HVAC airflow rates, room volumes, and outside
6.3 4.8 9.0 6.8 0.0 air percentages) should be investigated. Model calibration for
5.6 5.2 8.6 6.7 2.7 the synthetic building was deemed too arbitrary and was there-
fore only performed on the real building.
6.3 4.7 8.6 1.4
As mentioned above, this methodology identifies rooms
5.7 8.2 1.9 where it would be advantageous to place CO2 sensors during
testing (i.e., at least one sensor per macro-zone). However, the
5.8 9.1
macro-zones in this research were identified after the testing
5.7 had been performed by Sae Kow (2010). Therefore, there was
no guarantee that the sensor locations would match up with
identified macro-zones. The seven CO2 measurement loca-
varying conditions. There are several ways by which this could tion rooms used matched up with four of the macro-zones
be done. The method explored in this research was to set up a identified in this analysis. Thus, we were constrained to cali-
factorial sensitivity analysis where the response variable is the brating the rooms associated with these four macro-zones
706
ASHRAE Transactions
using the data collected by Sae Kow (2010). As a baseline, and the consequences of the well-mixed assumption. For the
graphs of the measured CO2 curves were plotted versus the factorial analysis to tune leakage parameters, depending on
predicted curves based on the final tuned model developed by the relationship between the predicted and measured curves,
Sae Kow (2010). Figure 5 (a) and (b) show example plots of one can determine the direction in which to change these
measured vs. predicted CO2 concentration curves for two parameters to reduce the number of simulations. This method
rooms on the third floor of the MBNA building. The graphs tunes one macro-zone at a time and does not provide a proce-
indicate that even after using the previously developed tuning dure for simultaneous calibration of all macro-zones. There-
algorithm, there are still some discrepancies between the fore there is some ambiguity in how to tune the flow
measured and predicted tracer gas behavior. Assuming the parameters for airflow paths which connect separate macro-
HVAC airflow rates were measured accurately, it was initially zones.
hypothesized that these differences must be attributed to Despite the lack of CO2 data needed to calibrate the entire
airflow into and out of the room via the exterior envelop or third floor of the MBNA building, one of the identified macro-
inter-zonal airflow paths. Consequently, we attempted to zones was analyzed as an example of model calibration. The
adjust the flow parameters (coefficients and exponents) for internal and external flow coefficients were varied ±15% and
these airflow paths. By first identifying macro-zones, the ±30%. After initial simulations, it became apparent that the
complexity of the model has been reduced and locations for leakage needed to be increased in order to reconcile measured
CO2 measurement have been identified. Reducing the and predicted behavior. Thus, simulations were performed
complexity of the model also reduces the complexity of these with the flow coefficient changed by +50% and +75%. A flow
flow parameter adjustments. The macro-zones identify coefficient increase beyond 75% would represent an unrealis-
groups of flow paths whose parameters can be changed by the tic leakage coefficient based on the data gathered from real
same amount. Other sources of discrepancies between buildings (Persily 1998). The flow exponent was varied from
measured and predicted behavior could include uncertainties the default of 0.65 to 0.5 and 0.7. This represents the range of
or errors in HVAC airflow rate and CO2 concentration typical flow exponents recommended by Walton and Dols
measurements, errors in the development of the PCW model, (2008). Figure 6 below shows that despite all of these changes
(a)
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
(b)
Figure 5 Measured vs. predicted CO2 curves for (a) Room 1 and (b) Room 15 of the MBNA building based on the previously
developed tuning algorithm.
©2012 ASHRAE
707
to the flow coefficient, there is no noticeable difference in the It was found that the tracer gas behavior in each room was
predicted concentrations for the measurement location of this much more sensitive to factors impacting air changes (i.e.,
macro-zone. The leakage exponent did have a more noticeable room volume and outside air percentage) than to changes in
impact in shifting the entire curve. Changing to the flow expo- flow parameters. Figure 7 shows the impact of changing room
nent to 0.5 provides a better match for the decay part of the volume and Figure 8 shows the impact of changing outside air
curve. However, the peak concentration is still far too high. percentage on the match between the measured and predicted
These changes fail to adequately provide a better match CO2 curves for the measurement location in macro-zone 3.
between the predicted curve to the measured curve. Increasing or decreasing room volume results in a decrease or
increase, respectively, in the peak concentration with a negli-
With the flow parameters having little impact, other gible change in the concentration decay rate. In Figure 7, it is
possible sources of the discrepancies between measured and clear that changing the ceiling height from 8 ft – 6 in [2.6 m]
predicted tracer gas curves could include incorrect supply or to 10 ft [3 m] provides a much better match between the
return duct volume estimates, room volumes, HVAC flow predicted and measured peak concentrations. The room
rates, outside air percentages, the well-mixed assumptions, volumes in the model were taken from architectural plans.
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
etc. Sae Kow (2010) discusses the impact of system volume However, when creating a multi-zone model, each zone is a
estimates on the predicted tracer gas curves. An increase in the volume of air. Therefore, the ceiling height may not be the best
system volume results in a decrease in peak concentrations. value to use when defining the zone. If the room has an acces-
However, not all measured locations have higher peak concen- sible ceiling with acoustic tile there may be significant air
trations than predicted. Therefore, rather than adjusting the transfer to the plenum space above which may need to be
system volume, there may be errors in individual room accounted for in the model.
volumes which were estimated from floor plans. Also, Sae In the field testing performed on the MBNA building, Sae
Kow (2010) notes that uncertainties in the measurement of the Kow (2010) noted difficulty in accurately measuring the
outside air percentage for AHU3 were very large since the air outside air percentage for AHU3. Due to physical constraints,
temperature fraction method was used. It is therefore likely the only practical option for measuring the outside air percent-
that the outside air percentage is incorrect. The well-mixed age was to use the air temperature fraction method. The
assumption of the multi-zone model may also be a significant outdoor air, return air, and mixed air temperatures were
source of error. Therefore, it is possible that the CO2 sensor measured using hot wire anemometers. However, the calcu-
was placed in a location where it recorded concentrations that lated outside air percentage based on these measurements had
are not representative of the room average. a very large uncertainty (approximately 70%). This was due to
Figure 6 Leakage parameter adjustments for the calibration of macro-zone 3.
708
ASHRAE Transactions
Figure 7 Impact of altering room volume on the match between measured and predicted CO2 concentration curves for macro-
zone 3.
Figure 8 Impact of altering outside air percentage on the match between measured and predicted CO2 concentration curves
for macro-zone 3.
the small temperature difference between the outside air and CO2 concentrations in buildings whose airflow dynamics are
the return air. Therefore, it is fair to assume that the outside air dominated by HVAC system airflow rates, one should focus on
percentage of 20% initially used in the model is not accurate. those factors which have significant impacts on room air
Figure 8 shows that changing the outdoor air fraction to 40% change rates first and model flow parameters second.
provides a much better match in the decay rate (justified since
the OA fraction measurement had a high measurement uncer- SUMMARY AND CONCLUSIONS
tainty). Therefore, accurate outside air measurement is signif- The advantages and insights provided by the proposed
icant. The outside air percentage, however, did not yield an methodology are listed below:
improvement in matching the peak concentration. Although The concept of macro-zone identification provides a
these changes are somewhat arbitrary, they help to illustrate robust manner of calibrating a complex model consistent
that in rectifying differences between measured and predicted with experimental data. Macro-zones reduce the number of
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
©2012 ASHRAE
709
individual flow parameters that need to be tuned during cali- nical support of A. Musser was also of great assistance during
bration. this work.
This paper provided a scientific means of identifying sets
of rooms or “macro-zones” which have similar airflow REFERENCES
dynamics and tracer gas behavior. Therefore, occupants in ASTM. 2003. D5157-97: Standard Guide for Statistical
these rooms are likely to be exposed in the same manner when Evaluation of Indoor Air Quality Models. The American
a contaminant release occurs. Society for Testing and Materials, Philadelphia, PA.
The identification of these “macro-zones” takes into
Bahnfleth, W.P., Sae Kow, P., Firrantello, J., and P.A. Kre-
account the effect of different climatic conditions and building
mer. 2012. Comparison of semi-quantitative and formal
flow characteristics (as against previous work which only
metrics from multi-zone airflow model quality assess-
considered one set of conditions).
ment. HVAC&R Research, 18(1-2): 252-263. American
The methodology proposed allows insights into experi- Society for Heating, Refrigerating, and Air Condition-
mental design; namely, suggests rooms where it would be ing Engineers, Atlanta, GA.
advantageous to place CO2 sensors and monitor concentration
Firrantello, J., Bahnfleth, W.P., Musser, A., Freihaut, J., and
dynamics during testing.
J. Jeong. 2005. Application of Sensitivity Analysis to
It also allows calibrating flow parameters for different
Multizone Airflow Model Tuning. Proceedings of the
macro-zones based on measured CO2 concentration data.
8th REHVA World Congress, CLIMA 2005, Lausanne,
Finally, it facilitates understanding of building airflow Switzerland.
behavior and tracer gas behavior under varying conditions via
Firrantello, J. 2007a. Development of a Rapid, Data Driven
experimental design techniques.
Method for Tuning Multizone Airflow Models. Master of
This methodology was implemented on a synthetic build-
Science Thesis, Department of Architectural Engineer-
ing as well as a real building for which previous research
ing, The Pennsylvania State University, State College,
efforts obtained measured tracer gas and airflow data. Various
PA.
graphical methods of analyzing simulation data were explored
Firrantello, J., Bahnfleth W., Jeong, J., Musser, M., and J.
(as described in Snyder 2011). CO2 concentration curves and
Freihaut. 2007b. Field Testing of a Data Driven Multi-
scatter plots proved to be the most useful in helping to under-
zone Model Calibration Procedure. Proceedings of the
stand building airflow dynamics and tracer gas behavior.
10th REHVA World Congress, CLIMA 2007, Helsinki,
Macro-zoning helps to simplify the model as well as identify
Finland.
locations where it would be advantageous to place tracer gas
sensors during testing. Both of the buildings analyzed where Persily, A.K. 1998. Airtightness of Commercial and Institu-
found to have airflow dynamics dominated by HVAC system tional Buildings: Blowing Holes in the Myth of Tight
airflow rates. Similarly, the identified macro-zones were Buildings. DOE/ASHRAE/ORNL/BETEC/NRCC/CIBSE
directly related to rooms with similar air change rates. Conference Thermal Performance of the Exterior Enve-
Altering the flow parameters in the airflow paths of lopes of Buildings VII, pp. 829-837.
macro-zones during calibration did not seem to significantly Sae Kow, P. 2010. Field Verification of a Semi-Empirical
improve the match between measured and predicted tracer gas Multizone Airflow Modeling Calibration Method. Mas-
curves. It was hypothesized that after all of the HVAC flows ter of Science Thesis, Department of Architectural Engi-
had been measured and entered into the model, the remaining neering, The Pennsylvania State University, State
discrepancies between predicted and measured curves would College, PA.
be the result of incorrect flow parameters in the models Snyder, S., 2011. Investigation of CO2 Tracer Gas-Based
describing the flow through the building envelope and Calibration of Multi-Zone Airflow Models, Master of
between internal zones. The fact that altering the flow param- Science Thesis, The Design School, Arizona State Uni-
eters did not appear to improve model prediction accuracy versity, Tempe, AZ, April.
may indicate errors either in measurements or in model devel- Turk, B.H., Grimsrud, D.T., Brown, J.T., Geisling-Sobotka,
opment. Due to the fact that airflow dynamics were found to K.L., Harrison, J. and R.J. Prill. 1989. Commercial
be HVAC dominated and since the identified macro-zones are Building Ventilation Rates and Particle Concentrations.
directly related to room air change rates, it is crucial that all ASHRAE Transactions, 95(1): 422-433. American Soci-
factors impacting room air changes be measured accurately. ety for Heating, Refrigerating, and Air Conditioning
Engineers, Atlanta, GA.
ACKNOWLEDGMENTS Vandemusser Design, LLC. 2007. PCW 1.0 Software User
This research was funded by the Technical Support Work- Manual. Vandemusser Design, LLC under subcontract
ing Group—Combating Terrorism Technical Support Office. to The Pennsylvania State University, State College, PA.
The authors acknowledge invaluable assistance of J. Firran- Walton, G.N., and W.S. Dols. 2008. CONTAM, Version 2.4c.
tello and P. Sae Kow for sharing the relevant data and software National Institute of Standards and Technology, Gaith-
files which were the starting point of this research. The tech- ersburg, MD.

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---

SOCIETY BUSINESS

June 23–27, 2012
San Antonio, Texas
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---

--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---

2011–2012 ASHRAE OFFICERS, DIRECTORS,
STAFF, AND COUNCIL AND COMMITTEE MEMBERS
Executive Committee Board of Directors Staff
Executive Vice President:

OFFICERS DIRECTORS Jeff H. Littleton
Executive Assistant to Board of Directors:
President July 2009–June 2012 Gloria F. Cofer
Spencer Morasch (I) Program Director of Government Affairs:
Ronald E. Jarnagin Douglas E. Read
Isabelle Lavoie, PEng
Manager of Government Affairs: Mark Ames
Guy Perreault, PE (II)
President-Elect Kevin W. Fallin, PE (III)
Manager of Communications: Jodi Scott
Director of Development: Bert Huffman
Thomas E. Watson, PE Erich Binder, PE (XI) Comptroller/Director of Administrative Services:
Michael J. Brandemuehl, PhD, PE* Cindy M. Simmons
Treasurer Manager, Accounting: Annmarie S. Wilhoit
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Samir R. Traboulsi, PE*
Thomas E. Werkema, Jr.* Manager, Human Resources: Sharon L. Priebe, PHR
William P. Bahnfleth, PhD, PE Manager, Information Technologies: Dana Suffes
Publisher/Director of Publications and Education:
July 2010-June 2013 W. Stephen Comstock
Vice Presidents T. Randall Jones (IV) Group Manager, Education and Certification:
Constantinos A. Balaras, PhD, PE R. Lee Millies, PE (V) Joyce Abrams
Tina M. Brueckner (VI) Manager, Professional Development: Karen Murray
William W. Malphus (XII) Group Manager, Publishing Services and Electronic
Ross D. Montgomery, PE Communications: David Soltis
Wichai Laksanakorn, PE (XIII)
Inventory and Subscription Manager, Kimberly Gates
T. David Underwood, PEng Hugh F. Crowther* Group Manager/Editor, Handbook
Eckhard A. Groll, PhD* and Special Publications: Mark Owen
Charles E. Gulledge III, PE* Manager, Electronic Communications
Timothy G. Wentz, PE and Applications: Joslyn Ratcliff
July 2011-June2014 Manager, Advertising Sales: Greg Martin
Secretary Jeff Gatlin, PE (VII) Editor, ASHRAE Journal: Fred P. Turner
John L. Harrod, PE (VIII) Director of Member Services: Carolyn Kettering
Jeff H. Littleton Manager, Membership and Customer Service:
Richard J. Bowman (IX) Ashley M. Pruett
Michael T. Burgess, PE (X) Manager, Region Activities: Vickie Grant
Ashok K. Virmani (RAL) Manager, Chapter Programs: Rosy Douglas
Roberto R. Aguilo, PE Manager, Conference Services: Judy Marshall
Patricia T. Graef, PE Manager, Conference Programs: Anthony Giometti
Stephen D. Kennedy Manager, Fundraising: Patricia Adelmann
Director of Technology:
*Director-at-Large
Claire B. Ramspeck
Manager, Research and Technical Services:
Michael R. Vaughn, PE
Manager, Special Projects: Lilas Pratt
Manager, Standards: Stephanie Reiniche
Manager, Standards and Codes: Steve Ferguson
ASHRAE ORGANIZATIONAL CHART
ASHRAE Membership
Board of Directors
Members Technology Publishing and

Council Council Education Council
Chapter Technology Transfer Environmental Health Certification
Chapters Regional Refrigeration Electronic Communications
Conferences and Expositions Research Administration Handbook
Honors and Awards Standards Historical
Membership Promotion Technical Activities Professional Development
Research Promotion Publications
Student Activities
Young Engineers in ASHRAE
Committees not reporting to a council: Advocacy, Building Energy Quotient, Finance, Nominating, Planning, President-Elect Advisory, and Society Rules.
© 2012 ASHRAE
713
2011–2012 COUNCILS AND COMMITTEES
(BOD ExO = ex officio Board of Directors, CO = coordinating officer, SLI = staff liaison, NVM = nonvoting member, RAL = Region-at-Large, DAL = director-at-large)
Advocacy David E. Knebel (VIII) Kevin W. Fallin (III)

Thomas E. Watson, Chair Tim J. McGinn (XI) T. Randall Jones (IV)
William P. Bahnfleth, Vice Chair William E. Murphy (VII) R. Lee Millies, Jr. (V)
Steven H. Bernhardt Garry N. Myers (I) Tina M. Brueckner (VI)
Darryl K. Boyce M. Ginger Scoggins (IV) Jeff W. Gatlin (VII)
Richard A. Charles Edward K. C. Tsui (XIII) John L. Harrod (VIII)
Robert G. Craddock Gloria F. Cofer, SLI Richard J. Bowman (IX)
Harvey M. Sachs Michael T. Burgess (X)
Planning
Scott E. Wayland Stephen D. Kennedy, Chair Erich Binder (XI)
Thomas E. Watson, CO Jason Z. Alphonso William W. Malphus (XII)
Douglas E. Read, SLI Tina M. Brueckner Wichai Laksanakorn (XIII)
Sheila J. Hayter Ashok K. Virmani (RAL)
Building Energy Quotient
Thomas H. Phoenix, Chair Richard D. Hermans Nonvoting Members
Amy B. Musser, Vice Chair Ross D. Montgomery Malcolm D. Knight
T. Randall Jones Bjarne W. Olesen Russell J. Lavitt
Kenneth M. Fulk Harvey M. Sachs Elbert G. Phillips
Karen C. Thrasher Samir R. Traboulsi Joel C. Primeau
Hywel Davies Ross D. Montgomery, CO Matt C. Rowe
Harry P. Misuriello Jeff H. Littleton C. Brian Wandling
Daniel H. Nall W. Stephen Comstock, Staff Director Dennis J. Wessel
Constantinos A. Balaras, Consultant Gloria F. Cofer, SLI Carolyn Kettering, SLI
Martha Brook, Consultant President-Elect Advisory Chapter Technology Transfer
Charles N. Eley, Consultant Thomas E. Watson, Chair/CO Elbert G. Phillips, Chair
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Lilas K. Pratt, SLI William P. Bahnfleth, Vice Chair Farooq Mehboob, Vice Chair
Roberto R. Aguilo Wei Sun, Vice Chair
Finance
William P. Bahnfleth, Chair Michael J. Brandemuehl Steven L. Rosen (I)
Constantinos A. Balaras Hugh F. Crowther Jean-Gabriel Joannette (II)
Mark A. Tome (III)
Hugh F. Crowther Patricia T. Graef
Harris M. Sheinman (IV)
Richard D. Hermans Eckhard A. Groll
Robert J. Laneve (V)
David E. Knebel Charles E. Gulledge, III
Corey B. Metzger (VI)
Wayne R. Reedy Stephen D. Kennedy
Chad E. Moore (VII)
M. Ginger Scoggins Samir R. Traboulsi
Keith H. Reihl (VIII)
Timothy G. Wentz Thomas E. Werkema, Jr.
Thomas G. Lewis (IX)
Cindy M. Simmons, SLI Gloria F. Cofer, SLI
Scott E. Wayland (X)
Nominating Society Rules N. Eileen Jensen (XI)
Gordon V. R. Holness, Chair Patricia T. Graef, Chair John P. Pennington (XII)
Kent W. Peterson, Vice Chair Steven T. Bushby, Vice Chair Yong Kong Ng (XIII)
Emery Otruba (I) Michael J. Brandemuehl Ashish Rakheja (RAL)
Douglas C. Cochrane (II) Arthur L. Giesler Richard D. DesLauriers, Mbr/Prog. Coord.
Charles E. Henck (III) Max H. Sherman Andrew L. Cochrane, Mbr/Comm. Coord.
Caroline C. Calloway (IV) Jack H. Zarour Thomas E. Werkema, Jr., Cons/Gov’t Act.
Frederick W. Betz (V) T. David Underwood, CO Donald A. Siller, Cons/Refrigeration
Charles E. Dale-Derks (VI) Gloria F. Cofer, SLI R. Christopher Mathis, Cons/Energy
Michael R. Bilderbeck (VII) John T. Dieckmann, Cons/Technology
Florentino J. Mendez (VIII) MEMBERS COUNCIL Thomas E. Werkema, Jr., BOD ExO
D. Larry Simmon (IX) Thomas E. Watson, CO
Saul V. Villanueva (X) Thomas E. Watson, Chair/President-Elect Rosy Douglas, SLI
Norm Grusnick (XI) William P. Bahnfleth, Vice Chair/Treasurer Conferences and Expositions
Wade H. Conlan (XII) Stephen D. Kennedy, Mbr/DAL Dennis J. Wessel, Chair
Wei-Dong Wang (XIII) Samir R. Traboulsi, Mbr/DAL Ben A. Leppard, Jr., Vice Chair
Pankaj R. Dharkar (RAL) Thomas E. Werkema, Jr., Mbr/DAL Pamela L. Androff
Darryl K. Boyce (II) Spencer Morasch (I) Robert G. Baker
Traci A. Hanegan (XI) Isabelle Lavoie (II) Walid Chakroun
714
ASHRAE Transactions
Jon J. Cohen Research Promotion Michael M. Collarin (IV)
Wade H. Conlan M. Dennis Knight, Chair Steve N. Spentzas (VI)
Kelley P. Cramm Walter W. Law, Jr., Vice Chair Chris M. Gray (VII)
K. William Dean Jeff Clarke, Vice Chair Bo A. Twumasi (VIII)
Charles E. Henck John A. Rieke, Vice Chair Tim J. Hughes (IX)
Mohammad H. Hosni Darcy A. Carbone (I) Cynthia L. Hyatt (X)
Yunho Hwang Daryl E. Somers (II) Tariq A. Amlani (XI)
Julia A. Keen James M. Grant (III) Megan Tosh (XII)
Dunstan L. Macauley, III Will Thomason (IV) Dennis L. Alejandro (XIII)
Sarah E. Maston Derek A. Crowe (V) Paraskevi Koutsogianni (RAL)
Michael J. McDermott
Craig R. Johnson (VI) Roger K. Jones, Liaison/MP
Keith C. Newcomer
Edward A. Dusch (VII) Francis Lacharite, Liaison/SA
Robert B. Risley
Rhamy L. Morrison (VIII) Nicolas Lemire, Consultant
Steven L. Rosen
Gary L. Cooper (IX) Samir R. Traboulsi, BOD ExO
Monte G. Troutman
Tiffany Bates (X) Thomas E. Watson, CO
A. Damon Gowan, Joint Expo
Ray P. Sieber (XI) Rhiannon M. Loomis, SLI
Stephen D. Kennedy, BOD ExO
Richard J. Ellison (XII)
Thomas E. Watson, CO
Edward Ka Cheung Tsui (XIII)
Anthony Giometti, SLI PUBLISHING AND
Mohammed B. Anbari (RAL)
Honors and Awards EDUCATION COUNCIL
Jeff Clarke, ASHRAE Research Canada Rep.
C. Brian Wandling, Chair
Daniel R. Rogers, Consultant
Raymond E. Patenaude, Vice Chair T. David Underwood, Chair/VP
Thomas E. Werkema, Jr., BOD ExO
William J. Coad Timothy G. Wentz, Vice Chair/VP
Julian R. de Bullet Thomas E. Watson, CO
Roberto R. Aguilo, Mbr/DAL
Victor W. Goldschmidt Patricia A. Adelmann, SLI
Michael J. Brandemuehl, Mbr/DAL
Essam E. Khalil Student Activities Hugh F. Crowther, Mbr/DAL
Wilfred Laman Joel C.Primeau, Chair
Bruce Billedeaux
J. Thomas Sobieski John T. Moore, Vice Chair
Walid Chakroun
Marija S. Todorovic Om Taneja (I)
Kimball E. Ferguson
Vincent K. C. Tse Francis Lacharite (II)
Larry J. Fisher
Ronald P. Vallort Jennifer E. Leach (III)
Kenneth M. Fulk
Samir R. Traboulsi, BOD ExO Peggy J. Fritz (IV)
Tim J. McGinn
Thomas E. Watson, CO Douglas F. Zentz (V)
Rosy Douglas, SLI Richard M. Rose
B. Andrew Price (VI)
Membership Promotion Benjamin T. Womble (VII) Nonvoting Members
Russell J. Lavitt, Chair Aruna D. Abhayagoonawardhana (VIII) William S. Fleming
Roger K. Jones, Vice Chair Kevin L. Ainsworth (IX) George E. Menzies
Richard E.Vehlow (I) Joseph A. Chin (X) Amy B. Musser
Catherine E. Godin (II) Stefan Lidington (XI) John H. Nix II
Gary C. Debes (III) Rex E. Noble
Florentino Roson Rodriguez, Sr. (XII)
Bryan M. Holcomb (IV) Heather L. Platt
Sam C. M. Hui (XIII)
Douglas D. Fick (V) W. Stephen Comstock, SLI
Nohad I. Boudani (RAL)
Mark F. Miller (VI)
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
David B. Meredith, Accred Rep/ABET BOD Certification

Frank H. Schambach (VII)
Amir Karimi, Accred Rep/ABET EAC Amy B. Musser, Chair
Randy C. Schrecengost (VIII)
Larraine A. Kapka, Accred Rep/ABET TAC David B. Meredith, Vice Chair
Gaylen V. Atkinson (IX)
Stephen D. Kennedy, BOD ExO William K. Klock
Marites D. Calad (X)
Thomas E. Watson, CO Brian M. Lynch
Murdoch D. MacPherson (XI)
Tarra Holman, SLI Matt Nelson
Jason Z. Alphonso (XII)
Eric M. Panzo (XIII) Young Engineers in ASHRAE Chee Sheng Ow
Dimitrios A. Charalambopoulos (RAL) Matt C. Rowe, Chair Shanta L. Tucker
James R. Fields, Consultant Carrie R. Kelty, Vice Chair James W. Bochat, Consultant
Samir R. Traboulsi, BOD ExO Cara S. Martin (I) Hugh F. Crowther, BOD ExO
Thomas E. Watson, CO Aaron R. Smith (II) T. David Underwood, CO
Ashley M. Pruett, SLI Christopher G. Albright (III) Joyce Abrams, SLI
© 2012 ASHRAE
715
Electronic Communications Gary E. Grange
Heather L. Platt, Chair Gerald C. Groff TECHNOLOGY COUNCIL
Gildardo Yanez, Vice Chair Alexander Weiss
Ross D. Montgomery, Chair/VP
Robert Coleman J. Richard Wright
Constantinos A. Balaras, Vice Chair/VP
Mark M. Hydeman Bernard A. Nagengast, Consultant
Patricia T. Graef, Mbr/DAL
Angela M. Lewis Michael J. Brandemuehl, BOD ExO
Eckhard A. Groll, Mbr/DAL
Lawrence C. Markel T. David Underwood, CO
Charles E. Gulledge, III, Mbr/DAL
Benjamin A. Skelton Sarah Foster, SLI
Michael F. Beda
Adrienne G. Thomle
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Regional Historians
Hugh F. Crowther, BOD ExO Donald M. Brundage
Alexander Weiss (I)
T. David Underwood, CO Arthur D. Hallstrom
George E. Menzies (II)
Joslyn Ratcliff, SLI Lawrence C. Markel
Renison W. Tisdale, Jr. (III)
Daniel C. Pettway
Handbook Pamela M. Immekus (IV)
Bodh R. Subherwal
Rex E. Noble, Chair Thomas J. Ferdelman (V)
Karen C. Thrasher
Cynthia A. Callaway, Vice Chair Paul K. Melton (VI)
Roberto R. Aguilo, BOD ExO Nonvoting Members
Thomas E. Hodge (VII)
T. David Underwood, CO Kenneth W. Cooper
Randy C. Schrecengost (VIII)
Mark Owen, SLI Gary E. Grange (IX) Charles H. Culp, III
Hugh I. Henderson, Jr.
Subcommittee for 2012 Handbook Clayton A. Lampman (X)
—HVAC Systems and Equipment K. William Dean (XI) Jaap J. Hogeling
Cynthia A. Callaway, Subcommittee Chair Cesar Luis D.L. Lim
Oswaldo D. S. Bueno (XII)
David A. John Carol E. Marriott
Low Han Guan Den (XIII)
Thomas M. Lawrence William F. McQuade
Timothy C. Dwyer (RAL)
Russell E. Lindemann Douglas C. Scott
Harris M. Sheinman Professional Development
Chandra Sekhar
John H. Nix II, Chair
Forrest S. Yount Jianshun S. Zhang
Thomas H. Phoenix, Vice Chair
Subcommittee for 2013 Handbook Cynthia A. Cogil, Vice Chair
Claire B. Ramspeck, SLI
—Fundamentals
Hassan M. Bagheri, Subcommittee Chair
Donald L. Brandt Environmental Health
Jorge E. Donoso Jianshun S. Zhang, Chair/Past RAC
James D. Aswegan
Nathan P. Hart Chandra Sekhar, Vice Chair
Jill A. Connell
Cameron R. Labunski Leon E. Alevantis
Peter Simmonds
Hugh D. McMillan III Wane A. Baker
Jeff J. Traylor
Kirk T. Mescher William J. Fisk
David P. Yuill
Darin W. Nutter Roger L. Hedrick
Subcommittee for 2014 Handbook Saunders C. Smith Mark C. Jackson
—Refrigeration
Lawrence B. Stoff Yuguo Li
Daniel J. Dettmers, Subcommittee Chair
Donald L. Fenton Michael J. Brandemuehl, BOD ExO Zuraimi M. Sultan
Teddy S. Hansen T. David Underwood, CO Pawel Wargocki
Frederick A. Lorch Joyce Abrams, SLI Andrew K. Persily, Mbr/Past Society Officer
Lewis Radonovich, Jr., Mbr/Health Profess.
Ramon Pons Publications
Phillip M. Trafton William S. Fleming, Chair David T. Grimsrud, Mbr/Past Standards
Francis A. Mills, Vice Chair Jan Sundell, Mbr/Physician
Subcommittee for 2015 Handbook
—HVAC Applications George W. Austin, Jr. Paul Francisco, Mbr/Past TC Chair
Christopher J. Ahne, Subcommittee Chair Michael R. Brambley Hal Levin, Consultant
Paula A. Hernandez Timothy C. Dwyer Patricia T. Graef, BOD ExO
Sheldon M. Jeter Vinod P. Gupta Ross D. Montgomery, CO
Francis A. Mills Pamela M. Immekus Steven J. Hammerling, SLI
George J. Sestak Ashish Rakheja Refrigeration
Gerald G. Williams Richard H. Rooley Douglas C. Scott, Chair
Historical Branislav B. Todorovic Cesar Luis D. L. Lim, Vice Chair
George E. Menzies, Chair Roberto R. Aguilo, BOD ExO Pradeep K. Bansal
Irene Reichert, Vice Chair T. David Underwood, CO Jeffrey B. Berge
David Arnold Mark Owen, SLI Marc Chasserot
716
ASHRAE Transactions
Teddy S. Hansen Chris Scruton, Cons/CEC-PIER James R. Tauby
Todd B. Jekel Chris Pyke, Cons/USGBC James K. Vallort
Dan M. Manole Richard S. Sweetser, Cons/GPIC William F. Walter
Andrew B. Pearson Patricia T. Graef, BOD ExO Michael W. Woodford
Richard R. Royal Ross D. Montgomery, CO Craig P. Wray
Christopher J. Seeton Michael R. Vaughn, SLI Eckhard A. Groll, BOD ExO
M. Kent Anderson, Consultant Standards Ross D. Montgomery, CO
Donald A. Siller, Consultant/CTTC Liaison Carol E. Marriott, Chair Stephanie C. Reiniche, SLI
Charles E. Gulledge III, BOD ExO Kenneth W. Cooper, Vice Chair Technical Activities
Ross D. Montgomery, CO Douglass S. Abramson Charles H. Culp III, Chair
Steven J. Hammerling, SLI Karim Amrane William F. McQuade, Vice Chair
Research Administration Charles S. Barnaby Eric W. Adams
Hugh I. Henderson, Jr., Chair Hoy R. Bohanon, Jr. Michael R. Bilderbeck
Jaap J. Hogeling, Vice Chair Steven F. Bruning Drury B. Crawley
T. Agami Reddy, RAS Chair David R. Conover Stephen W. Duda
John M. House, RPS Chair Steven J. Emmerich John F. Dunlap
Pradeep Kumar Bansal Allan B. Fraser Mark W. Fly
Donald B. Bivens Krishnan Gowri Brian A. Fricke
Piotr A. Domanski Maureen Grasso Walter T. Grondzik
Srinivas Garimella Cecily M. Grzywacz Thomas M. Lawrence
Arthur L. Giesler Richard L. Hall Guistino N. Mastro
Stephen S. Hancock Rita M. Harrold William E. Murphy
Carl F. Huber Adam W. Hinge Victor L. Penar
George A. Jackins Debra H. Kennoy Brian M. Silvetti
William F. McCoy Jay A. Kohler J. Thomas Sobieski
David A. Yashar Janice C. Peterson Charles E. Gulledge III, BOD ExO
Xudong Wang, Cons/AHRI Douglas T. Reindl Ross D. Montgomery, CO
Hywel Davies, Cons/CIBSE Boggarm S. Setty Michael R. Vaughn, SLI
© 2012 ASHRAE 717

--`,`,``,,`,`,```,`

2011–2012 ASHRAE CHAPTER OFFICERS
CHAPTER PRESIDENT PRESIDENT-ELECT VICE PRESIDENT SECRETARY TREASURER
REGION I Spencer Morasch, Director and Regional Chair
Bi-state Nicholas Salomone — Erica Ross — Dennis La Vopa
Boston Patrick Duffy Stephen Nicholas Teri Shannon Daniel Diorio William Garvey
Central New York Steven Sill William Walter — Charles Bertuch Brett Fero
Champlain Valley Michael Cook Thomas Dacres Nathan Mascolino Robert Favali Ken Secor
Connecticut Michael Sherber Francis Brancieri Ronald Beliveau W. Anderson Ronald Beliveau
Granite State Joshua Benoist Richard Ruel Christopher Saltmarsh Christopher Jacques John Grout
Long Island Carolyn Arote Brian Simkins Andrew Manos Charles Lesniak Thomas Fields
Maine Sanford Robinson Jonathan Kerbel Aaron Lewis Paul Reddington Jeffrey Lapierre
New Jersey Roger Shults James Sarno — John Tellefsen Scott Smith
New York Meraj Ramnarine Steven Baumgartner — Lorey Flick Gene Geyer
Niagara Frontier James Price Joshua Doroshenko Rick Robinson Stephen Yager Jenna Baran
Northeast Michael Carr Nicholas Casale Cara Martin Heather Maxwell Douglas Belokopitsky
Rhode Island Michael George — Frank Zaino Richard Mabie Anthony Vessella
Rochester Jeffrey Ellis Michelle Sommerman Robert Wind Edward Burns Christina Walter
Twin Tiers Christopher Kopec Glenn Roberts James Chesko Andrew Nice Brendan Carr
REGION II Isabelle Lavoie, Director and Regional Chair
Halifax Jeff MacNeil Mark Lawrence — Genevieve Houle James Beck
Hamilton Mike Krewski Reaz Usmanali — John Freeman Iain Hill
London (Canada) Jason Vandenberghe Karl Gilroy — Ibrahim Semhat Jamie Kruspel
Montreal Caroline Paquet Michel Lecompte Anthony Jonkov Jean-Gabriel Joannette Stanislaw Kajl
NB/PEI Kevin Clannon Dwight Scott Robert McEwen Philippe Bilodeau Jennifer Chapman
Ottawa Valley Stephen Lynch Donald Weekes — Steve Moons Roderic Potter
Quebec Benoit Lacasse David Gauvin — Solange Levesque Stephane Dufour
Toronto Michael Khaw David Benedetti Leonard Borges Alan Porter, P. Eng Glenn Kilmer
Ashish Saxena
Windsor J Snell Troy Manson — Sara Algudady Diane Beresh
REGION III Kevin Fallin, Director and Regional Chair
Anthracite Alfred Speicher Tracey Jumper Matthew Archey Jonathan Keller Richard Karns
Baltimore Cory Colassard Richard Goins — David Lawson Michael Lamdin
Central Pennsylvania Andrew Odell Alyssa Wingenfield David Bender — Michael Conchilla
Delaware Toby Pierce — — Francis McFadden Robert Schunke
Hampton Roads Robert Ainsley Paul Butler Richard Roberts Daniel Johnston Scott Almond
Johnstown Cory Eberhart David Anderson David Anderson Kyle Koval John Weiland
Lehigh Valley Donovan Denlinger Nicholas Kaintz Ahmad Mahmud Nicholas Kaintz Gabriel Badesso
National Capital Dunstan Macauley Stephanie Mages Cynthia Cogil Andrew Rhodes Scott Houghton
Philadelphia Robert Finkboner James Piscopo Jared Johnson Ashley Lester Michael Witkowski
Pittsburgh Dennis Kelly — — Bryan Proto Paul Parise
Richmond Michael Orr Jaideep Karnik Christopher DeSoto Brad Trevillian David Ziegler
Roanoke Kyle Longbrake Richard Hughes Wesley Southall Jessica Murphy —
REGION IV T Randall Jones, Director and Regional Chair
Atlanta Charles Fischer Charles Miltiades — Priscilla Sager Pamela Androff
Charleston Eric Crawford Michael Cladakis Michael Cladakis — Clarence Ham
Greenville Grant Wiegmann William Tonnsen Devon Cortright William Tonnsen Rodney Hinton
North Piedmont James Swofford Jeffrey Vestal — Maurice Harrison Jeffrey Vestal
South Carolina Brandon Keatley Michael Boone — — Donny Pittman
Southern Piedmont James McDaniel Michael Collarin Michael Collarin Tracie Monfort Jerry McAuliffe
Triangle Jerald Howard Erik Hahn Erik Hahn Adam Sippel Daniel Suchy
REGION V R Lee Millies, Director and Regional Chair
Akron/Canton Eric Wright — Jennifer Richardson — Gregory Drensky
Central Indiana Jack Kibort Timothy Anderson Brad Auger Michael East Kathryn Vawter
Cincinnati Michael Mattox Clinton Schwartz — — Nicholas McDonald
Cleveland Shawn Beres Dominic Cacolici — Scott Dusenbury Matthew Morgan
Columbus James Arnold Michael Hoffa Franklin Eisenhower Jared Keron Franklin Eisenhower
Dayton Steven Meier Jason Gillespie — Matthew Herbert Kyle Kurtz
Detroit Derek Crowe Mark Sedmak — Kishor Khankari Michael Daly
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Eastern Michigan Matthew Zeigler — — — Craig Beardslee

Evansville — — — Theresa Bailey Robert Derr
Fort Wayne Charles Ray William Bissmeyer Kirk Wing JP Heisserer Ross Fredrickson
Northern Indiana Joseph Bowers Krista Taylor — Brian Klusendorf R Lee Millies
Toledo Matthew Rasey Mario Picciuto John Homenik Melissa Cole Michael Green
Western Michigan Douglas Zentz Adam Doubblestein James Merlino James Harrison Scott Smith
718
ASHRAE Transactions
REGION VI Tina Brueckner, Director and Regional Chair
Cedar Valley Adam Bunnell Jared Ramthun — Arlo Wieczorek Dusty Beenen
Central Illinois Kevin Bannon David Muchow — Wayne Robinson Daniel Hose
Illinois Benjamin Skelton Laura Ludwig — Laura Michel Daniel Hansen
Iowa Kelly Crow Matthew Jesson Matthew Jesson Justin Barrett Timothy Conrad
La Crosse Area Chris Hsieh Michael Patterson — Christopher Schlegel Anthony Dix
Madison Manus McDevitt Michael Gall — Carol Mori Sharon Gould
Minnesota Gary Fischbach Tim Manz Tim Manz Vinod Gupta Daniel Chudecke
Mississippi Valley Jacob Vorac Nathan Jacques — — Cris Washburn
Northeast Wisconsin Ryan Macario Kevin Weber Kevin Weber Matthew Veriha —
St Louis Vincent Stanec Eric Rogers Robert Jones Robert Glunt Robert Glunt
Wisconsin Justin Patrick Timothy Pann Timothy Pann Jason Leffingwell Jason Nenonen
REGION VII Jeff Gatlin, Director and Regional Chair
Baton Rouge Jordan Fish Mark Quebedeaux Terry Grier Jordan Fish Robert Smith
Birmingham John Jeffcoat Walter Stockli Zachary Lang Chandler McGarrah Taylor Adams
Bluegrass Kevin Davies Dustin Meredith Carrie Kelty Brad Wallace —
East Tennessee Philip Newby Richard Biggs Jade Culbertson John Yust Jack Hopkins
Louisville Ross Wojcik Julian Donahue Julian Donahue Jeffrey Murphy Thomas Weber
Memphis Ryan Hertter James Fleck — Aubrey Tucker Gregory Brown
Mississippi Brad Bumgarner Steven Armour Jonathan New Troy Pike Alan Watts
Mobile Richard Geist Eric Korn Eric Korn Shannon Holderfield —
Nashville Matthew Fruetel Christopher Reeves Michael Coradini David Ballard PJ Mezera
New Orleans Michael Cooper Mike Bonano — Wendell Humphres Jacob Alter
North Alabama Tony Horton Thomas Meier — William Marshburn Laurence Paxton
Northwest Florida Jean Piecuch Jean Piecuch — Tony Rodriguez John Staples
Tennessee Valley Victor Vosen Samuel Delay Joseph Cooper — Patricia King
REGION VIII John Harrod, Director and Regional Chair
Alamo Jennifer King Rocky Ormand — — Shannon Novak
Arkansas William Simpson Eric Fullerton — Mark Mizell Marlon Joiner
Austin Andy Collier Mark Mikulin Justin Johnson Ahmad Hamiyeh Derrick Van West
Central Oklahoma Bryan Garcia Charles Stoabs — Brian Determan Samuel Sexton
Dallas W Dickson Shelley Zhao — Charles Crawford Gene Lowery
East Texas Richard Cawley Raymond Rite Don Schuster Roy Crawford Bill Stiles
Fort Worth Richard Watters Daniel Merkel — Anthony Finch Jay Sullivan
Guadalajara Victor Torres Nava Luis Miguel Fontes Connie Herrera Cuauhtemoc Aguirre Haro Jorge Cabrero
Bracamontes
Houston Paul Alexander Robert Neely — Michael Langton Elliot Millican
Mexico City Antonio Gonzalez Magdalena Castillo Aguilar Ramon Davila Gildardo Yanez —
Monterrey Carlos Cavazos Sofanor Alarcon — Carlos Anastas Roberto Gonzalez
Northeastern Oklahoma William Kirk Chad Smith — Bruce Harman Matthew Pollard
Shreveport Roger Barnes William Wilson — Aaron Storer Herbert Patrick
South Texas Stephen Lewis James Ward — — Joseph Hammitt
West Texas Terry Parham Rhett Graham Walter Brown Michael Ramirez Benjamin Hames
REGION IX Richard Bowman, Director and Regional Chair
Big Sky Ammon Palmer Jacob Witham Jacob Witham Tim Model —
Black Hills Michael Mueller Steven Freeland Steven Freeland Steven Freeland Judy Oakley
El Paso Warren Marquette — — — —
Idaho Cory Law Randy Reed — Dennis Fox Laura Rasmussen
Kansas City Blake Ellis James Van Hoecke Thomas Benassi Austin Allen Jarrod Foster
Nebraska Kimberly Cowman Renee Meyersick — Sarah Poursharafeddin Jeff Backemeyer
New Mexico Charles Scoggin Patrick Watkins — Morgan Royce Erin Coffman
Ozarks Allen Davis Jennifer Winslow — Michael Cavin Jennifer Winslow
Pikes Peak Patrick Lancaster David Simmons David Simmons Elaine Gutierrez Elaine Gutierrez
Rocky Mountain Adam Bishop Michelle Swanson Jon Rundquist Jessie Jones Michael Harrington
South Dakota Gregory Hintgen Robert Everson Rob Marshall Michael Thompson Ryan O'Connor
Utah Thomas Fujikawa Trenton Hunt — Keith Stippich Gary Anthon
Wichita Curtis Wondra Michael Miller Michael Miller Dahl Carmichael Ryan Haug
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
© 2012 ASHRAE
719
REGION X Michael Burgess, Director and Regional Chair
Central Arizona Shawn Masterson Tyson Padgett — Thomas Andrews Colin Laisure-Pool
Golden Gate Kristin DeMartini Thomas Weaver — Isabelle Lavedrine Elizabeth Foster Courtney
Hawaii Paul Scott Kevin Saito Barry Jim On Blake Araki Donna Kishi
Northern Nevada Christopher Little Bryan Tilton — Jason Bender Brian Bassi
Orange Empire Steve Kang Robert Woods Michael Huynh Jeff Conrad John Sawyer
Sacramento Valley Adam Davis Jake Garcia — Raymond Schouweiler Karl Rapp
San Diego Dennis Christie Rebecca Haehn — Heather Schopplein Hugh Egan
San Joaquin Keith Harris Sean Skiles — Chan Kim Justin Westmoreland
San Jose Krystal Mae Matthews Orion Moore — Michael Cleary Andrea Zarour
Sierra Delta Andrew Hatfield Corey Haskin — Donavan Chavez Tyler Anderson
Southern California Jeffrey Landreth Matthew Church — Scott Neithercut Edwin Hornquist
Southern Nevada Bryan Im Thomas Tetting — Matt Lisiewski Michael Aslanides
Tri County Jaime Lopez Erick Delgado — — Yung-Hsiang Lin
Tucson Robert Kunkel Thomas McGreal — Quinton Lamoreaux Burton Wright
REGION XI Erich Binder, Director and Regional Chair
Alaska Jessi Cherry Ginger Schaaf Ginger Schaaf Richard Creelman Gregory Jernstrom
British Columbia Christopher Collett Mihaela Spulber Kimberly Rosval — Stephen Kooiman
Inland Empire Neal Smith Michael Caffrey Gerald Van Guilder James Cannon George Augustini Michael
Caffrey
Manitoba Johann Baetsen Corey Nation Stephen Norsworthy Mark Windeatt Doug Castor
Mid-Columbia James Clifford Weimin Wang — Jian Zhang Heejin Cho
Northern Alberta Lindsay Austrom Robert Prybysh Tom Jacknisky Amisha Pope Geoffrey Hurly
Oregon William Fink Adam Carlson Adam Carlson Ajit George Mike Cundiff
Puget Sound Brian Wolford Jason Hall — Alan Burt Michel Sotura
Regina Jason Danyliw Carla Spriggs Carla Spriggs — Kristopher Pockett
Saskatoon Jeffrey Frie Christopher Watt Christopher Watt — Blake Erb
Southern Alberta Don Blacklock John Forrest Serpil Clancy Brad Bond Brad Bond
Vancouver Island Calvin Tripp Mark Stitt — David Neufeld Calvin Tripp Mark Watson
REGION XII William Malphus, Director and Regional Chair
Argentina Paula Hernandez Carlos Mitroga Carlos Brignone Pablo Sarfiel Miguel Corigliano
Brasil Diogo Prado Mario Ferreira Edison Guimaraes Miguel Ferreirós Enio Kato
Central Florida John Miller Megan Tosh Anik Patel Nathaniel Boyd Malia Powers
Colombia Fabio Clavijo Mauricio Baena — Gerardo Alfonso Jose Guerrero
Florida West Coast Shawn Jeffrey Jason Proctor Philippe Jean Joseph Souza Mike Costello Joseph Cox
Gold Coast Phillip Ivins Keith Giniger — Javier Jarquin Danny Denaro
Jacksonville Robert McClure Jonathan McDaniel Julius Harden Joe Nagy James Ward
Miami Renny Urdaneta Ivan Velez Ivan Velez Tulia Rios Renny Urdaneta
Puerto Rico Steven Mikasobe Steven Mikasobe — William Sanchez —
Southwest Florida Charles Spelman Autumn Spalding Michael Wilson Jason Hardman Michael Wilson
Spacecoast Kevin Messer Jeanne Duce — — Brian Blenis
REGION XIII Wichai Laksanakorn, Director and Regional Chair
CAR Ashrae Beijing Group — — — Edward Tsui —
CAR Ashrae Shanghai — — — Edward Tsui —
Group
Hong Kong Wan Fuk Lin Chan Kit Lo — Ching Man Pang Hei Kin Chan
Indonesia Deddy El Rashid Muhammad Alhamid Ari Pasek Totok Sulistiyanto Achmad Chaidir
Malaysia Dennis Lim Ainee Cha Ching Loon Ong Lim Seng Quek Irin Foon
Philippines Emmanuel Punsalan Fernando S Guevara Dennis Alejandro Tito Soledad David Farala
Singapore Gek Tan Francis Lee — Uma Maheswaran Cheyyar Chen Ik
Ramanathan
Taiwan Fu-Jen Wang Chia Chi Hsu Yu-Hao Kang Hui-Chung Chen Lai Chang
Thailand Vorasen Leewattanakit Ballang Sathorn Sumet Jiambutr — Chokewichit Laksanakorn
--`,`,``,,`,`,```,`,,``,`,``,,
720
ASHRAE Transactions
REGION-AT-LARGE Ashok Virmani, Director and Regional Chair
ASHRAE Bangalore Bhagavan Srinivasan Satish Iyengar Seval Venkatachalam Tanjore Rao C. Subramaniam
ASHRAE Deccan B.B.N. Prasad Murali Ramanathan Narayanan Chandrasekar Venkataraman Vijayaragha- Ramchandar Addhipalli
van
ASHRAE Falcon Ahmed Alaa Eldin Rami Alkhalil Khaled Al-Sallal — Ahmed Alaa Eldin
Mohamed Mohamed
ASHRAE Mumbai Vikram Murthy Adarsh Kaul Anurag Pareekh Roshini Easow A Ravi
ASHRAE Nigeria Babatunde Badru Olaniyi Kehinde Oluwasesan Daramola — Marcus Osindero
Bahrain Shawqi Ebrahim Tariq Malik — Olavo Rodrigues Adil Inam
Cairo Fouad Hamad Essam Eldin Khalil — Mohamed Manzalawi Emad El-Din Mokhtar
Central Pakistan Amer Abbas Syed Shezad Raza Kazmi Safdar Ali Tahir Raza Nisar Mohyud Din
Chennai Sriram Subramaniam Venkataramani Sridhar K.P.S. Ramesh Rajagopal Saravanan Vinod K B
Danube Ioan Boian Milorad Bojic Branislav Todorovic Ioan Dobosi Gratiela Maria Tarlea
Hellenic Dimitris Atsidaftis Dimitris Mantas Demetri Ioannidis Vasiliki Galiotou Dimitrios Charalambopoulos
India Rajnish Joshi Dinesh Gupta Kanwar Singh Ashwani Jain Sunil Kher
Kuwait Nawaf Aljuwayhel Nawaf Al-Mutawa — Salem Alhajraf Ali Alajmi
Lebanese Mohammad Al Tassi Michel Hayek — Riad Assaf Nohad Boudani
Northern Pakistan Pervez Iqbal Sarfraz Ahmad Mohammad Azad Naseer Ahmad Muhammad Nadeem Zuberi
Pakistan Tajammul Husayn Sajjad Ghani Khalid Gulzar Khawaja Iqbal Noor Afzal
Portugal Ernesto Peixeiro Ramos Eduardo Maldonado Carlos Lisboa Fernando Calado De Brito Serafin Grana
Pune Ravindra Kulkarni Sanjay Phadke Jayant Joshi Shailesh Phatak Gajanan Khot
Qatar Oryx Hassan Sultan Hani Hawamdeh Hisham Hajaj Subrata Kar Seenu Pillai Kinan Fahs
Saudi Arabia Samir Al-Umair Khalid Al-Mulhim Saad Al-Dbass — Khalid Al-Mulhim
Spain Enrique Yotti Adolfo Sanz Izquierdo Alejandro Cabetas Jose Antonio Palomero Juan Boeta
Sri Lankan Chandana Dalugoda Wimalachandra Jayakody Pandukabhaya Amarasinghe Wimala Goonaratne Kapila Rodrigo
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Niranjan Caldera
Hiulet Jayaweera
Clarence Wijeyesekera
Western India Yogesh Thakkar Chetan Shah Ashish Shah Ashok Patel Bankim Dave
© 2012 ASHRAE
721
2011–2012 ASHRAE TECHNICAL COMMITTEES,
TASK GROUPS, TECHNICAL RESOURCE GROUPS,
AND MULTIDISCIPLINARY TASK GROUPS
ASHRAE’s 99 technical committees, 3 task groups, 2 technical resource groups, and 6 multidisciplinary task groups are responsible for
revisions in the ASHRAE Handbook, for development of work statements for research and research project monitoring, and for sponsoring
technical programs at Society meetings. Names of technical committees and task groups and their members, as of June 2012, are as follows.
2011–2012 TECHNICAL  Jason Hugenroth 1.3 HEAT TRANSFER AND 

ACTIVITIES COMMITTEE Omar Huzayyin FLUID FLOW
Charles H. Culp, Chair Anthony Jacobi
Voting
William F. McQuade, Vice Chair Edward Keuper
Victor L. Penar Sanford Klein Samuel Yana-Motta, Chair
J. Thomas Sobieski Thomas Kuehn Amir Jokar, Vice-Chair
William E. Murphy Ness Lakdawala Justin Kauffman
Michael R. Bilderbeck Jason LeRoy Rick Couvillion
Giustino N. Mastro Liping Liu Axel Kriegsmann
John F. Dunlap Pedro Mago John Thome
Drury B. Crawley Thomas Meyer Ebrahim Al-Hajri
Stephen W. Duda Pavan Naicker Zahid Ayub
Thomas M. Lawrence Nicole Okamoto Steve Brown
Eric W. Adams Lorenzo Cremaschi
Stanislav Perencevic
Mark W. Fly Steven Eckels
Jose Perez-Galindo
Brian A. Fricke Kenneth Schultz
Eric Ratts Petur Thors
Walter T. Grondzik
Trilochan Singh
Brian Michael Silvetti Nonvoting
Robert Tozer
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Charles E. Gulledge
Ross D. Montgomery Florian Zink Jon Hartfield
Michael R. Vaughn, Staff Liaison Kurt Zoellick Joseph Huber
Satheesh Kulankara
SECTION 1—FUNDAMENTALS 1.2 INSTRUMENTS AND S.A. Sherif
AND GENERAL MEASUREMENTS Brian James
Arindom Joardar
Voting
Omar Abdelaziz
1.1 THERMODYNAMICS AND Andreas Beutler
PSYCHROMETRICS Stephen Idem, Chair
James Bogart
B. Beck, Vice-Chair
Voting Alberto Cavallini
Frank Spevak
Gordon Dong
S.M. Sami, Chair James Douglas Kevin Dunshee
S.A. Sherif, Vice-Chair Charles Wright William Fox
Laura Schaefer John Scott Srinivas Garimella
Roy Crawford Michael Wegenka Eckhard Groll
Vernon Peppers David Yuill Stephen Hancock
William Fox Nonvoting David Honeycutt
Hans-Joachim Kretzschmar Anthony Jacobi
James Schaefer B. Moncrief Mark Johnson
Omar Abdelaziz Patrick Collins Cesare Joppolo
Yongfang Zhong Leonard Damiano Mohammad Khan
Ron Nelson Chad Dorgan Holger Koenig
Mark Ahlers Rick Heiden Konstantinos Kontomaris
Vikrant Aute Mohammad Hosni Branislav Korenic
Chad Bowers Liping Liu
Xianghong Huang
Robert Braun William McQuade
Eric Lee
Steve Brown Josua Meyer
Arsen Melikov
Howard Cheung Bruce Nelson
Jonathan Douglas Jose Perez-Galindo Darin Nutter
Ellen Franconi Nathan Redmann Michael Ohadi
Donald Gatley Lauren Ronsse Michael Pate
Sivakumar Gopalnarayanan James Schaefer Olivier Pelletier
Sebastian Herrmann Richard Stonier Stanislav Perencevic
Anna Hueffed Yuanhui Zhang Gary Price
Edwin Huestis Xiaohui Zhou Eric Ratts
722
ASHRAE Transactions
Raymond Rite Angela Lewis Walter Grondzik
James Schaefer Carol Lomonaco Jeff Haberl
Jeb Schreiber Christopher Miller William Healy
Mirza Shah Michael Monahan Kristin Heinemeier
Mahesh Valiya-Naduvath Chad Moore Robert Hitchcock
Jun Wang George Naim LinJie Huang
Mustafa Yanik H. Newman Rick Larson
Yongfang Zhong Zachary Obert Kwang Ho Lee
Robert Old Brian Moura
1.4 CONTROL THEORY AND William Pienta Ron Nelson
Paul Pinkston Robert Old
APPLICATION
Ofer Pittel Zheng O'Neill
Voting Michael Pouchak Vernon Peppers
Michael Schell Mary Piette
Steven Linn, Chair Boggarm Setty Michael Pouchak
Gaylen Atkinson Dennis Stanke Daniel Rau
Charles Coward Jeffrey Stein Lawrence Schaefer
Raj Daswani Jim Tello Michael Schwedler
Philip Haves Min Chien Tsai Stephen Selkowitz
Damian Ljungquist David Underwood Paul Wacker
Nemat Lotfi Paul Wacker Wayne Webster
Jin Wen Xinlei Wang
Chariti Young Mashuri Warren 1.6 TERMINOLOGY
Nonvoting Michael Wetter
Verle Williams Voting
Kimberly Barker, Vice-Chair Shui Yuan
David Bornside, Chair
Alfred Garza John Zhou
Bruce Billedeaux, Vice-Chair
David Kahn Xiaohui Zhou
Roger Lautz
Frank Shadpour Wangda Zuo
Charles Dale-Derks
Steven Taylor
Jaap Hogeling
Barry Bridges 1.5 COMPUTER APPLICATIONS Lawrence Drake
James Coogan
Voting Arthur Hallstrom
Nicholas Gayeski
Jennifer Leach
Sean Graham Michael Galler, Chair
Jarrod McMains Nonvoting
Xiaohui Zhou, Vice-Chair
James Nietfeld Steven Rosen Birol Kilkis
Peter Armstrong John Kennedy Suzanne LeViseur
Don Bailey Norman Bourassa Ross Montgomery
Christopher Benson John Crosby
David Bornside Bruce Billedeaux
David Branson 1.7 BUSINESS, MANAGEMENT
Benjamin Welle
Steven Bushby AND GENERAL LEGAL
Nonvoting
Monte Cole EDUCATION
Robert Coleman Stephen Roth Voting
Leonard Damiano Alison Kwok
Lindell Davidson Mark Hydeman Duane Wolf, Chair
Sharon Dinges Pornsak Songkakul Michael Connor, Vice-Chair
Gregory Dobbs Mahabir Bhandari Wayne Webster
Larry Felker Vikrant Aute Farooq Mehboob
Larry Fisher Albert Black Richard Rooley
Richard Franseen David Bornside Leon Shapiro
James Gartner Michael Brambley William O'Donnell
Mike Gibson David Branson Norm Maxwell
Donald Hardin Rodney Butler George Jackins
P. Hart J. Carpenter Michael Bilderbeck
Gregor Henze Daniel Choiniere Warren Hahn
John House David Claridge Judith Maxwell
Mark Hydeman Robert Coleman Ronald Millies
Kevin Kerr Charles Culp Lawrence Spielvogel
John Kettler Timothy Dwyer Nonvoting
Kristopher Kinney Peter Ellis
Curtis Klaassen James Forester Rodney Lewis
Anthony Lee Krishnan Gowri Mark Allen
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
© 2012 ASHRAE
723
David Arnold Patrick Noonan Birol Kilkis
Ronald Bailey John Scott Nadar Jayaraman
Martin Barenbaum Dragos Paraschiv
Dale Cagwin 1.9 ELECTRICAL SYSTEMS Geoffrey Bares
Cynthia Callaway Gearoid Foley
Voting
Leslie Carmichael Lucas Hyman
John Castelvecchi Steven Faulkner, Chair Nonvoting
Richard Charles Frank Pucciano, Vice-Chair
John Clark Abdi Zaltash
Kristoffer Findley
William Coad Mark Davis
Armin Hauer
Donald Colliver
John Stoops Nathan Ho
Samuel Cummings
Michael Todd Thomas Schmeister
Raj Daswani
Nonvoting Shawn Wedewer
Mark Diamond
Michael Edwards Ainul Abedin
Russell Tiffany Rudy Acevedo
James Escamilla
Richard Evans Lawrence Markel G. Anand
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
William Frey Kenneth Luther John Andrepont
Jeff Gatlin Leo Stambaugh Alonzo Blalock
Michael Gerhardt Ronald Antinori Robert Bunn
Theresa Goetz Kirk Bond Hywel Davies
Harley Goodman Subhash Brahmavar Ronald Davis
David Governo Curtis Crane Gregory Dobbs
Paul Halyard Robert Crosby Michael Ellis
Richard Jones Alexander Domijan Donald Erickson
Frederick Kohloss Chad Dorgan Steven Fischer
William Kumpf Fred Dougherty Craig Fredeen
Carl Lawson Carlos Estrada Ersin Gercek
Jennifer Leach Charles Foster Patricia Graef
Todd Meling Norman Fowler
Rhett Graves
Francis Mills Charles Frazell
Alan Green
Ross Montgomery Joe Hague
Vinod Gupta
Richard Pearson Robert Helt
Dennis Halsey
Paul Pratt Rodney Lewis
Hugh Henderson
M. Scoggins Joseph Lunsford
Frank Jakob
Bonnieblue Valentin Janet Lynch
John Mentzer Ali Jalalzadeh-Azar
Joseph Vallort
Wayne Mitchell John Kraft
Filza Walters
John Nix Shawn Lee
Rex Noble Kurt Liebendorfer
1.8 MECHANICAL INSULATION Pedro Mago
SYSTEMS Janice Peterson
Thomas Piwinski William McAuliffe
Voting John Richardson Joe Millham
Steven Rosenstock Richie Mittal
Gordon Hart, Chair Drew Overmiller
Richard Rundus
Charles Cottrell, Vice-Chair Vikas Patnaik
Colleen Smith
Darrell Peil
Louis Sulfstede Victor Penar
Charles Petty
Paul Tanpitukponoje Thomas Pierson
Andre Desjarlais
Marvin Thedford David Robertson
Carsten Rode
Gary Throckmorton Dan Shapiro
Christopher Crall
William Miller Marija Todorovic Donald Shepherd
Robert Neely William Warren Brian Silvetti
Victor Penar Harold Smith
Michael Resetar 1.10 COGENERATION SYSTEMS Chang Sohn
James Young Sriram Somasundaram
Voting
Klaus Sommer
Nonvoting
Riyaz Papar, Chair Ian Spanswick
Kenneth Collier Timothy Wagner, Vice-Chair Marija Todorovic
Thomas Ponder Dharam Punwani Donald Warren
Jack Blundell Louay Chamra James Watts
Ricardo Gamboa Richard Sweetser Terrence Whitehead
Michael King Annette Dwyer Verle Williams
724
ASHRAE Transactions
1.11 ELECTRIC MOTORS AND Fitsum Tariku JoonHo Choi
MOTOR CONTROLS Fiona Aldous Richard De Dear
Mark Anderson Charles Dorgan
Voting
Gil Avery E. Curtis Eichelberger
Daryl Showalter, Chair Hollace Bailey Jeff Gatlin
Terry Davies, Vice-Chair Stephen Barefoot Andreas Holm
Robert Richard Marcus Bianchi Raymond Horstman
John Tolbert Douglas Bibee Mohammad Hosni
Rick Heiden James Cummings Glenn Hourahan
Leonard Damiano Michael Humphries
Preben Christiansen
Claudio Darras Madhavi Indraganti
Rainer Grosse-Kracht
Andre Desjarlais Daniel Int-Hout
Rick Bunch
Larry Elkin Byron Jones
Armin Hauer
Diana Fisler Alison Kwok
Robert Helt
Stanley Gatland Hal Levin
Alex Lifson
Manfred Gerber Baizhan Li
Michael Todd
Samuel Glass Chao-Hsin Lin
Kurt Vega
Michael Hayes Dennis Loveday
Derrick Vigil
Dale Herron Arsen Melikov
Nonvoting Hartwig Künzel William Murphy
Carl Lawson Amy Musser
Thomas Lowery Ian Nelson
Mark Lawton
Charles Culp Bjarne Olesen
Neil Leslie
Rupal Choksi Gwelen Paliaga
Rodney Lewis
Deep Gupta Kenneth Parsons
Him Ly
Paul Lin James Reardon
Wahid Maref
Colin Tait Douglas Reynolds
John Murphy
Rudy Chervil Julian Rimmer
Albert Pucino
Jim Chmielewski Frederick Rohles
Mikael Salonvaara
Robert Coleman Kenneth Roy
Elia Sterling
Mark Doty Glenn Schuyler
Davidge Warfield
Phillip Johnson Michael Sherber
Bede Wellford
Paul McCabe Jerry Sipes
Theresa Weston
Riyaz Papar Zuraimi Sultan
Donald Perry Alexandra Thompson
Hermann Renz SECTION 2—
Dale Tiller
James Rhinehart ENVIRONMENTAL QUALITY Robert Tinsley
Harold Schnetzka Jorn Toftum
William Tinsley 2.1 PHYSIOLOGY AND  William Tower
Peter Walter HUMAN ENVIRONMENT Ahmet Ugursal
Thomas Watson Keyur Vadodaria
Voting
James Woods
1.12 MOISTURE MANAGEMENT Paul Lebbin, Chair David Wyon
IN BUILDINGS John Stoops Xiaojiang Xu
James Bushnell Hui Zhang
Voting
Chandra Sekhar
Lewis Harriman, Chair Peter Simmonds 2.2 PLANT AND ANIMAL
Anton TenWolde Diotima Von Kempski ENVIRONMENT
Alexander McGowan Edward Arens Voting
Steve Cornick Jaap Hogeling
Jeff Traylor Shin-ichi Tanabe Henry Hays, Chair
Hugo Hens Nonvoting James Coogan, Vice-Chair
Andreas Holm K. Dean
Achilles Karagiozis Mark Jackson, Vice-Chair Daniel Berckmans
S. Kelter Eric Adams J. Carpenter
Lan Nguyen Thi Lily Wang Sheryll Jerez
Paul Shipp Ronald Westbrook Yuanhui Zhang
Raoul Webb Joy Altwies Nonvoting
Nonvoting Karin Arnold
Larry Berglund Lingying Zhao
Raymond Patenaude, Vice-Chair Gail Brager Larry Howlett
Ronald Bailey Eric Brodsky Louis Albright
Joseph Griner Robert Chase Tami Brown-Brandl
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
© 2012 ASHRAE
725
Michael Brugger Low Han Guan Den Thomas Kuehn
Mark Butler Daniel Dowiak Christopher Peart
Heather Camden Christopher Flaherty Christine Sun
Leslie Christianson Karin Foarde Nonvoting
Donald Colliver W. Freeman
Daniel Frasier Radha Ganesh Donald Thornburg, Vice-Chair
Jerry Gaffney Rakesh Goel Thomas Justice
Richard Gates Michael Handley Marilyn Listvan
Daniel Ghidoni James Hanley Kenneth Snider
Angela Green Eric Hanson Ronald Bailey
Thomas Greiner Robert Hendry Robert Baker
Albert Heber Ker-Ching Hsieh Russell Baldinger
Carl Lawson Catherine Jacquiod Charlene Bayer
Ronaldo Maghirang Jennifer Kaminski Morton Blatt
James Meyer Edward Karpenski Susan Brandes
Gerald Riskowski Thomas Kelly Steven Burgess
Christian Robbian H. Khalifa Christopher Campion
Luis Solarte Erik Kuiper Eric Chen
Richard Stakutis Marianne Lane Johnnie Copley
Nolen Stripling Rick Larson Robert Culbert
Don Van Noy Josephine Lau Low Han Guan Den
Xinlei Wang Chang-Seo Lee Ronald Domitrovic
Hongwei Xin Ashish Mathur Leonard Duello
Joseph Yoshida Brian Monk Harry Elinsky
Joseph Zulovich Marc Newman Joseph Enneking
Deepak Pahwa David Ensor
2.3 GASEOUS AIR Uri Rainisch Forrest Fencl
CONTAMINANTS AND  Mark Renn Michael Flom
David Schaaf Joe Fly
GAS CONTAMINANT 
Robert Scheir Karin Foarde
REMOVAL EQUIPMENT Paul Francisco
Scott Sherwood
Voting Stanley Slabinski Jon Fretthold
Chang Sohn Jean Girard
Matt Middlebrooks, Chair David Space Rakesh Goel
W. Brad Stanley, Vice-Chair Christine Sun Volker Haendler
Charlene Bayer Jeromy Tincher Kim Hagstrom
Kathleen Owen Philip Winters Michael Handley
Marilyn Listvan David Witham Eric Hanson
Cecily Grzywacz Terri Wytko Scott Hanson
Paolo Tronville Stephen Zitin Randall Haseman
H. Burroughs Ker-Ching Hsieh
Michael Flom Mark Jackson
Brian Hafendorfer
2.4 PARTICULATE AIR
Timothy Johnson
William Lull CONTAMINANTS AND
David Jones
Christopher Muller PARTICULATE CONTAMINANT Matt Kalis
Jeffrey Roseberry REMOVAL EQUIPMENT Jennifer Kaminski
Jianshun Zhang Voting Carolyn Kerr
Nonvoting Douglas Kosar
James Hanley, Chair Brian Krafthefer
Carolyn Kerr Eric Brodsky James Kubokawa
Dean Tompkins Stephen Nicholas Erik Kuiper
Nick Agopian Mark Renn Marianne Lane
Monroe A Britt Bruce McDonald Don Largent
Brian Krafthefer Jeffrey Siegel Rick Larson
Ann Louise Sumner Alan Veeck William Lawrence
Wane Baker H. Burroughs David Matier
Robert Burkhead Monroe A. Britt Stephan Mcdonagh
Christopher Campion Nader Barjesteh Todd McGrath
Eric Chen Robert Burkhead Peter McKinney
Wenhao Chen Kyung-Ju Choi William Mele
Kyung-Ju Choi Michael Corbat Matt Middlebrooks
Johnnie Copley Zied Driss Emily Moore
Michael Corbat Charles Kern Kathleen Owen
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
726
ASHRAE Transactions
David Pui 2.6 SOUND AND VIBRATION Dan LaForgia
John Robertson CONTROL Heng-Yi Lai
Leslye Sandberg Brian Landsberger
Voting
E.C. Schroeter Joshua Leasure
Charles Seyffer Patrick Oliver, Chair H. Leventhall
Peter Shipp John Gierzak, Vice-Chair Patrick Marks
Chang Sohn Dustin Meredith Charles Mattocks
David Space Chris Papadimos Duane McLennan
E. Curtis Eichelberger Alexander Michaud
Leslie Sparks
Robert Simmons Erik Miller-Klein
Greg Sproule
Matthew Stead Andrew Mitchell
Steven Szymurski
Jason Swan Ralph Muehleisen
Jeromy Tincher Jose Nepomuceno
Francis Babineau
Paolo Tronville John Pappas
Robert Hassler
Ali Vatine Karl Peterman
Siu-Kit Lau
Michael Waring Robert Lilkendey James Pooler
Philip Winters Jerry Lilly Raj Prime
James Woodward Matthew Murello Michael Resetar
Terri Wytko Kim Osborn William Rockwood
Richard Peppin Eric Rosenberg
2.5 GLOBAL CLIMATE CHANGE Douglas Reynolds Erica Ryherd
Kenneth Roy Benjamin Sachwald
Voting Ken Shook
Randal Zimmerman
Nonvoting Tim Simcoe
Donald Brundage, Chair
John Sofra
Denis Clodic, Vice-Chair
Steven Wise William Stewart
Gursaran Mathur
Lauren Ronsse Vijay Tripathi
Van Baxter Lily Wang Terence Tyson
William Walter Michael Schwob Zhiping Wang
Anthony Arguez Mark Schaffer Jonathan Weinstein
Larry Butz Rami Alkhalil
Steven Bernhardt Jeffrey Babich 2.7 SEISMIC AND WIND
James Crawford Joseph Bridger RESISTANT DESIGN
David Godwin Nicole Cuff
Thomas Lawrence Erroll Eaton Voting
Jinghao Liu
Nonvoting Karl Peterman, Chair
Paul Meisel
James Carlson
James Martin J. Michael Spencer
Richard Sherren
Monjur Mourshed Nicholas Sylvestre-Williams
Daniel Rau
Roman Wowk
Karim Amrane James Tauby
Daniel Abbate
M. Anderson John Giuliano
Mark Bastasch
Warren Beeton Norman Broner Eren Kalafat
Timothy Bernadowski Todd Busch Donald Warick
Rockford Boyer David Carroll Daniel Abbate
James Calm Zvirimumwoyo Chinoda Scott Campbell
Piotr Domanski John Dunlap Edward Fitts
Cynthia Gage Ronald Eligator Michael Froehlich
Roderick Jackson Mark Fly Michael Morby
Osami Kataoka Michael Froehlich Panos Papavizas
Manoj Khati Kevin Gaghan Nonvoting
Bruce Lindsay Radha Ganesh
Jason George Robert Simmons, Vice-Chair
Dan Manole
Lewis Goodfriend Patrick Marks
Michael Martin
Brian Guenther Rami Alkhalil
--`,`,``,,`,`,```,`,,``,`,``,,,-`-`,,`,,`,`,,`---
Mack McFarland Philip Caldwell

Arthur Hallstrom
Christoph Meurer Eric Rogers
Joseph Horesco
Suresh Modi Michael Keating Paul Selman
Michael Roth Reginald Keith J. Ayres
Brian Silvetti Manoj Khati Richard Batherman
Leonard Van Essen James Kline Tim Benkert
David Wilson Marvin Kloostra Soner Bicer
Robert Yost Glenn Kowald David Bloom
© 2012 ASHRAE
727
James Brownell Michael Deru Brittany Romig
Mark Bryant Malcolm Lewis Stephen Selkowitz
Philip Caldwell Janice Means Michael Sherber
Mark Capriotti Mitchell Swann Som Shrestha

ASHARE Transactions-2012 PDF

Caricato da

Informazioni sul documento

Titolo originale

Copyright

Formati disponibili

Condividi questo documento

Condividi o incorpora il documento

Opzioni di condivisione

Hai trovato utile questo documento?

Questo contenuto è inappropriato?

Copyright:

Formati disponibili

ASHARE Transactions-2012 PDF

Caricato da

Copyright:

Formati disponibili

ASHRAE

ASHRAE ANNUAL CONFERENCE

2012 Volume 118

Copyright American Society of Heating, Refrigerating and Air-Conditioning Engine

ASHRAE Research: Improving the Quality of Life

Product code: 82122

ISSN 0001-2505 ISBN 978-1-936504-35-0

Any updates/errata to this publication will be posted on the

Copyright American Society of Heating, Refrigerating and Air-Conditioning Engine

Copyright American Society of Heating, Refrigerating and Air-Conditioning Engine

Information for Authors

SA-12-C004 Measured Impact on Space Conditioning Energy Use in a Residence 

SA-12-C005 Thermal Storage Properties of a Hybrid Ground Source Heat Pump

SA-12-C006 Field Trials of a Waterless Home Heating and Humidification Technology

SA-12-C007 Environmental Noise Case Studies: Air-Cooled Refrigeration Chiller Installations 

SA-12-C008 Field-Test of the New ASHRAE/CIBSE/USGBC Performance Measurement Protocols: 

SA-12-C009 An Experimental and Simulation Study of Lighting Performance 

SA-12-C010 Impacts of Air Filters on Energy Consumption in Typical HVAC Systems

SA-12-C012 Validation Methodology to Allow Simulated Peak Reduction 

SA-12-C015 Occupant Perceptions of an Indoor Thermal Environment in a Naturally Ventilated Building

SA-12-C016 What We Have Learned From 20 Years of HVAC System Commissioning

SA-12-C021 Laboratory Measurements of HVAC Installation and Maintenance Faults

SA-12-C024 Calibration of a Building Energy Model Considering Parametric Uncertainty

SA-12-C026 Using Energy Models to Verify Utility Consumption

SA-12-C029 Laboratory Evaluation of Aftermarket Boiler Control System

SA-12-C031 A High-Efficiency, Reduced-Emissions, Combustion Control System 

SA-12-C033 Zero Refrigeration for Data Centres in the USA

SA-12-C037 Contaminant Removal Effectiveness of Displacement Ventilation Systems 

SA-12-C039 Modeling and Analysis for Mixed-Mode Cooling of Buildings

SA-12-C040 Contaminants in Hotel Room Exhaust Air

SA-12-C043 Determining the Cost-Effectiveness of Various Sub-Metering Strategies

SA-12-001 Simulations of Chilled Water Cooling Coil Delta-T Characteristics

SA-12-003 Bayesian Analysis of Savings from Retrofit Projects

SA-12-004 Energy Savings Potential of Variable Condenser Water Flow Systems

Zhiqin Zhang and William D. Turner.......................................................................................388

SA-12-006 ASHRAE/IES Standard 90.1 Metal Building U-factors—Part 5: Mathematical Modeling 

SA-12-007 Review and Economic Feasibility Study 

SA-12-008 Reliability Modeling for Data Center Cooling Systems

SA-12-009 The Development of Simplified Rack Boundary Conditions 

SA-12-012 Short-Term Curtailment of HVAC Loads in Buildings (RP-1390)

SA-12-013 Development of a Tool to Improve the Energy Efficiency 

SA-12-014 Laboratory Testing of a Fabric Air Dispersion System

SA-12-015 Ventilation Requirements for Refrigerating Machinery Rooms (RP-1448)

SA-12-018 Development of High Efficiency Home Heating and Humidification Technology

SA-12-019 Selection of Desiccant Equipment at Altitude (RP-1339)

SA-12-022 Parametric Investigation of PCM Thermal Properties on Temperature of Buildings in Toronto

SA-12-023 Field Measurements of Thermal Conditions During Surgical Procedures 

SA-12-025 Occupancy Simulation in Three Residential Research Houses

Study of the Degradation of Typical HVAC Materials, Filters and Components 

SA-12-028 Design of Roofs for Increased Solar Potential BIPV/T Systems 

SA-12-029 Achieving Net-Zero Energy Conditions in Retrofit—

SA-12-030 HVAC Systems Commissioning in a Manufacturing Plant

SA-12-031 Investigation of CO2 Tracer Gas-Based Calibration of Multi-Zone Airflow Models

Index of Conference and Technical Papers, Volume 118, Part 2..........................................................................803

SAN ANTONIO, TEXAS

ALAMO CHAPTER OFFICERS 2011–2012 ASHRAE CONFERENCES

Copyright American Society of Heating, Refrigerating and Air-Conditioning Engine

4. Using Energy Flow Analysis to Assess Seminar 1 (Intermediate)

Modeling the Performance of

Copyright American Society of Heating, Refrigerating and Air-Conditioning Engine

Tuesday with UFAD through the use of proper Cx

Federal, State and Local

Copyright American Society of Heating, Refrigerating and Air-Conditioning Engine

' = per cent difference