Helsinki University of Technology Laboratory of Steel Structures Publications 18

Teknillisen korkeakoulun terasrakennetekniikan laboratorion julkaisuja 18

Espoo 2000 TKI<-TER-18

PB2001-106397

1111111111111111111111111111111111

FIRE SAFETY DESIGN OF COMPOSITE SLIM FLOOR

STRUCTURES

Zhongcheng Ma

Dissertation for the degree of Doctor of Science in Technology to be presented with due permission for
public examination and debate in Auditorium Rl at Helsinki University of Technology CEspoo,
Finland) on the 7th of December, 2000, at 12 o'clock noon.

Helsinki University of Technology

Department of Civil and Environmental Engineering
Laboratory of Steel Structures

REPRODUCED BY: ~
u.s. Department of Commerce
National Technical Information Service
Springfield, Virginia 22161
Teknillinen korkeakoulu
Rakennus- ja ymparistotekniikan osasto
Terasrakenneteknii kan laboratorin
Distribution:
Helsinki University of Technology
Laboratory of Steel Structures
P.O. Box 2100
FIN-02015 HUT
Tel. +358-9-451 3701
Fax. +358-9-451 3826
E-mail: sinikka.rahikainen@huUi

© Teknillinen korkeakoulu
NTiS is authr.rized t~ reproduce and se1! tl1~s
report. Permission 1m IUlihG; r"pro~lIc!ion
must be obtained from tMp. cepyriU[1t OWilGf.
ISBN 951-22-5222-l
ISS N 1456-4327

Otamedia Oy 2000
Preface

This study has been carried out in the Laboratory of Steel Structures at Helsinki University
of Technology. I wish to express my gratitude to my superior, Professor Pentti Miikeliiinen,
who has enabled my study and provided the fmancial support for the research work.

My cordial thanks to assistant Lie. Sc. (Tech.) J. Kesti, who arranged the facilities for the
research work. Sincere thanks also to my colleagues Misters M. Malaska, J. Outinen and O.
Kaitila.

Special thanks are due to Mrs Sinikka Rahikainen, our secetary with a very kind heart.

Thanks to Mr. N. Kimberley for the language checking of the manuscript.

Finally I would like to thank my wife, Wei Lu, for the 'joint' research life. Uncountable
misses to my Mom and Dad, who have almost lost their son since the start of my overseas
studying life. Many thanks to my parents-in-law, who have encouraged me to study abroad, to
seek being a man with open insight.

This work has been financially supported by the Centre for International Mobility (CIMO),
the National Technology Agency of Finland (TEKES), The Finnish Constructional Steelwork
Association Ltd., and the Finnish steel company - Rautaruukki Oyj, which I wish to
acknowledge with gratitude.

Espoo, October 2000

Zhongcheng Ma
Abstract

A structural fIre safety design method for composite steel-concrete slim floor structures
was developed, including the characterization of fully-developed compartment fIre curves and
the equivalent fIre exposure, temperature analysis of the composite structures, and the
structural response analysis of composite slim floor structures subjected to fIre. Special
interest was given to the structural responses of the slim floor beam both as an isolated
member and a part of the frame structure. The mechanical interaction behind the global
deformation response of the frame was emphasized.

Although the analyzed objects concentrated on composite slim floor structures, the
according fIre safety design method and the explored structural response mechanism in fIre
are equally applicable to other types of building structures.

Keywords: Fire safety design, parametric fIre curves, equivalent fIre exposure, fIre resistance,
slim floor, composite frame, numerical modelling
Contents

Preface 2

Abstract 3

List of Publications 7

Author's Contribution 7

1 Introduction 8
1.1 General. ....................................................................................................................................... 8
1.2 Structural Fire Safety Design Methods .................................................................................. 9
1.2.1 Methods Related to Fire Resistance Only ................................................................... 9
1.2.2 Methods Related to Global Fire Safety ..................................................................... 10
1.3 Development of Composite Slim Floor Structures ............................................................ 11
1.4 .Aim of the Present Study ....................................................................................................... 12

2 Parametric Temperature-Time Curves for Structural Design 13

2.1 Behaviour of Fires in Buildings ............................................................................................. 13
2.1.1 Introduction ................................................................................................................ 13
2.1.2 Regimes of Burning ................................................................................................... 14
2.1.3 Rates of Burning ........................................................................................................ 15
2.1.4 Maximum Average Compartment Temperature and Fire Duration .................. 16
2.2 Fire Models for Structural Design ........................................................................................ 17
2.2.1 Parametric Fire Models for Structural Design ....................................................... 17
2.2.2 Parametric Fire Curves based on Heat Balance .................................................... 17
2.2.3 Parametric Fire Exposure in Eurocode .................................................................. 18
2.2.4 Parametric Fire Curves based on Experimental Approximation........................ 19
2.3 Equivalent Fire Exposure ...................................................................................................... 19
2.3.1 Equal Area Hypothesis ............................................................................................. 20
2.3.2 Equal Maximum Temperature of Critical Structural Part ................................... 20
2.3.3 Equal Load-bearing Capacity ................................................................................... 21
2.3.4 Equal Heat Load ........................................................................................................ 21
2.3.5 Equivalent Fire Exposure on Exposed Steelwork ................................................ 21

3 Temperature Analysis of Building Structures Exposed to Fire 23

3.1 Formulation of Heat Transfer ............................................................................................... 23
3.2 Computer Implementation .................................................................................................... 25
3.3 Thermal Response of Composite Slim Floor Structures to Fire ...................................... 26

4 Structural Behaviour of Slim Floor Construction in Fire 28

4.1 Introduction ............................................................................................................................. 28
4.2 Fire Resistance Analysis of Isolated Beams - Eurocode Methods ................................... 29
4.3 Numerical Analysis of Isolated Beams ................................................................................. 30
4.4 Numerical Analysis of Frame Structures ............................................................................. 31
4.4.1 Effects of Axial and Rotational Restraints on Beam ............................................... 32
 4.4.2 Effects of Fire Protection on Columns and Heat Rate on the Beam ................... 34
4.4.3 Effect of Composite Floor Slab ................................................................................. 35
4.4.4 Effects of Sway and Non-sway Frames ..................................................................... 36
4.4.5 Global Deformational and Mechanical Responses of Frame Structures ............. 36

5 Conclusions 40

Bibliography 42

Appendix A Parametric Temperature-time Curves of Medium Compartment

Fires for Structural Design 53

Appendix B Fire Resistance of Composite Slim Floor Beams 73

Appendix C Behaviour of Composite Slim Floor Structures in Fire 95

Appendix D Temperature Analysis of Steel-Concrete Composite Slim Floor

Structures Exposed to Fire 115

Index of the Author's Publications 126

List of Publications

This dissertation consists of an overview, lastest research results on the parametric flre
temperature - time model and its equivalent flre exposure, the structural behaviour of the
composite slim floor frame in flre, and the following publications:.

A. Ma, Z. and Makelainen, P. Parametric Temperature-time Curves of Medium Compartment

Fires for Structural Design. Fire Safety Journal, vol. 34, no. 4, pp.361-375,June 2000.

B. Makelainen, P. and Ma, Z. Fire Resistance of Composite Slim Floor Beams. Journal of
Constructional Steel Research, vol. 54, no. 3, pp.345-363, June 2000.

C. Ma, Z. and Makelainen, P. Behaviour of Composite Slim Floor Structures in Fire. Journal
of Structural Engineering, ASCE, vol. 126, no.7, pp.830-837,July 2000.

D. Ma, Z. and Makelainen, P. Temperature Analysis of Steel-Concrete Composite Slim Floor

Structures Exposed to Fire. The 6th • ASCCS International Conference on Steel-Concrete
Composite Structures - the past and the coming century, March 22-24 2000, Los Angeles,
volume I, pp.263-270.

Author's Contribution
Some of the research results on the parametric flre temperature - time model and its
equivalent flre exposure, and the structural behaviour of the composite slim floor frame in flre
which are presented in the context in more details, are published at the flrst time by the
author. The other research related to the topic by the author can be found in publications 109-
117 and 187-192. The research work and preparation of the manuscripts in all papers A, B, C
and D were carried out alone by Zhongcheng Ma. The other author, Professor Pentti
Makelainen, has acted only as the supervisor of the research project.
 8

Chapter 1

Introduction

1.1 General

Fire safety design is an important aspect of building design. A properly designed

building system greatly reduces the hazards to life and limits property loss. The research
on fIre safety design started as long ago as almost 80 years (Inberg 1928). Today the
concept of fIre safety design has been improved signifIcantly and become more rational.
Current fIre safety concepts are optimally integrated packages of passive structural,
active technical and active organizational fIre precaution measures, which allow wel1-
defIned objectives agreed by the owner, the fIre authority and the designer to be fulfllled.
Passive structural measures include fIre-resistant load-bearing structures, the fIre
protection of the main load-bearing elements and fIre compartimentation by floors and
walls. Active technical measures include automatic fIre detection, automatic fIre alarm
transmission devices, fIre-fIghting installations and automatic fIre suppression devices
such as sprinklers, etc. Active organizational measures include water supply, public and
on-site fIre fIghting services, installations for means of escape, training, emergency plans
and so on.
The fIre safety objectives in the present European Fire Codes are explicitly based on
the life safety objective [35]. However, protecting properties from loss due to the careless
behaviour of a neighbour is another important aim. According to the Construction
Product Directive (CPD) [34] in the clause concerning 'Safety in case of fIre', the
following essential requirements have to be fulfilled in the event of an outbreak of fIre
the load-bearing capacity of the construction can be assumed for a specifIc period of
time
the generation and spread of fIre and smoke within the works are limited
the spread of fIre to neighbouring construction works is limited
occupants can leave the works or be rescued by other means
the safety of rescue teams is taken into consideration
The central objective of fIre safety in the current Fire Codes is to confIne the fIre
within the compartment in which it started. They consist of a collection of requirements,
only or mostly related to the structural fIre resistance of load-bearing elements and to
walls and slabs necessary to guarantee the compartmentation.
It should be noted that the objectives of fIre safety are a historical concept, in which
the content can change with the development of fIre science. Besides, the additional
objectives (e.g. reparability) can also be implemented if the client or authorities require a
particular building or project.
 9

1.2 Structural Fire Safety Design Methods

Structural flre safety design contains two important aspects, including the
characterization of fully developed compartment flres and veriflcation of the ability of
structures to resist and contain flres. Currently, the design methods may be classifled into
two classes:
Methods related to flre resistance only
Methods related to global flre safety
The flrst category of methods concerns the veriflcation methods of flre resistance.
The structural Eurocodes are for the time being strictly limited to this category. The
second category is based on the flre risk assessment technology, which is being
developed for particular buildings, important structures or individual projects, in which
the flre safety objectives are required by the owner, authorities and designer.
Generally, the flrst category of methods is simple to design, but the safety margin for
individual building is uncertain. The second category of methods is complex to design,
but the safety margin is known and it is cost-effective as well.

1.2.1 Methods Related to Fire Resistance only

The methods related to flre resistance only are combined by fIre models and structural
models. The flre models deflne the evolution of air temperature, the convective and
radiative boundary conditions, and the spreading of flre in a flre-affected room if
possible. The structural models deflne elements or parts of the structures, thus allowing
the prediction of the temperature increase in the structure or in elements ensuring
compartmentation, of the collapse temperature or the collapse time for a given load.
Zone modelling and fleld modelling (Computational Fluid Dynamics) are two
prevalent modelling techniques used to simulate the flre scenario in a building
compartment. Several commercial software packages such as CFAST by NIST,
HAZARD by Harvard, JASMINE by BRE, and SOFIE by a consortium of European
flre research laboratories, have been developed for the modelling of flre in a building
compartment. However, in view of the structural flre safety design, only the post-flash-
over flre has a signiflcant influence on the structural stability and integrity. These
modelling techniques offer a high degree of accuracy in calculating the temperature
evolution of a compartment flre but are extremely sophisticated. They are not very
convenient tools for structural flre safety design.
An approximate method to describe the temperature evolution of a fully developed
compartment flre has been considered desirable. The current Eurocode 1 gives a group
of nominal flre curves including a standard flre curve (ISO 834), an external fIre curve
and a hydrocarbon fIre curve. These curves are only time-dependent. As an alternative,
Eurocode 1 also gives a set of parametric flre curves for relatively small flre
compartments. These curves depend on the flre load, opening factor, thermal properties
of boundaries and the geometry of the fIre compartment.
In the current codes, three kinds of structural models are normally used
Simple calculation on elements like column and beam, with simple support
conditions;
Taking account of a sub-system with appropriate boundary conditions;
Analyzing the global structural response to flre.
 10

Anyone of the three models can be taken for the thermal and structural calculation of
flre resistance. However, in the flrst model, the buildings are treated as a series of
individual members, and the continuities and interaction between these members are
assumed to be negligible. Throughout the 1990s, following the investigation of the fue
event in Broadgate (1991, UK), flre tests in William Street (1992, Australia), and full-scale
flre tests on an 8-storey composite steel-framed building in Cardington (1995, 1996, UK),
it was found that the structural member in the frame exhibited significantly better
behaviour in fire than that in the standard fire-resistance test. The first model is very
conservative by disregarding the interaction between members. The fire event and tests
also highlighted that the first model, although conservative, did not address the true
behaviour of the building structure in fire, since the building was not acting as a series of
individual members. However, this model is simple to design.
The third model is an advanced analysis method by considering the building structure
as a whole. The restraints by the surrounding parts and interaction between the structural
members can be taken into account. However, the analysis is complicated and time-
consuming. For the time being this model is mostly used for research purpose or for the
individual project. The second model is in between the other two.

1.2.2 Methods Related to Global Fire Safety

As far as this category is concerned, the probabilistic and statistical methods are used
to determine the probability of occurrence and the severity of fire in a building. The
active fue measures such as heat/smoke detection, fire alarm, sprinklers, automatic
closure of doors leading to compartmentation and flre-flghting services, in conjunction
with the risk of fire occurring in the building - such as in a painting workshop or a
swimming pool -, and the mobility of people in the building, will affect the severity of the
fire design (including intensity and duration).
There are several different fire risk assessment methods developed respectively by
Switzerland, Australia, the USA, Denmark, Canada and the ECCS. The ECCS method is
one based on the logic tree analysis, which concerns with the probabilistic evaluation of
the occurrence and spreading of a fire scenario in a building. In this method, the event
'fue'is divided into sub-events, which are easier to analyze. The occurrence probability of
a fully developed fue can be evaluated, and thus the probabilistic fire curve can be
defined.
The structural response to the possible fire can be analyzed by the advanced structural
analysis techniques described above. The stability of the structures and integrity of the
compartmention can be checked according to the criteria of the codes.

1.3 Development of Composite Slim Floor Structures

In recent years, increasing interest has been shown throughout Europe in developing
and designing shallow floor systems in steel-framed buildings. In the
 [J
11

,p.
(<I) (b)

I (c) (d)
Fig. 1.1 Different Types of Slim Floor Beams: (a) Thor-Beam; (b) Delta-Beam; (c) Slimflor
Beam; (d) Asymmetric Slim Floor Beam.

shallow floor system, the steel beam is contained within the depth of the pre-cast
concrete floor or composite slab with proftled steel decks. This form of construction
achieves a minimum depth of building and the flat floor is beneficial because the building
services can be run in any direction. The shallow floor can be designed using various
forms of steel beams comprising either rolled or welded sections, which are called 'slim
floor beams'.

The key features of the composite slim floor construction are the steel beam and the
type of steel deck. One of the original slim floor concepts developed in Scandinavia was
the 'Thor-beam' (Fig. 1.1 a), which consists of two channel sections welded to a flat plate.
Additional angles are welded to the top flanges to provide the shear connection. Another
concept is the 'Delta-beam' (Fig. 1.1b). The slim floor system consisting of the 'Delta-
beam' and prefabricated concrete slab is very popular in Finland. In 1991, British Steel
and The Steel Construction Institute (SCI) developed a Slimflor® beam (Fig. 1.1 c),
which consists of a universal column section welded to a steel plate. Recently, interest
has been concentrated on the asymmetric hot-rolled steel beam (Fig. 1.1d) in the UK,
and on the asymmetric welded steel beam in Finland.

The slim floor beam offers good fire resistance because most of the steel beam is
encased within the depth of the concrete slab. Several fire tests performed by SCI (Steel
Construction Institute) in the UK [95,127,131] have shown that the composite slim floor
beam can achieve 60 minutes' fire resistance without any additional fire protective
measures if the load ratio is less than 0.5. To achieve 90 minutes' fire resistance, only a
very thin protection underneath the lower flange is needed [B].

Generally, the conventional steel-concrete composite beam is well established for

longer spans (> 9 m), but the slim floor beam creates more opportunities for steel in
spans of 5-9m. It also achieves a slab depth of 300 mm or so, which is much less than
that in conventional steel construction. This issue increases the competitiveness of
composite slim floor construction with the concrete flat slab system. On the other hand,
compared with the conventional composite frame system that has a primary-secondary-
beam system, the new slim floor frame has a rather precise structural form
 12

Primary Beam Slim Floor Beam

~ H
I
H,I
I
I
,
I
,I
I
I I
I
I ,
I

Secondary I
,I
I

a Beam a I
I
I
I
c- oo : Tie Tie:,
~ ~ ..;.
11
I
'0 I
I
Concrete Slab Concrete Slab I
I
I
I
I
I
I
I
I
I I

~ ~ H H
Ie 6- 9m
;001 Ie 5- 9m
~I

(a) (b)

Fig. 1.2 Composite Frame Systems: (a) Traditional Primary - Secondary Beams System; (b) Slim
Floor System.

(Fig. 1.2). In slim floor construction, the slab is supported directly by the primary beam,
and forms a part of the composite beam to work together with steel beam. Between the
rows of the single frame, the tie members are employed to link them together and
maintain the out-of-plane stability of the frame.

1.4 Aims of the Present Study

The aim of the present research project is to investigate fIre resistance of the
composite asymmetric slim floor beam both as an isolated member and as a part of the
frame using numerical analysis methods. As an isolated element, the research includes the
fIre resistance of the new beam and the effects of enhanced measures such as additional
reinforcements and fIre paints underneath the bottom flange. In the frame analysis, the
research focuses on the realistic behaviour of the slim floor frame as a whole in fIre. The
effects of frame continuity and its quantifIcation, mechanical interaction between the
structural members and the role of floor system on the stability of the whole building in
fIre have been investigated.
Another aim of the study is to develop a new parametric temperature-time curve for
structural design, or improve the present parametric curve in Eurocode, to consider the
situation in the fuel-controlled region and so on. The according method of equivalent fIre
exposure will also be developed.
 13

Chapter 2
Parametric Temperature-Time Curves for
Structural Design

2.1 Behaviour of Fires in Buildings

2.1.1 Introduction

The temperature-time curves in a fIre compartment designate the characteristics and

intensity of fIre developing process. A typical temperature development in the
compartment is shown in Fig. 2.1, which includes three phases: fIre growth, full
development and decay.

In the fIre growth phase, the room temperature is low and the fIre is local in the
compartment. The period is important for evacuation and fIre fIghting. Usually, it is not
of signifIcant influences on the structures. After flashover, the fIre enters into the fully
developed phase, in which the temperature of the compartment increases rapidly and the
overall compartment is engulfed in fIre. The highest temperature, highest rate of heating
and largest flame occur during this phase, which gives rise to the most structural damage
and much of the fIre spread in buildings. In the decaying period, the temperature
decreases gradually. It is worth pointing out that this period is also important to the
structural fIre engineering because for the R.C. structures, composite structures, insulated
steel structures and unprotected steel structures of low section factor, the internal
temperature of cross-section will still increase signifIcantly even though it is the decaying
period.

The efforts to quantify fIre characteristics can be seen as far back as 1928 in Inberge's
study [77]. From efforts spanning almost 80 years, considerable knowledge about the fIre
has been obtained. The numerical modelling of fIre scenarios for medium and large
compartments has also produced many successful experiences. A one-zone model, multi-
zone model and computational fIeld model (CFD) have been developed and commercial
software is available. However, the overview in this section will only focus on those
relevant to the characterization of fIre temperature evolution in a compartment.

The factors that have a signifIcant influence on compartment fIres are fIre load and
type, ventilation opening, and the geometry of fIre compartment. The thermal properties
of the compartment boundary also have some effect on the fIre, and the variation on the
maximum temperature is of the order of 10% [76,94].
 14

flash-over:

gr:awth phase
1
to to +7: TilDe

Fig. 2.1 Typical Temperature Development in a Compartment.

2.1.2 Regimes of Burning

The fIrst systematic study of the behaviour of the fully developed compartment fue
was carried out in Japan in the 1940s by Kawagoe et al. [81,82]. In the study, the burning
rate of wood cribs was found to depend strongly on the size and shape of the ventilation
opening, which was called 'ventilation-controlled fIres'. However, Thomas et al. [170]
found that, if the ventilation opening was enlarged, a condition would be reached beyond
which the burning rate would be independent of the size of the opening and would be
determined instead by the surface and burning characteristics of the fuel, which was
called 'fuel-controlled fues'.

Harmathy [73] analyzed data from a large number of wood crib fues in a
compartment and plotted the results as R/ A fuel versus P.g!/'~!/'/ A fuel ' This plot shows a
clear distinction between the 'ventilation-controlled regime' and a 'fuel-controlled regime'
in which the burning rate R is independent of the ventilation factor (Fig. 2.2). Harmathy
recommended the following conditional equation

!/'A .Jh
P ag IV = 0.235 ~ 0.290 (2.1)
A/uel

where Pa is the density of air (kg/ m~; g is the acceleration due to gravity (m/s~; ~ and h
are the area (m~ and height (m) of the opening respectively; Afue1 is the surface area of the
fuel (m2).

2.1.3 Rate of Burning

The rate of heat release, which is the product of the rate of burning and the calorifIc
value of fuel, is one of the most important characteristics of a fIre. In some one-zone and
multi-zone modelling, the rate of heat release of the fuel has to be input as the basic
given variable. It represents the released heat by the fuel per unit of time
 15

Sea I e

~ full

PagX,A""hY.!/A'fue! kg/ m2 s

Fig. 2.2 Identification of the Transition Region between Ventilation-controlled and Fuel-
controlled Fires for Wood Cribs [73].

In fact, in most experiments the rate of weight loss of the fuel was measured, rather than
heat release; and although the rate varies during the course of a fIre, it was found that
during the fully developed period it had a steady value R. Generally, R was defIned as the
average burning rate of fuel beginning when the fuel mass had dropped to 80% of its
original value and ending at 30%. This period is also known as a 'primary burning'.

The earlier experimental results in Japan in the 1940s and later in the former USSR,
the USA and the UK showed that R was mainly controlled by the rate of airflow that
enters the compartment. This led to the well-known equation R=O.1A.J1'1, (kg/s). Later,
this equation was verifIed by the theoretical derivation of heat balance in a fIre
compartment, under the specifIed restraint conditions [3]. Harmathy analyzed the
experimental results for wood cribs and found that the rate of burning was well
correlated to the surface area of fuel for fuel- controlled fIres (Fig. 2.2). Further analysis
on CIB fIre tests by Law et al. [91,92] showed that, for ventilation-controlled fIres, other
features should be taken into account, notably the area of the enclosing surfaces of the
compartment and the depth D of the compartment in relation to the width W of the wall
containing the window (Fig. 2.3).

In general, the rate of burning of wood-based compartment fIres can be expressed as:

For ventilation-controlled fIres,

R = 10.8· (1- e.xp(-0.036 .7]))' Aw Jh. ~W / D (2.2)

and for fuel-controlled fIres as,

R = 0.372· A fllel (2.3)

where R is in kg/ s.
 16

0.15 "

"'e
~
=-
--
~

0\>:: 0.10

~I\
n<:(

0.05

large-scale data:
o BoreharnwoOC1
+ MetZ
C Carteret

Fig. 2.3 Rate of Burning for Ventilation Controlled fIres [92, 168].

2.1.4 Maximum Average Compartment Temperature and Fire Duration

The air temperature and fIre duration are two important factors in characterizing flre
severity. The maximum average air temperature represents the intensity of a
compartment flre. Ma and Miikelainen [A, 114] analyzed the data from fully developed
compartment fue tests for wood-cribs flres, and found that for fuel-controlled flres, the
fue intensity had a good correlation with the ratio of opening factor AJl% / A.. (where A..
is the total surface area of compartment boundary, m2) and the critical opening factor
(determined by Eq. 2.1). The following equations were given
For ventilation-controlled flres,

Tgmax =1240 -11'11 (2.4)

and for fuel-controlled flres,

(2.5)

where Tgmax is the maximum average air temperature in a compartment (0C); 11 = A..
/ AJl'!2 , the inverse of opening factor (m-'!,); 11cr is the value of 11 in the critical region,
which can be obtained from Eq. 2.1. T gmcr is the maximum temperature in the critical
region (0C), which can be obtained by substituting 11cr into Eq. 2.4.

Thomas et al. [172] gave an expression on the maximum air temperature in a

compartment flre, after the analysis of data from CIB small-scale flre tests. However, the
critical regime of two kinds of flres was ill deflned. In the expression, the critical regime
was dependent only on a constant of opening factor (11=12.5 m-'!;, which was not
correlated to the fuel.
 17

--.
~
Q,l
1200

....= 1000
.
eo:
Q,l
Q. 800
e
Q,l
f-< 600
II)
eo:
t.:l
e 400
=
e
.~ 200
eo:
~ 0
0 10 20 30 40 50

I nverse 0 ·
f opening f actor '1 (Dl-'h)
Fig. 2.4 Maximum Average Gas Temperature in a Compartment [A]. (The other parameters:
the ratio of floor area to the total surface area enclosing the compartment is 3/14, and
the surface area of unit weight of fuel (equavilent woods) is 0.15 m 2 /kg).

Although fIre duration is an extensively used phrase, it is not clearly defIned up to

today. Intuitively, the fIre duration could be defIned as

r=GjR (2.6)

where G is the total weight of fuel (equivalent woods, kg).

2.2 Parametric Fire Models for Structural Design

2.2.1 Nominal Fire Curves

The nominal fIre curves are a set of temperature-time relationships, in which no

physical parameters are taken into account. The main purpose of the prescription of the
nominal fIre curves was to make the fIre resistance tests reproducible. The ability of fIre
resistance of building elements can be evaluated under the same heating curve. The
current Eurocode gives three nominal fIre curves: ISO standard fIre curve, external fIre
curve and hydrocarbon fIre curve. The fIrst two are for cellulosic fIre and the third is for
hydrocarbon based fIre.

2.2.2 Parametric Fire Curves based on Heat Balance

Several groups (for instance, Kawagoe's, Pettersson's and Babrauskas's) have

attempted to predict the temperature-time history in a compartment fIre, for the purpose
of structural fIre resistance design, based on the heat and mass balances of the
compartment. The earlier curves by Kawagoe et al. [82] were derived only from the heat
balance equations. The constant rate of burning was used and it was assumed that there
was an unlimited supply of fuel. Pettersson et al. [142] complemented the mass equation,
and used the realistic rate of burning derived from the measured data. Nevertheless, the
rate of burning in his analysis was still only ventilation controlled. Using this model, a set
of temperature-time curves was developed for the convenience of the designer, which
were in the parameters of fIre load density, opening factor and thermal properties of
compartment boundary.
 18

Franssen etal [64] attempted to take account of the case for fuel-controlled fires as
well. A one-zone model was used and the combustion model could defIne the burning
regimes. The computer program, OZONE} was coded and a set of temperature-time
curves in the parameters of fIre load density and opening factor were calculated. These
curves were stored on the PC disk in the form of differentiated data.

Another attempt made by Babrauskas et al. [5] gave an approximate mathematical

expression by fitting the formulae with the calculated results using the COMPF2
computer code. This method requires the rate of burning as input, which is still difficult
to determine nowadays.

2.2.3 Parametric Fire Exposure in Eurocode

With the thinking that the methods (Kawagoe's, Pettersson's and Babrauskas's) were
sophisticated and that the assumptions and uncertainties associated with the models were
easy to make them premature, Lie (1974) suggested that it is only necessary flnd a fIre
temperature-time curve 'whose effect, with reasonable probability, will not be exceeded
during the use of the building', if the objective is to develop a method of calculating fire
resistance. He developed an expression based on a series of temperature-time curves
computed by Kawagoe et al. It certainly inherited the assumption of unlimited fuel
supply made in Kawagoe's computation, before the fIre duration was reached.

Having a similar idea to Lie's, Wickstrom developed an expression by 'applying the

standard fire curve for expressing natural fires for design purposes'. The current Annex B
of Eurocode 1 [55] is based on his work. Compared with Lie's expression, which is only
dependent on the opening factor, the method in Eurocode takes account of all the four
signifIcant fIre parameters. However, it is very poor for fuel-controlled fIres [114].

The original temperature-time equation in Annex B of Eurocode 1 is given by

Tg =1325· (1- 0.324e- O2

. 1" - 0.204e -1.71" - 0.472e -19/* ) (2.7)

where t*=r·t, t is time in hours; r =(0/ b)2 /(0.04/1160)2,0 is the opening factor (mY'),
b is the thermal properties of enclosing boundary (= ~pA,C, Jm2sY'K). The time
corresponding to maXlffium temperature is td = 0.13 X 10-3 • q
1 /0, ill which
ql = q' A f At' q is the fIre load density per unit floor area (MJ/m~. After the
/

maximum temperature is reached, the temperature declines linearly and the slope
depends on the value of td* (=r·tJ.
 19

2.2.4 Parametric Fire Curves based on Experimental Approximation

An attempt has been made by the authors [A, 114, 116] to 'produce a simple
calculation procedure which will, to a suitable degree of precision, estimate the
temperature history in a fully developed compartment ftre'. TIlls model should take
account of both kinds of ftres (fuel and ventilation controlled).

It has been recognized that the major effect of ftre on a building structure depends on
the combination of ftre intensity (maximum average air temperature) and ftre duration
[46, 73, 76]. It was expected that a general shape function in the parameters of the ftre
intensity and duration could describe the temperature history in a compartment ftre. The
physical parameters such as ftre load and opening factor would only have an influence on
the ftre intensity and duration.

Based on the analysis of the measured ftre curves for fully developed cellulosic ftres,
the shape function was given by

(2.8)

where Tg and Tgmax are the air temperature and the maximum air temperature,
respectively; To is the room temperature; t and t max are the time and the time when the
maximum temperature is reached, respectively; 8 is the shape constant of curve, usually
0.5 for the ascending phase and 1.0 for the descending phase; t",ax is a function of ftre
duration 'to T gmax and 't can be obtained from Eq. 2.1-2.6.

Comparisons with the measured curves showed that this model ftts well for both fuel-
controlled and ventilation-controlled ftres.

2.3 Equivalent Fire Exposure

Due to the large variety of possible temperature-time curves in buildings, the

assessment would be very expensive if the building elements were tested for each
particular ftre curve. Therefore, a nominal standard curve is necessary in some sense. The
concept of equivalent ftre exposure is a bridge between the realistic ftre curve and the
standard ftre curve.

The equivalent ftre exposure is the time during which a specifted compartment or
structure is submitted to ISO standard ftre in order to obtain the same severity (effect) as
the real ftre curve. TIlls concept has been developed into several different methods and
the key difference lies in the deftnition of 'severity' or 'effect'. It is primarily intended as a
measure for the rating of ftre resistance for a compartment.
The development of equivalent ftre exposure surrounds the selection of the 'indicator'
for the equivalent, which represents the ftre 'severity' or 'effect'. So far, there exist four
indicators, including the area of ftre curve, the maximum temperature of the critical part
of the structures, load-bearing capacity and heat load.
 20

2.3.1 Equal Area Hypothesis

The basic idea of the equivalent area hypothesis is that the area of the temperature-
time curve under standard fIre above a certain baseline (lS0 or 300°C) corresponding to
equivalent fIre exposure should be equal to that under real ftre.
According to this principle, Inberge obtained the first quantified relationship of
equivalent fire exposure and fire load, by burning office furniture and papers in a room
and measuring the temperature attained. Inberge also realized the importance of
ventilation; he did not quantify its effects, but adjusted it to give what he deemed to be a
severe condition. His work formed the starting point of the current regulations of fIre
gradings:

te = k j • q (min) (2.9)

where q is the fIre load density of equivalent woods per unit floor area (kg/ m\ and kl is
the coefftcient, which is of the order of unity.
However, Inberg's original hypothesis cannot be justifted theoretically. Actually, the
product of temperature and time is not the heat as expected. Because the heat transfer
occurred during fIre is dominated by radiative heat, the dependence of radiative flux on
~4 makes simple scaling very diffIcult (where ~ is the gas temperature of the ftre
compartment) .

2.3.2 Equal Maximum Temperature of Critical Part in Structural Elements

Law (1973) [91] attempted to seek another way to express the severity of a
compartment fIre. She simplifIed the real fIre as a constant temperature, which equalled
the maximum air temperature, sustained for a period of time as ftre duration (Eq. 2.6).
Using the simplifIed natural fIres, she analyzed the maximum temperature of insulated
steel columns exposed to standard ftre and real fIres. The equivalent fIre exposure was
given by

(2.10)

where q is the fIre load density per unit floor area (kg/m~, k2 is a coeffIcient.
Equation 2.10 can also be written as

Ie = c· q' W (2.11)

where w is related to the ventilation conditions, and c is a constant coeffIcient. This type
of equation has been extensively used by the Eurocode, CIB W14, Din 18 230 (1998) and
the Code by New Zealand, except for using a different value of c and expressions for w.

2.3.3 Equal Load-Bearing Capacity

In this method, the residual sectional capacity of elements is taken as the equivalent
indicator. Provided that the lowest sectional capacity under an overall fIre process is R,
then the time te corresponding to R under the standard fIre curve, is defIned as the
equivalent fIre exposure.
 21

Based on this defmition, the authors [109,114, 116] carried out extensive parametric
studies on the RC columns, in which the parameters are the characteristic ratio of steel
a = p fy / fern (0.05-0.5), width of the cross-section B (0.3-0.6 m), the aspect ratio H/B
(1-3), the thickness of protection d for reinforcements (10-70 mm). The following
relationship was obtained

(2.12)

where T gmax is the maximum gas temperature (0C) and 'l' is the fire duration (min), both
of which are ventilation and flre load dependent. k3 is of the order of 8x1 0-6•
The equivalent time by Eq. 2.12 is related only to the flre behaviour itself. In this
equation, te is the function of the product of flre duration and Tg4max , not the usual T gmax.
It is quite rational in such a case that the heat transfer under flre is dominated by
radiation.

2.3.4 Equal Heat Load

Since the damage on the structures by flre is caused by the absorbed heat, it is also
natural to use the heat load during the flre duration as an indicator. This idea has been
developed since the earliest study, but was changed into the version of equal area
hypothesis, for the purpose of simplicity. The author [109] made a study on the R.c.
structures by equating the absorbed heat load under real flres and standard flre. The
following equation was given
3 , /3
t e =k 4 . (r·Tgmax (2.13)

where k4 is of the order of 4x10-6•

Theoretically, Equation 2.13 is also suitable for composite structures and steelwork
although they are developed from R.C. structures, because their basis is not strongly
related with the materials of elements, according to the theory of NHL (Normalized
Heat Load) uniformity [64,74]. The problem is that they are perhaps not suitable for
exposed steelwork, since exposed steelwork is more sensitive to the maXlmum gas
temperature than to the heat when the duration reaches a certain value.

2.3.5 Equivalent Fire Exposure on Exposed Steelwork

Since exposed steelwork is more sensitive to the maximum flre temperature than
other building materials, the principle of 'equal maximum temperature of the critical part
in structural members' (section 2.3.2) is reasonable to use. The method in Eurocode 1
Part 2.2 employs this principle. Figure 2.4(a) shows the comparison between numerical
analysis [114] and direct calculation using Annex E of Eurocode 1 Part 2.2. In the
numerical analysis, the parametric flre curves in Eurocode have been used as input. It can
be seen that there is a large discrepancy. This reflects the inaccuracy in the current
formulae for equal flre exposure on exposed steelwork in Eurocode.
 22

90.-------~~--~--~------~ 90

60
•
1
__________ :6 __ d-~_g_o_-2.--C-iJ.---
,
p
0 D 1:.
-------------- I
';:!; 60
o
__ 0- ___ ..9.

°0
N
0' c:
Q
L· ,
-~ -- - - - -- - - -~- of /V=30 ~
c F /V=60
6-
r.l
AF /V=90
....
.c

0
30 60 90 30 60 90
by Numerical calculation (min) by Numerical calculation (min)

(a) (b)

Fig. 2.5 Equivalent Fire Exposure for Exposed Steelwork - Comparisons between Equation
Calculations and Numerical Analysis: (a) Eurocode Method; (b) Proposed Method
by the Author [114].

A study by the author using the same principle has been carried out to obtain a high
accuracy for exposed steelwork. In the study, the parametric flre curves according to Eq.
2.8 were used as the input of natural flre curves. The results showed that the index of the
fue duration is section factor dependent. The following equation was given

te = k 5 . -r a . T1.3 gmax
(2.14)

where a = 0.267 + 5.87 . ~ , in which V IF

is the inverse of the section factor; V is the
F
volume of unit length, and F is the surface area of the steel of unit length. k5 is of the
order of 1.25x1O-3•

Equation 2.14 is quite similar to Eq. 2.12 in the case of a lower section factor. The
equation here especially takes account of thin-walled steel members, whose inaccuracy in
Eq. 2.12 is evident due to the coverage of the analyzed structures. On average, the
resultant variable in Eq. 2.14 is the product of fIre duration and max , where b is in T:
between 2 - 4, dependent on the section factor. This indicates that for steelwork of a low
section factor, the temperature of the steelwork is radiation-dominated, and for steelwork
of a high section factor, the temperature is dependent both on the radiation and
convection. Figure 2.5 illustrates the comparison between the analysis results and the
results using Eq.2.14.
 23

Chapter 3
Temperature Analysis of Composite Slim Floor
Structures Exposed to Fire

3.1 Formulation of Heat Transfer

The transient heat transfer equation based on Fourier Law is given as

_VT (AVT)+e=O (3.1)

where T is the temperature, A is the thermal conductivity matrix, and e= de / dt is the

rate of specific volumetric enthalpy change. The gradient operator V is defined as

a a a]T (3.2)
V = [axay az
where x,y and Z are Cartesian coordinates. Superscript T is the transfer operator of the
vector.
In principle, the above equations can be derived as follows, by using the finite
difference method

(3.3)

where pc is the volumetric specific heat, Pwcw is the volumetric specific heat of water,
and cp is the volume ratio of moisture content. T; is the temperature of the ith element in
the local system (Fig. 3.1). R)s the thermal resistance term, which can be given by

1 1 1
R.=-+-+- (3.4)
l Ki Ko K int

where Ki and Ko are the thermal conduction terms. K,itt is the interface resistance
coefficient. The sensitivity analysis in [110] has shown that the value K int =50 W / m2K
could be used for the interface coefficient between steel and concrete. In the case of
perfect contact. Kin! = 00 can be presumed.
 24

L::..X
Y
1- -I
2 •

1. • ·3
~
of. •

Fig. 3.1 Local Mesh of Control Volume for Finite Difference Analysis.

A generalization of Eq. 3.3 for the fInite difference analysis was presented in [D].
Besides, Eq. 3.3 was derived for the internal elements. The complete formulation of heat
transfer also includes the equations for the boundary elements and the auxiliary equations
[D,110].
The heat input rate q is expressed as

(3.5)

where Tg is the air temperature of fIre and Ts is the surface temperature of the structures.
K is the coeffIcient of total heat transfer. According to ECCS Recommendations [26], it
can be determined by

1 1 d
-=-+- (3.6)
K a A

where d is the thickness of insulation. A is the conductivity of insulation material; and a

is the resultant convective factor, which can be given by

(3.7)

where a c is the coefficient of heat transfer due to convection from fire to the exposed
surface of the member; a r is the coefficient of heat transfer due to radiation; a is the
Stephan-Boltzmann constant, and E is the resultant emissivity coefficient.

The effect of moisture in concrete was modelled by the modified specific heat of
concrete before its temperature reaches to 120°C. After that the absorbed heat is totally
used to evaporate the water until the water content of the element becomes zero. The
formulation of the volume change of water within the elements has been given in [D,
110].
 25

3.2 Computer Implementation

According to the above fonnulation of heat transfer, a temperature analysis program

(TACS-FIR) for composite structures using the fInite difference method was written by
the authors [110]. The temperature distribution and history of both composite and
concrete structures can be analyzed. The fIre curves can be ISO standard flre or any type
of natural flres. The program has been validated by the fIre tests on composite slim floor
beams (280ASB and 300ASB) by BRI [95], concrete-filled tubular columns by VTI [137]
and concrete columns by HUCEA [121]. A comparison of the analysis results between
this program and the TASEF commercial software was also made. Using this program,
the temperature of composite structures in flre can be evaluated without the necessity of
testing.

The computer program was coded in Fortran 77. The geometrical divisions of the
current version are limited to the rectangular grids. The material thermal properties at
elevated temperature can be selected by the user inputting the material ID according to
the different codes or researchers, which were summarized in [110]. The material
properties can also be input in the form of differentiated data deflned by the user. In that
case it is assumed to vary piece-wise linearly with temperature between the neighbouring
data points. The conductivity of concrete can be specifIed to remain approximately as
that at maximum temperature instead of the current temperature in the cooling phase.
The critical time increment is usually very short, which is mainly dependent on the
minimum dimension of the cells in the steel web. The theoretical criterion was derived in
[D, 110]. A warning message will occur and the proposed value will be given on the
screen if the inappropriate time increment was given.

The flre temperature history may be defIned as a number of differentiated points on a

temperature-time curve. Temperature between these points is obtained by internal
interpolation. Several flre models described in Chapter 2, which are of the mathematical
expression, are provided by the program and can easily be selected by inputting the
according ID. These flre models are ISO standard fIre (lD=2) , parametric flre curves
according to Annex B of Eurocode 1 Part 2.2 (lD=3) and the proposed flre model by
the authors (lD=4).

The explicit forward difference scheme described in the previous section is used. The
temperature at the central point represents the temperature of the elements. The
maximum temperature of each element is specially stored in a vector for use in
calculating the material properties in the cooling phase if necessary (for concrete only).
The element temperature is output at the specifled times.

The major structure of this program is illustrated in Fig.3.2 and the major routines are
shown in Table 3.1. The invoking relationship of these routines and the main variables
are described in [110].
 26

TABLE 3.1 Functions of Subroutines in TACS-FIR.

No. Subroutines Description oljunctions

1 INPUT Input the geometric, physical and other parameters of the object for
analysis
2 FIR_CAS Handle the fIre temperature-time curves. Including a small collections
of several fIre curves
3 TEMPER Analyze the temperature of internal elements
4 THER_RES Calculate the thermal resistance items of internal elements
5 MATERIAL Calculate the thermal properties of steel and concrete at elevated
temperature
6 WATER Calculate the evaporation of free water content in concrete for
internal elements
7 BOUNDARY Analyze the temperature of boundary elements and handle all
boundary conditions
8 B_RES Calculate the thermal resistance items of boundary elements
9 B_WATER Calculate the evaporation of free water content in concrete of
boundary elements

TElVIPER IvIATEFlAL

WATER

'---~ BOUNDARY t-----I1Jo-I l\o'lA TEFlAL

Fig. 3.2 Structure of the Computer Program TACS-FIR [110].

3.3 Thermal Response of Composite Slim Floor Beam to Fire

Figure 3.3 shows the temperature distribution of the Finnish composite slim floor
beam with and without flre protection underneath the bottom steel flange at 60 minutes'
fIre exposure.
 27

49 48 47 44 40 35 30 28 29 31 33 34 49 48 47 44 40 34 29 26 24 25 25 25
57 56 55 52 46 39 33 30 31 34 37 39 57 56 54 51 46 38 32 27 26 26 27 27
77 77 74 70 61 50 39 34 77 76 74 69 61 49 38 31
113 112 109 102 89 68 50 41 113 112 109 102 88 68 49 37
172 170 166 156 134 99 67 48 172 170 166 156 134 98 66 45
265 263 258 244 210 146 91 59 265 264 258 244 210 145 90 56
418 416 411 393 341 218 124 74 417 416 410 392 341 217 123 71
687 685 681 663 594 325 165 92 686 685 680 662 593 323 163 88
481 210 111 479 207 106
563 246 131 560 241 123
605 274 151 600 265 139
631 299 174 622 285 155
653 328 206 636 303 174
679 369 253 649 325 201
714 432 327 659 356 241
7

(a) (b)
Fig. 3.3 Temperature Distribution of Composite Slim Floor Beam at 60 minutes' ISO
Fire Exposure: (a) without Fire Protection; (b) with Fire Protection
(intumescent fIre paint of 250 11m).
 28

Chapter 4
Structural Behaviour of Slim Floor Construction
in Fire

4.1 Introduction

When subjected to ftre, steel and composite structures lose their loading capacity and
stiffness. To ensure safety to the life and prevent property loss, the indispensable fire
resistance of the building is required by the authorities [36]. Traditionally, the fire
resistance ability of the structural members was tested using the isolated element heated
by the ISO standard fire. In this methodology, buildings were treated as a series of
individual members, and the continuities and interaction between these members were
assumed to be negligible. Consequendy, most of the structural members need to be
protected by the insulation materials, such as intumescent paints and fire protection
boards, in order to achieve the required fire resistance.
Throughout the 1990s, following the investigation of the fire event in Broadgate
(1991, UK), fire tests in William Street (1992, Australia), and full-scale fire tests on an 8-
storey composite steel-framed building in Cardington (1995, 1996, UK) [79,162], it was
found that the structural member in the frame exhibited significandy better behaviour in
fire than that in the standard ftre-resistance test. The standard fire test was very
conservative by disregarding the interaction between members. The fire event and tests
also highlighted that the current Codes, although conservative, did not address the true
behaviour of the building structure in ftre, since the building was not acting as a series of
individual members [8-9,52-53, 131, 151].
In recent years, increasing interest has also been shown throughout Europe to
develop and design a shallow floor system in steel-framed buildings [95, 128, B, C, 110,
111, 112]. In the shallow floor system, the steel beam is contained within the depth of the
pre-cast concrete floor or composite slab with proftled steel decks. Recendy, interest has
been concentrated on asymmetric hot-rolled steel in UK [130] and on the asymmetric
welded steel beam in Finland [B, C, 110-113].
The aim of this research is to investigate the fire resistant behaviour of the Finnish
asymmetric slim floor beam both as an isolated member and as a part of frame. The
research results have been published in [B, C, 110-112]. The thermal analysis of the
composite slim floor beam was included in publications [D, 110]. The fire resistance of
the composite slim floor beam as an isolated member, with varied enhanced measures
such as ftre protection additional reinforcements, was described in publications [B, C,
110, 111]. The effect of various axial restraints on the fire resistance of isolated members
 29

200

400

Fig. 4.1 Section Shape of the Composite Finnish Asymmetric Slim Floor Beam.

has been published in [111]. Effects of the axial and rotational restraints in two-
dimensional subframes and the effects of composite floor slab were preliminarily studied
in [C, 111]. Both the deformation and mechanical responses to fire have been explored.
For further research on the behaviour of the slim floor frame in ftre, the overview in
this context (Section 4.4) concentrates on the latest research results, which have not yet
been published due to time issues. The content includes the global and local deformation
response of the whole frame to ftre, and the force and moment distribution and variation
with time of the frame members. The mechanism of the beneftcial and detrimental
effects of the slim floor beam as a part of frame was explored. The effect of columns
being protected and unprotected, and effect of a sway and non-sway frame were also
investigated. The effect of the composite floor slab was further studied by three-
dimensional frame analysis.
Most of the research conclusions for the frame analysis were also applicable for the
other types of composite frame structures in ftre conditions, even though the analyses
were made on the composite slim floor structures in this context.

4.2 Fire Resistance Analysis of Isolated Beams - Eurocode Methods

Figure 4.1 shows the section shape of the Finnish asymmetric slim floor beam. The
slim floor steel beam consists of three plates welded together. The steel deck is a Finnish
product, Rannila 153.
Using the temperature distribution of the beam section in ftre described in Chapter 3,
the plastic moment capacity of the slim floor beam at elevated temperatures was analyzed
by the moment capacity method. The mechanical properties of steel and concrete under
high temperatures are given according to Eurocode 3 and Eurocode 4. The studies have
shown that the composite slim floor beam has a 60-minute' ftre resistance if the load
ratio is lower than 0.47. In practice, the load ratio of 0.5-0.55 is the most common when
an offtce or residential building is subjected to ftre. To achieve the desired ftre resistance,
some enhanced measures such as additional reinforcements and ftre protection on the
underside of the bottom steel flange must be taken. The 60-minute' ftre resistance can be
obtained if rebars 2d32 are added in the concrete close to the bottom flange and the
according load ratio is 0.57. Fire protection such as insulation board or intumescent fIre
 30

paint underneath the bottom steel flange is very efficient. When the intumescent flre
paint Nulliflre® and an allowed minimum thickness of 250 ~m are used, the fIre
resistance can be up to 90 minutes under the load ratio of 0.67.
Resistance analysis under natural fIres according to Eurocode 1 Part 1.2 was also
made in [111]. The critical load ratios were given for different natural flre curves, which
were in the parameters of opening factor and fire load density. After the analysis, it was
concluded that the composite slim floor beam could be used without additional measures
2
if the flre load density of the apartment is lower than 1100 MJ/m , which is rarely
exceeded for offlce and residential buildings.

The contributions from each part of the composite beam section have also been
analyzed. In the analysis, the beam section was divided into flve parts: bottom steel
flange, lower part of steel web, upper part of steel web, top steel flange and compressive
concrete slab. The contribution analysis showed that at 60 minutes' fire exposure, the
moment capacity contribution of the bottom flange is reduced to 31 % from the original
value of 44% at ambient temperature. The lower part of the web contributes 53% of the
total moment capacity at 60 minutes' flre exposure while this value is 6% at ambient
temperature. The steel web contributes a major part to the total moment capacity in flre
while the bottom flange has a major contribution at ambient temperature.
The fIre resistance of the composite slim floor beam using a hybrid steel beam was
studied in [112]. In the hybrid steel beam, the bottom steel flange was substituted by
high-strength steel or fIre-resistant steel to achieve the desired behaviour of the
composite beam.

4.3 Numerical Analysis of Isolated Beams

The numerical analysis of the beams at elevated temperature was performed using the
general flnite element program, ABAQUS/Standard [1]. The material properties of steel
and concrete were taken from Burocode 4 Part 1.2. The temperature distribution and
history described in Chapter 3 were used. Uniformly four-point loading was applied and
the corresponding load ratio was varied from 0.35 - 0.7. The parameters included
additional reinforcements, fIre protection on the bottom steel flange, beam spans and
axial restraint stiffeners.
The displacement response of the beam to flre was investigated. First of all, two fIre
tests carried out by Warrington Fire Research Center (WFRC) [95] were analyzed, in
order to validate the fInite element model. Then, the parametric analysis was carried out.
The displacement response of the Finnish slim floor beam under an applied load ratio of
0.53 is shown in Fig. 4.2.

The parametric analysis has shown that a 60-minute' fire resistance can be achieved if the
load ratio is lower than 0.53. Higher flre resistance can be obtained by reducing the load
ratio, employing additional reinforcements or flre protection on the bottom
 31

ISO Fire Exposure Time (min)

0
....;::
..,
e.., -100
u
til

Q
=-
.lil
-200

7-
.~
e -300
... £
>;::
til
-400
;..
'f -500

~ -600

Fig. 4.2 Displacement Response of the Composite Slim Floor Beam under ISO Fire.

flange. Intumescent ftre paint of 250 Jlm can increase the fIre resistance up to 105
minutes when the applied load ratio is 0.53. The studies also showed that the effect of
the beam span on ftre resistance is negligible if the applied load ratios are the same. The
beams with constant axial restraints have a larger displacement at the earlier phase of ISO
ftre due to the secondary P-d effect caused by the thermal expansion. However, a more
stable deformation response can be achieved thereafter (see Fig. 4.2). If the non-linear
force-displacement curves with varied temperature are used for the axial springs, the
deformation curve of the beam is in between that with constant stiffness and that
without axial restraints.

4.4 Numerical Analysis of Frames

Figure 4.3 shows a typical frame structure using a composite slim floor system. Three
types of frames, which were extracted from this structure, were analyzed. These frames
include one-bay two-storey 2-d subframes, two-bay four-storey 2-d frames and two-bay
two-span two-storey 3-d frames. The parametric studies on the 2-d subframes aimed to
investigate the effect of semi-rigid connection (rotational restraints) and its
quantiftcation, the effect of the axial restraint provided by the columns, and the axial
force and moment variation of the beams and columns in ftre. In the study, the steel
columns were filled in between the steel flanges with the aerated concrete.
Analysis of the 2-d five-storey frames aimed to investigate the structural behaviour of
the composite slim floor frame as a whole in ftre conditions. Both the deformation
behaviour of the structural members and the mechanical interaction between the
members were studied. The additional lateral deformation of the side-column caused by
the thermal expansion and the catenary action in the beam in the different ftre phases
was highlighted. The moment variation in the columns during ftre and the variation of
the axial force in the heated beam were also investigated.
Analysis of the 3-d frames aimed to investigate the effect of composite floor slab on
the behaviour of frame in ftre. Comparison between the deformation behaviour of the
heated beam in the plane frame and the ~spatial frame indicated the excellent effects of
the concrete floor slab on the stability of the frame structures in ftre.
 32

<D ® ®, @) ® @)

©._-- -
0
0
C)

'"
@----
C)
0
0
'"
@----.
l
4
5 x 4500 = 22500
1
l
Fig. 4.3 Layout of the Analyzed Frame Building using Composite Slim Floor System.

4.4.1 Effects of Axial and Rotational Restraints on Beams

The structural behaviour of a beam or column in a plane frame can be resolved into
an individual member with both rotational and axial restraints at two ends, as shown in
Fig. 4.4. In the figure, K", 1<.. and ~ are the stiffness of the rotational, axial and support
restraints respectively, provided by the surrounding columns and beams. Usually, the
support restraint can be simplified as rigid. In a 3-d frame, the restraint caused by the
floor slab should be included, on condition that the floor slab was compositely
connected with the beams.
In [C], a method employing the 'modified load ratio' was put forward to quantify the
effect of rotational restraint on the fire-resistant behaviour of the beam. The method was
generalized to consider the rotational restraints provided by the semi-rigid column-beam
connections.
Two major schemes have been carried out to investigate the effect of axial restraint
on the fire-resistant behaviour of the beam. In the first scheme, the individual beams
were subjected to fire with axial restraint spring of linear elastic properties and nonlinear
temperature-dependent properties. Only the deformation behaviour of the beam in fire
was investigated in this scheme. In the second scheme, the beam in the frame with
hinged connection, was subjected to fire. Both the deformation and mechanical
responses were analyzed in this scheme.
The effect of axial restraint on the mid-span vertical displacement for a simply
supported beam in fire is shown in Fig. 4.2. It can be seen that the beam with axial
restraint achieves a stable deformation response in the later phase and the run-away point
is significandy detained, whereas before that, it causes larger deformation than without
axial restraints. The mechanism is interpreted later.
 33

Fa Kr
t"'w~
Beam in l,. Kv
the frame

Fig. 4.4 Restraints of the Beam in a Frame Structure.

ISO Fire Exposure Time (min)

o 20 40 60 80 100

E
S -100
E~
E~ -200
u
.:
-a.
is'" -300
(;l
.~ -400
1::
~
>
-500

Fig. 4.5 Vertical Mid-span Displacement of the Heated Beam with Hinged Beam-Column
Connections.

For the cases in the second scheme, the deformation responses of the beam were
similar to the results in the first scheme (see Fig. 4.5). In the figure, the columns in the
frame were left unprotected. The according analysis results have been described in detail
in [115]. Figure 4.6 shows the axial force variation of the heated beam during the fire
exposure. It can be seen that the axial force in the beam is compressive and increases
rapidly at the earlier ISO heating phase, which is due to the axial restraints by the
surrounding structure on the thermal expansion of the beam. However, with the
increasing deformation of the beam, catenary action comes into effect. Meanwhile, the
heating rate of the beam becomes slow with time and consequently the catenary action
occupies the leading role. At around 60 minutes' fire exposure, a transient balance
between the thermal expansion and catenary action occurs, and the axial force becomes
zero. After that, the beam starts to be in tension. Another important factor, which
contributes to the development of the axial force in the beam, is the decreased restraints
by the column due to the material softness of steel at elevated temperature. This point is
clearly reflected by the comparison between the column being heavily protected or
unprotected, which is highlighted in Fig. 4.6.

An interesting point can be seen by looking through Figs. 4.5 and 4.6. The frame
beam still maintains its stability at 62 minutes' fire exposure whilst the simply supported
beam starts to run away. This corresponds to the time that the axial force in the beam
starts to become tensile (Fig. 4.6). The catenary action starts to take effects to maintain
the stability of the frame beam when the simply supported beam starts to run away under
the modified load ratio.
 34

.-,
Z 500 •••••..•••••••••••• y ••• • .. ••••••••••••••••···•·••• .. T····· ............. ·········1····· .......................... '1" ••••
e 400
I
-----------~------------~------
_ columns unprot~~~~?____ !
~
~ 300
,_.'------;.
- - - - - - - - - - - ~ -r-'"- - - - - - - - -
··· .. ···columnspro,tected
-:-.= . . ""'-_.- ------ ----------- ---------- --j
~ ~
,
~ /, I I I

:::7":)::~;:::::;:: . :::: ::'

-=
~ 200
~
~
100
...
lI:
0 0
I I "' I

: : :.~:
.
III
\I
0
-100
,

- - - ----- - -21) - ---- - - -- -48- --- -- ----61} -- -";

I
I

I I
,''''

...
I

--SD----- ---
~
I I I ' ' ____ --.....,.

-; -200 -----------~-----------

~ -300
ISO Fire Exposure Time (min)

Fig. 4.6 Axial Force-time Curves of the Slim Floor Beam in the Frame under ISO Fire.

4.4.2 Effects of Fire Protection on Columns and Heat Rate on the Beam

In some practices, the columns are fIre protected to avoid their possible collapse
during fIre. In fact, fIre-protected columns provide a larger restraint on the heated beam
in the course of fIre attack. In publication [115], the effect of fIre protection measures on
the column are presented. The studies showed that the fIre protection measures
produced better deformation behaviour for the heated beam and a smaller lateral
displacement in the column end after 60 minutes' ISO fIre exposure. However, the
protection measures also caused a bigger axial force (both compressive and tensile) in the
beam throughout the fIre exposure.
In reality, the heating rate of the structural members may be different in each fIre
event, because of the different fIre protection on the members or the different fuel types
in the fIre room, and so on. Three kinds of heating rates for the bottom flange of the
slim floor beam were used in this study. Case 1 represents a linear heating rate of the
bottom flange. Case 2 represents the heating rate caused by the ISO standard fIre curve,
and Case 3 represents a faster temperature rise in the earlier stage of fIre. For all cases,
the bottom steel flanges reach the same temperature at 90 minutes.
The analysis results show that the structural responses of the frame corresponding to
the same temperature of the bottom steel flange are very close to each other, if the
columns in the analyzed frames are heavily protected. However, if the heated columns
are unprotected, the differences between the different heating rates are rather signifIcant.
This can be attributed to the interaction between the beam and the surrounding columns
in fIre. The analysis results are listed in Table 4.1. Actually, corresponding to the same
temperature in the bottom flange of the beam, the fIre exposure time is different for each
of the different cases of heating rate. The according temperature distributions in the
columns are hence different, which cause the different axial constraints on the heated
beam.
 35

Table 4.1 Structural response to different heating rates - exposed columns

Beam mid-span Axialforce in Lateral displacement oj Additional moment

Heating rate displacement (mm) beam (kN) side-column (1 st floor) ojside-column
(mm) (kNm)
Case 1 -140 104 -12.8 126
T=600 °C Case 2 -153 164 -12.2 248
Case 3 -192 226 -9.8 407
Case 1 -418 -193 22.0 -163
T=890 °C Case 2 -432 -177 22.1 -130
Case 3 -454 -156 32.9 -82

ISO Fire ExposureTime (min)

o , , ,
'0
__
6p
_________ L _________
8p
________ _
100
-100 -------- -:-, - - ~ ~

,-,-,-_- ----
,

-200 ---------i --------- ~ -~~~~~+,- i- ---------

-300
-.--~ plane framl
- - - 3-D frame case 1
______ ~~~~ ~,~'_~ t~---~
,
' "
.~:
1--'
____ _
-400
------ 3-D frame case 2 -.<
- - Simplely supported-be~affi - - - - - -,---.::...::...::: - - - --

-500

Fig. 4.7 Vertical Displacement of the Heated Beam in Plane Frame and 3-D Frame. (In the case
1 of 3-D frame, the beams along axis 1 and 3 were assumed to be well-protected and
only the beam along axis 2 was heated; in the case 2, the beams along axis 1, 2 and 3
were heated. It was assumed that the columns C/2 and B/2 were subjected to flre.)

4.4.3 Effects of Composite Floor Slab

In order to identify the influence of the composite floor slab on the flre-resistant
behaviour of the heated beam and the whole frame, several three-dimensional frames
were analyzed. In the analyzed frames, the beam was connected by a hinge to the
columns so that the rotational restraints were removed and only the axial restraints and
the restraint of the composite floor slab existed. Then, the deformation and mechanical
responses of the three-dimensional frame were compared with the plane frame which
was similar to the three-dimensional frame but only the composite slab of a width of 1/8
beam span was included. In this way, the influence of the composite floor slab was
expected to be identified.
Figure 4.7 shows the deformation response of the heated beams in an isolated case,
in a plane frame and in a three-dimensional frame. It can be seen that the beam in the
three- dimensional frame has a significantly better deformation behaviour after 60
minutes' flre exposure, whereas the deformation curves of the beams in the three-
dimensional frame and plane frame are very close to each other before that. In this
figure, the deformation response of the heated beam with the surrounding beams also
heated, was shown. When the surrounding beams were also heated, the same as the
 36

central beam, the restraint on the composite floor slab was weakened and the heated
central beam had a larger displacement.

Figure 4.8 shows the deformation proflle of the floor slab at 90 minutes' ISO flre
exposure. It can be seen that the floor slab seems to support the beam, not the beam
supporting the floor slab. This phenomenon is similar to that in the Cardington fIre tests,
which can be attributed to the 'membrane action' in the floor slab that becomes
signiflcant in the case of large displacement.

4.4.4 Effects of Sway and Non-sway Frame

For a sway frame, research has shown that the frame collapsed at a very early phase
of ISO fIre exposure (20 minutes in the studied case), if the columns were left
unprotected. This was mainly caused by the signiflcant P-o effect and the large thermal
expansion of the heated beam. However, if the columns were flre-protected, the frame
kept its stability up to 90 minutes' flre exposure, which was close to the non-sway frame.

4.4.5 Global Deformational and Mechanical Responses of the Frame in Fire

The global deformation of the frame is shown in Fig. 4.10. As interpreted later, the
lateral deformation of the frame is largely caused by the variation of axial force in the
heated beam. In the earlier phase, the thermal expansion of the heated beam produced a
push to the surrounding structures. In the later phase, the catenary action took effect and
a pull to the surrounding structures was produced. The lateral deformation of the frame
follows the variation of the axial force within the heated beam.
Study of the mechanical response to flre of the frame members lends itself to
understanding the mechanisms behind the deformation response. Studies in this scheme
showed that three kinds of restraints had the major influence on the behaviour of the
frame structure in flre, which are axial restraints, rotational restraints and the restraints by
the composite floor slab. The later, causes a phenomenon that has been described as 'the
floor slab supports the beam' in flre, which is actually produced by the catenary action in
 37

the slab in case of large deformation. In this scheme, the plan frame was investigated and
the variation of the axial force in the heated beam and the moment variation in the
columns were tracked. The research results are presented in publications [B,C,lll, 115].
Generally, the variation of the axial force in the heated beam can be divided into two
phases (see Fig. 4.6). At the early ISO ftre exposure, the axial force is compressive and
increases rapidly up to several times of that before heating. This is attributed to the axial
restraints caused by the surrounding structures on the thermal expansion of the beam.
However, with the increasing deformation of the beam, the catenary action starts to take
effects and gradually takes the leading role. This is the second phase. After a period of
time, a transient balance between the thermal expansion and catenary action takes place,
and the axial force becomes zero. After that, the beam starts to be in tension.
All factors that affect the degree of the axial restraints in the heated beam contributes
to the development of the axial force in the beam, for example, the material softness of
steel columns at elevated temperature. A comparison of the axial force response between
the frames with the columns being ftre-exposed and ftre-protected is shown in [115].
The variation of the axial force in the heated beam causes the signiftcant additional
moment in the surrounding frame members, especially in the columns. Fig. 4.11
illustrates the moment distribution of the columns at 30 and 90 minutes' ftre exposure,
respectively. In normal conditions, the moments in the columns are very small for the
studied frame. It can be seen that the additional moments in the columns under flre are
surprisingly large.
 38

-0 .14 u.uu -0.42 ..-----..",.0.""'21:-11---',

-0.16
-0 .30 -0.65 \----,.,....,...-1+--'.,

-0.33 -007
-0.8 0

-11.2 ! 1.11 26.1 6.0

\ 77r 77

30 min 90 min

(a)

30
, /~

20
- - One-sidE-heated colurrms
- - - - Protected colurrms :
)(----------
___________ J, ____________ ,, ____________ t, ___________ .JI ___________ _
I I • I

........ ---:
I , , ,~

:
I
:
I
:I ~/ 1

o ..', _: : ,,", : .
"- '<::~
-----------~---~---:---'---'---~------------:
40 _./<~\J 8,b 100
,, "
"
, "
-20 ----- -- ----~--- -- - ------:- ---- - -- ----~ - - - --------~- --- - ---- - --'
, , , ,
,, ,, ,,
I I I I

.......................................<.............................1
-30 ••••••••••••••••••••••••••••• 1•••••••

ISO Fire Exposure Time (min)

(b)

Fig. 4.10 Global Deformation of the Frame in Fire: (a) Deformation proftle of the frame at 30
and 90 minutes' fIre exposure; (b) Lateral displacement - time curves of the side -
column e/2 in the fIrst storey.
 39

3.52 1t-_ _ _ _ _ -t:r1.,;;;6;.;..7...J,

259.0F====3-----~""'-Br----i,

419.4F="I-------f=p,

30 min 90 min

(a)

'"<'l
'tl

:t"
500

400 ___ --~~._+_-_-_,',~ _; _~ ______ -_~ __

'"
,

I ....
I

I
1

I
3

'8'
2. - - __ :___ - - ____ .

JaZ ;f:J?S+~:sJ,:: :::~+ :~:::::

300

zoo
o 100
U,:.:
'E~ 1(1 : : I .... ~/ :

d ........ I I _ .... - . . . . I

" ·100 -->",-:,-.-2fL ------- ~- =:.:_"C":,"?_ - - - - --:-::~~.:: - - - _100

S
0
.zoo - - - - - - - :~:-.:=~.:~:.-.~~ - - - - - - - ~ - - - - - - - -
;;;: : 3 : :
_1- __ _ _ _ _ _ _ .

:
. 300 ........ ! ......................... !. ......................1..... . ............1... .•.

ISOFire EXposure 'Time (min)

(b)

Fig. 4.11 Moment Distribution and Variation with Time of the Columns in Fire: (a) Moment
distribution at 30 and 90 minutes' flre exposure; (b) Moment variation with time of
the side-column Cj2 in the 1st and 2nd storey.
 40

Chapter 5

Conclusions

Structural fire safety design is an important aspect in the design of steel and
composite steel-concrete structures. Development of the structural aspect of modern fIre
safety design includes the characterization of fully developed fire curves, the temperature
analysis and the structural response analysis. The probabilistic aspect of the fIre curves
and the action on the structures in fire are also very important issues. The probabilistic
aspect of the fire curves could be considered using the probabilistic fIre load and
ventilation, in which the effects of sprinklers and fire-fighting by the fire brigades can
also be included in a convenient way. The action (applied loads) on the structures in fire
was developed by survey and probability analysis [42]. This dissertation is concerned with
the characterization of fully-developed fire curves, the temperature analysis and the
structural response analysis.
A method to estimate the temperature-time curve in a fully developed fIre
compartment for structural design purposes, without using computer code, has been
developed, on the basis of extensive test results and increasing knowledge of fire
engineering science. This method can provide a quick calculation of natural fIre curves
for use in the structural fire-resistance design, with a rational precision. Accordingly, the
equivalent fire exposure methods have been developed. The equivalent fIre exposure was
found to be in the function of the product of fire duration 't and the maximum gas
temperature Tb gmax, where b is in between 2 - 4, according to the equivalent principle.
This composition indicates the fact that fire is radiation-dominated.
A heat transfer model has been developed to analyze the temperature distribution
and variation with time when the composite structures are exposed to fIre. In the model,
the effect of moisture in the concrete and fIre protection on the steel beam can be taken
into account. A computer code using the finite difference method (fACS-FIR) has been
written.
In recent years, increasing interest has been shown throughout Europe to develop
and design slim floor systems in steel-framed buildings. This dissertation presents the fire
resistance behaviour of the composite asymmetric slim floor beam both as an isolated
member and as a part of the frame using numerical analysis methods. In general, three
schemes have been investigated, including isolated beams, plane frames with semi-rigid
and hinged beam-to-column connections and the three-dimensional slim-floor frame
system. The fIrst scheme aimed to explore the fire resistance of the beams according to
standard fire-testing methodology. The objective of the second scheme was to reveal the
effects of frame continuity on the fire resistance of the slim floor beam and the
mechanical interaction between the frame elements. The third scheme aimed to identify
the influence of the composite slab on the beam behaviour in fire. The analyzed results
indicate that the axial restraints provided by the surrounding parts cause a major
 41

influence on the global deformational and mechanical responses of the frame to fIre. The
rotational restraints essentially cause the change in the applied load ratio in fIre, which
can be quantifIed using the 'modifIed load ratio' proposed in the context.
 42

Bibliography

[1] ABAQUS/Standard. User's Manual (version 5.7). Hibbitt, Karlsson & Sorensen Inc., 1997.

[2] Ali, F.A., Shepherd, P., Randall, M., Simms, lW., O'Connor, D.]. and Burgess, I. The Effect of
Axial Restraint on the Fire Resistance of Steel Columns. Journal of Constructional Steel Research
1998; 46 (1-3): Paper No.177.

[3] Babrauskas, V. and Williamson, R.B. Post-flashover Compartment Fires, Fire Research Group
Report No. UCB FRG 75-1, University of California, Berkeley, 1975.

[4] Babrauskas, V. and Williamson, RB. The Historical Basis of Fire Resistance Testing, part I. Fire
Technology 1978; 14 (3): 184-194.

[5] Babrauskas, V. and Williamson, R.B. The Historical Basis of Fire Resistance Testing, part II. Fire
Technology 1978; 14 (4): 304-316.

[6] Baily, CG. The Behaviour of Asymmetric Slim Floor Steel Beams in Fire. Journal of Constructional
Steel Research 1999; 50: 235-257.

[7] Baily, CG., Burgess, I. and Plank, R Bridging and Restraint Effects of Localised Fires in Composite
Frame Structures." Proc. of the International Conf. Composite Construction, Innsbruck, Austria,
1997: 379-384.

[8] Baily, CG., Lennon, T. and Moore, D.B. The Behaviour of Full-Scale Steel-Framed Building
Subjected to Compartment Fires. The Structural Engineer 1999; 77 (8): 15-21.

[9] Bailey, C Computer Modelling of the Corner Compartment Fire Test on the Large-Scale Cardington
Test Frame. Journal of Constructional Steel Research 1998; 48: 27-45.

[10] Baldwin, R. and Nonth, M.A. The Number of Sprinkler Heads Opening in Fires. Joint Fire research
Organisation Fire Research Note 886, Borehamwood, 1971.

[11] Babrauskas, V. A Close-form Approximation for Post-flashover Compartment Fire Temperature.

Fire Safety Journal 1981; 4: 63-73.

[12] Barnett, J.R Structures Subjected to Accidental Fires.

http://www.wpi.edu/ Academics/Depts/Fire/ papers/ chicago/ chicago.htrnl. WPI Center for
Firesafety Studies, Worcester, USA, 1995.

[13] Becker, W. and Hertel, H. Fire Safety Concept. Fire Safety Journal 1991; 17 (2): 103-111.

[14] Bennetts, I.D., Proe, D.]., Lewins, RR and Thomas, I.R. Open-deck Car-park Fire Tests. BHP
Melbourne Research Laboratories, Report MRL/PS69/85/001, Australia, 1985.

[15] Bennetts, I.D., Proe, D.]., Lewins, R.R and Thomas, I.R Fire and Unprotected Steel in Closed Car-
parks. BHP Melbourne Research Laboratories, Report MRL/PS98/87/001, Australia, 1987

[16] Bernuzzi, C, Gadotti, F. and Zandonini, R Semi-continuity in Slim Floor Steel-Concrete Composite
Systems. 1st European Conference on Steel Structures, Athens, Greece, May 18-20, 1995.
 43

[17] Bernuzzi, C, Gadotti, F. and Zandonini, R. Joint Action in Slim Floor Systems, PSSC95, 4th Pacific
Structural Steel Conference, Volume 3: Steel-Concrete Composite Structures, Singapore, 1995.

[18] Bjorkman,]. and Keski-Rahkonen, O. Simulation of the Steckler Room Fire Experiment using
SOFIE CFD-model. Research Center of Finland, VIT publications 265, 1996.

[19] Black, N.S. Ultimate Strength Study of Two-way Concrete Slabs. Journal of Structural Division,
Proceedings of American Society of Civil Engineers 1975; 101 (ST1): 311-324.

[20] Bohm, B. Calculated Thermal Exposure of Steel Structures in Fire Test Furnaces - a non-Eres
approach. Laboratory of Heating and Air Conditioning, Tech. University Denmark, CIB
W14/82/73 (DK).

[21] Bolwin, R., Law, M. and Giffiths, L.G. Survey of Fire Loads in Modern Office Buildings- Some
Preliminary Results. Fire Research Note No.808,Joint Fire Research Organization, England, 1970.

[22] Both, C, Fellinger, ].H.H. and Twilt, L. Shallow Floor Construction with Deep Composite Deck:
From Fire Tests to Simple Calculation Rules. Heron 1997; 42 (3): 145-158.

[23] Both, C Fire-Exposed Continuous Span Composite Steel- Concrete Slabs. Heron 1996; 41 (3): 187-
199.

[24] Both, C, Twilt, L. and Stark, J.W.B. Numerical Simulation of Thermal and Structural Response of
Composite Steel/Concrete Slabs to Fire. 4th Pacific Structural Conference, Singapore, Volume 3,
October 1995.

[25] Both, C and Twilt, L. The Real Behaviour of Modern Largely Unprotected Steel Framed Buildings
under Natural Fire Conditions. Diana World 1997; 1: 4-8.

[26] Bruls, A., Cajot, L.G. and Franssen ].M. Characterization of an Insulating Material With Regard to
ECCS Recommendations for the Fire Safety of Steel Structures. Journal of Constructional Steel
Research 1988; 9: 111-135.

[27] BS5950, Structural Use of Steelworks in Building: Part 8, Code of Practice for Fire Resistance
Design. British Standards Institution, London, UK, 1990.

[28] Burgess, LW., Olawale, A.O. and Plank, R.]. Failure of Steel Columns in Fire. Fire Safety Journal
1992; 18: 183-201.

[29] Buress, I.W. and Najjar, S.R. A Simple Approach to the Behaviour of Steel Columns in Fire. Journal
of Constructional Steel Research 1994; 31: 115-134.

[30] Burgess, R.A. Coatings for Fire Protection. Modern Steel Construction, December 1998: 64-70.

[31] Butcher, E.G. and Chitty, T.B. The Temperature Attained by Steel in Building Fires. JFRO, Fire
Research Technical Paper No.l5, London, HMSO, 1966.

[32] Butcher, E.G., Bedford, G.K. and Fardell, P.G. Further Experiments on Temperatures Reached by
Steel in Building Fires. Proceedings of the Symposium held at the Fire Research Station, Boreham
Wood,January, 1967: 45-56.

[33] Cajot L.G., Schleich, ].B. Global Fire Safety Concept for Buildings. International Conference on
Composite Construction - Conventional & Innovative, Innsbruck, Australia, Sep. 16-18, 1994, pp
415-20.

[34] Cajot, L.G., Franssen , ].M. and Schleich, ].B. Computer Model for the Fire Resistance of
Composite Structures. IABSE Symposium: Mixed Structures, including New Materials, Brussels
1990: 395-400.

[35] CEC, Construction Products Directive. Commission of European Community, N°L40/12,

Luxembourg, December 1988.
 44

[36] CEC, Interpretative Document No 2 /Essential Requirement on Safety in case of Fire. Commission
of European Community, October 1993.

[37] CIB W14, A Conceptual Approach towards a Probability Based Design Guide for Structural Fire
Safety. Fire Safety Journal 1983; 6 (1).

[38] CIB W14, Design-Guide - Structural Fire Safety. Fire Safety Journal 1986; 9: 77-136.

[39] Collier, P. Fire in a Residential Building: Comparisons between Experimental and Fire Zone Model.
Fire Technology 1996; 32(3).

[40] Collier, P. Fire Modelling of Large Spaces. http://www.branz.org.nz/Branz/PandS/ftreResearch/

largespaces.htm, September 1998.

[41] Cooke, G.M.E., Lawson, R.M. and Newman, G.M. Fire Resistance of Composite Deck Slabs. The
Structural Engineer 1988; 66 (16).

[42] Corotis, R.B. Load Combinations for Building Exposed to Fires. Proc. of National Steel
Construction Conference, American Institute of Steel Construction, Missouri, March 14-17 1990:
8.1-8.19.

[43] Corradi, L., Poggi, C. and Setri, P. Interaction Domains for Steel Beam-Columns in Fire Conditions.
Journal of Constructional Steel Research 1990; 17: 217-235.

[44] Culver, e.G. Survey Results for Fire Loads in Offtce Buildings. NBS, Building Science Series 85,
National Burea of Standards, Washington D.C., 1976.

[45] Dhima, D., Fouguet, G., etc., Experimental and Numerical Assessment of the Behaviour of External
Heavy Steel Columns Under Natural Fire. Journal of Constructional Steel Research 1998; 46 (1-3).

[46] Drysdale, D. An Introduction to Fire Dynamics. John Wiley & Sons Ltd., 1987.

[47] ECCS, Fire Safety in Constructional Steelwork. European Convention for Constructional Steelwork,
CECM-III-74-2E,1974.

[48] ECCS TC3, Fire Safety of Steel Structures. European Recommendations for the Fire Safety of Steel
Structures, Elsevier, 1983.

[49] ECCS TC3, Fire Safety in Open Car Parks. European Convention for Constructional Steelwork,
Technical Note No.75, Brussels, Belgium, 1993.

[50] Edwards, M. The Performance in Fire of Concrete Filled SHS Columns Protected by Intumescent
Paint. Proceedings of International Conference on Nordic Steel Construction, Bergen, Norway,
1998: 27-40.

[51] El-Houssieny, O.M., Salam, S.A., Atria, G.A.M. and Saad, A.M. Behaviour of Extended End Plate
Connections at High Temperature. Journal of Constructional Steel Research 1998; 46: 1-3.

[52] El-Rimawi, J.A., Burgess, I.W. and Plank, R.J. The Influence of Connection Stiffness on the
Behaviour of Steel Beams in Fire. Journal of Constructional Steel Research 1997; 43(1-3): 1-15.

[53] El-Rimawi J.A., Burgess, LW. and Plank, R.J. Studies of the Behaviour of Steel Subframes with
Semi-Rigid Connections in Fire. Journal of Constructional Steel Research 1999; 49: 83-98.

[54] Emmons, H.W. Prediction of Fires in Building. 17 th International Symposium on Combustion, the
Combustion Institute, Pittsburgh, 1978.

[55] ENV 1991-2-2, Eurocode 1, Basis of Design and Actions on Structures, Part 2.2, Actions on
Structures Exposed to Fire. Final edited version/CET/TC250/SC1, Brussels, November 1994.
 45

[56] ENV 1993-1-2, Eurocode 3, Design of Steel Structures, Part 1.2: General Rules - Structural Fire
Design. CEN/TC250/SC3, Brussels, September 1995.

[57] ENV 1994-1-1, Eurocode 4, Design of Composite Steel and Concrete Structures, Part 1.1: General
Rules and Rules for Buildings. Final edited version/CEN/TC250/SC4, Brussels, November 1993.

[58] ENV 1994-1-2, Eurocode 4, Design of Composite Steel and Concrete Structures, Part 1.2: Structural
Fire Design. Final edited version/CEN/TC250/SC4, Brussels, November 1993.

[59] Favre,J.P., Fontana, M. and Hass, R. Optimal Fire Safety through Modern Engineering Approaches.
Seminar held in Helsinki University of Technolgy, Espoo, Finland,1995.

[60] FIP, FIP /CEB Report on Methods of Assessment of the Fire Resistance of Concrete Structural
Members. FIP Commission on the Fire Resistance of Prestressed Concrete Structures, 1978.

[61] Fire Tests of the 140 William Street Office Building/The Effect of Fire in the Building at 140
William Street.

[62] Fitzgerald, R.W. and Lie, T.T. Structural Fire Protection. ASCE Manuals & Reports on Engineering
Pratice, No.78, 1992.

[63] Fitzgerald, R.W. Building Fire Safety Engineering Method, Workbook, Worcester Polytechnic
Institute, Worcester, Mass, 1993.

[64] Franssen, J.M., Cajot, L.G. and Schweppe, H. Connection between Parametric Temperature-Time
Curves and Equivalent Time of Fire Exposure (Annex B and E of ENV-1991-2-2). IABSE
Colloquium, Basis of Design and Actions on Structures, Delft, 1996: 407-417.

[65] Franssen, J.M., Cooke, G.M.E. and Latham, D.J. Numerical Simulation of Full Scale Fire Test on a
Loaded Steel Framework. Journal of Constructional Steel Research 1995; 35: 377-408.

[66] Franssen, J.M., Talamona, D., Kruppa, J. and Cajot, L.G. Stability of Steel Columns in case of Fire:
Experimental Evaluation. Journal of Structural Engineering 1998; 124(2): 158-163.

[67] Franssen, J.M. and Dotreppe, J.c. Fire Resistance of Columns in Steel Frames. Fire Safety Journal
1992; 19: 159-175.

[68] Franssen, J.M., Schleich, J.B., Cajot, L.G. and Azpiazu, W. A Simple Model for the Fire Resistance
of Axially Loaded Members - Comparison with Experimental Results. Journal of Constructional
Steel Research 1996; 37 (3): 175-204.

[69] Hagen, E. and Haksever, A. A Contribution for the Investigation of Natural Fires in Large
Compartment. Fire Safety Science- Proceedings of the First International Symposium, Berkeley,
Califonia, November 1985: 149-158.

[70] Haksever, A. Fire Engineering Design of Steel Framed Car-park Buildings. Department of Structural
Engineering, Helsinki University of Technology, Report No.112, Espoo, Finland, 1990.

[71] Hamerlinck, R. and Twilt, L., Fire Resistance of Composite Slabs. Journal of Constructional Steel
Research, Special Issue on Fire and Steel 1995; 33 (1-2): 71-86.

[72] Harda, K. and Terai, T. Dependence of Thermal Response of Composite Slabs Subjected to Fire on
Cross Sectional Shapes. Fire Safety Science - Proceedings of the Fourth International Symposium,
International Association for Fire Safety Science, 1995: 1159-1170.

[73] Harmathy, T.Z. A New Look at Compartment Fire. Fire Technology 1972; 8 (3): 197-217; 8 (4):
327-351.

[74] Harmathy, T.Z. Fire Severity: Basis of Fire Safety Design. Presented at the International Simposium
of Fire Safety of Concrete Structures, Fall Convention of the American Concrete Institute, San Juan,
September 21-26,1980.
 46

[75] Harmathy, T.Z. and Mehaffey, J.R. Normalized Heat Load: A Key Parameter in Fire Safety Design.
Fire and Materials 1982; 6 (1): 27-31.

[76] Heselden, A.J.M. Parameters Determining the Severity of Fire. Proceedings of the Symposium held
at the Fire Research Station, Boreham Wood,January, 1967: 19-28.

[77] Ingberg, S.H. Test of Severity of Building Fires. Quarterly, National Fire Association, vol.22, 1928.

[78] ISO, Fire Resistance Tests - Elements of Building Construction. International Organisation for
Standardisation, ISO-834 Part 1, Switzerland, 1992.

[79] Johnson, P.P. International Developments in Fire Engineering of Steel Structures. Journal of
Constructional Steel Research 1998; 46 (1-3).

[80] Jones, W.W., Forney, G.P., Baily, J.L. and Tatem, P.A. Comparison of CFAST Prediction to Real-
Scale-Fire Tests. National Institute of Standards and Technology, Naval Technology Center for
Safety and Surivivability, 1996.

[81] Kawagoe, K Fire Behavior in Rooms. Building research Institute, Japan, Report No.27, 1958.

[82] Kawagoe, K and Sekine, T. Estimation for Fire Temperature-time Curve in Rooms. BRI Occasional
Report No.11, Building Research Institute, Japan, 1963.

[83] Kemp, KO. Yield of a Square Reinforced Concrete Slab on Simple Supports, Allowing for
Membrane Forces. The Structural Engineer 1967; 45 (7): 235-240.

[84] Kirby, B.R. Recent Developments and Applications in Structural Fire Engineering Design- A
Review. Fire Safety Journal 1986; 11: 141-179.

[85] Kirby, B.R. British Steel Technical European Fire Test Programme - Design, Construction and
Results. Proceedings of the 2nd Cardington Conference, 1996: 111-126.

[86] Kordina, K Natural Fires: Research and Application in Consulting. Fire Safety Journal 1991; 17:
115-130.

[87] Kouhi, J. Euroopan Hiili-ja Tersunionin (Ecsq T&K-toirninta. Presented in Terasrakenteiden

Tutkimus- Ja Kehityspiiivat, August 25-26, 1998.

[88] Kruppa, J. and Zhao, B. Fire Resistance of Composite Beams to Eurocode 4 part 1.2. Journal of
Constructional Steel Research, Special Issue on Fire and Steel 1995; 33 (1-2): 51-70.

[89] Law, M. A Basis for the Design of Fire Protection of Building Structures. Structural Engineer, 1983;
61A(I):25-33.

[90] Law, M. and Arnault, P. Fire Loads, Natural Fires and Standard Fires. International Conference on
Planning and Design of Tall Buildings Preprints: Reports vol. 1-8, 1972, ASCE-IABSE.

[91] Law, M. Prediction of Fire Resistance. Paper No.2 of Symposium No.5: Fire Resistance
Requirements for Buildings - A New Approach, London, HMSO, 1973

[92] Law, M. Design for Fire Safety of Bare Exterior Structural Steel: 1. Theory and Validation. 2. State
of the Art, Ove Arup and Partners, London, 1977.

[93] Law, M. A Basis for the Design of Fire Protection of Building Structures. The Structural Engineer
1983; 61A (1): 25-33.

[94] Law, M. Fire Grading and Fire Behaviour. Fire Safety Journal 1991; 17 (2): 147-153.

[95] Lawson, R.M., Mullett, D.L. and Rackham, J.W. Design of Asymmetric Slimflor Beams using Deep
Composite Decking. The Steel Construction Institute, Publication P175, 1997.
 47

[96] Lawson, RM., and Blaffart, H. Fire Safe Design of Composite Deck Slabs. Structural Engineering
International 1996; 2: 119-122.

[97] Lawson, R.M. Fire Resistance of Slim Floor Construction. Proceedings of International Conference
on Steel and Composite Structures, Delft, The Netrherland 1999: 9.1-9.14.

[98] Lennon, T. Large Compartment Fire Tests. Fire, Static and Dynamic Tests of Building Structures,
Proceedings of the Second Cardington Conference, England, March 1996: 45-53.

[99] Li, G.Q., Jiang, S.C, Lin, G.x. and Yu, L.H. An Experimental Research on Behaviour of Steel
Frames Subjected to Fires. Proceedings of the 3rd Sino-Japan Conference on Structural Technology,
Shen Zhen, P.RChina, November 4-8, 1997.

[100] Lie, T.T. and Chabot, M. A Method to Predict the Fire Resistance of Concrete Filled Hollow Steel
Columns. Journal of Fire Protection Engineering 1990; 2 (4); 111-126.

[101] Lie, T.T. and Irwin, R.J. Fire Resistance of Rectangular Steel Columns Filled with Bar-Reinforced
Concrete. Journal of Structural Engineering 1995; 121 (5): 797-805.

[102] Liew, ].Y.R, Tang, L.K., Holmas, T. and Choo, Y.S. Advanced Analysis for the Assessment of Steel
Frames in Fire. Journal of Constructional Steel Research 1998; 47 (1-2): 19-45

[103] Liu, T.CH. Effect of Connection Flexibility on the Fire Resistance of Steel Beams. Journal of
Constructional Steel Research 1998; 45 (1): 99-118.

[104] Liu, T.CH. Finite Element Modelling of Behaviours of Steel Beams and Connections in Fire.
Journal of Constructional Steel Research 1996; 36 (3): 181-199.

[105] Lu, X. and Miikelainen, P. Strength and Stiffness of Composite Slim Floor Beams. Structural Steel,
PSSC'95, 4th Pacific Structural Steel Conference, Volume 3, Steel-Concrete Composite Structures,
Singapore, 1995: 93-100.

[106] Lu, X. Study on the Structural Behaviour and Composite Action in Composite Slim Floor Beams.
Acta Polytechnica Scandinavica, Civil Engineering and Building Construction Series No. 103,
Helsinki,1995.

[107] Lu, X. and Miikelainen, P. Slim Floor Development in Sweden and Finland. Structural Engineering
International 1996; 2: 127-129.

[108] Loikkanen, P. and Mangs,]. Characterization of the Fire Behaviour of a Burning Passenger Car. Fire
Technology Laboratory VTT, Reports PAL 00455/90 and PAL 0045SA/90, Finland, 1992.

[109] Ma, Z. Damage Evaluation and Aseismic Repair of RC Structures After Fire. Research Report
No.3, Steel Structures in Helsinki University of Technology, Espoo, Finland, 1998.

[110] Ma, Z. and Miikelainen, P. Temperature Analysis of Steel-Concrete Composite Slim Floor Structures
Exposed to Fires. Laboratory of Steel Structures Publications 10, Helsinki University of Technology,
Espoo, 1999.

[111] Ma, Z. and Miikelainen, P. Numerical Analysis of Steel-Concrete Composite Slim Floor Structures in
Fire. Laboratory of Steel Structures Publications 11, Helsinki University of Technology, Espoo,
1999.

[112] Ma, Z. and Miikelainen, P. Fire Resistance of Hybrid Steel Beam in Composite Slim Floor Structures
Construction. Teriisrakenteiden Tutkimus- ja Kehityspaiviit, Teknillinen Korkeakoulu, Tammikuu
2000, pp.S3-60.

[113] Ma, Z. and Makelainen, P. Fire Safety Design of Composite Floor System. Teriisrakenteiden
Tutkimus- ja Kehityspaiviit, Lappeerannan Teknillinen Korkeakoulu, August 1998, Finland.
 48

[114] Ma, Z. Fire Engineering Design of Composite Floor System. Seminar on Steel Structures: Design of
Steel-Concrete Composite Structures. Laboratory of Steel Structures Helsinki University of
Technology, TKK-TER-4, 1998.

[115] Ma, Z. Modern Fire Safety Design Methods. p.78 (unpublished).

[116] Ma, Z. Structural Behaviour of Composite Slim Floor Frame in Fire Conditions. P.23 (unpublished).

[117] Ma, Z., Wu, B. and Ou,J. Fully-Developed Compartment Fire Model and Its Application. Journal of
Harbin University of Civil Engineering and Architecture 1998; 31 (1): 16-24.

[118] Ma, Z., Wu, B. and Ou, J. Literature Review of Fire Engineering Design ofR.e. Structures. Building
Structures (Chinese Journal) 1997; 11: 48-52.

[119] Ma, Z., Wu, B. and Ou, J. Seismic Characteristics of Re. Columns after Fire. Chinese Journal of
Computational Mechanics 1997; 14 (4): 443-453.

[120] Ma, Z. and Wu, B. Evaluation Technology of Fire- Damaged Re. Structures. Institute of
Engineering Theory and Its Application, Harbin University of Civil Engineering and Architecture,
Research Report, 1997, 126p.

[121] Magns, J. and Keski-Rahkonen, O. Characterization of the Fire Behaviour of a Burning Passenger
Car. Part I: Car Fire Experiments. Fire Safety Journal 1994; 23: 17-35.

[122] Magns, J. and Keski-Rahkonen, 0. Characterization of the Fire Behaviour of a Burning Passenger
Car. Part II: Parametrization of Measured Rate of Heat Release Curves. Fire Safety Journal 1994; 23:
37-49.

[123] Magnusson, S.E. and Thelandersson, S. Temperature-time Curve for the Complete Process of Fire
Development, A Theoretical Study on Wood Fuel Fires in Enclosed Space. Acta Polytechnic
Scandinavia, Civil Engineering and Building Construction Series No.65, 1970.

[124] Magnusson, S.B. and Thelandersson, S. Comments on Rate of Gas Flow and Rate of Burning for
Fires in Enclosures, Division of Structural Mechanics and Concrete Construction, Lund Institute of
Technology, Sweden, Bulletin 19, 1971.

[125] Malhotra, H.L. Design of Fire-Resisting Structures. Surrey University Press, 1982.

[126] Mehaffey, J.R. and Harmathy, T.Z. Thermal Response of Compartment Boundaries to Fire. Fire
Safety Science-Proceedings of the First International Symposium, Berkley, California, November
1985.

[127] McCaffery, B.]., Quintiere, J.G. and Harklerroad. Estimating Room Temperatures and the
Likelihood of Flashover Using Test Data Correlation. Fire Technology 1981; 17 (2): 98-119.

[128] Mullett, D.L., Lawson, R.M. Slim Floor Construction Using Deep Decking. SCI P127, The Steel
Construction Institute, Berkshire, United Kingdom, 1993.

[129] Myllymiiki, J. and Kokkala, M. Fire Tests of a Steel Beam Under Ceiling Exposed to a Localized
Fire. Fifth Finnish Steel Structures R&D Days, 18-19 January 2000: 146-153.

[130] Neves, I.e. The Critical Temperature of Steel Columns with Restrained Thermal Elongation. Fire
Safety Journal 1995; 24: 211-227.

[131] Newman, G.M. Fire Resistance of Slim Floor Beams. Journal of Constructional Steel Research 1995;
33 (1-2): 87-100.

[132] Newman, G.M. Design Implications of the Cardington Fire Research Programme. Proceedings of
the Second Cardington Conference, Cardington, England, 1996; 161-168.

[133] Newman, G.M. Fire Resistance of Slim Floor Beams. Journal of Constructional Steel Research 1995;
33 (1-2): 87-100.
 49

[134] Newman, G.M. and Walker, H.B., Design Recommendation for Composite Floors and Beams
Section 2: Fire Resistance. Steel Construction Institute, 1983.

[135] NFPA/SFPE, Handbook of Fire Protection Engineering (2nd Edition). Society of Fire Protection
Engineers, National Fire Protection Association, 1995.

[136] Nilsson, L. Time Curve of Heat Release for Compartment Fires with Fuel of Wooden Cribs.
Division of Structural Mechanics and Concrete Construction, Lund Institute of Technology,
Sweden, Bulletin 36, 1974

[137] NullifIre S607 - Fire Painting, CertifIed Directions for Use (in Finnish), TRY-18-1996, Helsinki,
1996.

[138] Oksanen, T. and Ala-Outinen, T. Determination of Fire Resistancs of Concrete Filled Hollow Steel
Sections. Technical Research Center of Finland , Research Notes 1287, Espoo, 1991.

[139] Ove Arup & Partners, Design Guide for Fire Safety of Bare Exterior Structural Steel. Technical
Reports, American Iron and Steel Institute, Constrado, London, 1977.

[140] O'Connor, M.A. Numerical Modelling of Composite Structures Subjected to Thermal Loading.
Proc. of the Second Cardington Conference, Cardington, England, 1996: 127-131.

[141] Paulsen, O.R. and Hadvig, S. Heat Transfer in Fire Test Furnaces. Journal of Fire and Flammability,
vol.8, 423-442

[142] Pettersson, O. The Possibilities of Predicting the Fire Behaviour of Structures on the Basis of Data
from Standard Fire-Resistance Tests. Lund Institute of Technology, Lund, Sweden, 1971, 43p.

[143] Pettersson, 0., Magnusson, S.E. and Thor,]. Fire Engineering Design of Steel Structures. Swedish
Institute of Steel Construction, Publication 50, Stockholm, Sweden, 1976.

[144] Pettersson, O. Rational Structural Fire Engineering Design, Based on Simulated Real Fire Exposure.
Fire Safety Science, Proceedings of the Fourth International Symposium, June 1994: 1-24.

[145] Plank, R., Burgess, I. and Bailey, e. Modelling the Behaviour of Steel-framed Building Structures by
Computer. Proceedings of the 2nd Cardington Conference, 1996: 145-159.

[146] Poh, KW. and Bennets, I.D. Behaviour of Steel Columns at Elevated Temperatures. Journal of
Structural Engineering 1995; 121 (4): 676-684.

[147] Poh, KW. and Bennets, I.D. Analysis of Structural Members under Elevated Temperature
Conditions. Journal of Structural Engineering 1995; 121 (4): 664-675.

[148] Purkiss, ].A. Developments in the Fire Safety Design of Structural Steelwork. Journal of
Constructional Steel Research 1988; 11: 149-173

[149] Ramachandran, G. Economic Value of Automatic Fire Dectors. BRE Information Paper 1. pages
27 -80, Garston, 1980.

[150] Robinson,]. Fire - A Techincal Challenge and a Market Opportunity. Journal of Constructional Steel
Research 1998; 46 (1-3): Paper No.179.

[151] Robinson,]. and Newman, G.M. Cardington Fire Tests: First Results. New Steel Construction 1997;
5 (3): 23-27.

[152] Rocket, J.A. Modelling of NBS Test with the Harvard Mark V Fire Simulation. NBSIR, 81-2240,
1980.

[153] Rose, P.S., Bailey, e.G., Burgess, I.W. and Plank, R.J. The Influence of Floor Slabs on the Structural
Performance of the Cardington Frame in Fire. Journal of Constructional Steel Research 1998; 46 (1-
3): Paper No.181.
 50

[154] Schleich, J.B. and Scherer, M. Compartment Temperature Curves in Function of Opening Factor
and Fire Load. ECSC research 721O-SA/112, Activity Cl/ RPS Report No 08/90, Luxembourg,
1991.

[155] Schleich, J.B. and Cajot, L.G. Natural Fire in Large Compartments. ECSC Research 7210-
SA/517etc., B_E_F_L_NL, 1993-96.

[156] SchleichJ.B. Competive Steel Buildings through Natural Fire Safety Concept. ECSC Research 7210-
SA/522 betc, B_D_E_F_CL_NL_UK&ECCS, 1994-98.

[157] Schleich, J.B. A Natural Fire Safety Concept For Buildings -1. Fire, Static and Dynamic Tests of
Building Structures, Proceedings of the 2nd Cardington Conference, Cardington, England, 12-14
March 1996: 79-104.

[158] Schleich,J.B. Eurocode l/partZ.2: Actions on Structures Exposed to Fire. IABSE Colloquium, Basis
of Design and Actions on Structures, Delft, 1996: 71-81.

[159] Schleich, J.B. Global Fire Safety Concept According to the Eurocodes. Proceedings of Nordic Steel
Construction Conference '95, volume II, Malmo, Sweden, June 19-21, 1995: 667-76.

[160] Schleich, J-B. Global Fire Safety Concept for Buildings: Connection to ECSC Research, ECCS
Developments and CEN Standardization. In: Proceedings of International Conference on Steel and
Composite Structures, Delft, The Netherlands, 1999:8.1-8.26.

[161] Schleich, J.B. Fire Safety, Design of Composite Columns. Presented in the international conference
on Fire Safe Steel Construction: Practical Design, Luxembourg, April 1984.

[162] Schleich, J.B. Computer Assisted Analysis of the Fire Resistance of Steel and Composite Concrete-
Steel Structures. REFAO-CAFIR, Luxembourg, 1987, 293p.

[163] SCI, Structural Fire Engineering: Investigation of Broadgate Phase 8 Fire. Steel Construction
Institute, UK, 1991. -

[164] Sha, W. Fire Resistance of Floors Constructed with Fire-Resistant Steels. Journal of Structural
Engineering 1998;124 (6): 664-670.

[165] Tanaka, T., Sato M. and Wakamatsu, T. Simple Formula for Ventilation Controlled Fire
Temperatures. Fire Science and Technology, Special Issue for Kawagoe Memorial Symposium 1997;
17: 15-27.

[166] Tanaka,T. presented in the Proceedings ofJanpanese Institute of Fire, 1979; 29 (2), 1980; 30 (1,2).

[167] Tanaka, T. A Model of Multiroom Fire Spread. Fire Science and Technology 1983; 3 (2): 105-121.

[168] Thomas, P.B. and Webster, C.T. Some Experiments on Buoyant Diffusion Flames. Combustion
and Flame 1961; 5: 359p.

[169] Thomas, P.H. Fire in Model Rooms: CIB Research Programmes, BRE current paper CP32/74,
Borehamwood,1974.

[170] Thomas, P.B. Some Problems Aspects of Fully Developed Room Fires. Fire Standards and Safety,
ASTM, SP-614, Aprile 5-6, 1976, pages 112-130.

[171] Thomas, P.B. and Heselden, A.J.M. Fully-developed Compartment Fires, Two Kinds of Behaviour.
Fire Research Technology, Paper No.18,Joint Fire Research Organization, England, 1967.

[172] Thomas, P.H. and Law, M. Fire Prevention Science and Technology 1974; 10: 19-26.

[173] Thomas, P.B. and Heselden, A.J.M. Fully Developed Fire in Single Compartments. A Cooperative
Research Programme of the Conseil Internationale du Batiment. CIB Report No.20, Fire Research
Note No.923, 1972.
 51

[174] Thomson, G. and Preston, R.R. Towards Harmonized Standard Fire Resistance Testing. Fire Safety
Journal 1996; 27: 91-112.

[175] Translated by Sun, J.x., etc. Global Fire Safety Design for Building Structures. By Japanese
Constructional Ministry (1983). Scie. & Techn. Translation Ltd in TianJin, China, 1994.

[176] Tschemmernegg, F. Composite Frames with Slim Floors. Journal of Constructional Steel Research
1998; 46 (1-3).

[177] Twilt, L. and Witteveen, J. Calculation Methods for Fire Engineering Design of Steel and Composite
Structures. Proceedings of International Conference on Design of Structures Against Fire, held at
Aston University, Birminham, 1986: 155-176.

[178] Twilt, L., Cajot, L.G. etc. Input Data for the Natural Fire Design of Building Structures. IABSE
Colloquium, Basis of Design and Actions on Structures, Deft 1996: 395-405.

[179] Underwriters Laboratory Inc. Fire Severity at the Exterior of a Burning Building. Technical Report,
American Iron and Steel Institute, Washington, D.C., April 1975.

[180] Wang, Y., Cooke, G. and Moore, D. Large Compartment Fire Tests at Cardington and the
Assessment of Eurocode 1. IABSE Colloquium, Basis of Design and Actions on Structures, Delft
1996: 419-427.

[181] Wang, Y.C Some Considerations in the Design of Unprotected Concrete-Filled Steel Tubular
Columns Under Fire Conditions. Journal of Structural Steel Research 1997; 44 (3): 203-223.

[182] Wang, Y.C Tensile Membrane Action in Slabs and Its Application to the Cardington Fire Tests.
Proceedings of the 2nd Cardington Conference, 1996: 55-67.

[183] Weatherford, w.o. Scaling of Flames Above Free-Burning Structural Models. Combustion and
Flame, Vol.14, 1970, 21p.

[184] Webster, CT. and Smith, P.G. The Burning of Well Ventilated Compartment Fires, Part IV, Brick
Compartment, 2.4 m Cube. JFRO, Fire research Note No.578, 1964

[185] Witteveen, J. and Twilt, L. A Critical View on the Results of Standard Fire Resistance Tests on Steel
Columns. Fire Safety Journal 1981/82; 4.

[186] Williamson, RB. and Dembsey, N.A. Advances in Assessment Methods for Fire Safety. Fire Safety
Journal 1993; 20: 15-38.

[187] Wong, M.B. and Patterson, N. Unit Load Factor Method for Limiting Temperature Analysis of Steel
Frames with Elastic Buckling Failure Mode. Fire Safety Journal 1996; 27: 113-122.

[188] Wu, B., Ma, Z. and Ou, J. Experimental Study on Seismic Properties of RC Columns after High
Temperature. China Civil Engineering Journal 1999; 32 (2): 53-58.

[189] Wu, B., Ma, Z. and Ou, J. Seismic Behavior of Fire-Damaged R.C Columns. Earthquake
Engineering and Engineering Vibration (Chinese Journal, English version), Special Issue for 30th
World Conference of Geology 1996; 16: 76-86.

[190] Wu, B., Ma, Z. and Ou, J. Stress-Strain Curves of Fire-Damaged Concrete under Cyclic Axial
Loading. Earthquake Engineering and Engineering Vibration (in Chinese) 1997; 17 (3).

[191] Wu, B., Ma, Z. and Ou, J. Experimental Study on Deformation and Constitutive Relationship of
Concrete under Axial Loading and High Temperature. Chinese Journal of Building Structures 1999;
20 (5); 42-49.

[192] Wu, B., Ma, Z. and Ou, J. A Study on Seismic Behavior of Fire-Damaged RC. Compression-
Flexural Components. Earthquake Engineering and Engineering Vibration (Chinese Journal) 1994;
14 (4): 24-34.
 52

[193] Wu, B., Li, H. and Ma, Z. Damage Analysis of Fire -Damaged R.c. Structures under Low-Cyclic
Loading. J oumal of Harbin University of Civil Engineering and Architecture 1997; 30 (3): 7-12.

[194] Wu, B., Ma, Z., etc. Fire Disaster and Its Retrofit of Shenyang Department Store. Chinese Journal of
Catastrophe 1996; 11 (4): 48-53.

[195] Wu B. and Ma, Z. Damage Assessment and Seismic Retrofit of R.c. Structures after Fire. Theories
and Practices of Structural Engineering, Seismological Press - Proceedings of the 5th International
Symposium on Structural Engineering for Young Experts, Shenyang, P.R. of China, 1998: 599-607.

[196] Yokoi, S. Study of the Prevention of Fire Spread Caused by Hot Upward Currents. Building
Research Institute, Report No.34, Tokyo.

[197] Odeen, K. Theoretical Study of Fire Characteristics in Enclosed Spaces. Royal Institute of
Technology Division of Building Construction, Bulletin, N 0.11, 1963.

[198] Odeen, K. Temperature Development in Externally Fire Exposed Rooms. Fire Safety Joumal1991;
17: 165-170.
 53

Appendix A

Parametric Temperature-time Curves of Medium

Compartment Fires for Structural Design.

Ma, Z. and MakeEiinen, P.,

Fire Safety Journal, vol. 34, no. 4, pp.361-375, June 2000.

54
 55

Parametric temperature-time curves of medium

compartment fires for structural design

Zhong cheng Ma, Pentti Makelainen *

Laboratory of Steel Structures, Department of Civil and Environmental Engineering,
Helsinki University of Technology, P.D.Box 2100, FIN-02015 HUT, Finland

Abstract

A parametric temperature-time curve for structural fire design purposes is developed

for small and medium compartment fires. This method may be used, with reasonable
precision, to estimate the temperature history of a fully developed compartment fire. It
concerns mainly cellulosic (wood-based) fires. This method has three main
characteristics: (1) the temperature-time curves are determined by two key parameters of
fire severity: maximum gas temperature (intensity) and fire duration; (2) the maximum
gas temperature and fire duration are determined by the fire load density (or surface area
of fuel), opening factor, geometric shape and thermal properties of compartment
boundaries; (3) both ventilation-controlled fire and fuel-controlled fire are explicitly
distinguished in this method. Comparisons of the predicted curves with the test results
and computer simulation show the applicability of this method.

Notation
Af floor area of fire compartment, m 2
Afuel surface area of the fuel bed, m 2
At total surface area of compartment boundaries (openings included), m2
Aw area of openings, m 2
Aw-Jh ventilation factor, m 5/2
b =.J pAc , thermal property of compartment boundary, Jim2 sy,2K
c specific heat of boundary material, J/kgK
D depth of fire compartment, m
g acceleration due to gravity, =9.81 mls2
Go total mass of fuel, kg of wood
 56

h height of the opening, m

kf conversion factor
o = Aw.,j h I At' opening factor, m l12
q fire load per unit floor area, kg of wood 1m2
R rate of burning, kg of wood Imin
t time, min
to time corresponding to the flashover, min
tm time corresponding to the maximum gas temperature, min
To reference temperature, C 0

Tg fire temperature, C0

T gm maximum fire temperature, C 0

Tgmcr maximum fire temperature in the critical region, C 0

W width of the wall which contains the opening, m

() shape constant of the temperature-time curve
11 = At / Aw -Jh , inverse of opening factor, m- 1I2
llcr the value of 11 in the critical region, m- l12
<p surface area ratio of fuel, m 2/kg
K = A/At, ratio of floor area (AI) to the total surface area (At)
A thermal conductivity of compartment boundary, WImK
P density of the compartment boundaries, kg/m3
Pa density of air, kg/m3
't fire duration, min

1. Introduction

In modem fire safety engineering, the design of structures is moving from the
traditional prescriptive method to the performance-based methodology [1,2]. In
performance-based structural fire-resistant design methods, natural fire curves are used
instead of the standard fire curve [2,3]. The performance of structures will be studied
under natural fires.
A typical compartment temperature evolution is illustrated in Fig.l. From the point
of structural fire resistant design, most structural damage occurs during the fully
developed fire phase. So the phase of fire growth is often neglected and only the fully
developed fire phase and the decaying phase are taken into account. It is convenient to
regard to as the origin of the temperature-time coordinate system, corresponding to the
point of flashover.
There are numerous ways to model the fire scenario in a building compartment, such
as zone-modelling and CFD (Computer Fluid Dynamic) modelling. These models have a
high accuracy in calculating the temperature history. However, these models are
sophisticated and the appropriate software package is very large. They are not very
convenient tools for structural design purposes, which need only the temperature history
of the fully-developed stage.
 57

Natural Fire Development

Fig. 1. Typical compartment evolution.

An approximate method to describe the temperature evolution of a fully developed

compartment fire has been considered desirable since the earliest studies. Kawagoe and
Sekine [4] in their pioneering work presented an iterative, graphical technique for
obtaining approximate solutions. Pettersson et al. [5] developed a series of design curves
for ventilation-controlled fires, based on the conventional mass and energy balance
equations, using the heat release rate as input and solving the equations iteratively by
computer code. Babrauskas [6] gave an approximate method by fitting the formulae with
the calculated results using computer code COMPF2. This method requires detailed
information of the fuel as input. Lie (1974) presented an expression based on a series of
temperature-time curves computed by Kawagoe and Sekine, which could be used as an
approximation for the most severe fire that is likely to occur in a particular compartment
[7]. His work formed the basis of Annex B of Eurocode 1 Part 2.2.
The purpose of the present research is to produce a simple calculation procedure
which will, to a suitable degree of precision, estimate the temperature history in a fully
developed compartment fire. It is this kind of procedure which becomes useful for
making quick fire design calculations, without using computer code.

2. Basic idea

The past theoretical and experimental studies have revealed that the following
variables determine the fire temperature
Fire loads;
Ventilation opening;
Geometric size and shape of the fire compartment;
Thermal properties of compartment boundary.
The primary effect of fire on a structure depends on the fire intensity (maximum
temperature) and fire duration. These two parameters may be used as the controlling
parameters of the fully developed temperature-time curve. Then a simple and general
 58

shape function can be selected to define the temperature history. The maximum
temperature and fire duration are dependent on the fire loads, ventilation conditions,
geometry and material properties of the compartment. To identify fuel-controlled and
ventilation-controlled fires, also the conditional equation of the burning region has to be
determined.
Keeping the above in mind, a shape function will flrst be established, after which the
expressions for the maximum gas temperature and fIre duration will be presented.

3. Development of this model

To determine a general shape function of the temperature-time curves, some measured

results from the past flre tests are collected. A total of 25 complete gas temperature- time
curves are collected, including BISF/JFRO (1966-1967)(11) [8,9,10], Underwriters
Laboratory Inc.(1967) (7) [11], BSC/FRS (3) [12], Research Institute of Japanese
Constructional Ministry (2) [13], Sonderforschungsbereich148 (SFB 148) (1) [14], and BRE
flre tests in Cardington (1993)(1) [15,16]. The data set from CIB small-scale flre tests [17,18]
will be used to determine the maximum gas temperature for ventilation-controlled flre. The
data mainly from BISF/JFRO will be used to determine the maximum gas temperature and
fIre duration for fuel-controlled flre. The curve of the recent BRE flre test is not used in the
derivation. It is used only for comparison purposes.
Due to the fact that the published or measured data of past fire tests are insufficient in
a single fire test, the comprehensive information will be used in the following derivation
of the new parametric equations.

3.1 The shape/unction

The 25 complete gas temperature- time curves collected from different laboratories are
non-dimensionalized by the maximum gas temperature and the time to reach the
maximum temperature, which are illustrated in Fig. 2. The time is measured from to
instead of zero, where to is the time corresponding to the flashover. For all of these
curves, the fire load density is 10 - 40 kg/m2 , the ventilation factor Aw -J h is 5 - 16 m5/2
and the thermal properties of compartment boundary .J pAc is 555 - 1800 J / m 2 SI/2 K .
The floor area of the compartment is less than 100 m2 and the maximum height is 4.5 m.
The shape ofthe compartment (ratio of depth to width) for the full-scale tests is 0.5 - 2.
It can be seen that these temperature-time curves are of similar shape. A non-
dimensional expression based on curve fitting procedure can be given by the equation

(1)

where Tg = gas temperature (C); Tgm = maximum gas temperature (C); To = room
temperature, (20°C); t = time (min); tm = time corresponding to the maximum gas
 59

1.2
~
~
1
h; 0.8
<:>

..
.~
(':i
0.6
..
III

...=
..
(':i
III
0.4

0..
e
III
0.2
~
o
o 0.5 1 1.5 2 2.5 3
Time ratio tit",

Fig. 2. Non-dimensionalized temperature-time curves.

temperature (min); 0 = shape constant of curve. For the ascending phase and the
decaying phase, 0 has different values. The constant for the descending phase is
approximately twice that for the ascending phase. Figure 2 shows that the shape constant
0= (0.8,1.6) has the best fit with the data, and 0 = (0.5,1.0) gives a fuller curve shape.
Since choosing 0 = 0.5 for the ascending phase and 1.0 for the descending phase gives an
approximate upper envelope for these measured curves, these values could be used for
design purposes.

3.2 Critical region for ventilation-controlled andfuel-controlled fires

For a fully developed compartment fire, the rate of burning is dependent on the
opening size if the air supply from the openings is insufficient (ventilation-controlled);
the rate of burning is dependent instead on the surface area and burning characteristics of
the fuel if the air supply is ample (fuel-controlled). It is found that the transition from
ventilation- to fuel- control depends predominantly on the surface area of fuel and the
opening size [7].
Harmathy [19] analysed many experimental data and developed a semi-empirical
equation to determine the critical region of compartment fires, which can be written as

pag'h Aw Jh =0.263 (2)

A fuel

where Pais the density of air (kglm 3), g is the acceleration due to gravity (m/s2). Aw is the
area of windows (m2) and h is the height of the windows (m). Ajilel is the surface area of
the fuel bed (m2). The surface area of the fuel bed is a variable which is somewhat
 60

--
U
~
'-'
1400
1200
~
~
... •
1000
+ ... + -
+ •
• +
I\) + +

~ 800 + +
+ +
'"
I\) 600 + CI8 •
8
....
I\) 400 •
...
Japan
JFRO

~ 200 -Eq.6
~
0
0 10 20 30 40 50

Inverse of opening factor 7J (m1f2)

Fig. 3. Maximum temperature Tgm versus 11 for ventilation-controlled fire.

difficult to use in practical applications at this moment, because different fuels exist in
the actual compartments. For this reason, Ajuel is expressed as

Ajuel = q>Go (3)

where Go is the total fire load of equivalent woods (kg); <p is the surface area ratio of fuel
(m2/kg). For general furniture, the fuel surface area ratio is between 0.1 and 0.4 m 2/kg.
The most common value of <p is between 0.12 and 0.18 m 2/kg.
Taking Pa=1.205 kg/m3 as the density of air and g=9.81 mls 2 as acceleration due to
gravity, Eq. 2 can be expressed in parameters of the fire load density and the opening
factor as

0/ q =O.071<.xp (4)

where 0 is the opening factor, = Aw.J h / At (m'l2); q is the fire load density per unit floor
area (kg of wood 1m2); and K is the ratio of floor area (Af) to the total surface area (At) in
a fire compartment.
Equation 2 for the critical region can also be written as

11 cr = 14.34/
/lCq>q (5)

where llcr is the inverse of the opening factor in the critical region (m-'l2).
 61

1
§
.S
....
..
to 0.8
~ ... 0 0

....to::I ~ 0.6 8
.. E-<
[ ~" §
E ~ 0.4
2 E-< ! o
Eq.7!
JFRO
K
to 0.2
::g
0
0 0.2 0.4 0.6 0.8 1

Ventilation ratio T)/T) cr

Fig. 4. Non-dimensionalized maximum temperature for fuel-controlled ftre.

3.3 Determination a/the maximum gas temperature

For ventilation-controlled fire, it is well known that the compartment temperature

depends mainly on ventilation conditions. CIB fire tests provided about 25 data points of
the maximum temperature [17]. Data points from the Research Institute of Japanese
Constructional Ministry (briefly referred to as Japan in the figures) (2 points) and
BISF/JFRO (2 points) are also proved to be ventilation-controlled according to Eq. 5, and
are shown in Fig. 3.
Using the least-squares method, the maximum temperature for a ventilation-
controlled fire can be given by the expression

Tgm = 1240 - 11 . 1] (6)

where 11 = AI 1Aw.,j h , inverse of opening factor (m-v,).

For fuel-controlled fire, the main complete data are from BISF/JFRO tests. The main
results are listed in Table 1. The surface area of fuel bed Ajuel is about 0.131 m2/kg. The
heat loss through the part of the floor covered by the fuel bed is thought to be negligible
due to the shielding of the fuel bed. The area of the floor was thus not included here in
the calculation of the total surface area of the compartment.
From Fig. 4 it can be seen that the ratio of Tg,,/Tgmcr has a good correlation with the
ventilation ratio 11/11cr . Using the least-squares method, the maximum gas temperature of
fuel-controlled fire can be expressed as

(7)
 62

Table 1. Results and analysis offrre tests by BISF/JFRO

Reference Inverse of Total Fire load Critical Critical Max.value

letter of opening fuel per unit value 11 cr max. of measured
Tgmcr
1I2
tests factor Go floor area 1]cr(m- ) temp. average gas
1] (m- 1I2) (kg) q (kglm2) Twn cr temp. ( C)
A 12.75 218 7.5 48.6 0.262 705 315 0.447
B 12.75 218 7.5 48.6 0.262 705 340 0.482
D 6.40 436 15 24.3 0.262 973 465 0.478
E 6.40 872 30 12.2 0.524 1106 755 0.683
F 6.40 872 30 12.2 0.524 1106 800 0.723
G 12.75 436 15 24.3 0.524 973 700 0.720
H* 12.75 872 30 12.2 1.04 1106 1040 0.940
r* 12.75 872 30 12.2 1.04 1106 1035 0.936
K 6.40 218 7.5 48.6 0.13 705 275 0.390
L* 6.40 1744 60 6.1 1.04 1173 1055 0.900
p# 6.40 218 7.5 48.6 0.13 705 272 0.386
0# 12.75 218 7.5 48.6 0.262 705 488 0.692
N 6.40 872 30 12.2 0.524 1106 645 0.583
R# 6.40 872 30 12.2 0.524 1106 816 0.738
M* 12.75 872 30 12.2 1.04 1106 1027 0.929
Q*# 12.75 872 30 12.2 1.04 1106 1060 0.958
X 12.75 436 15 24.3 0.524 973 580 0.596
if 12.75 1744 60 6.1 2.10 1173 1160 0.989
y+ 27.5 1744 60 6.1 4.20 1173 1115 0.951

Notes The side dimension of the square cribs is 45 mm, the length is l.1m and the density is 500kglm3;
Size of compartment is 7.7x3.7x 3 m; A}Uel :0.131 m2/kg.
* : frre in critical region; +: ventilation-controlled fire; #: with linings of mineral wool slab

where Tgmcr is the maximum temperature in the critical region, which can be obtained by
substituting 11cr into Eq. 6. 11cr is the value of 11 in the critical region, which can be
determined by Eq. 5.
Figure 5 shows the possible maximum fire temperature in a compartment in the
whole region. It can be seen that the maximum fire temperature declines with increasing
11 (i.e. decreasing ventilation) for ventilation-controlled fire, while it increases with 11 for
fuel-controlled fire. The peak value of the maximum fire temperature in the case of
specified fire loads occurs in the intersection of fuel and ventilation controlled regions.
 63

1200

--
U
0,-,
1000
~
}-I 800
<U

~I...
600
<U

...8"
400
<U

~ 200
:?::
0
0 10 20 30 40 50
Inverse of opening factor T} (m-"")

Fig_ 5_ Max. temperature of fully developed fire versus ventilation and fire load.

3.4 Fire duration and time to reach the maximum temperature

Assuming a constant burning rate R during the entire period of primary burning, the
fire duration can be given by
(8)

where R is the burning rate (kg of wood /min), which can be given by

R = 10.8(1- exp(-0.036 '1])) . Aw.Jh . W / D (9)

for ventilation controlled fire [20], and

R =0.372cpG o (10)

for the fuel-controlled fire [19],

where W is the width of the wall which contains opening (m) and D is the depth of the
fire compartment (m).
Equations (8) and (10) show that for fuel-controlled fire, the fire duration is dependent
only on the fuel surface area ratio. For a general office building, <p is 0.12 - 0.18 m 2/kg
and the corresponding duration of fuel-controlled fire is about 15 - 22 min.
The time to reach the maximum temperature can be obtained from the fire duration 't,
which is illustrated in Fig. 6. It can be seen that 't/tm is mostly between 1.4 and 2.0,
mainly distributed between 1.4 and 1.7. On the basis of Fig. 6 and the sensitivity analysis
in section 4.2, the following formula is approximately valid:

r = 1.6· t m or t m = 0.63 . r (12)

 64

0.5

0.4
~
...=
~

""
0.3
0'"
~

'"' 0.2
~

0.1
0
1.45 1.60 1.75 1.90 1/tm

Fig. 6. Distribution diagram of r / t m .

3.5 Influences o/thermal properties of the fire compartment

The heat loss through the compartment boundaries (such as walls, ceiling, floor) is
about 1O~30% ofthe heat release, according to the BISF/JFRO fire tests [10]. It is also an
important factor influencing the gas temperature-time history of the compartment. An
approximate way to consider the influence of the thermal properties of compartment
boundaries is to convert the actual fire load and opening factor into the equivalent fire
load and the equivalent opening factor, respectively [5]. These equivalents can be
obtained by multiplying the actual fire load density q and the opening factor 0 by the
factor kj, respectively. Reference [5] gives the proposed values for several common types
of compartment boundaries.

4. Sensitivity analysis

4.1 Surface area ratio offuel

To show the effect of the surface area ratio of fuel in this method, some calculations
are performed. Since the surface area ratio of fuel only has effects on fuel-controlled
fires, it makes sense that the calculation parameters for fuel-controlled fires are desirable.
The calculation results are shown in Table 2. The numbers in brackets represent the
calculated maximum fire temperature (C) and duration 't (min). It can be seen that the
fuel surface ratio has a significant influence on the fire curve. If the fuel has a low value
of surface area ratio, the fire has a lower maximum temperature and a longer fire
duration; accordingly, for the fuel with a high value of surface area ratio, the fire has a
higher maximum temperature and a shorter duration.
 65

0.5
'@j 0.4
~. "'.'''.
1
Q.
0.3
.,.....
...

.. ..... ~- .........
"
c ....................-..".
0.2
..
',d
l':I
l':I
-~---.--~-----------------
III 0.1
~
III
'.l
l':I 0
'i
~ 0 10 20 30 40 50 60
Size of square wood crihs (mm)

5 10 15 20 25 30
Thicknes of wood slah (mm)

Fig. 7. Relations between surface area ratio <p and the size of wood fuel.

Figure 7 shows the relationship between the surface area ratio of wood and the size
of wood pieces. For general office buildings, as mentioned in section 3.2, the most
common value of fuel surface ratio is between 0.12 - 0.18 m2 /kg, which corresponds to an
average furniture thickness of 15 - 25 mm. Figure 7 can be used as a direction for the
evaluation of <p .

Table 2 Sensitivity study: effect of fuel surface on fuel-controlled fires.

~ 0.09 0.12 0.15 0.18 0.21

11=5 q=15 (221,28) (305,21) (374,17) (434, 14) (488, 12)

q=30 (834,28) (894,21) (950,17) (1002, 14) (1053, 12)
11 =8 q=15 (279,28) (385,21) (473,17) (550, 14) (617,12)
q=30 (1055,28) (1130,21)
11=16, q=15 (395,28) (545,21) (669,17) (777, 14) (873,12)

Table 3 Variables of the sensitivity analysis for t"l"

I~
1m/" = 0.5,0.6,0.7, q =15,60 kg of wood
11 =16 m-l", <p=0.15 m2/kg
 66

4.2 Ratio o/tm andfire duration

The ratio of the time to reach the maximum temperature tm and the fire duration 't is
between 0.5 - 0.7 (Fig. 6). Two values of fire load density q are taken so that fuel-
controlled and ventilation-controlled fires are produced. The other parameters are listed
in Table 3. The calculated temperature-time curves are illustrated in Fig. 8. It can be seen
that the resulting difference appears mainly in the decaying phase, and it is also not so
significant. An intermediate value of 0.63 is desirable.

II 16 =
800 q = 15
:p= 0.15
..-.. 700 ~~....:~--.:.-..;:-._. k= 311 4
U
CI
'-'
600 -----==;~
....... .::..:" ...........
-.. " _-. , .. W ID = 1/3
.... .. -. -.
~ ........
500
... ~

~\00
4D0 "'----..
....
.........
-

~
300 ----,~ - ... 0.7
S-
~
200 05 0,6
}-I
100
0
0 5 10 15 20 25 30

'lime (min)

Fig. 8 (a). Sensitivity analysis: influence oftn/'t (fuel-controlled fIre).

11=16
q=60
1200
cp=O.15
.-
U 1000 k=3/14
--=.
0

IIJ

....t':
800
..
....
W!D=ll3
- ...........
... ...... -.:.-. 0.7
..
IIJ
600 .....
0.6
=-e 400 0.5
~ 200
0
0 20 40 60 so 100 120
'Iime(min)
Fig. 8(b). Sensitivity analysis: influence oftm/'t (ventilation-controlled fIre).
 67

5. Comparisons with fire tests and computer simulation

5.1 JFRO fire tests

Relatively complete data about the fire curves were published in BISF/JFRO fire tests
[9], which will be compared with the new procedure. It may be argued that since the data
from these tests have been used in the previous derivation of this method, the
comparisons may not be valid. As noted above, the comprehensive information is used in
the derivation of this method. No single piece of test data was used through the whole
derivation. In fact, the data from the BISF/JFRO tests were only used in the derivation of
Eq.7.
The curves calculated by this method and the measured curves are compared in Fig.
9. In the figure, the dashed lines show the predictions by the new method and the solid
lines show the average measured curves. A good agreement between the measured curves
and the predictions by this method can be seen. The related parameters, such as the
inverse of opening factor 11, the fire load Go and the size of compartment are presented in
Table 1.

5.2 BRE fire tests

One of the fire curves from recent Cardington fire tests by BRE (Building Research
Establishment) in the UK was published in reference [15]. The BRE fire test was
conducted in one comer of the building, simulating the dimensions of a typical office
room. This room measured 9 m long, 6 m wide and 4.185 m high. The floor of the fire
compartment was constructed of concrete with corrugated steel decking, which is covered
by sand. The external end wall was made of lightweight concrete blocks. The remaining
internal walls of the fire compartment were formed using plasterboard. The window is 9
m long and 2.685 m high with double-glazing. The fire load consisted of 40 kg/m2 of
timber distributed uniformly over the compartment floor.
The comparisons between measured and calculated curves are made in Fig. 10. The
measured curve is treated according to Fig. 1 so that only the fully developed phase and
the decaying phase are presented. It can be seen that the calculated curve from the
proposed model agrees well with the measured curve.

5.3 Computer simulation

A comparison between the predictions by this method and the results of computer
simulation in reference [21] is also made (Fig. 11). The one-zone model is used in the
simulation and the combustion model used in this program (Ozone) can independently
define the burning region (fuel or ventilation - controlled). In Fig. 11, the fire load is 33
kg/m2 and the values of the opening factor represent three specified fires: 0.02 for
ventilation controlled fire, 0.06 for fire in the critical region and 0.2 for fuel-controlled
 68

fire. In this comparison, the fuel surface ratio <p is taken as 0.15 m 2/kg in the calculation
of fuel-controlled fire and the ratio of width to depth of compartment (WID) is 112.
The thermal property b of the compartment boundary is 1500 J/m2 sYzK, which is
considered to be close to compartment type A defined by [5]. According to [5], the value
of Iv is taken as 1.0. Consequently, the equivalent opening factors and fire loads are the
same as the actual values. Figure 11 indicates a fairly good agreement between the curves
calculated by this method and the computer simulation.

6. Conclusions

A new method to estimate the temperature-time curve in a fully developed fire

compartment for structural design purposes, without using computer code, is developed
in this paper, on the basis of extensive test results and increasing knowledge of fire
engineering science. This method can provide a quick calculation of natural fire curves
for use in structural fire-resistance design.
A general shape function was derived, so that the two fire severity parameters can
describe the temperature-time curve in combination with this function. The fire loads,
opening factor, geometry and thermal properties of the compartment determine the
maximum gas temperature and fire duration. A sensitivity analysis was also carried out to
investigate the effects of the surface area ratio of fuel and the ratio of the time to reach
the maximum fire temperature and the fire duration in this method. The results show that
the surface area ratio of fuel cp has a significant influence on the temperature-time curve.
A method to estimate the value of fuel surface area ratio is also suggested. The ratio of
the time to reach the maximum fire temperature and the fire duration 1m /7: does not have
a significant influence on the fire curves. The proposed value 1m /7: = 0.63 provides
sufficient precision for practical use.
The predicted results are compared with experimental and simulated results. The good
agreements show the applicability of this method.
 69

1400

1200
..-
U
'L- 1000
OJ

~.... 800
OJ

S'
..,
600
..
OJ

~
400
D
200

0
0 20 40 60 80

Time (min)

Fig. 9(a). Comparisons with JFRO tests - fuel-controlled fires.

1400

..- 1200
..u
'-" 1000
'Il

~~
800
'Il
600
S-
....'Il
400
'"CI:I
d
200

0
o 10 20 30 40

Time (min)

Fig. 9(b). Comparsions with JFRO tests - ventilation-controlled fires.

 70

1200 -+- BRE test curves

•....•. <p = 0.15 m2/kg, calculated
j 1000 -
2
<p = 0.18 m /kg, calculated
::s
.jJ
as 800
~
Q)
600

.jJ
~ 400 '0=0.17 m'Y' i

iCJ=::~g~~(~~i
III
to!)
as 200

0
0 10 20 30 40 50

T:in e tn:in)

Fig. 10. Comparsions with BRE fIre tests.

............................,
2
'q = 33 kg/m
1200 - - computer modeling [21]· 2
iAt= 360 m
- - Drediction bv this method i 2 i
-. 1000 A,=lOOm i

0
U .b.=1.5illL!b.!{~~K_c
--- ~
800
':::s"'
~
'"' ~ 600 ".

S"
....
~ 400
en
tIS
t.::l 200 '.

0
0 20 40 60 80 100

Time (min)

Fig. 11. Comparisons with fIre curves by computer modelling in ref.[21].

 71

References

[1] Johnson, P.F. International Developments in Fire Engineering of Steel Structures. Journal of
Constructional Steel Research 1998; 46(1-3).
[2] Pettersson, 0. Rational Structural Fire Engineering Design, Based on Simulated Real Fire
Exposure. Fire Safety Science, Proceedings of the Fourth International Symposium, June
1994: 1-24.
[3] Schleich, J.B. A Natural Fire Safety Concept For Buildings -1. Fire, Static and Dynamic
Tests of Building Structures, Proceedings of the 2nd Cardington Conference, Cardington,
England, March 1996: 79-104.
[4] Kawagoe, K. and Sekine, T. Estimation of Fire Temperature-time Curve in Rooms. BRI
Occasional Report 11, Building Research Institute, Tokyo, 1963.
[5] Pettersson, 0., Magnusson, S.E. and Thor, J. Fire Engineering Design of Steel Structures.
Swedish Institute of Steel Construction, Publication 50, Stockholm, Sweden, 1976.
[6] Babrauskas, V. A Close-form Approximation for Post-flashover Compartment Fire
Temperature. Fire Safety Journal 1981; 4: 63-73.
[7] Drysdale, D. An Introduction to Fire Dynamics. John Wiley & Sons Ltd., 1987.
[8] Butcher, E.G., Chitty, T.B. The Temperature Attained by Steel in Building Fires. JFRO, Fire
Research Technical Paper No.15, London, HMSO, 1966.
[9] Butcher, E.G., Bedford, G.K. and Fardell, P.G. Further Experiments on Temperatures
Reached by Steel in Building Fires. Proceedings of the Symposium held at the Fire Research
Station, Boreham Wood, January, 1967: 45-56.
[10] Heselden, A.J.M. Parameters Determining the Severity of Fire. Proceedings of the
Symposium held at the Fire Research Station, Boreham Wood, January, 1967: 19-28.
[11] Underwriters Laboratory Inc. Fire Severity at the Exterior of a Burning Building. Technical
Report, American Iron and Steel Institute, Washington, D.C., April 1975.
[12] Kirby, B.R. Recent Developments and Applications in Structural Fire Engineering Design-
A Review. Fire Safety Journal 1986; 11: 141-179.
[13] Translated by Sun, J.x., etc. Global Fire Safety Design for Building Structures. By Japanese
Constructional Ministry (1983). Scie. & Techn. Translation Ltd in Tian Jin, China, 1994.
[14] Hagen, E. and Haksever, A. A Contribution for the Investigation of Natural Fires in Large
Compartment. Fire Safety Science- Proceedings of the First International Symposium,
Berkeley, Califonia, November 1985: 149-158.
[15] Wang, Y., Cooke, G. and Moore, D. Large Compartment Fire Tests at Cardington and the
Assessment of Eurocode 1. IABSE Colloquium, Basis of Design and Actions on Structures,
Delft 1996: 419-427.
[16] Lennon, T. Large Compartment Fire Tests. Fire, Static and Dynamic Tests of Building
Structures, Proceedings of the Second Cardington Conference, England, March 1996: 45-53.
[17] Thomas, P.H. Fire in Model Rooms: crn Research Programmes, BRE current paper
CP32174, Borehamwood, 1974.
[18] Law, M. Design For Fire Safety of Bare Exterior Structural Steel. 1. Theory and Validation.
2. State of the Art. Ove Arup and Partners, London, 1977.
 72

[19] Hannathy, T.Z. A New Look at Compartment Fire. Fire Technology 1972; 8(3-4): 197-217,
327-351.
[20] Law, M. A basis for the design offtre protection of building structures. Structural Engineer,
1983; 61A(1):25-33.
[21] Schleich, J.B. A Natural Fire Safety Concept for Buildings. Fire, Static and Dynamic Tests
of Building Structures, Proceedings of the Second Cardington Conference, Cardington,
England, March 1996: 79-104.
 73

AppendixB

Fire Resistance of Composite Slim Floor Beams

MakeHiinen, P. and Ma, Z,

Journal of Constructional Steel Research, vol. 54, no. 3, pp.345-363, June 2000.
74
 75

Fire resistance of composite slim floor beams

Pentti Makelainen and Zhongcheng Ma

Laboratory ofSteel Structures, Helsinki University of Technology, P. O.Box 2100, FIN-020 15
HUT, Finland

Abstract

There is increasing interest in steel-concrete composite slim floor construction in the Nordic
Countries and the UK. The slim floor beams, which are partly contained within the floor slab,
have inherent good performance under fire. At the moment, a new type of slim floor beam is
under development in Finland. In this paper, the thermal and structural performance of this new
slim floor beam is investigated under fire by using numerical analysis programs. The temperature
analysis and calculations of the load-bearing capacity are carried out under fire conditions (both
ISO standard fire and natural fire) according to Eurocode 1 Part 2.2. The structural fire resistance
behaviour of this new slim floor beam is investigated and the relationship between the fire
resistance time and load ratio under ISO standard fire is discussed. Moreover, the minimum load
ratio of the beam under a natural fire is analyzed. The critical relationship between the minimum
load ratio and the fire parameters (opening factor and fire load density) is also established,
provided that the structural failure is caused by a reduction in the flexural capacity of the fire-
exposed beam.

Keywords: Temperature analysis; Fire resistance; Slim floor beam

Notation

Af floor area of enclosure (m2)

At total surface area of enclosure (m2)
c specific heat (J/kgK)
di centroid distance (mm)
e specific volumetric enthalpy (J/m 3)
h distance of plastic neutral axis to the top side of floor slab (mm)
Kint interface thermal resistance coefficient (W/m2K)
Ko thermal conduction term (W/m2K)
Ki thermal conduction term (WIm2K)
Mu plastic bending moment capacity of composite cross-section (kNm)
Ms sagging bending moment capacity of composite beam (kNm)
 76

Mh hogging bending moment capacity of composite beam (kNm)

M, applied bending moment of simply supported beam (kNm)
o opening factor (mY:,)
q heat input rate per unit area (W1m2)
qf fire load density per unit floor area (MJ/m2)
R load ratio
Rer critical load ratio
Rj thermal resistance term (m2K/W)
T temperature (OC)
Tg fire temperature (OC)
Ts surface temperature of structural component (OC)
.,j pAc thermal property of enclosure boundary (J/m 2 sY:,K)
ex convection coefficient (WIm2K)
E resultant emissivity coefficient
<p volume ratio of moisture content in concrete
<l> diameter of reinforcement bar (mm)
"A thermal conductivity (WImK)
P material density (kg/m3)
0' Stephan-Boltzmann constant, a=5.67xlO- 8 W/m 2K4

1. Introduction

In recent years, increasing interest has been shown in the Nordic Countries and the
UK in developing and designing shallow floor systems in steel-framed buildings. In the
shallow floor system, the steel beam is contained within the depth of the pre-cast concrete
floor or composite slab with profiled steel decks. This form of construction achieves a
minimum depth of building and the flat floor is beneficial because the building services
can be run in any direction. Structurally, the shallow floor system has inherent good fire
resistance by virtue of the partial encasement of the steel section. The shallow floor can
be designed using various forms of steel beams comprising either rolled or welded
sections. The word 'slim floor' is used in the Nordic Countries and the UK for this type
of construction. Some examples of shallow floor beam sections are shown in Fig. 1.

One of the original slim floor concepts developed in Scandinavia was the 'Thor-beam'
(Fig. la), which consists of two channel sections welded to a flat plate. Additional angles
are welded to the top flanges to provide the shear connection. British Steel pIc and the
Steel Construction Institute (SCI) have developed a Slimflor® beam (Fig. 1c), which
consists of a universal column section welded to a steel plate. Recently, interest has been
concentrated on the asymmetric hot-rolled steel beam (Fig. Id) in the UK, and on the
asymmetric welded steel beam in Finland.
 77

rl
[J ,I
.
\
. -\
(a) Thor-beam (b) Delta-beam

R
(c) Slimflor beam (d) Asymmetric slim floor beam
Fig. 1. Different types of slim floor beams.

In Finland, a new asymmetric slim floor beam is under development. The objective is
to develop a new section shape for the slim floor beam, which is suitable to apply to the
present Finnish composite deck. In this paper, the fire resistance of the new slim floor
beam is investigated both in ISO standard fire and natural fire situations.

2. Section shape of the new slim floor beam

The section shape of the slim floor beam is illustrated in Fig. 2. The depth of the floor
slab is 300 mm and the height of the profiled steel deck is 117 mm (Rannila 120). The
steel beam consists of three plates welded together. The web plate is welded to the
bottom and top flange by using the submerged arc welding method. The new asymmetric
steel beam has a relatively thin top flange with a thickness of 10 mm and a web with a
thickness of 20 mm. The bottom steel flange is 18 mm thick and a 400 mm wide. The
height of the steel beam is 258 mm. The new floor system has a concrete slab depth of
183 mm over the steel deck, which is demanded by the sound insulation requirements.
The section gives an efficient fire resistance to the floor system. This is presented in
Section 4.

3. Temperature analysis

The thermal response of the new slim floor beam in fire is analyzed using the computer
program, TACS-FIR (Temperature Analysis of Composite Structures Exposed to FIRe),
which is a two-dimensional finite difference program developed in the Laboratory of
Steel Structures at Helsinki University of Technology [1]. The moisture content of
concrete can be analyzed and the interface resistance of heat transfer between concrete
and steel can be
 78

200
1-< ~I

I.
r, I
~_ _ _ _ _ _ _ _~40~0_ _ _ _ _ _ _ _~>~1

Fig. 2. Section shape of the new slim floor beam.

modelled with this program. The material properties of concrete and steel (thermal and
mechanical) are temperature dependent. The fire temperature-time curves can be chosen
according to the ISO standard fire or any kind of natural fire. There are several
collections of fire temperature-time models and temperature-dependent material
properties in this program.

In this paper, the thermal and mechanical properties of the steel and concrete are given
in accordance with Eurocode 3 Part 1.2 and Eurocode 4 Part 1.2 [2,3], and the fire
temperature-time curve for natural fire is according to Annex B of Eurocode 1 Part 2.2
[4].

3.1 Heat transfer model

The transient heat transfer equation based on Fourier Law is given as

(1)

where T is the temperature (OC), A is the thermal conductivity matrix (WlmK), e=de / dt
is the rate of specific volumetric enthalpy change. The gradient operator V is defined as

[ a a a]T
V=axayaz (2)

where x, y and z are Cartesian co-ordinates. Superscript T is the transfer operator of the
vector.
 79

.2

• 1 .0 .3

.4

I Ax I
Fig. 3. Local system of an internal element.

By some intermediate transformations, the above equations can be derived as follows

(3)

where pc is the volumetric specific heat (J I m 3 K ), p wCw is the volumetric specific heat
of water, qJ is the volume ratio of moisture content. Ti is the temperature of the ith
element in the local system (Fig. 3). Ri is the thermal resistance term, which can be given
by

1 1 1
R.=-+-+- (4)
K
1 Ko Kint
j

where Ki and Ko are the thermal conduction terms. K int is the interface resistance
coefficient. In [5] the value K int =50 W Im 2K is proposed for the interface coefficient
between steel and concrete. This value is also taken for this study. In the case of perfect
contact, Kin! = 00can be presumed. It is assumed in the program that the interface
resistance is caused by the evaporation of water [1].

The heat input rate q is expressed as

(5)

where (J' is the Stephan-Boltzmann constant, (J' =5.67xlO-8 Wlm2IC; e is the resultant
emissivity coefficient. Following the work in [6] and [7], e=0.6 is adopted for the bottom
steel flange and e =0.3 for the composite floor. a is the convection factor, which is
assumed as 25 WI m 2 K for the exposed side and as 8 WI m 2 K for the unexposed side
[8].
 80

,,'···.0=0.15 0=0.06 mY, ISO standard fir

f ·~
\
\~F1.1~? •• '-OFllOO M.lLi~-~:"':':';""";;'--
..
f ./"I
,/'( '.,
'/ \
'.
\
"\.
'"
0=0.06
"
.. , . .
, \ • 'ClF550 " •
\ \" 'Ito',
0=0.15 \\ '. "" ' " ,,
qF550 \ ,
\. " '.
\ \... . ""-

Heating time (min)

Fig. 4. Fire temperature-time curves.

3.2 Temperature-time curves offires

The Eurocodes allow the structures to be calculated either under the ISO standard fire
or under specified parametric fires. In this paper, the temperature distributions of the new
slim floor beam are analyzed both under the ISO standard fire and under the parametric
natural fires as defined by Eurocode 1 Part 2.2.

The ISO standard fire curve is an idealized temperature-time curve, which has only
the ascending phase shown by the solid line in Fig. 4. In reality, the temperature-time
curves of fires can vary widely. The two most important factors that determine the fire
temperature-time curves are the fire load and the size of openings through which the air
can enter. Usually, the fire load is characterized by the fire load density per unit floor
area, and the size of opening by the opening factor. Fig. 4 shows how the fire load and
the opening factor affect the temperature-time curves, primarily in relation to the
intensity and the duration of fire.

3.3 Temperature distribution

The analysis program is validated extensively by the two Slimflor® test results of the
SCI, some composite columns and other types of composite beams [1]. The cross-section
of the beam is modelled using a rectangular grid. The division of the slim floor section
into cells is shown in Fig. 5.
 81

The temperature distribution of the new slim floor beam at 60 minutes under ISO
standard fire is illustrated in Fig. 6. It shows that the bottom steel flange has a high
temperature (about 750 °e on average) and there exists a large temperature difference,
around 200 °e, between the end and the central point. This phenomenon is caused by the
two-way heat input at the end of the bottom steel flange. Within the web, an extremely
non-uniform temperature distribution exists due to the encasement of the concrete. The

,-','~"'"-'-.-"--~'--~--'--~,
~' .. ,".",~.---.-~.-~---~,--~
'~.-;""---~' -~--~-------~~-~--~'~.
c,7"~., _-==----===concrete
c-:>,:·~,·~,~----'-'elements
symmetric ~
elements steel ~ symmetric

},_:---~ ..~--..-~....._~._..._~._~~el~e~_en_t_s~_._ '" elements

interface between
steel and concrete
----._--/----
/

Fig. 5. Division of section into cells for temperature analysis.

Time=60 min ts=0. 6 tc=0.3 CXL =25 W/m2 K <Xu = 8 W/m2 K Kint =50 W/m2 K

29 29 29 28 27 25 24 26 29 33 36
36 36 35 34 32 29 27 28 34 41 46
58 58 57 55 50 41 34 36 48 60 78
109 108 107 102 91 71 50 40 .' 100
212 212 210 203 181 132 80 51 49 76 130
431 431 429 420 383 257 129 72 62 91 172
721 721 720 713 678 409 173 91 75 102 206
584 217 113 92 119 245
679 277 158 132 159 317
727 349 235 207 233 410
,: 530
867 832 795 756 719 653

Figures in bold and black shade are for the steel parts.
The others are for the concrete parts of the section.

Fig, 6. Temperature distribution in the new slim floor beam under ISO fIre (at 60 minutes).
 82

------ .-.- .-
1000
fire temperature • _ • - -- .-

0
-.
U
'-'
CD
800
,....
."" --
."". . ..--"""

600
~
.I'

~
400
I
I point 2 _--- ---
! 200
",,,,..-"-
..-"-
point3
--_.--- - - -...- .. - ... - .. -- ...- _.- ... - ....
.."."/
,...,-
0
~-:--.-- ....
0 10 20 30 40 50 60 70 80 90
Heating time (min)

Fig. 7. Temperature rise in the steel section lIDder the ISO fire.

1000

-. 800
~
600 .
I
.
. ",..- ------~~ _ _
pomt2
,./ "
200 ,~ "
.."."""/ POin~: ___ .__ .______ .______ ._
.. -- .... .. -- .. - ... -
o
o 10 20 30 40 50 60 70 80 90
Heating time (min)

Fig. 8. Temperature rise in the steel section lIDder a natural fire.

temperature of the steel beam above one quarter of the web depth from the bottom flange is
lower than 400°C and the full strength can be expected. On the top side of the floor, the
temperature is less than 50°C and the insulation criterion is certainly satisfied.
 83

The temperature development of the steel section is shown in Fig. 7. The curve of
point 1 represents the average temperature rise of the bottom steel flange. Within the first
40 minutes, the rate of temperature rise is in the range of 12 to 18°C/min. From 40 to 60
minutes, the average rate of temperature rise is about 8°C/min. The curve of point 2
represents the temperature rise at one quarter of the web depth. Within the 60 minutes'
duration of ISO fire, the heating rate is in the range of 3 to 8°C/min, which is much
lower than that of the bottom steel flange.

Fig. 8 illustrates the temperature rise of the steel section under natural fire. The natural
fire curve is calculated according to Annex B of Eurocode 1 Part 2.2 (Fig. 4). The
opening factor 0 is 0.02 mY, and the fire load density qj is 1 100 MJ/m2 . In practical
cases, this curve defines a severe fire scenario. It shows that the average heating rate at
the bottom steel flange reaches a value of 20°C/min in the ascending phase and a value
of 6°C/min in the cooling phase. At point 2, the average heating rate is 8 °C/min and the
cooling rate is 2°C/min, respectively.

4. Resistance analysis for the ISO standard fire case

The plastic moment capacity of the slim floor beam at elevated temperatures was
analyzed using the moment capacity method. The division of elements was the same as
that in the temperature analysis case. The mechanical properties of steel and concrete
under high temperatures are given according to Eurocode 3 and Eurocode 4. The moment
capacity Mu can be calculated by

Mu = iifijihijl1Yijdi (6)
i=! j=!

where m and n are the numbers of divisions in the x and y directions, respectively; /y is
the concrete or steel strength of the element (i,j); ihij and l1y ij are the element
dimensions; di is the centroid distance of the elements (i,j) from the plastic neutral axis.
The position of the plastic neutral axis (p.n.a.) can be obtained by

iifijihijl1Yij =0 (7)
i=! j=!

where fij has a positive value in the tensile zone and a negative value in the compressive
zone. The tensile strength of the concrete is ignored.

The reduced moment capacity of the composite section under fire is described using
the concept of load ratio, which is defined as
 84

1.2

~ Hogging region
.-
0
~
I-<
0.8
0.6
,
...
"~
-~
Sagging region '.~..•
]
0
....:l
0.4
Normal-weight concrete
-- ---. - .
0.2 Light-weight concrete

0
0 10 20 30 40 50 60 70 80 90

Heating time (min)

Fig. 9. Influence of concrete types on load ratio in the sagging and hogging region.

Moment capacity in fire

R =-----~-~-~-- (8)
Ultimate moment capacity at 20°C

Fire is regarded in the design philosophy of the Eurocodes as an accidental situation or an

accidental action. Therefore, the characteristic values of relevant material properties are
adopted in fIfe design, and the partial safety factors are accordingly assumed as unity (1.0)
when the moment capacity under fIfe conditions is analyzed. The ultimate moment capacity
at 20°C is calculated using the partial safety factors presented in Eurocode 4 Part 1.1 [10].

The following factors affecting the fIfe resistance of the slim floor beam are investigated
in this section:
• Type of concrete: normal-weight and light-weight
• Enhanced reinforcement
• Fire protection of the bottom steel flange
• Fire resistance of continuous beam using the plastic hinge analysis method

There are two important factors that affect the plastic moment capacity of a slim floor
beam: the resistant force of each part of the beam section and the position of the plastic
neutral axis. The calculations were also carried out to analyze the contributions of the section
parts to the plastic moment capacity in fIfe. Changing of the plastic neutral axis position in
the cross-section under fIfe with heating time is investigated further.

In all the analyses, the effective width of the concrete floor slab is defmed as one quarter
of the beam span in the sagging region and as triple the width of the upper steel
 85

flange in the hogging region. In the hogging moment calculation, the reinforcement bars
are taken as <1>16 c/c 150. Analyses show that the change of hogging reinforcement bars
does not have any significant influence on the load ratio in the hogging region. For
instance, if the reinforcement bars <1>16 c/c 50 are used instead of <1>16 c/c 150, the load
ratio is slightly lower, i.e. about 0.05. In this context, the 'sagging region' means the beam
region where the cross-section is subjected to a sagging (positive) bending moment, and
the 'hogging region' means the region where the cross-section is subjected to a hogging
(negative) bending moment.

4.1 lrifluence of concrete types

Fig. 9 represents the relationship between load ratio R and heating time under the ISO
fire for normal weight and lightweight concrete. It shows that the load ratio for
lightweight concrete is slightly lower than that for normal-weight concrete. In the sagging
region, the load ratio for lightweight concrete and normal-weight concrete at 60 minutes
is 0.44 and 0.47, respectively. This small difference is due to the fact that the slim floor
beam with a lightweight concrete slab has a higher temperature in the bottom flange of
the steel section. This is caused by the bad heat-sinking effect of the lightweight concrete.
Fig. 9 also shows the variation of load ratio against heating time both in the sagging and
in the hogging region. It shows that the load ratio in the sagging region is much lower
than that in hogging region. At 60 minutes, the load ratio in the sagging region and in the
hogging region (normal-weight concrete) is 0.47 and 0.73, respectively.

4.2 lrifluences of enhanced reinforcement and fire protection

In the practical cases, the applied load ratio of the beam under fire rarely exceeds 0.6
and values in the range of 0.5 to 0.55 are most common [6]. Analysis of the previous
section indicates that the applied load ratio should be lower than 0.47 if a 60-minute fire
resistance is desired. Therefore, for the new slim floor beam, some other measures have
to be found to obtain a 60-minute fire resistance. Applications of an enhanced
reinforcement and fire protection of the bottom steel flange are the efficient measures
(Fig. 10) commonly used in practice. Fig. 11 illustrates the influence of enhanced
reinforcement on the load ratio (for details of reinforcement see Fig. lOa). At 60 minutes,
the load ratio is 0.57 when a reinforcement 2<1>32 is added.

Fig. 11 also illustrates the effect of fire protection in the bottom flange of the steel
section. The intumescent fire painting, Nullifire S607 [9], with an allowed minimum
thickness of 250 pm was used in the analysis. The details of the fire painting are
illustrated in Fig. lOb. As indicated in Fig. 11, the full strength can be obtained at 60
minutes with a load ratio of 0.98, if the fIre painting is used, and a load ratio of 0.67 at 90
minutes.
 86

<1 <1
<1
4
4
<1"

100 concrete
<1

• "Li

P1?z~~~~~~~@--steel
tL:
section ==l~~~~~~~~~
fire painting

(a) enhanced reinforcement (b) fire painting

Fig. 10. Enhanced reinforcement and fire protection in the composite section.

1.2

~ 0.8
.....0
..
~
0.6
"tj

~
d
0
0.4
- the new slim floor beam
--
0.2 _ _ with reinforcement
with fIre painting
•••• I

0
0 10 20 30 40 50 60 70 80 90
Heating time (min)
Fig. 11. Influence of enhanced reinforcement and fire painting on load ratio in the sagging region.

4.3 Plastic hinge analysis ofcontinuous slim floor beam

In common practice, there is interest in designing a moment-resistant connection between

the slim floor beam and the column. In this case, the bending resistance of the slim floor
beam under fIre depends on both the sagging and the hogging moment resistances of the
sections. It is assumed that the plastic resistance can be developed in the section and the
bending moment capacity can be determined from the following expressions:

Internal span: Ms +Mh "C.M[ (9)

End span: M s + 0.45M h "C. M [ (10)

In Eqs. (9) and (10), Ms is the sagging moment resistance, Mh is the hogging moment
resistance and Ml is the applied bending moment in a simply supported beam.
 87

1.2
-.-----.

~ -. Internal span
.-"'ai
0
0.8
···
7 . -- . . - . - . - -..
.-.
;... 0.6
"0 .........
t'\I
0
0.4 "-
~
- - Plastic hinge analysis of continuous beam ---
0.2 Sagging moment capacity
........... Hogging moment capacity
0
0 10 20 30 40 50 60 70 80 90
Heating time (min)

Fig. 12. Influence of the slim floor beam continuity on load ratio.

350
,,-...,
E 300
[53~p.n.a
E
'-'
...t:: 250
(!)
<) 200
a
.-
......
tI.l
"CS
150

cd 100
Sagging region
~
0.. 50
0
0 10 20 30 40 50 60 70 80 90
Heating time (min)

Fig. 13. Changing of plastic neutral axis position with heating time.

The plastic moment resistance of the continuous slim floor beam is illustrated in Fig. 12. It
can be seen that a load ratio value of 0.52 for the end span and that of 0.58 for the internal
span can be reached at 60 minutes under an ISO fIre. Therefore, a 6O-minute fIre resistance
can be achieved in most cases if the continuity of the slim floor beam is taken into
consideration.
 88

4.4 Resistance analysis: contributions of the composite cross-section parts

To investigate the contributions of the beam section parts to the fIre resistance, the beam
cross-section is divided into five parts: bottom steel flange, top steel flange, lower part of
web, upper part of web and compression part of the concrete slab cross-section. Here the
lower part of the web means the part of web below the plastic neutral axis and the upper part
of web is that above the plastic neutral axis. The yield strength of both structural steel and
reinforcement steel is 355 N/mm2 and the cube strength of concrete is 30 N/mm2 • For the
calculations of moment capacity at ambient temperatures, partial safety factors of the
material strength of 1.1 for structural steel, 1.15 for reinforcement steel and 1.5 for concrete
[10] are used. For the moment capacity calculations under fire, the partial safety factor of the
material strength is taken as unity (1.0) for both steel and concrete.

In Table 1, the contributions of the beam section parts to the total sagging moment
capacity at 60 minutes are compared with those at ambient temperatures. Due to the elevated
temperature, the moment capacity contribution of the bottom flange is reduced to 26% from
the original value of 59% at ambient temperature. The lower part of web contributes 55% of
the total moment capacity under fIre while this value is 14% at ambient temperature. It can be
seen that the web contributes a major part to the total moment capacity in fire while the
bottom flange has a major contribution at ambient temperature. The top flange makes a slight
contribution to the moment capacity both at ambient and at elevated temperatures because of
its position close to the plastic neutral axis. At 60 minutes under the ISO fIre, the concrete
above the top steel flange still has a very low temperature (~60 °C) (Fig. 6) and the full
strength can be expected. The contribution by this part to the total moment capacity is
approximately 20%.

In Table 2, the contributions of the composite section parts to the total moment capacity in
the hogging region is analyzed in a similar way to that in the sagging region. In the hogging
region, the hogging reinforcement, top steel flange and upper part of the web are in tension.
The bottom steel flange, lower part of the web and the concrete near the bottom flange are in
compression. At ambient temperature, the compressive parts contribute only 8% to the total
moment capacity and the contribution by tensile parts is 92%. However, in fIre, the
compressive parts contribute 42% and the tensile parts 58% to the total moment capacity,
respectively. This is caused by the significant change of the plastic neutral axis position in
fIre even if the full strength in the tensile parts remains after 60 minutes' heating.

Fig. 13 illustrates the changing of the plastic neutral axis position during fIre. The position
ofplastic neutral axis (p.n.a.) is referred to as the distance h from the p.n.a. to the top side of
the floor slab. In the sagging region, the p.n.a. distance h has a stable value of 100 mm for the
fIrst 25 minutes, and then it declines rapidly to 65 mm at 40 minutes. After
 89

Table 1
Sagging moment capacity contributions for the section parts under ISO fIre (60 minutes)
(Numbers in brackets are those for ambient temperature)
Section parts Temperature Centroid Axial Moment Percentage of
roC) Distance from Resistance Resistance Total Moment
p.n.a. (mm) (kN) (kNm) Capacity (%)
Bottom flange 650-870 227 (189) 399 (2324) 90 (439) 26 (59)
Lower part of web 85 -580 121 (90 ) 1555 (1162) 188(105) 55 (14)
Upper part of web -- (25 ) -- (323) -- (8 ) -- (1 )
Top flange 50 -80 0.6 (55 ) 710 (646 ) 0.4 (36) o (5)
Concrete slab ~100 32 (60) 2034 (2550) 66 (153) 19 (21)
Total Moment Capacity (kNm) 344 (741)

Table 2
Hogging moment capacity contributions for the section parts under ISO fIre (60 minutes)
(Numbers in brackets are those for ambient temperature)
Section parts Temperature Centroid Axial Moment Percentage of
roC) Distance from Resistance Resistance Total Moment
p.n.a. (mm) (kN) (kNm) Capacity (%)
Re-bars 30-40 150 (260) 428 (372) 64 (97 ) 20 (22)
Upper flange 50-80 115 (225) 710 (646) 82 (145) 25 (33)
Upper part of web 85 -250 55 (113) 781 (1452) 43 (163) 13 (37)
Lower part of web 250-580 49 (5 ) 775 (32 ) 38 (0.2) 12 (0)
Bottom flange 650-870 129 (14) 399 (2324) 51 (33 ) 15 (8)
Concrete 100-800 63 (5 ) 795 (33 ) 50 (0.2) 15 (0)
Total Moment Capacity (kNm) 328 (438)

that, the p.n.a. distance h varies slightly with the heating time. This can be interpreted
clearly by referring both to Fig. 7 and to the strength reduction of steel at elevated
temperatures. After 25 minutes' heating, the average temperature of the bottom flange
rose to 400°C, where the strength reduction begins. From 25 to 40 minutes, the average
temperature of the bottom flange rises to 600 °C at a heating rate of 18 °C/min while the
material strength of the steel sharply drops to 45% of its original value. Therefore, a rapid
decline in the p.n.a. position occurs between 25 and 40 minutes. After 40 minutes'
heating, the average heating rate of the bottom steel flange is less than 8 °C/min.
 90

:o:O~o6;~--1
1.2
• .~- -.- -in..=550
c=._._ .._._.MJ/m2i
__.__._.J
,.
r~~~.06~~·· .. J

~ 0.8 • "lqrll00 MJ/m2

....-;s0 1# '
,
: . - . _ - _ •• _ . _ - - - - - _ . _ - '

0.6
"tj """
d
0 0.4
~
rISO fire
............. ..i
0.2

0
0 10 20 30 40 50 60· 70 80 90
Heating time (min)

Fig. 14. Load ratio vs. heating time under two natural fires.

1.4
I I I I
6 1.2 _________ I_________ ~---------~---------~---------~---

I I I
_____ ~

,
_ _ _ _ _ _ _ _ _ _ _ _ _ • ___ _

~ • • ..
"
"
,
I

....«i0 1.0
I
I
I
I
- ... ~

I
I
.....
I

I
.'...
I .... I
,

1
I
I

-+.•., '-..---- r ----- ----:- ---------:- -----. -- --- --- ---

I
I
I
I I I I " .... I I
""" 0.8 ---------~--------i---------7---- ----~---------;.-.~-:--~--------- ---------
-g : : : I : --:_ • • • __

--........
I I I r I I
0 0.6 ---------~--------~---------t---------~--- ---~--------~--------- ----.----

-
I I I ' I
I I I I
d ,
I "I I '-~
I
0 0.4 ---------~----------------------------r---------~----- ---~---------
,, ,,, ---------
_ : Sagging region ,, ,,,
,,
U""" ...............:'Hogg'Ing region
.
_ _ _ _ _ _ _ _ _ 1.

0.2 .-----~---------~--------~--------- ---------

,,, ,,
,
,
,,
0.0
0 200 400 600 800 1000 1200 1400 1600

Fig. 15. Lowest critical load ratio vs. fire load density.

In the hogging region, changing the p.n.a. distance h is similar to that in the sagging
region. The original p.n.a. distance h is about 290 nun, i.e. very close to the upper side of the
bottom flange. A rapid change in the p.n.a. distance h occurs between 25 and 40 minutes and
after that the change becomes slower. Generally, the p.n.a distance h in the hogging region
varies more significantly than that in the sagging region. One important reason for this is the
significant reduction in the compression part resistance of the concrete and bottom steel
flange at elevated temperature.
 91

5. Resistance analysis for the natural fire cases

The Eurocodes give an alternative way of analyzing the fire resistance of structures
under specified natural fires. In this section, the structural fire resistance of the new slim
floor beam will be investigated under the parametric natural fires defined by Annex B of
Eurocode 1 Part 2.2.

The major controlling parameters of the natural fires in Euorocde 1 are the opening
factor 0, fire load density ql and the thermal property of enclosure boundary ..J pAc .
Generally, the effect of the thermal property of the enclosure boundary on the fire
temperature is in the range of 1O~20%. Here the thermal property of the enclosure
boundary is taken as 1 160 (J Im 2 s 'hK), which represents a standard compartment.
Another factor is the ratio of the floor area AI to the total surface area At. Calculations
show that the effect of this factor on the fire temperature according to Annex B of
Eurocode 1 Part 2.2 is within the range of 10%, if the ratio A/At varies between 1:3~1:8.
Here the ratio A/At is taken as 1:4, which represents a relatively severe situation in the
fire compartment.

A parametric study was performed to investigate the minimum load ratio R with the
major controlling parameters of natural fires. These parameters were as follows:

• Opening factor 0 (mYz): 0.02, 0.06, 0.10, 0.15, 0.20

• Fire load density per unit floor area ql(MJ/m2 ): 275, 550, 825, 1 100, 1 375
• Thermal property ..J pAc of the enclosure boundary (J 1m 2 s 'h K ): 1 160
• Ratio of floor area to total surface area of enclosure (AI: At): 1:4

Due to the fact that a natural fire curve has both an ascending and a cooling phase, the
irreversibility of the material properties of the concrete in the cooling phase has to be
taken into account. The strength of the concrete parts depends on the maximum
temperature ever attained, and it cannot be recovered during or after the cooling phase.
For the steel parts, the strength is dependent on the temperature at the current time. In the
cooling phase, the strength of steel can be recovered.

Fig. 14 illustrates the variation in load ratio both during the natural fires and an ISO
fire in the sagging region. It shows that the load ratio curve has a recovering part in a
natural fire while that curve always declines in an ISO fire.

There is a minimum load ratio for each parametric natural fire (Fig. 14). The
relationship between the minimum load ratio and the fire parameters can be obtained
when the moment capacity analyses under different parametric fire curves are carried out.
Table 3 illustrates the analysis results of the minimum load ratio and the fire parameters.
The minimum load ratio in Table 3 is called 'critical load ratio' (Rer) because the slim
floor beam is unsafe during fire if the applied load ratio is higher than the corresponding
value in Table 3.
 92

In practical cases, the opening factor has more significant uncertainty and the fire load
density is taken as the major fire parameter in ranking the fire severity in buildings. From
Table 3 it can be seen that for a specified fire load density there exists a lowest load ratio
corresponding to the most severe fire for all opening factors.

Fig. 15 illustrates the relationship between the lowest critical load ratio and the fire
load density. Generally, the slim floor beam can be used without applied fire protection if
the fire load density is lower than 1 100 MJ/m2 • When the fire load density is 1 375
MJ/m2 , the lowest critical load ratio is 0.43 for the sagging region.

Table 3
Critical load ratio of the slim floor beam in the sagging region under natural fIres given by opening factor
and fIre load density
y,
Openingfactor 0 (m ')
Load Ratio 0.02 0.06 0.10 0.15 0.20

275 1.16 1.17 1.17 1.17 1.17

Fire load density 550 0.94 0.99 1.06 1.12 1.13
qj (MJ/m2) 825 0.70 0.66 0.70 0.75 0.87
1100 0.56 0.51 0.52 0.57 0.63
1375 0.48 0.43 0.44 0.47 0.50

6. Conclusions

There is increasing interest in slim floor construction in the Nordic Countries and the
UK. At the moment, a new asymmetric slim floor beam is under development in Finland.
In this paper, the fire resistance behaviour of the new slim floor beam is investigated. The
following conclusions can be drawn on the basis of this study:

• A 60-minute fire resistance can be achieved if the applied load ratio for a simply
supported beam is lower than 0.47. Otherwise, some additional measures such as
enhanced reinforcement and/or fire protection of the bottom steel flange are required.
• Generally, the new slim floor beam can achieve a 60-minute fire resistance without
any additional measures, if the moment-resistant beam-to-column connection is
reliably designed.
• The fire resistance of the slim floor beam is provided mainly by the lower part of the
steel web, while the bottom flange loses its resistance under fire.
• The fire resistance behaviour of the slim floor beam is also investigated under natural
fires defined by Eurocode 1. The results show that the new slim floor beam can be
used without any additional measures if the fire load density is less than 1 100 MJ/m2,
which is rarely exceeded in office and residential buildings [11,12].
 93

Acknowledgements

The financial support provided for this research project by the Technology
Development Centre of Finland (TEKES), the Finnish Constructional Steelwork
Association Ltd (FCSA) and the Finnish companies, Rautaruuki Oyj, PPTH Teras Oy and
Deltatek Oy is gratefully acknowledged by the authors.

References

[1] Ma, Z. and MiikeHiinen, P. Temperature Analysis of Composite Structures Exposed to Fires - TACS-
FIR. Laboratory of Steel Structures Publications 10, Helsinki University of Technology, Espoo,
1999.
[2] ENV 1993-1-2, Eurocode 3, Design of Steel Structures, Part 1.2: General Rules - Structural Fire
Design. CEN/TC250/SC3, Brussels, September 1995.
[3] ENV 1994-1-2, Eurocode 4, Design of Composite Steel and Concrete Structures, Part 1.2: Structural
Fire Design. Final edited versioniCENITC250/SC4, Brussels, November 1993.
[4] ENV 1991-2-2, Eurocode 1, Basis of Design and Actions on Structures, Part 2.2, Actions on
Structures Exposed to Fire. Final edited versioniCET/TC250/SCl, Brussels, November 1994.
[5] Newman, G.M. Fire Resistance of Slim Floor Beams. Journal of Constructional Steel Research 1995;
33(1-2):87 -100.
[6] Lawson, R.M. Mullett, D.L. and Rackham, lW., Design of Asymmetric Slimflor Beams using Deep
Composite Decking. The Steel Construction Institute, Publication P175, 1997.
[7] Schleich, J.B. Fire Safety, Design of Composite Columns. Presented in the international conference
on Fire Safe Steel Construction: Practical Design, Luxembourg, April 1984.
[8] Both, C., Fellinger, J.H.H. and Twilt, L. Shallow Floor Construction with Deep Composite Deck:
From Fire Tests to Simple Calculation Rules. Heron 1997;42(3):145-158.
[9] Nullifire S607 - fire painting, Certified directions for use (in Finnish), TRY-18-1996, Helsinki,
1996.
[10] ENV 1994-1-1, Eurocode 4, Design of Composite Steel and Concrete Structures, Part 1.1: General
Rules and Rules for Buildings. Final edited versioniCENITC250/SC4, Brussels, November 1993.
[11] CIB W14, Design Guide - Structural Fire Safety. Fire Safety Journal 1986;9:77-136.
[12] Schleich, J-B. Global Fire Safety Concept for Buildings: Connection to ECSC Research, ECCS
Developments and CEN Standardization. In: Proceedings of International Conference on Steel and
Composite Structures, Delft, The Netherlands, 1999:8.1-8.26.
94
 95

Appendix C

Behaviour of Composite Slim Floor Structures in Fire

Ma, Z. and MakeHiinen, P.

Journal of Structural Engineering, ASCE, voL 126, no.7, pp.830 - 837 ,July 2000.
96
 97

BEHAVIOUR OF COMPOSITE SLIM FLOOR

STRUCTURES IN FIRE

By Zhongcheng Mal and Pentti MakeUiinen2

ABSTRACT In recent years, increasing interest has been shown throughout Europe in developing and
designing slim floor systems in steel-framed buildings. This paper presents the fire resistance behaviour of
the composite asymmetric slim floor beam as an isolated member and as a part of the frame using
numerical analysis methods. Three schemes were investigated, including isolated beams, a plane sub frame
with semi-rigid beam-to-column connections and a three-dimensional slim-floor frame system. The first
scheme aimed to explore the fire resistance of the beams according to standard fire-testing methodology.
The objective of the second scheme was to reveal the effect of frame continuity on the fire resistance of the
slim floor beam and the mechanical interaction between the frame elements. The third scheme was to
preliminarily identify the influence of the composite slab on the beam behaviour in fire. The investigations
show that the isolated slim floor beam has a 6O-minute standard fire resistance without any additional fire
protection, if the load ratio is less than 0.5. As a part of the frame, the beam still keeps its stability even
when the temperature of the bottom steel flange of the beam reaches up to 900°C (90 minutes' ISO fire
exposure). The analyzed results indicate that the axial restraints provided by the surrounding parts cause a
larger deformation of the beam in the earlier ISO heating phase and, however, a more stable behaviour
thereafter. The rotational restraints essentially cause the change in the applied load ratio in fire, which can
be quantified using the 'modified load ratio' proposed in this paper.

1 Res. Fellow, Ph.D.(Eng.), Dept. ofCiv. and Envir. Engrg., Helsinki Univ. ofTechnol., P.O.Box 2100, FIN-02015
HUT, Finland.
2 Prof., D.Sc.(Tech.), Head of the Steel Struct. Lab., Dept. of Civ. and Envir. Engrg., Helsinki Univ. of Technol.,
P.O.Box 2100, FIN-02015 HUT, Finland.

Keywords: Slim floor, asymmetric steel beam, composite frame, fire resistance, numerical modelling
 98

INTRODUCTION

In recent years, increasing interest has been shown throughout Europe in developing and
designing shallow floor systems in steel-framed buildings. In the shallow floor system, the steel
beam is contained within the depth of the pre-cast concrete floor or composite slab with profiled
steel decks. This form of construction achieves a minimum depth of building and the flat floor is
beneficial because the building services can be run in any direction. The key feature of this
system is the steel beam, which can be either rolled or welded sections (see Fig. 1). One of the
original slim floor concepts developed in Scandinavia was the 'Thor-beam', which consists of
two channel sections welded to a flat plate. The slim floor system using a 'Delta-beam', which
consists of a welded beam section and a prefabricated concrete slab, is popular in the
Scandinavian Countries (Lu and Makelainen 1996). British Steel, in collaboration with the Steel
Construction Institute (UK) developed a Slimflor® beam, which consists of a universal column
section welded to a steel plate (Newman 1995; Mullet and Lawson 1993). Recently, interest has
been concentrated on the asymmetric hot-rolled steel beam in the UK (Lawson and Mullet 1997),
and on the asymmetric welded steel beam in Finland (Ma and Makelainen 1999; Malaska and
Makelainen 1999) (see Fig. 2).
Generally, the slim floor beam provides more opportunities for steel in spans of 5 - 9 m.
It achieves a slab depth of 300 mm or so, which is much less than in conventional steel
construction. This system also has an inherently good fire resistance, due to the partial
encasement of the steel beam within the concrete slab. On the other hand, compared with the
conventional composite frame system, which has a primary-secondary-beam system, the slim
floor frame has a rather precise structural form and therefore offers significant savings in
construction cost. In slim floor construction, the slab is sustained directly by the primary beam,
forming a part of the composite beam that works together with the steel beam. Between the rows
of a single frame, the tie members are employed to link them together to keep the out-of-plane
stability of the frame.

[J n .,
I
/.
\

(a) (b)

I (c)
I (d)

FIG. 1. Different Types of Slim Floor Steel Beams: {a)Thor-Beam; (b) Delta-Beam;
{c)Slimflor® Beam; (d) Asymmetric Slim Floor Beam
 99

k 183 (198)
[- '"I

<.'

(a) (b)

FIG. 2. Composite Sections of Asymmetric Slim Floor Beams: (a) British 280(300) ASB
Beams; (b) Finnish Asymmetric Slim Floor Beam

Traditionally, the fire resistance of constructional steelwork has been investigated by

standard fire tests. The additional fire resistance was achieved by applying fire protection to the
steel members. Since the late 1980s, fire engineering has rapidly developed, especially in
Europe, to the point where buildings are designed with significant built-in fire resistance.
Following the investigation of a series of fire events, such as the Broadgate Fire (UK), William
Street (Australia) and Washington (USA), as well as the full-scale fire tests in Cardington (UK),
it was concluded that the traditional standard fire test on isolated structural components was very
conservative. Future research into the structural behaviour of steel-framed buildings subject to
fire should consider the structures as a complete entity and not as a collection of isolated
members (Johnson 1998; Robinson 1998; Bailey 1997; Newman 1996).
The aim of this paper is to explore the fire resistance of the composite asymmetric slim
floor beam both as an isolated member and as a part of the frame using numerical analysis
methods. The effect of frame continuity on the behaviour of a fire-exposed beam, and the
moment redistribution and axial force variation among the frame members are investigated. The
role of the floor slab in fire is also preliminarily identified. The procedure presented here can
also be used in fire engineering design of other steelwork products.

BEHAVIOUR OF ASYMMETRIC SLIM FLOOR BEAMS IN FIRE

Model Description

The modelling of fire-exposed structures includes both the thermal and structural aspects.
The modelling of thermal response here is based on the computer program, TACS-FIR
(Temperature Analysis of Composite Structures exposed to FIRe) (Ma and Makelainen 1999),
which is purposely developed for the temperature analysis of composite structures in fire. The
 100

explicit forward difference method was used in this program. The major assumptions used in the
analysis are:
•The thermal properties of steel and concrete are taken from Eurocode 4 Part 1.2 ("Eurocode
4" 1993);
•The resultant emissivity for exposed steel is 0.6 and OJ for the composite slab with steel
decking;
.The convection factor is 25 W/m2K for the exposed side and 8 W/m 2K for the unexposed
side;
•The interface resistance between the concrete and steel is 50 WIm2K.
The structural response was modelled by the general finite element program, ABAQUS
(ABAQUSIStandard 1997). The composite slim floor beam was represented by combining a shell
element (S8R) and a beam element (B32). The asymmetric steel beam was modelled by 3-node
beam elements with 6 degrees of freedom at each node. The concrete slab was modelled using 8-
node thick shell elements, also with 6 degrees of freedom at each node. The reinforcements were
modelled using REBAR elements encased in the concrete shell elements. In this study, the
concrete was considered to be an elastic-plastic material, which has a plastic plateau after
reaching the compressive and tensile strength. No descending phase in compression and in
tension was taken into account. The tensile strength of the concrete was taken as 10 percent of its
compressive strength.

TABLE 1. Material Properties for Structural Modelling

Steel Concrete
Yield Thermal Cubic Thermal
strength expansion strength expansion
(N/mm2) (oC"1) (N/mm2) (oC"1)
(1 ) (2) (3) (4)
355 1.4x10"o 30 1.8x10"5

The Newton-Raphson iterative technique was used for the solution process. Both physical
and geometrical nonlinearity were included in the modelling. The temperature distribution and
variation with time of the composite section were introduced according to the thermal modelling.
The temperature-dependent stress-strain curves of the steel and concrete were taken from
Eurocode 4 Part 1.2 ("Eurocode 4" 1993). The material properties for the structural modelling
are shown in Table 1.

British ASB SLlMDECK Beams

A SLIMDECK system using an asymmetric Slimflor beam (ASB) was developed in UK

and two standard fire tests were conducted by the Warrington Fire Research Center (WFRC)
(Lawson and Mullet 1997). These tests consisted of two loaded beams spanning 4.5 m, which
were heated according to the ISO standard fire curve. The section size is shown in Fig. 2(a) and
 101

""'-

562.5 J 1125 J.
4500
1125 J 1125 J
-
562.5
Tirre(nin)

~ 84.6kN -~ 84.6 kN t 84.6 kN ~84.6 kN 100 120

A .J;l
Fire tests details: 280ASB profile JO

- 4500
I
;0

1==1
1470 .1.520 .1. 520 J 520 1470
JO

;0

Fire tests details: 300ASB profile

FIG. 3. Load Arrangement of SLIMDECK Beams FIG. 4. Mid-Span Vertical Displace-

Tested by WFRC (Lawson and Mullet 1997) ment of British SLlMDECK Beams in
ISO Fire

the loads and structural details are shown in Fig. 3. The steel grade is S355 and the concrete
grade is NWC C30/25. The load ratio is calculated as 0.36 for the 280 ASB beam and 0.43 for
the 300 ASB beam.
Fig. 4 shows the displacement-time curves of the tested slim floor beam. Both of the tests
were terminated when the load-bearing criteria specified in BS 5950: Part 8 were exceeded. In
the 280 ASB test, this occurred at a deflection of span/20 after 107 minutes. In the 300 ASB test,
this occurred at span/30 after 75 minutes, as the allowable rate of deflection was exceeded before
a deflection of span/20 had been reached. The predicted curves by FE analysis are also illustrated
in this figure. It can be seen that there is a good correlation between the predicted and the
measured curves.
Table 2 summarizes the measured and predicted results. It can be seen that the calculated
bottom-flange temperature is very close to the measured value, with only an average difference
of less than 30°C. The predicted fire resistance by FE analysis is also very close to the measured
value. However, the predicted fire resistance according to the general rules in Eurocode 4
(numbers in brackets) is lower than the testing time.

TABLE 2. British SLIMDECK Beams: Measured and Predicted Results

Test and Calculated Data 280ASB 300ASB

(1 ) Test Predicted Predicted Test Predicted Predicted
(2) (3) (4) (5) (6) (7)
Applied Moment (kNm) 200 310
Design Moment Resistance at ULS (kNm) 557 714
Applied Load Ratio 0.36 0.36 0.5 0.43 0.43 0.5
Fire Resistance (min) 107 109(93) 66(60) 75 74(65) 61(55)
Bottom Flange Temp. at Failure (OC) 869-987 841-1004 650-856 727-861 735-908 675-825
 102

1000

0
°
800][
6000
-- ...
I!!~ 600

---
III
1200 J. 1200 .1. 1200 J 1200 J 1200 II) 400 point2 _ - - - - - -

t90 kN ~ 90 kN + 90 kN +90 kN
a.
E
~
200
//
/-
....
point 3
........ ----
o --~--::-- ...... _'..
:0: o w ~ ~ W ~ ~ M ~ 00
lime (nin)

FIG. 5. Load Arrangement of the Finnish Slim FIG. 6. Temperature Variation of Finnish
Floor Beam Beam in ISO Fire

In practicality, the most common load ratio is between 0.5-0.6 when buildings are
subject to a fire attack. Because the test load ratios were less than this, a further investigation
was carried out to obtain the allowed maximum load ratio for 60 minutes' fire resistance. The FE
analysis results indicate that the 60 minutes' fire resistance can be obtained where the load ratio
is less than 0.5 (see also Table 2).

Finnish Asymmetric Slim Floor Beam

The Finnish asymmetric slim floor steel beam uses the welded section. The web plate is
welded to the bottom and top flange using the submerged arc-welding method. The section shape
of the composite beam is shown in Fig. 2(b). The load arrangement of the investigated beam is
shown in Fig. 5. This beam is uniformly four-point loaded and simply supported. The applied
load ratio is 0.53.
The temperature variation of the steel beam with time is shown in Fig. 6. It can be seen
that there exists a large temperature gradient in the beam section due to the encasement of
concrete. The average temperature in the bottom flange at 60 minutes is 760°C and only 50°C
in the upper flange. In this context, the structural response is described against the ISO fire
exposure time. Certainly, it is also reasonable for this to be given against the bottom flange
temperature. This transformation can be easily made according to Table 3.

TABLE 3. Relationship Between ISO Fire Exposure Time

and Temperature of Bottom Flange

ISO Time (min) o 15 30 45 60 75 90

Bottom Flange
20 220 500 650 760 830 890
Temp. (0C)
 103

Time (min)
Time (min)
0
0
40 50 60 70 80 90
E 60 70 80 90 E -100

-
.5.
c
III
E
-100
.5.
t:G> -zoo
E
G>
U
III
Ci.
-zoo ..
III
u
Ci.
-300 ks-lOkN/mm
••.•.. ks=50 kN/mm
III
-Loadratio=0.53 " ". .!!! _. _. ks=100 kN/mm
C C -400
- .. - ks=500 kN/mm
iii -300 ..... Load ratio = 037 ' "" iii - - - ks=IOOO kN/mm
.E - . _. Load ratio = 0.69 -----\---.---••.•.. __••_--..•_. ..., "" u
1:: t: -500 -ks=IOOOO kN/mm .
-ks=O
III
> \
\
~
-400 -600

FIG. 7. Mid-5pan Vertical Displacement of FIG. 8. Effect of Constant Axial Restraints

Finnish Beams in ISO Fire on the Beam Behaviour in Fire

TABLE 4. Analysis Results of Finnish Composite Slim Floor Beam

Load Ratio in Analysis

Numerical Analysis Data 0.37 0.53 0.69
(1) (2) (3) (4)
Applied Moment (kNm) 226 324 422
Design Moment Resistance at ULS (kNm) 618
Moment Resistance at ULS (FE analysis) (kNm) 705
Predicted Fire Resistance by FE Analysis (min) 86 62 47
Predicted Fire Resistance by Eurocode 4 (min) 80 55 43
Bottom Flange Temperature at Failure (0C) 740-950 660-880 550-740

Fig. 7 shows the vertical mid-span displacement versus the heating time. According to
BS 5950: Part 8, the 'failure' point corresponds to the deflection reaching up to 1120 span length,
or to the critical deflection rate at a deflection of span/30 ("BS 5950" 1990). From Fig. 7, it can
be seen that 60-minutes' fire resistance can be achieved if the load ratio is less than 0.53. In this
figure, the analyzed vertical displacement responses under varying load ratio (0.37 and 0.69
respectively) are also illustrated.
Table 4 summarizes the relevant results from the modelling. In this table, the design
moment capacity of the beam at ultimate limit state (ULS) at ambient temperature was calculated
with the material safety factor specified by Eurocode 4 Part 1.1 ("Eurocode 4" 1992).
Nevertheless, since fire is regarded as an accidental action in the Eurocodes, the moment
resistance in fIre was calculated using the characteristic values of the material properties. The
effective width of the concrete slab was taken as 118 span length in calculating the sectional
moment capacity, which was proposed by Lawson and Mullet (1997).
The bending-moment capacity method was used to investigate the contributions of the
beam section parts to the fire resistance. The beam cross-section is divided into five parts:
bottom steel flange, top steel flange, lower part of web, upper part of web and compression part
of the concrete slab cross-section. Here, the lower part of web means the part of web below the
plastic neutral axis (p.n.a), and the upper part of web is that part above the plastic neutral axis. In
 104
Table 5, the contributions of the beam section parts to the total moment capacity at 60 minutes
are compared with those at ambient temperatures. Due to the elevated temperature, the moment
capacity contribution of the bottom flange is reduced to 31% from the original value of 44% at
ambient temperature. The lower part of the web contributes 53% of the total moment capacity at
60 minutes fire while this value is 6% at ambient temperature. It can be seen that the web
contributes a major part to the total moment capacity in fire while the bottom flange has a major
contribution at ambient temperature. The top flange has a slight contribution to the moment
capacity both at ambient and at elevated temperatures because of its position close to the p.n.a.
At 60 minutes under ISO fire, the concrete above the top steel flange still has a very low
temperature (:::;70 °C) and the full strength can be expected. The contribution by this part to the
total moment capacity is approximately 14% at 60 minutes' fire and 30% at ambient temperature.

TABLE 5. Moment Capacity Contributions for the Section Parts (at 60 minutes under ISO Fire)
(Numbers in Brackets are at Ambient Temperature)

Section Parts Temperature Centroid Axial Moment Percentage of

eC) Distance from Resistance Resistance Total Moment
p.n.a. (mm) (kN) (kNm) Capacity (%)
(1 ) (21 (3) (4) (5) (6)
Bottom Flange 650-870 235 (118) 399 (2324) 94 (274) 31 (44)
Lower Part of Web 85 -650 115 (54.5) 1411 (702) 162 (38 ) 53 (6 )
Upper Part of Web 65-85 2 (60.5) 28 (783 ) -- (48) - (8)
Top Flange 45-65 9 (116 ) 710 (646) 7 (75 ) 2 (12)
Concrete Slab ~100 37 (106) 1110 (1730) 41 (183) 14 (30)
Total Moment CapaCIty (kNm) 304 (618)

Effect of Axial Restraints

Fig. 8 shows the effect of axial restraints on the behaviour of the beam in fIre. For simplicity,
the spring is assumed to be elastic and well protected to keep a constant stiffness in fIre. Two
issues can be seen from the results:
• Before a certain time (say, 60 minutes in this example), the increasing axial restraint
stiffness causes the larger mid-span vertical displacement.
• After that, the displacement becomes more stable and the post-behaviour improves with the
increasing axial stiffness.
This phenomenon is caused by the secondary p-~ effect and the catenary action at large
displacement. During the early ISO fire phase, a signillcant axial compressive force in the beam
occurs due to the rapid thermal expansion. The stiffer the axial restraint, the bigger the axial
force. Meanwhile, the significant thermal bowing caused by the large temperature gradient
within the beam section enlarges the beam deflection. However, with the increasing deflection,
the catenary action begins to take effect, which leads to a rather stable deformation behaviour
afterwards.
It can also be seen that the fire resistance time happens to be almost the same for the
investigated axial stiffness at a displacement of span/20. However, it is evident that the beam
 105

with an axial spring of higher stiffness has more stable deformation behaviour versus the ISO
fire exposure time. As we know, the failure criterion for isolated members in the standard fire
tests is defined in terms of displacement, and is set at values which prevent damage to the
furnace during testing ("BS 5950" 1990). Generally, there is no conflict with the philosophy in
the Eurocodes for isolated members at this point. In the Eurocodes, fire is regarded as an
accidental action and the fire engineering design criterion is to keep the structural stability and
integrity during and after fire attack. However, the conflicts are obvious when the members are
considered a part of the structures, and it is therefore arguable whether the displacement criterion
is still valid when the members are considered a part of the building structures.

MECHANICAL INTERACTION BETWEEN SLIM FLOOR SUBFRAME ELEMENTS

Analyzed Subframe and Loads

The analyzed one-bay subframe is shown in Fig. 9. The span is 6 m and the level height is 3
m. The beam was the Finnish composite slim floor beam. The concrete slab was 1.5 m wide and 147
mm thick. The yield strength of the steel was 355 N/mm2 and the concrete grade was NWC 30/25.
The anti-crack reinforcement mesh A142 was incorporated into the concrete slab and the yield
strength was 235 N/mm2 • The universal steel columns were filled between the steel flanges by the
aerated concrete blocks, which were considered to be non-structural. The composite beam was
connected with the columns via the steel beam.

The fire was located in the lower level of the frame and followed the ISO standard curve.
The beam was heated from the lower side and the columns in the lower level were assumed to be
heated on all sides. The temperature distribution and variation with time were calculated using
computer program, TACS-FIR. The columns in the upper level were not attacked by fire.

j 690 kN
f4 "" .p-i 690kN
1
11 .f:· ,........:; ,I H
.~;'. .'.o;.I~ 'n

1
•. t'
column '-

~
II <'.~ !:;~~i,;>'1 3000

119 kN 119kN 119kN 119kN

1 1 I ;
,-

3000

.,.,rr ..,,.,.

J
1200
I'
1200
I'
1200
6000
~
\
1200
•
I
1200
t
FIG. 9. Analyzed One-Bay Subframe and Loads
 106

The applied load ratio for the beam was 0.69 in the case of the pinned beam-to-column
connection. The columns were also loaded to simulate a realistic situation. The applied axial force
was 690 kN and the according load ratio for the lower columns was around 0.2.

Deformation Behaviour of Frame Beam and Modified Load Ratio

The sub frame was studied under the same load value with different connection stiffness.
Three cases were studied: pinned steel beam-to-column connection, semi-rigid connections with
rotational stiffnesses being 600 and 2 000 kNm/rad, respectively. Fig. 10 shows the deformation
behaviour of the beam with different rotational stiffness in flre.

Compared with the isolated simply supported beam, the beam in the subframe with the
pinned beam-to-column connection has more stable behaviour during flre. In the earlier phase, the
beam in the frame has a larger deflection than the isolated beam. This is caused by the axial restraint.
Since the thermal expansion is restrained by the column, a compressive axial force is initiated in the
beam. Apparently, the second-order P-Ll effect enlarges the deflection of the beam. However, the
behaviour of the beam becomes more stable with heating time.

With the increasing rotational connection stiffness, the vertical deflection of the beam
becomes smaller. In essence, the semi-rigid connection changes the practical load ratio applied to
the beam. For the semi-rigid beam-to-column connection, the formulae of the load ratio calculation
for the simple beam should be modified. According to the deflnition of load ratio, the following
equation is suggested:

R = ____M--.::.::sta=::ti"---c_ __ (1)
M p+ + min( M p_ ,Mcon)

Where Mstatie is the isostatic (free) bending moment in the beam, ~+ and ~_ are the plastic
moments of the sagging section and hogging section, respectively, and Meon is the minimum value of
the plastic bending moment of the connection and the applied moment to the connection at the
limit rotation Slim' For a beam in the frame, Meon can be given by
1
M con ="2 (M con,left + M con,right ) (2)

where Meon, left and Meon, right are the bending moment capacity of the connections in the left end and in
the right end of the beam, respectively.

The feature in Eq. 1 is that it accounts for the realistic rotational restraints in beam-ends
when the beam is a part of a structure. It also covers the cases for pinned (Meon =0) and fIxed
(Meon>~) end restraints. Using Eq. 1, the modified load ratio of the investigated beam is shown in
Table 6. In the study, Meon means the applied moment on the connection at the limit state at ambient
temperature, which was obtained here by FE analysis. In the practical design, Meon and Slim can be
calculated according to the design codes when the connection details are given.
 107

TABLE 6. Modified Load Ratio for Semi-Rigid Connection

Rotational stiffness of connection (kNm/rad)

Analyzed Data O(pinned) 600 2000
(2) (3) (4)
(1 )
Isostatic bending moment 427 427 427
Sagging plastic moment Mp+ 618 618 618
Hogging plastic moment Mp_ 345 345 345
Connection moment Mean 0 84 178
Modified load ratio 0.69 0.60 0.53

Fig. 10 also shows the mid-span displacement of the beam (with pinned connection) using
the modified load ratio. The solid lines in the figure represent the displacement of the beam with
specific connection conditions, and the dashed lines represent the displacement of the beam with a
pinned connection under the modified load by Eq. 1. It can be seen that there is a good similarity
between these two results. This similarity indicates that the deformation behaviour of the beam with
rotational restraints in fire is similar to that with pinned restraints under the modified load. The
effect of rotational restraints on the beam fire resistance is consequently quantified. This method
provides an easy way to calculate the fire resistance of the beam which has the rotational restraints in
the beam-ends.

Moment Redistribution and Axial Force Variation

The stiffness of the frame component decreases during fire because the elastic modulus of
structural steel and concrete reduce significantly at elevated temperatures. In most cases, the fire is
limited within one or several rooms or on one level such that only a relatively small part of the
whole structure is attacked by fire with the according deterioration of stiffness. This local
deterioration of component stiffness changes the original stiffness ratio and results in the moment
redistribution in the structural components. Another notable factor that affects the moment and
force in the structural components is the thermal expansion of the heated structural components.

It is important to recognize the variation of moment and axial force in the frame
components during fire. It is helpful to understand the deformation behaviour and failure
mechanism of a fire-exposed structure. Moreover, the design details can therefore be enhanced and
the potential premature failure of elements may be avoided. Fig. 11 (a) shows the moment variations
of a semi-rigid joint in fire. In this study, the rotational stiffness of the beam-to-column connection
is 600 kNm/rad. It can be seen that the end moments of the upper and lower columns increase
rapidly, by up to around three times the original values in the first 15 minutes, before starting to
decline, while the beam-end moment continues to increase during fire.

Fig. 11 (b) shows the change in the beam axial force in fire. In the first 15 minutes, the axial
force increases rapidly due to thermal expansion. After that, the axial force begins to
 108

Time (min)

90
E
.§. -100
~
C
CD
iu -200
CIS
Q.
is -300
- -Pinned frame beam
iii - -K = 600 kNmlrad
u
t! -400 --K=2000 kNmlrad

~ - - 06- - Pinned, modified R=0.53

- - G- - Pinned, modified R=0.6
-500

FIG. 10. Mid-Span Displacement of Beam with Semi-Rigid Beam-to-Column Connection

200
- 300

-
E
Z
150
100
Z
-e
~
250
CD 200

~
~
c 50 u.0 150
.... -- .... - .. 100
CD ii
.;(
E 0 50
0 c(
:1E -50 10 20 30 40 50 0
-100 0 10 20 30 40 50 60 70 80 90
Time (min)
lime (nin)
(a) (b)

FIG. 11. (a) Moment Variation of Beam and Columns; (b) Axial Force Variation of Frame
Beam

decrease. This decrease may be caused by the reduction in axial restraint stiffness with time
(softening of columns), the reduction in the axial beam stiffness and the large displacement in the
beam (catenary action). Up to 90 minutes, the axial force in the·beam becomes very small.
Comparing Figs. Il(a) and (b), it can be seen that the variation in column-end moments
during fire is very similar to that of the axial force in the beam. It can be surmised that the
change in the column-end moments are mainly caused by the axial force in the beam, or the
thermal expansion of the beam. The large push-force by the thermal expansion of the beam
initiates the significant moment variation in the column-ends. Another factor affecting the
moment variation is the thermal bowing in the beam. As mentioned before, for the slim floor
beam, a large temperature gradient occurs in the beam section in fIre. The beam tends to bend
 109

towards the fire-exposed side due to the greater expansion of the lower part of the beam. Because
of the flexural continuity at the beam-ends, this deformation is resisted and the end moments are
developed in the beam-ends. The third factor contributing to the moment variation is the
variation in the stiffuess ratio among the structural members during fire. This variation might
lead to a moment redistribution.
Although the axial restraints are somewhat advantageous to the stability of a fire-exposed
beam, the large push-force due to the thermal expansion is detrimental to the stability of the
columns. The p-~ effect can reduce the fire resistance of the column, especially the side-
columns. Moreover, the large axial force in the beam, in combination with the hogging moment,
is prone to cause local buckling in the beam-ends. This buckling will further lead to a larger
vertical displacement of the beam. Nevertheless, this will not occur for the slim floor beam
because most parts of the steel beam are encased within concrete.

STRUCTURAL RESPONSE OF A THREE-DIMENSIONAL SLIM FLOOR FRAME

Frame and Loads

The analyzed composite steel-frame building was designed to resemble one part of a typical
office development. The frame covered an area of 9x12 meters with a level height of 3 m (see Fig.
12). There were two equally spaced bays of 6 m along the length of the building, and of 4.5 m across
the width of the frame. The structure is assumed to be a braced frame. The beams were designed as
pin-connected with the steel column via the steel beam, acting together with the floor slab. The
overall minimum depth of the slab was 147 mm. The Finnish slim floor beam was used for both the
central beams and edge beams (see Fig. 2(b)). The composite floor system consisted of the Ranila
153 steel deck with normal-weight concrete of grade C30/25 and A142 anti-crack mesh. The
hogging reinforcement is <1>16 c/ c 150. The universal column of grade 355 was protected by aerated
concrete blocks in between the steel flanges. The uniform design load on the floor was 2.5 kN/m2 in
fire and the according load ratio was 0.15. The beams were four-point loaded and the nominal load
ratio was 0.69. The frame was connected together by the tie members across the width.

The fire followed the ISO standard curve and was set in the compartment of 1 ~2 axis and
A~C axis. The temperature distribution of the heated columns (Al, Bl, Cl and A2, B2, C2) was
similar to that in the previous section.

Four types of elements were used in this modelling:

.Beam element (B32), to represent the steel beams and columns;
.Shell elements (S8R), to represent the continuous floor slab;
•Truss elements (T3D2), to represent the tie members;
.Joint (or Spring) element, to represent the steel beam-to-column connection;
Both physical nonlinearity and geometric nonlinearity were taken into account in this study.
 110

-<; Slim floor beam

---rIB EJ El
3000

E Rannila 153 109kN 109 kN 109 kN l09kN 109kN 109 kN 109 kN 109 kN
It)
.,.: ,composite slab Tie member

B
-*- B
E
It)
Slim floor beam

q=2.S kN/m2
E3
H-steel column
with areated
concrete
B
Tie member
>-----6000
:.
• i
6000
i
3000

.,.:
(b)

). ~ El Slim floor beam D B

6m 6m
-
-- =:

t ~ .3
(a)

FIG. 12. Analyzed Frame: (a) Layout; (b) Elevation and Loads

Structural Response of the Floor System

The analyzed deformation mode at 90 minutes of ISO standard ftre is shown in Fig. 13. The
vertical deflection profile of the composite floor in 1 ~2 and A~C axes at 90 minutes is shown in
Fig. 14. The maximum deflection in the slab at 90 minutes is 318 mm and that in the central beam
(B and 1 ~2 axes) is 302 mm. The floor still maintained the stability at 90 minutes' ISO heating.
From Fig. 14(a), it can be seen that the maximum relative displacement between the beam and
concrete floor slab is small.
Fig. 14(b) shows the displacement of the composite beam with the heating time. The
isolated simply supported beams with a load ratio of 0.69 and 0.54 (modifted according to Eq. 1, see
Table 7) are also represented in this ftgure. It can be seen that the beam in the frame exhibits far
better behaviour in ftre than the isolated beam. When applying the modifted load on the isolated
simply supported beam, the deformation behaviour is close to that of the beam in the frame.
However, the beam in the frame maintains stability even after 60 minutes, and no run-away point
occurs. It can also be concluded that the effect of the composite slab is not signiftcant in this
example, because the little restraints were provided by the surrounding structures.

TABLE 7. Modified Load Ratio of Frame Beam

Calculation of Modified Load Ratio

Free Bending Moment Mstatic (kNm) 442
Sagging Design Moment Capacity Mp+ (kNm) 618
Hogging Design Moment Capacity Mp- (kNm) 399
Modified Load Ratio (by Eq. 1) 0.54
 111

FIG. 13. Deformation Mode of Frame at 90 min under ISO Fire

Time (min)
-..... ~
..
cG)
0
, .......
-.
50 60 70 80 90
S '\
-'00 G) -100 \ ...
u
-200 --
'"CoSS
.!! -200
"-,,-
"- "-
"-
"-
-300 Q-
Ci
u
-frame beam
Isolated beam (R\O.54,
",
1: -300 ----. modified)

~
-.-. Isolated beam (R=~9) ;

z 2 -400 \ \

k-:x @
Jl. ().--Q)

(a) (b)
FIG. 14. Deformation of Floor System (at 90 min): (a) Floor Slab; (b) Beam

TIJ1'e (rrin)

ro 7080~

-'00

-200 .. ~
- ...... .. - ...

-300

-fraTeb!ern
_•••• fraTe b!ern(ED;Je beEmv.e1 ~

2
(~r~· 0---"0
(a) (b)
FIG. 15. Deformation of Floor System (at 90 min, Edge Beam Well-Protected): (a) Floor Slab; (b)
Beam
 112

Effect of Edge Beams

In practicality, the lower flanges of the edge beams are often well protected by the
surrounding walls. When the edge beams are well protected, more restraints are provided for the
composite slab. Fig. lS(a) shows the obtained deflection proftle of the composite floor at 90
minutes' heating, where the edge beams were protected. The maximum vertical displacement of the
floor is 230 mm, which is 88 mm lower than when the edge beams are left unprotected. Fig. 15(b)
compares the mid-span displacement of the beam (B and 1 ~2 axes) with and without the protected
edge beams. It can be seen that the protected edge beams reduce the vertical displacement of the
beam (B and 1 ~2 axes). At 90 minutes, the reduced displacement is up to 83 mm. This effect may
be attributed to the composite slab. Since the additional vertical restraint provided by the edge
beams is negligible, the only factor is the tensile membrane action in the slab. Due to the edge
beams being protected, the slab can utilize the tensile membrane action at large displacement to
maintain the floor's stability.

CONCLUSIONS

The fire-resistant behaviour of the composite asymmetric slim floor beam both as an
isolated member and as a part of a frame with surrounding restraints was investigated using
numerical methods.
The investigated simply supported asymmetric slim floor beam has a 60-minute fire
resistance if the load ratio is less than 0.5, without any additional fire protection. The steel web
plays an important role in the fire resistance of the beam, as the bottom steel flange does at
ambient temperature. The investigation also shows that the axial restraint causes a larger vertical
displacement of the beam in the early fire phase and a more stable behaviour in the later phase.
The behaviour of the beam in the frame exposed to fire is significantly more complicated than
that of the isolated beam. The effect of frame continuity can be identified as the axial and
rotational restraints. The effect of rotational restraints can be quantified using the modified load
ratio (Eq. 1). The influence of the axial restraints on the beam is difficult to quantify. However, it
appears that the axial restraints keep the beam standing for up to 90 minutes under ISO standard
fire in this investigation (Fig. 10). The moment variation in the side-columns is largely due to the
thermal expansion of the beam. This might be detrimental to the columns due to the significant
p-~ effect. Thermal bowing and material softening (stiffness degradation) are the other factors
which contribute to the moment variation in the beam.
The behaviour of the composite slim floor system exposed to fire was also investigated.
The effect of the concrete slab on the beam deformation in fire was preliminarily identified. With
increasing fire exposure, the surrounding restraints and the large displacement make the tensile
membrane action take effect. The investigated slim floor frame still maintained its stability under
a modified load ratio of 0.54 up to 90 minutes' ISO fire exposure.
 113

ACKNOWLEDGEMENTS

The work in this paper was financially supported by the National Technology Agency of Finland
(TEKES), the Finnish Constructional Steelwork Association Ltd. and the Finnish steel company
Rautaruukki Oyj.

APPEND~I.REFERENCES

ABAQUS/Standard. (1997). "User's manual (version 5.7)." Hibbitt, Karlsson & Sorensen Inc ..
Baily, C.G., Lennon, T. and Moore, D.B. (1999). "The behaviour of full-scale steel-framed building
subjected to compartment fires." The Struct. Engr., 77(8), l5-2l.
Baily, C.G. (1999). "The behaviour of asymmetric slim floor steel beams in fire." J Constr. Steel Res.,
50, 235-257.
Baily, C.G., Burgess, L and Plank, R. (1997). "Bridging and restraint effects of localised fires in
composite frame structures." Proc. of the International Conj Composite Construction, Innsbruck,
Austria, 379-384.
"BS 5950: Part 8. Code of practice for fire resistance design." (1990). Structural use of steelwork in
buildings, British Standards Institute (BSI), United Kingdom.
El-Rimawi J.A., Burgess, LW. and Plank, R.J. (1999). "Studies of the behaviour of steel sub frames with
semi-rigid connections in fire." J Constr. Steel Res., 49,83-98.
"Eurocode 4, design of composite steel and concrete structures: Part 1.1-- General rules and rules for
buildings." (1992). ENV 1994-1-1, Commission of the European Communities, Brussels, Belgium.
"Eurocode 4, design of composite steel and concrete structures, Part 1.2: Structural fire design." (1994).
ENV-1994-1-2. Commission of the European Communities, Brussels, Belgium.
Johnson, P.F. (1998). "International developments in fire engineering of steel structures." J Constr. Steel
Res., 46(1-3), Paper No.415.
Lawson, R.M., Mullett D.L. and Rackham J.W. (1997). "Design of asymmetric 'Slimflor' beams using
deep composite decking." SCI P 175, The Steel Construction Institute, Berkshire, United Kingdom.
Lu, X. and Makelainen, P. (1996). "Slim floor development in Sweden and Finland." Struct. Engrg.
International,2, 127-129.
Ma, Z. and Makelainen, P. (1999). "Temperature analysis of composite steel-concrete slim floor
structures exposed to fire." TKK-TER-10, Laboratory of Steel Structures, Helsinki University of
Technology, Espoo, Finland.
Ma, Z. and Makelainen, P. (1999). "Numerical analysis of steel-concrete composite slim floor structures
in fire." TKK-TER-JI, Laboratory of Steel Structures, Helsinki University of Technology, Espoo,
Finland.
Malaska, M. and Makelainen, P. (1999). "Study on composite slim floor beams." TKK-TER-9, Laboratory
of Steel Structures, Helsinki University of Technology, Espoo, Finland.
Mullett, D.L., Lawson, R.M. (1993). "Slim floor construction using deep decking." SCI P127, The Steel
Construction Institute, Berkshire, United Kingdom.
Newman, G.M.(1996). "Design implications of the Cardington fire research programme." Proc. of the
Second Cardington Conj, Cardington, England, 161-168.
 114

Newman, G.M. (1995). "Fire resistance of slim floor beams." J. Constr. Steel Res., 33(1-2), 87-100.
O'Connor, M.A.(1996). "Numerical modelling of composite structures subjected to thermal loading."
Proc. of the Second Cardington Conj., Cardington, England, 127-l31.
Robinson, J.(1998). "Fire - A techincal challenge and a market opportunity." J. Constr. Steel Res., 46(1-
3), Paper No.1 79.
Robinson, J. and Newman, G.M. (1997). "Cardington fire tests: first results." New Steel Construction.
5(3),23-27.

APPENDIX II. NOTATION

The following symbols are used in this paper:

R = load ratio, defined as a ratio between the load on the beam in fire conditions to the failure load
under normal conditions;
Mean = bending moment capacity of connection at ambient temperature (kNm);
Meon,left = bending moment capacity of left connection at ambient temperature (kNm);
Meon,right = bending moment capacity of right connection at ambient temperature (kNm);
Mp+ = plastic moment capacity of the sagging beam section (kNm);
Mp_ = plastic moment capacity of the hogging beam section (kNm);
Mstatie = isostatic (free) bending moment in the beam (kNm);

Mu = plastic bending moment capacity of composite cross-section (kNm);

T = temperature (0C);

Slim = design value oflimited rotation of beam-to-column connection (rad).

 115

AppendixD

Temperature Analysis of Steel-Concrete Composite

Slim Floor Structures Exposed to Fire
Ma, Z. and Makelainen, P.,

Presented in The 6th • ASCCS International Conference on Steel-Concrete Composite

Structures - the past and the coming century, March 22-24 2000, Los Angeles, volume
I, pp.263-270.
116
 117

Temperature Analysis of Steel - Concrete Composite Slim

Floor Structures Exposed to Fire

Zhongcheng MA, Ph.D.

Researcher

Pentti MA.KELA.INEN, D.Sc.(Tech.)

Professor, Head of the Laboratory
Laboratory of Steel Structures, Department of Civil and Environmental Engineering
Helsinki University of Technology, P.O.Box 2100, FIN-02015 HUT, Finland

Abstract

This paper presents a heat transfer model to calculate the temperature of composite slim
floor beams exposed to fire. The interface thermal resistance between concrete and steel and the
effect of fire protection can be accounted for in this model. Using this model, the fire resistance
of the composite slim floor beams can be calculated and the effects of the significant parameters
such as load ratio, section size of the beam, additional reinforcement and fire protection can be
evaluated without the necessity of testing.

Introduction

In recent years, increasing interest has been shown throughout Europe in developing and
designing slim floor systems in steel-framed buildings. This system provides more opportunities
for steel in spans of 5 - 9 m. The usage of composite slim floor system has several benefits:
A slab depth of 300 mm or so, which is much less than that in the conventional steel
construction;
An almost flat underside floor which causes no restrictions for the service layout;
An inherently good fire resistance due to the encasement of concrete.
The key features of this system are the steel beam and the floor slab. There are several types
of steel beams such as 'Thor-beam', 'Delta-beam', 'Slimflor' and Asymmetric slim floor beam,
etc. As for the asymmetric slim floor beam, it can be rolled or welded. The floor slab can be a
pre-cast concrete unit or a composite slab.
Since 1991, in the Laboratory of Steel Structures at Helsinki University of Technology,
studies have been conducted to develop the design methods for the slim floor system. Recently,
the interest has been moved to develop the composite frame system with welded asymmetric
slim floor beams. Investigation of the fire resistant performance of this system was one
important part of this project.
 118

A heat transfer model and a successful program TACS-FIR for the temperature prediction of
the composite structures have been developed. This paper will focus on the temperature analysis
of the asymmetric slim floor beam, including the mathematical model and its applications. The
fire resistance of the composite asymmetric slim floor beams will also be preliminarily studied.

Heat Transfer Model

Division of Cross Section into Elements

The cross-section of the beam is subdivided into a number of rectangular elements (Fig. 1).
In view of symmetry, only half of the section needs to be modelled.

I I

Concrete
I I Steel

~.
Symmetric -
Concrete-steel'- f---
elements interface elerren I--

Fig. 1. Mesh of composite slim floor beam for thermal analysis.

Equations for Internal Elements

The coordinate system and the local element mesh are illustrated in Fig. 2(a). Based on the
governing equations of heat conduction, the following heat equilibrium equation can be derived:
At time t = (j + l)At,
TJ+l
m,n
.=.
TJ +
m,n ( )j (
At
) j
[Tj
m-l,n
-Tj

R
m,n
R
j
+ Tjm,n+l - Tm,n + Tjm+l,n - Tm,n
R
j
+ Tjm,n-l
R
-Tj
m,n
1 (1)
pc m,n + PwC w m,nqJm,n m-l,n m,n+l m+l,n m,n-l

where T is the temperature (OC). Superscripts j+ 1 and j represent the time (j + l)Llt and jLlt,
respectively. Subscript (m,n) represents the (m,n) element, and subscript w represents the water
content. Llt is the time increment. cp is the volume ratio of the water content in the concrete
element, P is the density (kglm3), and c is the specific heat (J/kgK). Thermal resistance items R
can be given by

R = Axm-l,n ·Axm,n + Ax·Ax

m,n m,n +~
Ax
(2a)
m-l n
'
A
2 m-l,n
A
2 m,n K int
R - A:y m,n+l . A:y m,n + A:y·
m,n
All
~ m,n
All
~
(2b)
m,n+l - 21 21 + K
A.m ,n+l Am,n int

R = Axm+l,n ·Axm,n + Ax·Ax

m,n
Ax·
m,n + ~
(2c)
m+l,n 2Am+1,n 2Am ,n K int
 119
R = L1Ym,n_1 ·L1Ym,n + L1Ym,n ·L1Ym,n + L1Ym,n
m,n-I 21 1 (2d)
A.m,n_1 2A.m,n K int

where Ax, Lly are the dimensions of element (m). ;t is the thermal conductivity (W/mK). Kin! is
the interface thermal resistance coefficient (W/m2K). As for the slim floor beam, the value
K;nt=50 WlnrK is proposed for steel to concrete interface. In the case of perfect contact,

Kin! =00 can be presumed (Newman 1995; Ma and Makelainen 1999).

It is assumed in this study that the interface thermal resistance is caused by the water
evaporation. The interface thermal resistance is not fonned until the temperature of the adjacent
concrete elements reaches the temperature of water evaporation (120 °C).

y
~ y

~
·
(m,n+I) •
\
· .
(1,n+I)

.
(m-I,n) (m,n) (m+-I,n)
I · .
(I,n) (2,0)

·
(m,n-I)
T.
·
(l,n.l)

~------------------~--x
I---m=1
~------------------~~x
I
I

(a) Internal elements (b) Boundary elements

Fig. 2. Local element mesh.

Equations for Boundary Elements

There are two kinds of boundaries for the slim floor beam:
• boundary exposed to fire
• boundary unexposed to fire
At time t = (j + 1)M , along the boundary exposed to fire, (see Fig. 2b)

.=.
T,J+I T,J +
M [TJ2,n -TJ T/ -TJ
l,n + l,n+1
T,1 -T/
I,n + I,n-I I,n
I,n I,n (
pc
)J
I,n + ( )
PwCw I,ntpl,n
J R2,n R l,n+1 Rl,n-I
(3)
+ &a
Llxl,n
[(Tj +273/ -(~~ +273/ ~~(Tj-~~)
Llxl,n
1
where Tg is the fire temperature (OC), which may follow the ISO standard fire, or some natural
fIre curve. The number of fire-exposed elements is set as m=1 here for convenience of
expressions. & is the resultant emissivity coefficient, G' is Stephan-Boltzmann constant, G' = 5.67
xlO·8 W/m2K4, a is the convection factor (W/m2K).
 120

Along the boundary unexposed to fire,

(4)

where Tgo is the air temperature of unexposed sides. The element number at the unexposed side
is set as m=M, for convenience of expressions.

Auxiliary Equations
Along the symmetric boundaries (set n=1 and n=N),
Tj+l - Tj+l and Tj+l = Tj+l (5)
m,l - m,2 m,N m,N+1

Effect of Moisture in Concrete

The modelling of water movement in concrete in fire is still very difficult so far. The moisture
content is modelled here in the following approximate way:
Before the element temperature rises to 120 cC, the effect of the moisture content is taken into
account by modifying the specific heat of concrete. When the element temperature rises up to
120 cC, the absorbed heat energy of the calculated element is totally used to evaporate the water
until the water content of the element becomes zero. Meanwhile, the element temperature
maintains 120 cC. After that the temperature ofthe element resumes to rise.
For the internal elements, the volume of water content is:
(6)

The change LlVm,n in water content in the internal elements under fire is derived as

(7)

along the boundary exposed to fire

LlY, = Llt·tu l,n Lly l,n [Tj

2,n -TJ
l,n + Tjl,n+l -TJl,n + Tjl,n-l -TJl,n
l,n Pwhw R 2,n Rl,n+l Rl,n-l
(8)
+ea / Llxl,n ~Tl +273l-(T/n +273l}ra / Llx ,n(Tl- T/n)
I ]

and along the boundary unexposed to fire

LlV = M·tu M,n Lly M,n [TjM-J,n -TjM,n + TjM,n+l -TjM,n + TjM,n-l -TjM,n
M,n Pwhw RM-J,n RM,n+1 RM,n-1
(9)
+ea / LlxM,JTlo +273l-(T'&,n +273l ]+a / LlxM,n(Tio -T'&,n) ] ,

where hw is the potential heat for the water evaporation, hw=2.3x10 6 J/kg.
 121

Effect of Fire Protection

The modelling of insulation material is based on the method for the temperature calculation
of steel members by the ECCS Recommendations (ECCS TC3 1983). In the recommendations,
the thermal conductivity of insulation material is measured based on the assumption of the linear
temperature distribution inside the insulation under fire. The conductivity of insulation material
is determined experimentally, based upon the equivalent of the measured steel temperature and
the calculated temperature by simplified heat transfer equation (Bruls et al. 1988; Nullifire S607
1996). The advantage in using this method in the modelling is that almost all kinds of fire
protection can be modelled with the only change in the heat input.
The heat flow transferred through the insulation to the steel elements is:
(10)
where Tg is the gas temperature (OC), Ts is the temperature of steel (OC), /),.t is the time interval
(s). A is the surface area of steel member per unit length (m). K is the coefficient of total heat
transfer (WIm2K) , which can be determined by
lId
-=-+- (11)
K a A
where d is the thickness of insulation (m). A is the conductivity of insulation material. a is the
resultant convective factor, which can be given by
(12)
where a c is the coefficient of heat transfer due to convection from fire to the exposed surface of
the member (W/m 2 °C). a r is the coefficient of heat transfer due to radiation.
It can be seen that heat input equation (Eq. 10) is the same as that used in Eq. 3, but the heat
transfer coefficient K has an additional term d.fA.

Stability Criterion
To insure the numerical stability of the temperature analysis, the critical time increment has
to satisfy the following equation

(sec) (13)

The minimum values of the volume specific heat (pc)min are 3.34x106 (J/m 3K) for steel and
2x10 6 (J/m3K) for normal-weight concrete, respectively. The maximum values of the thermal
conductivity Amax are 54 W/mK for steel and 2 W/mK for normal-weight concrete, respectively
(Eurocode 4 1994). So, the above equation can be simplified as
(14)
For the temperature analysis of steel-concrete composite structures, the minimum division
sizes of lOx10 mm are the most common and accordingly the critical time increment is
approximately 1.5 seconds.

Proposed Values for Thermal Constants

Table 1 lists the proposed values for the thermal constants which appear in Eqs. 1-12, for the
thermal modelling of the slim floor beam. Those values are based on the previous studies on
 122

composite structures (Schleich 1987; Eurocode 1 1994; Eurocode 4 1994). The sensitivity study
by Ma and MakeHiinen (1999) has shown the validity of these values for slim floor beam
modelling.
Table 1 Thermal constants used in the modelling.
Resultant Emissivity Convection Factor Interface Resistance
Coefficient E a (W/m2K) K int (W/m2K)
0.6 for steel 25 for fire-exposed side 50 for steel-concrete contact
0.3 for composite slab 8 for unexposed side oc for perfect contact

Temperature Analysis of Composite Slim Floor Beams

A computer program TACS-FIR has been written to analyze the temperature rise for the
composite slim floor structures using the previously described model. Many fire tests have been
simulated using this program, including slim floor beams, concrete-filled steel tubular columns
and concrete columns (Ma and MakeHiinen 1999). The results show a high accuracy of this
model.

Section Shape of the Beam

Since 1997, the Laboratory of Steel Structures at Helsinki University of Technology, under
the support of the Finnish steel industry, has developed a new type of asymmetric slim floor
beam (Fig. 3). This beam has a welded section, which consists of three steel plates. A steel deck
of 153 mm depth (Rannila Steel) is adopted for the composite slab. The composite beam has a
total depth of 318 mm. zoo
1< ~I
....... " " .,'. . :' ~., , . ' " .
•• '
.
. : ' ,4,'.
.'
_

Fig. 3. Composite section of asymmetric slim floor beam.

Temperature of the Beam During Fire Exposure

In this analysis, the thermal properties of concrete and steel were taken from Eurocode 4 Part
1.2 (Eurocode 4 1994). The beam with the fire protection underneath the bottom steel flange was
also modelled. Nullifire® S607, an intumescent fire paint, was used in this study ("Nullifire
S607" 1996). A nominal thickness of 250 Jlm was adopted, which is the permitted minimum
thickness in practice. The temperature distribution of these beams at 60 minutes' ISO fire
exposure is shown in Fig. 4. It can be seen that there is a large temperature gradient within the
 123
steel web due to the encasement of concrete. In the case of the protected bottom flange, the
average temperature of the bottom flange reduces from the unprotected value of 760°C to 500
°C. The temperature rise vs. time for the steel beam without fire protection is shown in Fig. 5.

C
4.9 48 ':'4744' ':4cf-'35 494$-47'44'40 ',;34
57 '56"55'52' 46.:39 57' 56 ,54 '51',46' :38:
77 '7774"'10 .6i<50 ,77167.4,69,:61' '49
H3112 l09 '102'8968 1l3ii2io9fo2. "&868
172170 ·16'6.'i56134 ',99 172170166156.134:'98
265263258244 2l,O.14.6.' 265 .2,64, 25S244'2:1b145:90.
4113416'.4:11393 341"21,8 417 '4'16.410.392,3:4'1' '.21:7 3:23
68,7 685.681'663.'594 . 686 ,685:680662 '. 593 '323>163,88:
47.9207106'
560Z4ii23
600'Z65,139 • 8,9
62228'5,'104
6.36"3'03, ""'''',.L:''.~:,
6;49,325 .

(a) without fire protection (b) with fire protection

Fig. 4. Temperature distribution of composite slim floor beam at 60 minutes' ISO fire exposure.

1.4
1000
1.2
800 c:::
6'
~
~ 600
.Q
~ 0.8
'.
::l

-------
lii JXlint2 _-- ~ 0.6
Q)
Q.
400 o
...J 0.4 --Wrt:lDut prctection
--- ...
E
Q) ..,....,..---- •. - - .. WIth iItuIrelcellt paint
I- 200 ..,.. 0.2 - - _. WIthre1:ms2d32
o ....-...-.--::-.~... --
c:~~~~~~ __ ~~
_._ .. -_ .. _--_ ...
_-----_ ..__
... - .. - ...
~~--J o ~~--~~--~~~~--~~~
o 10 20 30 40 50 60 70 80 90 o 10 20 30 40 50 60 70 80 90
Rre exposure time (min) Fire expc:st.Je tirre (mn)

Fig. 5. Temperature variation in the steel beam Fig. 6. Influence of enhanced reinforcement
under ISO fire. and fire paint on load ratio R.

Fire Resistance Analysis

After the temperature is known, the bending-moment capacity method is used to analyze the
fIre resistance of the composite beam. The reduction factors of the yield strength for structural
steel and compressive strength for concrete are taken according to Eurocodes. The resistant load
ratio vs. time is calculated according to the ISO fire exposure. The resistant load ratio is defIned
as
R = Moment capacity at fire lim it state
(15)
Ultimate moment capacity at 20°C
In Eurocodes, fire is regarded as an accidental action so that the structural capacity will be
calculated using the characteristic values of material strength. However, in the denominator of
the right side ofEq. 15, the ultimate capacity should be calculated with the appropriate factored
 124

material strength. Grade S355 (yield strength 355 N/mm2) for structural steel and grade
NWC30/25 for concrete are used in this analysis. The yield strength of the reinforcement is 355
N/mm2. Full composite action between steel and concrete is assumed.
Fig. 6 shows that the resistant load ratio of the beam section decreases under the ISO fire
exposure. Three cases were analyzed: the beam without any additional measures, the bottom
steel flange with 250 /lm-thick Nullifire® intumescent fire paint and the beam with a
reinforcement 2<1>32 close to the bottom flange. The reinforcement is located in the concrete 50
mm away from the bottom flange. It can be seen that the beam has 50 minutes' fire resistance if
the load ratio is less than 0.6, without any additional fire protection. A fire resistance of 60
minutes could be reached if a reinforcement 2<1>32 was added. In the case of the intumescent fire
paint (250 JlID-thick Nullifire®), the fire resistance time of 105 minutes can be achieved.

Conclusions

A heat transfer model for the temperature analysis of composite slim floor beams has been
presented. Using this model, the temperature development of the Finnish asymmetric slim floor
beam was analyzed and its fire resistance was investigated.
U sing this model, the temperature development and fire resistance of the slim floor beam can
be evaluated with the different significant parameters such as the section size of the beam, fire
protection underneath the bottom flange and additional reinforcement, without the necessity of
testing. This model can also be used for the temperature prediction of other composite
components having the rectangular sections exposed to fire, with only few changes necessary for
the modelling.

Acknowledgements

The authors gratefully acknowledge the financial support to this research project from the National
Technology Agency of Finland (TEKES), the Finnish Constructional Steelwork Association Ltd (FCSA)
and the Finnish companies, Rautaruukki Oyj and PPTH Teras Oy.

References

ECCS TC3, 1983, "Fire safety of steel structures". European Recommendations for the Fire Safety of
Steel Structures, Elsevier.
Eurocode 4, 1994, "Design of composite steel and concrete structures, Part 1.2: Structural fire design".
ENV 1994-1-2, Brussels.
Ma, Z. and Makelainen, P., 1999, "Temperature analysis of steel-concrete composite slim floor structures
exposed to fire". Laboratory of Steel Structures, Helsinki University of Technology, TKK-TER-I0,
Espoo, 47p.
Newman, G.M., 1995, "Fire resistance of slim floor beams". Journal of Constructional Steel Research,
33(1-2): 87-100.
Nullifire S607, 1996, "Fire paint, certified directions for use (in Finnish)". TRY-18-1996, Helsinki.
Schleich, J.B., 1987, "Computer assisted analysis of the fire resistance of steel and composite concrete-
steel structures". REF AO-CAFIR, Luxembourg, 293p.
125
 126

Index of the Author's Publications

1. Articles in scientific journals

1. Ma, Z. and Makelainen, P., Behaviour of Composite Slim Floor Structures in Fire.
Journal of Structural Engineering, ASCE, Vol. 126, No.7, pp.830-837, July 2000.
2. Ma, Z. and Makelainen, P., Parametric Temperature-time Curves of Medium
Compartment Fires for Structural Design. Journal of Fire Safety, Vol. 34, No.4, pp.361-
375, June 2000.
3. Makelainen, P. and Ma, Z., Fire Resistance of Composite Slim Floor Beams. Journal of
Constructional Steel Research, Vol. 54, No.3, pp.345-363, June 2000.
4. Wu, B., Ma, Z. and Ou,]., Experimental Study on Seismic Properties of R.C. Columns
after High Temperature. China Civil Engineering Journal, Vol.32, No.2, 1999, pp53-58.
5. Ma, Z., Wu, B. and Ou, J., Seismic Characteristics of R.C. Columns after Fire. Chinese
Journal of Computational Mechanics, Vol.14, No.4, 1997, pp443-453.
6. Wu, B., Ma, Z. and Ou, J., Seismic Behavior of Fire-Damaged R.C. Columns.
Earthquake Engineering and Engineering Vibration (Chinese Journal, English version),
Special Issue for 30th World Conference of Geology, Vol.16, 1996, pp 76-86.
7. Wu, B., Ma, Z. and Ou,]., Stress-Strain Curves of Fire-Damaged Concrete under Cyclic
Axial Loading. Earthquake Engineering and Engineering Vibration (in Chinese), Vol.17,
No.3,1997.
8. Wu, B., Ma, Z. and Ou,]., Stress-Strain Curves of Fire-Damaged Concrete Under Static
Axial Loading. Chinese Journal of Building Structures 1999; 20 (5); 42-49.
9. Wu, B., Ma, Z. and Ou, J., A Study on Seismic Behavior of Fire-Damaged R.C.
Compression-Flexural Components. Earthquake Engineering and Engineering
Vibration (Chinese Journal), Vol.14, No.4, 1994, pp24-34
10. Ma, Z., Wu, B. and Ou, J., Fully-Developed Compartment Fire Model and Its
Application. Journal of Harbin University of Civil Engineering and Architecture, V01.31,
No.1, 1998, pp16-24
11. Wu, B., Li, H. and Ma, Z., Damage Analysis of Fire -Damaged R.C. Structures under
Low-Cyclic Loading. Journal of Harbin University of Civil Engineering and
Architecture, Vol.30, No.3, 1997, p7-12
12. Ma, Z., Wu, B. and Ou, ]., Literature Review of Fire Engineering Design of R.c.
Structures. Building Structures (Chinese Journal), No.ll, 1997, pp48-52
 127

13. Wu, B., Ma, Z., etc., Fire Disaster and Its Retrofit of Shenyang Department Store.
Chinese Journal of Catastrophe, Vol.11, No.4, 1996, pp48-53

14. Ma, Z., Wang, L. and Wu, Z., Aseismic Behavior of Composite Steel-Concrete Frame-
State of the Art. Journal of Harbin University of Civil Engineering and Architecture,
Vo1.29, No.6, 1995, pp132-138
15. Ma, Z. and Wu, Z., Elastic Stiffness Matrix of Fully Connected Steel-Concrete
Composite Beams. Journal of Harbin University of Civil Engineering and Architecture,
Vo1.28, No.5, 1995, pp235-241
16. Ma, Z. and Wu, Z., A Mechanical Model on the Stud Connectors in Composite Beams.
Journal of Harbin University of Civil Engineering and Architecture, Vo1.28, No.5, 1995,
pp242-248
17. Ma, Z. and Wu, Z., A Parametric Study on the Effective Width of Steel-Concrete
Composite Beams - A New View. Journal of Harbin University of Civil Engineering
and Architecture, Vo1.28, No.5, 1995, pp249-255

2. Articles in scientific conference proceedings

18. Ma, Z. and Makelainen, P., Temperature Analysis of Steel-Concrete Composite Slim
Floor Structures Exposed to Fire. The 6th • ASCCS International Conference on Steel-
Concrete Composite Structures - The Past and the Coming Century, March 22-24, 2000,
Los Angeles, Volume I, pp.263-270.
19. Wu B. and Ma, Z., Damage Assessment and Seismic Retrofit of R.C. Structures after
Fire. Theories and Practices of Structural Engineering, Seismological Press -
Proceedings of the 5th International Symposium on Structural Engineering for Young
Experts, Shenyang, P.R. China, 1998, pp599-607
20. Ma, Z. and Makelainen, P., Fire Resistance of Hybrid Steel Beam in Composite Slim
Floor Construction. Terasrakenteiden Tutkimus- ja Kehityspaivat, Teknillinen
Korkeakoulu, Tammikuu 2000.
21. Ma, Z. and Makelainen, P., Fire Safety Design of Composite Floor System.
Terasrakenteiden Tutkimus- Ja Kehityspaivat, Lappeerannan teknillinen korkeakoulu,
Aug. 1998, Finland.

3. University publications

22. Malaska M., Ma, Z. and Makelainen P., Steel-Concrete Composite Slim Floor Frame
System. TKK-TER-14, Laboratory of Steel Structures Publications 14, Helsinki
University of Technology, Espoo, 2000, p53.

23. Ma, Z. and Makelainen, P., Numerical Analysis of Composite Steel-Concrete Slim
Floor Structures in Fire. TKK-TER-ll, Laboratory of Steel Structures Publications 11,
Helsinki University of Technology, Espoo, 1999, p47.
 128

24. Ma, Z. and Makeliiinen, P., Temperature Analysis of Steel-Concrete Composite Slim
Floor Structures Exposed to Fire. TKK-TER-l0, Laboratory of Steel Structures
Publications 10, Helsinki University of Technology, Espoo, 1999, p47.
25. Ma, Z., Fire Engineering Design of Composite Floor System. Seminar on Steel
Structures: Design of Steel-Concrete Composite Structures, Mikko Malaska etal.
Laboratory of Steel Structures, Helsinki University of Technology, Report 4, Espoo,
1998, pp88-104
26. Ma, Z., Damage Evaluation and Aseismic Repair of R.C. Structures after Fire. TKK-
TER-3, Laboratory of Steel Structures Publications 3, Helsinki University of
Technology, Espoo, 1998, p75

27. Ma, Z. and Wu, B., Evaluation Technology of Fire- Damaged R.C. Structures. Institute
of Engineering Theory and Its Application, Harbin University of Civil Engineering and
Architecture, Research Report, 1997, P 126
28. Ou, J., Duan, Z., Ma, Z., et al. Studies on the Ice-Vibrated Controls of Offshore
Structures Using Viscous-Elastic and Hydraulic Dampers - Numerical Analysis. Internal
Research Report, Institute of Engineering Theory and Its Application, Harbin
University of Civil Engineering and Architecture, July, 1997, p420.
29. Makelainen, P. and Ma, Z., Vertical Deflection of Buried Corrugated Steel Pipes.
TeRT-99-05, Laboratory of Steel Structures, Helsinki University of Technology, Espoo,
1999.

4. Computer programs

Ma, Z., Temperature Analysis of .Gomposite ,Structures Exposed to FIRe (fACS-FIR).

Laboratory of Steel Structures, Helsinki University of Technology, Nov., 1998.
Ma, Z., Finite Element Analysis of Steel-Concrete Composite Beam - for some special
problems (FEA-SCCB). Department of Structural Engineering, Harbin University of
Civil Engineering and Architecture, 1994.
IS BN 951-22-5222-8
ISSN 1456-4327