Sei sulla pagina 1di 322

lindustria

italiana
del Cemento

The 3rd International fib Congress


Washington DC, May 29 June 2, 2010

Italian National Report


Research and Construction
ISSUE 30,00
ISSN 0019-7637

ASSOCIAZIONE ITALIANA ANNO / YEAR LXXX


SPECIAL ISSUE
TECNICO ECONOMICA
DEL CEMENTO
APRILE / APRIL 2010
854
RIVISTA DELLASSOCIAZIONE ITALIANA TECNICO
ECONOMICA DEL CEMENTO (AITEC) lindustria italiana
del Cemento
Anno/Year LXXX
854 Aprile/April 2010 Contents

Direttore responsabile
3 Foreword (M. Menegotto)
Managing Editor
Laura Negri RESEARCH
SEISMIC BEHAVIOR
Collaboratori
Assistants
6 Earthquake Engineering of Reinforced Concrete Structures: The Italian
Marco Veronesi State-of-the-art
L. Ascione, E. Cosenza, G. Mancini, G. Manfredi, G. Monti, P. E. Pinto
Grafica e Impaginazione
Design & Editing
Studio Mariano - Roma
74 Experimental Research on Seismic Behavior of Precast Structures
F. Biondini, G. Toniolo
Editore
Publisher CONCRETE
80 State-of-the-art on Research on Structural Concrete in Italy
M. Collepardi
Direzione e redazione:
Piazza Guglielmo Marconi, 25 - 00144 Roma - Tel. 06/54210237 -
Telefax 06/5915408 - E-mail: iic@aitecweb.com LAQUILA EARTHQUAKE
Autorizzazione del Tribunale di Roma n. 301 del 24 Ottobre 1950.

Concessionaria per la pubblicit: 88 Damages of LAquila earthquake


Idra S.A. G. Manfredi
Strada Cardio, 4 - 47891 Dogana (RSM)
Tel. 0549-909090, fax 0549-909096
e-mail: info@idrabeton.com www.idrabeton.com 124 Reconstruction between temporary and definitive: the CASE project
Amministrazione: G.M. Calvi, V. Spaziante
PUBBLICEMENTO s.r.l. - Sede legale:Viale Ettore Franceschini, 37 -
00155 Roma. Sede amministrativa e operativa:
Piazza Guglielmo Marconi, 25 - 00144 Roma
Tel. 06/54210237 - Telefax 06/5915408 CONSTRUCTION
Fascicolo/Issue 30 CIVIL ENGINEERING WORKS
Tutti i diritti di riproduzione sono riservati. Nessuna parte di questa 154 Italian High-Speed Network. A special focus on concrete structures
Rivista pu essere riprodotta in nessuna forma. La Rivista non
assume la responsabilit delle tesi sostenute dagli Autori e delle HS railway cable-stayed bridge over Po river
attribuzioni relative alla partecipazione nella progettazione ed Piacenza viaduct
esecuzione delle opere segnalate dagli stessi Autori.
All rights of reproduction are reserved. No part of this Magazine may be Modena system viaducts
reproduced in any form whatsoever.The Magazine assumes no
responsibility for the theses put forward by the authors or for the authors' Savena viaduct
indications of their responsibilities regarding participation in the design
and construction of the works.
Caivano variation structures
Tunnels in the Florence-Bologna stretch of High-Speed Line
New stations for Italian High-Speed Network
188 Colletta cable-stayed bridge
192 Rio SAdde viaduct
196 Bridge over Vajont creek
200 Cesare Cant cable-stayed bridge
204 Bridge between La Maddalena and Caprera Islands
208 Don Bosco bridge. Architecture, white as light
212 Viaduct for State Road (SS) 23
216 Roccaprebalza viaduct
220 Sandro Pertini bridge upgrade
224 Cable-stayed footbridge over the Frodolfo river
228 Bridge over Mazzocco creek
232 Bridge over the Sacco river
236 Bridge over the Santa Caterina channel
Associata allUnione Stampa Periodica Italiana.
240 Bridge over the Cimadolmo branch
ISSN 0019-7637 242 Isola della Scala bridge
Stampa/Printed by: Grafica Ripoli snc - Via Paterno - Villa Adriana (Tivoli) 244 The Strada dei Parchi
In copertina 248 A24 Completion of the motorway Roma-LAquila-Teramo
On front cover
Pantheon dome, Rome. Inside view 250 S. Antonio viaduct
254 Viaduct for the Algeria East-West Motorway
lindustria italiana
del Cemento

Contents

BUILDINGS
258 MAXXI Center for the contemporary arts
262 Olympic Palavela
266 New Bocconi University
270 Milanofiori 2000 Corporate Center
274 Acqua minerale San Benedetto plant
278 New Sky Italia Headquarters
282 Light Pavilion
286 Agenzia Spaziale Italiana new headquarters
290 Altra Sede for the Regione Lombardia
294 New SantAnna Hospital
298 Verdi Theatre
302 Banca Lombarda Center
306 Boglietti Palace
310 Cuore immacolato di Maria parish complex
314 San Giovanni Battista parish complex
318 Somada Business Center

Edited by A.I.C.A.P. (Associazione Italiana Calcestruzzo Armato Precompres-


so) and lindustria italiana del Cemento ii C
The 3rd International fib Congress
Italian National Report
Research and Construction
Foreword

he 3rd fib International Congress takes place during (hopefully towards the end of) a global downturn in the economy.
T The impact was felt in all fields, including construction industry, and in all countries.
Crises are opportunities for rethinking activities and habits. The need to face high costs of materials and production stimulates
the search for improvements. This parallels the awareness of further needs, not of immediate return, such as saving resources,
building more durable constructions, recover for reuse or recycling at least of materials, or better yet of whole structures or parts
of them.
Design and construction usually take account of safety and economy but more and more they consider too service life
requirements, products life cycle assessment, resource saving and other environmental issues, in one word, sustainable
development.
CEB and FIP, now fib, were accompanying and leading the progress of structural concrete design and construction during the
past half century and have been a great partner in cultural interchange for our country.
AICAP, the national association for structural concrete, mirrors fib and takes advantage also of its actions to contribute in its
turn to the dissemination of knowledge and the improvement of practice. Among other initiatives directed toward encouraging
engagement in better design and execution, AICAP launched for the first time a national award for the best concrete structures
for buildings and civil engineering works, that will be given every second year, and edits this National Report, where a
selection of works using structural concrete and completed in Italy in the last four years is illustrated, among which both 2009
award winners. Romes Pantheon is the exception, with its record 43.4 m dome shown on the cover, cast altogether 20 centuries
ago, using a lightweight concrete quite similar to the modern one.
Research is playing a decisive role in advancing techniques to fulfil new requirements and has progressed in Italy during these
years, promoted also by the industry, more sensitive to it while under pressure of the crisis. Therefore, this Report also describes
some results related to the national state of the art.
Italy is a highly seismic country, thus most researchers in recent past have worked in the field of seismic engineering, aiming at
better design of new structures as well as at assessing and retrofitting older ones.
Unfortunately, one year ago this science was called to prove its ability after the LAquila earthquake. The authors of the articles
in this Report were personally involved also in the field operations. Their contributions helped the surveys in the immediate
post-event days, the directives for the repairs and the implementation of the CASE project. The latter has represented quite a
successful system response, up-to-date and praised worldwide, that provided a large population of homeless with sets of
permanent houses, built in few months on seismically isolated platforms, located in sites purposely selected and equipped in the
city neighbourhoods.
AICAP is then glad to present the 2010 National Report, which comes in to being too thanks to the renowned magazine ii C,
which is publishing this special issue, in spite of the difficult times.

Marco Menegotto
Head of the Italian delegation to fib

iiC4/2010 3
RESEARCH
Earthquake Engineering of Reinforced
Concrete Structures:
the Italian State-of-the-art

T his extended paper summarizes part of the results, those related to


reinforced concrete (RC) structures, of the largest research program
on earthquake engineering ever held in Italy. The ReLUIS project fund-
2FC: Calibration of Confidence Factors
3IRREG: Assessment of the Nonlinear Behavior of Buildings, with
Emphasis on Irregular Ones
ed by the Department of the National Civil Protection granted 4MIX: Assessment and Strengthening of Mixed-type (Masonry/RC)
15.000.000 and involved more than 600 researchers all over the coun- Buildings
try between 2005 2008. The ten research tasks (lines) ranged from the 5TAMP: Influence of Infills on Structural Response
seismic risk of existing structures to new design paradigms, including 6SCALE: Behavior and Strengthening of Stairs
geotechnical earthquake engineering issues and innovative approaches to 7NODI: Behavior and Strengthening of Beam-Column Joints
seismic risk reduction as earthquake early warning systems as well as 8BIAX: Behavior and Strengthening of Columns under Combined
emergency management. Axial Load and Biaxial Bending and Shear
In the following the ReLUIS research lines regarding: (I) assessment of 9PREFAB: Behavior and Strengthening of Prefabricated Industrial
existing RC buildings (Line 2); (II) bridges (Line 3); and (III) retrofit of Structures
RC structures via innovative materials (Line 8), coordinated by the For the sake of clarity of exposition, given the large variety of different
authors, are described in their development and findings. The three sec- subjects, the structure of the paper, in each of its sections, follows the
tion of the paper are structured as stand-alone, including their own intro- Task organization above.
duction, conclusions, vision and references, for readability purposes. More
details, references and products of the project may be found in the 1.1 MND: Non-Destructive Methods for the Knowledge of Existing
research section of the ReLUIS website (http://www.reluis.it/). Structures
Because of timeliness of the ReLUIS project and the involved scientists, Task MND specifically focuses on the knowledge of the constituent
the work described in the following is likely to express the state of the art material properties of Reinforced Concrete (RC) existing structures.
of earthquake engineering concerning reinforced concrete structures in Particularly, Task MND is devoted to the estimation of the in-situ con-
Italy. crete strength by using destructive and non destructive methods. Data
on material properties from several in-situ and laboratory investigations
were collected and analysed with the major objective of defining reli-
able as well as not very expensive procedures and criteria for the esti-
I - ASSESSMENT AND REDUCTION OF THE mation of the in-situ concrete strength. Further, methods for the treat-
VULNERABILITY OF EXISTING REINFORCED ment of the uncertainty that characterizes experimental data obtained
CONCRETE BUILDINGS through in-situ and laboratory investigations were analysed and some
theoretical simulations of the influence of material properties on the
1. INTRODUCTION seismic capacity of existing buildings were carried out.

Research Line 2 focuses on the assessment of the seismic performance 1.2 FC: Calibration of Confidence Factors
of existing reinforced concrete buildings, covering a wide spectrum of A fundamental phase in the assessment of existing reinforced concrete
problems, each one treated within a single Task. These aspects span buildings and in their strengthening design is the knowledge process
from those related to the preliminary knowledge phase, to the use of that one has to follow to acquire the necessary information. This is
nonlinear assessment methods, while placing emphasis on peculiar based on the collection of different kinds of information regarding: a)
modeling problems, such as those related to the presence of stairs, the structural system configuration, b) the materials strength, c) the
infills, beam-column joints, and biaxial behavior of the elements. The reinforcing steel details, and d) the conditions of the structural ele-
Research Line is also devoted to the study of mixed-type (masonry/RC) ments.
buildings and of prefabricated industrial buildings. The Italian Code (OPCM 3431, 03-05-05, Annex 2) as well as the most
The following Task list collects and organizes the entire scientific activ- advanced International Codes (FEMA 356, EC8 Part 3) specifies data
ity: collection procedures about the configuration of the structural system,
as well as material strength and condition of the structural elements
1MND: Non-Destructive Methods for the Knowledge of Existing comprising the building, and ensuing Confidence Factors (CF) to apply
Structures to the mean values materials properties, based on the quantity and

6 RESEARCH - Seismic behavior


Luigi Ascione1, Edoardo Cosenza2, Giuseppe Mancini3, Gaetano Manfredi2, Giorgio Monti4 and Paolo E. Pinto4
1 Universit degli Studi di Salerno, Fisciano (SA), Italy
2 Universit degli Studi Federico II, Naples, Italy
3 Politecnico di Torino, Turin, Italy
4 Sapienza Universit di Roma, Rome, Italy

quality of the information gathered (the so called Knowledge Level). In masonry infills can modify substantially the expected seismic response
the current approach, the CFs are given through tables. of framed structures although special devices connecting the infill pan-
Aim of the Task has been: a) evaluation of CF effects on the assessment els with the surrounding meshes of frame are not applied.
of buildings seismic performances; b) development of a procedure for In spite of that, most seismic codes (the more recent too) give some pro-
the evaluation of concrete and steel strength, to be reliably used in visions in order that the resisting elements of frame bear the
assessing members capacity; c) new definition of CF, evaluated by a unfavourable effects of a non-uniform infill distribution in plan or/and
closed-form equation as a function of number, kind and reliability of in elevation, but they do not suggest any procedure to quantify these
each testing method employed, and of the reliability of prior informa- effects or the favourable lateral stiffness and resistance contributions
tion. that infills give when they are uniformly located.
This gap occurs since the influence of the masonry infills on the seis-
1.3 IRREG: Assessment of the Nonlinear Behavior of Buildings, with mic response of framed buildings is a still open research topic, where
Emphasis on Irregular Ones univocal and general results do have not been achieved. The present
Task IRREG deals with problems related to the definition of plan and study refers to this subject with the following main objectives:
elevation irregularity and the effects of irregularity on the structural - mechanical characterization of the masonry infill kinds that are com-
behavior and its prediction though different methods of analysis pro- monly utilized in the Italian country by means of experimental tests on
vided by the current design codes. their components (resisting elements and mortar) and masonry samples;
More specifically, the main focuses of this tasks are: - experimental investigation on infilled meshes of RC frames, with the
a. Study of the definition and effects of plan irregularity on the response aim of calibrating a pin-jointed equivalent diagonal strut model.
of RC buildings up to the Ultimate and Collapse limit states; A companion study has also been devoted to verifying the influence of
b. Comparison and calibration of different linear and nonlinear meth- the infills on the seismic response of RC framed structures; for this pur-
ods for reinforced concrete structural members, with emphasis on pose, shaking table tests on a 1:2 scaled 3D building and numerical
pushover and nonlinear dynamic analyses; nonlinear analyses of multi-storey frames have been carried out.
c. Comparison between research-oriented and professional-oriented
structural analysis software, in order to identify analytical tools that sat- 1.6 SCALE: Behavior and Strengthening of Stairs
isfy both modeling precision and computational speed. The main objectives of this Task are the following: Identification of the
main stairs typologies used in the past construction practices;
1.4 MIX: Assessment and Strengthening of Mixed-type (Masonry/RC) Numerical investigation of the influence of the stair substructure on the
Buildings structural seismic response. In particular, both global and local seismic
The work has been developed on the evaluation of the seismic response performance have to be investigated with reference to frame and stairs
of mixed-type buildings behaving as parallel systems, with regard to members connections; Construction of building sub-assemblages,
both local (interaction between masonry and RC elements) and global including a stair substructure, for experimental tests execution, specif-
features, by performing a series of non linear numerical analyses. ically targeted at understanding their seismic performance.
The research activity has been focused on:
a. the classification of the main geometrical characteristics of such kind 1.7 NODI: Behavior and Strengthening of Beam-Column Joints
of buildings and the study of their response behaving as parallel sys- This Task aims at investigating the experimental behaviour of RC struc-
tems subjected to horizontal forces; tural members, particularly beam-column joints without or with
b. the problems concerning the modelling of mixed-type buildings and strengthening, thus providing a contribution to a more reliable evalua-
on the distribution of the seismic action between masonry and rein- tion of the seismic vulnerability of RC existing buildings. In particular
forced concrete elements by performing a series of numerical analyses of great interest is the understanding and the validation of capacity
obtaining the capacity curves of individuals resistance elements and models relevant to the joint panel zone in beam-column sub-assem-
the building as a whole. blages reported in the literature and in seismic codes. Further, there is
a need of knowledge in the field of strengthening and retrofit systems
1.5 TAMP: Influence of Infills on Structural Response that can be used taking into account the actual geometry of joints: e.g.
Several theoretical and numerical analyses, and, above all, the damage presence of slab and other framing elements that could prevent an
distributions on buildings that have suffered an earthquake show that effective arrangement of the retrofitting system. To this purposes, wide

RESEARCH - Seismic behavior 7


bibliographic research on the experimental investigations on beam-col- gaps in the code, related to certain procedural and methodological
umn joints and on different repairing/strengthening techniques as well aspects in the seismic assessment of existing buildings.
as experimental researches on different joint specimens have been car- Specifically, for Task MND, it is noted that, in the case of non-destruc-
ried out. tive testing, a lack of clearness exists about the relative importance of
such tests with respect to destructive ones for the evaluation of materi-
1.8 BIAX: Behavior and Strengthening of Columns under Combined al properties.
Axial Load and Biaxial Bending and Shear The data acquisition modality has immediate consequences on the cal-
Modern approach to safety assessment of existing reinforced concrete ibration of Confidence Factors, treated in Task FC, which may assume
structures and design of strengthening interventions, in particular those different values from those given in the code, in case one accepted to
aimed at increasing ductility of columns, are based on enhanced and include results from non-destructive tests in addition to or even in
complex methods for structural analysis (seismic demand), but also on substitution of destructive ones.
the availability of data concerning performances of members at failure Moving to the level of analysis methods for seismic safety evaluation,
(seismic capacity). the need of a deeper insight into the usual assessment techniques is
On the other hand, common constructions are not necessarily affected recognized, with particular emphasis to their predictive capacity when
by regular shapes and/or regular distribution of seismic resistant sub- dealing with irregular buildings, dealt with in Task IRREG. It would be
structures, so that seismic actions result in complex deformation paths expedient to identify, for example, a synthetic parameter capable of
on columns and in general on compressed resisting members. quantifying the level of irregularity and, possibly, an associated applic-
This is the reason why Task BIAX research activity has been devoted ability threshold that helped selecting the most appropriate assessment
to provide an insight on the response of r.c. members subjected to biax- method, be either of simplified nature, such as pushover analyses, or
ial bending and axial load. In particular, some aspects have been more refined, such as nonlinear dynamic analyses.
analysed in detail. In compliance with the overall objectives of the For mixed-type (masonry/RC) buildings, the lack of code provisions,
research programme as a whole, Task BIAX duties were the definition which could guide the designer towards the assessment of the com-
of a set of reliable and well documented data and procedures concern- pound behavior in a unitary manner also accounting for interface
ing: (a) rotation capacity of r.c. members subjected to generalised bend- actions between different constructive typologies , is absolutely strik-
ing and axial forces; (b) development of simplified methods of analysis ing and appropriate methods and provisions should be identified in
for general r.c. cross sections for design safety checks; (c) development Task MIX.
of refined methods for assessment of generalized moment-curvature A different remark is needed for the influence of infills on the structur-
relationships of cross sections; (d) extension of results to r.c. members al response, where the motivation for the research carried out by Task
reinforced with FRP materials. TAMP stems from the absence of code provisions to account for the
meaningful interactions that develop between infills and structure, with
1.9 PREFAB: Behavior and Strengthening of Prefabricated Industrial significant effects, both, at the global level (behavior factors), and at the
Structures local level (collapse mechanisms induced by the presence of localized
The assessment and reduction of seismic vulnerability of a widespread forces).
category of precast structures typically used for industrial buildings is The following three Tasks SCALE, NODI and BIAX, related to stairs,
a topic of high importance. The production of these structures starts beam-column joints, and biaxial behavior of columns, respectively, deal
since from the years 50s of last century with elements and construction with three aspects where the need of providing the designers with oper-
solutions which had a relevant evolution through the subsequent times. ational tools is imperative, especially for as regards the assessment of
It is a social important interest to know the state of this wide building the capacity of such elements. In the first case, the motivation is to
heritage with respect to its seismic vulnerability so to address, follow- obtain a deeper insight about the influence of stiffening elements the
ing rational criteria, possible interventions of upgrading of inadequate stairs on the structural response. Generally, when modeling the resist-
structures. ing system, these elements are either neglected or modeled with unac-
ceptable simplifications. In the second case, that relative to beam-col-
2. BACKGROUND AND MOTIVATION umn joints, it is necessary to develop more accurate capacity models,
accounting for the joint panel behavior, but also for the presence of sec-
The essential motivation for each Task stems from recognizing some ondary phenomena significantly modifying the resisting mechanism,

8 RESEARCH - Seismic behavior


such as concentrated forces ensuing from hook-bent bars, or bond-slip rials strengths.
in rebars. In the third case, that of biaxial behavior of columns, the Difference in the knowledge procedure about the single structural para-
motivation for research stems from the awareness that the capacity meters and the actual possibility of propagation to the structure as a
equations currently available in the code are calibrated on the mono- whole of information gathered on single members unlikely can be
axial behavior, besides, without interaction with shear. accounted for by a single CF to be applied to mean materials strength
Finally, for the prefabricated structures studied in Task PREFAB, the values.
intention is to provide the normative framework with more complete Material strength is characterized by, both, an intrinsic spatial variabil-
indications than those currently available, with the objective of bridg- ity and an epistemic uncertainty, caused by either workmanship (for
ing the current information gap through the proposition of specific instance not compliance with the original project, execution of struc-
guidelines for the seismic assessment and strengthening of such build- tural elements in different times with different materials strength), or
ings. reliability of testing methods, or degradation of material properties with
The above considerations are expanded in the following sections. time, or a combination of the former. On the other hand, amount and
detailing of reinforcement, defective detailing, etc., neglecting the
2.1 MND: Non-Destructive Methods for the Knowledge of Existing Structures intrinsic uncertainties, are characterized by epistemic uncertainties
Modern seismic codes require that a knowledge level (KL) is defined only, mainly due to lack of the original project and/or not compliance
(e.g. 3 KLs in EC8 part 3: limited, normal and full knowledge) in order with it; collected data on one structural element are certain but do not
to choose the admissible type of analysis and the appropriate confi- allow to eliminate uncertainties about other elements.
dence factor values in the evaluation. Among the factors determining Objectives of recent studies (Franchin et al. 2008, Jalayer et al. 2007)
the KL, there are the mechanical properties of the structural materials. have been the evaluation of the effect of CF on the assessment of the
In RC structures, the compressive strength of concrete has a crucial role structural reliability and new proposals for calibration of a CF.
on the seismic performance and is usually difficult and expensive to
estimate. Reliable procedures to take into account the factors influenc- 2.3 IRREG: Assessment of the Nonlinear Behavior of Buildings, with
ing the estimation of in-situ concrete strength, particularly in case of Emphasis on Irregular Ones
poor quality concrete, are not currently available. According to various Many of the existing RC structures were built without accounting for
codes (e.g. in Europe EC8-3, in Italy NTC 2008) estimation of the in- seismic actions, thus much attention has been paid in recent years to
situ strength has to be mainly based on cores drilled from the structure. the development of reliable methods of analysis and assessment. Linear
However, non-destructive tests (NDTs) can effectively supplement cor- methods seem inappropriate in most cases; many current seismic codes
ing thus permitting more economical and representative evaluation of and guidelines include provisions for nonlinear analysis (Eurocode 8,
the concrete properties throughout the whole structure under examina- 2003a, EuroCode 8, 2003b, FEMA 356, 2000, ATC-40, 1996), which
tion. The critical step is to establish reliable relationships between NDT seems to be the natural choice for existing structures subjected to mod-
results and concrete strength. The approach suggested in most codes erate and strong design earthquakes. This is obviously a big issue in
(e.g. in EC8-3) is to correlate the results of in-situ NDTs carried out at Italy, a seismically active country where many buildings were erected
selected locations with the strength of corresponding cores. Thus, NDTs in the 60s, 70s and 80s usually accounting for only gravitational
can strongly reduce the total amount of coring needed to evaluate the actions. Furthermore, the new seismic zonation classifies areas previ-
concrete strength in an entire structure. ously considered non-seismic as seismic, thus new assessment are
needed even on recently built structures.
2.2 FC: Calibration of Confidence Factors Following the publication of the most recent Italian Seismic Codes, the
Data collected for the assessment of a building are obtained from avail- ReLUIS program of the Italian Department of Civil Protection intends
able drawings, specifications, and other documents for the existing con- to validate and improve the new code, to propose alternate procedures
struction, and must be supplemented and verified by on-site investiga- when deemed necessary, and to provide practical examples to practic-
tions, including destructive and nondestructive examination and testing ing engineers. These activities are particularly important for new
of building materials and components. methodologies, such as nonlinear methods of analysis. Focus of these
As a function of the completeness of as-built information on buildings studies is not only the application of the nonlinear methods of analysis,
(Knowledge Level) the Italian Code specifies different analysis methods but also the use of the results of the nonlinear analyses to assess the
and Confidence Factors (CF) to be applied to the mean values of mate- seismic vulnerability of structure.

RESEARCH - Seismic behavior 9


2.4 MIX: Assessment and Strengthening of Mixed-type (Masonry/RC)
Buildings
From the early 20th-century the combined RC-masonry buildings wide-
ly spread in European, Mediterranean and Southern America countries.
Despite the diffusion of this combined building typology, the interna-
tional guidelines have not followed building evolutions; nowadays,
international guidelines are not exhaustive to deal with specific prob- a b
lems of this building typology, such as: horizontal loads repartition, con- Fig. 1- a) Main published papers ; b) Ratios w/d proposed by different authors.

nections between different technology elements and over strength fac-


tor. mechanical properties of the two sub-systems (frame and infill). In the
The Argentinean guideline (NAA-80) points out the fundamental role figure, w denotes the height of the section and d is the diagonal length
performed by slab, on the base of the own relative stiffness, for sharing of the infilled mesh (Refs. 3-7). Further discordant results can be found
seismic action between vertical different technology resistant elements. considering models including the post-elastic hysteretic behaviour.
During the years, the Italian guidelines have provided discordant indi- Therefore, the main motivation of the present study lies in the non-
cations. availability of an univocal approach, able to define an appropriate infill
The Italian guideline (D.M. 1996) suggested to assign the total seismic model depending on the properties of the masonry utilized.
action to masonry walls in the case of new buildings, while for existing
buildings, the combined RC-masonry buildings should be considered 2.6 SCALE: Behavior and Strengthening of Stairs
as structural elements typology that prevalently supports horizontal In general the presence of a stair creates a discontinuity in a regular
loads, generally masonry walls. reinforced concrete skeleton frame made of beams and columns; in fact,
Regarding masonry buildings, the Italian guideline (O.P.C.M. 3431) from the geometrical point of view, a stair is composed by inclined ele-
allows to employ different technology elements to support gravity loads, ments (beams and slabs) and by short (squat) columns. These elements
only if the seismic action is fully supported by elements with the same contribute to increase the stiffness of the stair due to the elastic behav-
technology. In the need to consider the collaboration of masonry walls iour of inclined elements and of squat columns. For these reasons the
and different technology systems to sustain the seismic action, a non- elements that constitutes the stair are often characterized by a high seis-
linear analysis should be carried out according to O.P.C.M. 3431. The mic demand: the squat columns are subjected to high shear force that
latest Italian code (D.M. 2008) confirms the instructions provided by can lead to a premature brittle failure; the inclined beams, differently
O.P.C.M. 3431 by which the real structural system should be consid- from the horizontal beam, are defined by high variation in axial forces
ered with particular attention to, both, stiffness and strength of the that can modify the resistance and deformability of all these elements.
slabs, and the connections effectiveness between the structural ele- Although this is well known, no studies have been conducted by
ments. researchers to evidence the role of stairs on the seismic capacity of
existing RC buildings; the identification of the weakest elements of the
2.5 TAMP: Influence of Infills on Structural Response structure and the failure type considering the presence of the stairs are
The very numerous papers that concern the behaviour of infilled frames of particular interest. In this way, the knowledge of structural solutions
are quite uniformly distributed within the last forty years (Figure 1a). and design practice of stairs is an important step in order to define their
This subject has kept topical mainly because of the following reasons: real geometric definitions and to understand their seismic perfor-
- different materials that can be utilized for the infill panels; - difficul- mances.
ty in modelling the frame-infill interactions; - high number of parame-
ters governing the lateral response of an infilled mesh of frame. 2.7 NODI: Behavior and Strengthening of Beam-Column Joints
It follows that the models that have been proposed by different Observation of the damage caused by strong earthquakes on RC build-
researchers are strongly related to the kind of masonry infills that were ings designed to resist only to gravity loads showed that the main mech-
examined and to the experimental tests validating the models them- anisms that characterize structural collapses are, beyond the yielding of
selves. As regards this, Figure 1b shows how the section of the equiva- primary elements such as column and beams, slippage of longitudinal
lent diagonal strut by different authors is differently related to the same bars in columns and beams and joint failures (e.g. Braga et al. 2001).
synthetic parameter lh, which depends on the geometrical and Based on these observations and on the results of extensive experimen-

10 RESEARCH - Seismic behavior


tal campaigns, some provisions were inserted in the Italian technical rational basis of the research and above all the actual usefulness of its
regulations imposing performance criteria for the design of new RC results.
structures placed in seismic zones. The capacity design approach pro-
vided by current Italian and European codes (NTC 2008, CEN 2004) 2.9 PREFAB: Behavior and Strengthening of Prefabricated Industrial
aims at preventing brittle failure mechanisms in beam, column and Structures
joint members as well as at ensuring a weak beam-strong column glob- Precast structures passed through the check of weak and strong earth-
al collapse, being more favourable in terms of overall ductility. For quakes and have been submitted to a wide specific experimental and
existing RC structures, designed without anti-seismic criteria, there is theoretical investigation performed in the main international research
the problem of a reliable assessment of their seismic resistance also in centres. From these experiences some key aspects turned out to be
order to identify the more appropriate strengthening intervention sys- determinant for the good seismic behaviour of precast structures. These
tems. Improving knowledge on capacity models, particularly as for typ- key aspects are listed below:
ical Italian building structures, is the principal thrust for the research - dry friction supports, not suitable to avoid the loss of bearing;
activity of task 7 (NODI) in the framework of DPC-ReLUIS 2005-2008 - diaphragm action, important to avoid joint distortions;
Project. - lateral supports, necessary to avoid the overturning of beams;
The workflow in terms of literature review, experimental testing and - 2nd order effects, to be considered to avoid early collapses;
numerical analysis performed by the RUs involved in the task is, then, A positive condition of the existing buildings of concern is the possible
finalized to the analysis and validation of the provisions of Italian and presence of a bridge-crane which required a structural design with rel-
European codes in order to improve them and make them more adher- evant horizontal forces and a proportioning of the columns which could
ent to the reality of the Italian existing RC building stock. be adequate also for seismic action even in the presumption of low duc-
tility.
2.8 BIAX: Behavior and Strengthening of Columns under Combined The regulation in force for the design of structures in seismic zones at
Axial Load and Biaxial Bending and Shear the time of construction is obviously a conditioning aspect which affects
Modeling of reinforced concrete members is really a traditional topic of the seismic capacity of existing buildings. Actions and rules for design
structural engineering, but some aspects need further development have been taken from that regulation which may result inadequate on
when seismic assessment of existing constructions is concerned. In fact, the base of the today knowledge. The problem concerns the seismic
well-established results for modern concrete structures do not cover a zoning on one hand and the design criteria on the other.
large population of members built with obsolete materials and structur-
al details like smooth bars. This is actually a relevant issue, since bond 3. RESEARCH STRUCTURE
between steel bars and the surrounding concrete is poor and anchoring
mechanical devices can play a relevant role in the development of plas- In the following sections, the objectives pursued in each Task are
tic deformation and therefore of the drift capacity. described.
A number of models characterized by different models of complexity
can be found in the National and International technical literature (fib, 3.1 MND: Non-Destructive Methods for the Knowledge of Existing Structures
2003; Panagiotakos and Fardis, 2001; Park and Paulay, 1975) and pro- Research has been mainly focused on the evaluation of the role of the
vide an estimation of the rotation capacity at yielding and at failure of main factors affecting the estimation of the in-situ concrete strength
columns member. However, they generally are able to well represent through destructive and non-destructive tests, on the determination of
response of r.c. members where deformed bars are used. Based on such the design concrete strength, on the evaluation of the possible damage
a background, the research on columns subjected to biaxial bending on core specimens due to drilling and, finally, on the load bearing
and axial force has been conceived to cover the lack of knowledge at the capacity of structural members subjected to drilling before and after
time of proposal. In fact, advances in seismic Codes and increasing restoration interventions. Other activities were mainly devoted to ana-
need of data for design purposes can be addressed among the primary lyze the correlations between the various methods and the possible spa-
motivations of Task BIAX. On the other hand, since tools for the esti- tial variability of concrete properties throughout the surveyed members.
mation of strength and deformation of bare cross sections were not so Regarding data collection, a large amount of experimental data from
consolidated, a specific focus on columns strengthened with FRP mate- destructive and non destructive in-situ investigations on real strucures
rials is certainly of applicative interest. This circumstances confirm the were collected and analysed, also through treatment of uncertain vari-

RESEARCH - Seismic behavior 11


ables with different mathematical nature. In addition to the above top-
ics, the seismic behavior assessment of buildings with structure com-
posed of unidirectional RC frames was carried out by means of non-
destructive in situ tests, with the objective of estimating their horizon-
tal load-carrying and dissipative capacity. Finally, another important
objective was the evaluation of the dispersion of experimental results
from non-destructive measurements based on a critical review of data
reported in the literature.

3.2 FC: Calibration of Confidence Factors


This Task had two objectives. The first one was to propose a methodol-
ogy for the calibration of the CF for materials strength, taking into
account the uncertainties characterizing existing building and the
effects on the reliability of the assessed structural performance. A
methodology was also sought for the evaluation of material strength by
destructive and non destructive in situ testing methods taking into
Fig. 2- Tested buildings.
account the relevant reliability. The procedures were meant to be based
on the application of the Bayesian method. The proposed methodology the nonlinear methods of analysis of the European seismic codes: a
and the equation developed for FC have been validated on several sim- description of the main sources of nonlinearities in existing RC build-
ulated cases and on tests made on several buildings. The second objec- ings: the application of different modeling techniques to the seismic
tive was to develop a probabilistic methodology for seismic assessment
vulnerability assessment of the three building mentioned above. The
of existing buildings taking into account explicitly the uncertainties in
document is intended to be a primer for practicing engineers who want
the material properties and the structural detailing parameters and
to use nonlinear methods of analysis.
implementing the available test and inspection results. This methodol-
ogy may be used for determination of confidence factors.
3.4 MIX: Assessment and Strengthening of Mixed-type (Masonry/RC)
Buildings
3.3 IRREG: Assessment of the Nonlinear Behavior of Buildings, with
With reference to the first goal, technical literature and international
Emphasis on Irregular Ones
guidelines have been studied in order to define the classification of the
The main objectives are:
main geometrical characteristics of such kind of buildings and the study
Validation of available modelling alternatives for RC buildings, mainly
of the response of mixed-type buildings behaving as parallel systems
lumped-plasticity and distributed plasticity models, both in commercial
and research software. subjected to horizontal forces.
Validation of current methods of analysis for the seismic assessment of With reference to the second goal, a series of numerical analyses based
existing RC buildings, with emphasis on nonlinear methods and their on different and progressively refined modelling assumptions have been
applicability to plan-irregular buildings. performed in order to investigate the seismic action distribution bet-
The above validations were carried out through the analysis of several ween different technology elements, changing the size and then the
buildings selected by the different research units. Three buildings stiffness of the RC elements, but retaining the geometry of the building
(shown in Figure 2) were selected as common tested structures: one is and comparing the seismic behaviour of the mixed-type building with
a doubly symmetric rectangular building, one is an L-shaped building, the original masonry one. Pushover analysis have been performed by
and the third is a rectangular building with an internal court. These using a lumped model to evidence critical zones and possible failures.
buildings are representative of the structural buildings commonly found
in Italy. Several commercial and research programs were used for the 3.5 TAMP: Influence of Infills on Structural Response
nonlinear analyses, including SAP2000, OpenSees, Midas, etc. A first experimental investigation was devoted to determining the
The final objective of this task is the compilation of a document that mechanical properties of three typical kinds of resisting elements, com-
contains: an introduction to nonlinear modeling of RC buildings and to monly used for infill masonry, and of the mortar utilized for their assem-

12 RESEARCH - Seismic behavior


bly. Then, several infill samples were subjected to compressive tests by on different and progressively refined modelling assumptions and crite-
assuming orthogonal or parallel loading directions with respect to the ria has been performed in order to investigate the principal failure
mortar layers. Further results were obtained under diagonal compres- modes. A critical study has been conducted on the different shear
sive loading, to determine shear modulus and resistance. At the end of strength formulations present in literature (Biskinis et al.2004; Sezen et
this phase, the experimental values of elastic moduli and resistances al. 2004; Zhu et al., 2007), in order to simulate potential shear failure
were compared with the values that Italian M.D. 87 provides by link- in squat columns, which can be easily found in most buildings. The
ing the mechanical properties of masonry elements to those of their pushover analysis by using a lumped model has been performed to evi-
components. dence critical zones and possible failures.
A second phase of the experimental research was devoted to acquiring With reference to the third point, a test set-up has been defined in order
the response of square infilled meshes of RC frames subjected to a to investigate the experimental behaviour of a building sub-assem-
cyclically varying lateral forces. Two 1:2 scaled samples were tested for blages, including a stair substructure.
each of the three kinds of infill that had been mechanically character-
ized previously. The results of these tests have made it possible to cal- 3.7 NODI: Behavior and Strengthening of Beam-Column Joints
ibrate the hysteretic model of pin-jointed diagonal strut proposed in A wide experimental campaign on beam-column joints representative
Cavaleri et al., 2005. of typical members present in Italian existing buildings was planned,
A further experimental investigation was carried out by means of shak- designed and carried out. In particular, the research activities were
ing table tests on a 1:2 scaled 3D infilled RC frame, reproducing an devoted to outline the influence of some parameters on the mechanical
actual non-infilled building, previously subjected to pseudo-dynamic behaviour and the failure mechanism of the joints, such as axial force,
tests at the ELSA-JRC-Ispra. These tests had the following objectives: amount of reinforcing steel and earthquake design level. Furthermore,
- to quantify the lateral stiffness and resistance contributions that infills the research focused on the code expressions for the evaluation of the
can provide; - to verify the influence of the infills on the crack distrib- ultimate rotation of RC elements in order to highlight possible discrep-
ution and the collapse mechanism. The same objectives were pursued ancies between the theoretical and experimental results. Another
by nonlinear numerical analyses on multi-storey RC frames subjected research objective was the analytical modelling of beam-column joints
to natural seismic accelerograms. These analyses also showed the neg- by using DIANA software to analyze the main parameters affecting their
ative effects of non-uniform infill distribution along the height. seismic performance and, specifically, the analytical modelling of
Another experimental campaign on infilled r.c. frames (1:1/2 scale), on experimental tests conducted on external joints. Other experimental
materials (concrete, steel, blocks and mortar), and on subassemblages tests on beam-column joints relevant to existing buildings were per-
(small panels) has been performed. These tests had the objective of cal- formed as well, following an experimental program complementary to
ibrating equivalent strut models, through comparison of experimental the above-mentioned one, that is, in this case specimens reinforced with
results on bare and infilled frames, in order to evaluate the infill con- smooth bars were tested and, in some cases, after the first test, joints
tribution as well as its uniaxial force-displacement relationship. The were retrofitted to evaluate the effectiveness of some retrofit systems.
constitutive models for masonry infills have been also calibrated in Finally, supplemental activities followed two main branches: on one
order to predict the cyclic response of infilled frames. hand, a series of either reinforced or unreinforced base joints were test-
ed thus evaluating the performance of several strengthening systems,
3.6 SCALE: Behavior and Strengthening of Stairs and, on the other hand, a wide database of tests on beam-column joints
With reference to the first goal, several available manuals and books at was built and analyzed.
the time of construction have been studied in order to define the typol-
ogy classification and the corresponding evolution of this classification 3.8 BIAX: Behavior and Strengthening of Columns under Combined
during the years with the increasing knowledge on the use of the mate- Axial Load and Biaxial Bending and Shear
rials and of computational machines. An analysis of the codes used The activities have been planned to cover four main objectives: (1)
from 1909 to the 1980ies has been conducted in details with a critical review of technical literature with specific reference to available exper-
judgement based on the actual knowledge. Examples of stairs designed imental data; (2) development of refined and simplified models for bare
for only gravitational loads have been studied with reference to differ- and FRP reinforced members; (3) experimental activity on columns
ent typologies. subjected to cyclic actions; (4) drafting of a technical report summariz-
With reference to the second point a series of numerical analyses based ing the main applicative aspects of the work.

RESEARCH - Seismic behavior 13


3.9 PREFAB: Behavior and Strengthening of Prefabricated Industrial
Structures
The study envisages a preliminary classification of the industrial prefab-
ricated building typologies existing in Italy, from which the most frequent
characteristics of element-to-element connection types will emerge. This
first cataloguing phase is then followed by a purely experimental phase in
which some connections, identified as more vulnerable (e.g., friction con-
nections), are subjected to a series of cyclic tests to simulate seismic con-
a
ditions. Results and information obtained from the experimental tests will
serve as a basis to develop practical models for assessing the capacity of
such connection zones and to orient towards the definition of criteria and
techniques for strengthening interventions.
The main results obtained by each Task are summarized in the follow-
ing sections.

3.10 MND: Non-Destructive Methods for the Knowledge of Existing Structures b


The results obtained during the Project are mainly made up by the exe-
cution and analysis of experimental investigations either on in situ real Fig. 3- Role of past applied loads on in-situ measurements: (a) qualitative bending moment due to vertical loads (a), and (b)
test results along the lower part of the extracted beam (rebound number S, direct velocity V, surface velocity Vs, core strength
structures or on laboratory specimens, by the implementation of some fcore).

procedures to estimate the in situ concrete strength and by the uncer-


tainty treatment of the structural characteristics of existing strucures. tional to the core strength as provided by the compression test. Finally,
A wide experimental program was carried out, comprising more than 20 some important results regarding the effect of core drilling on the struc-
RC beam and column members, several hundreds of non destructive tural members, performing tests before and after a possible restoration.
tests (NDTs) and more than 50 destructive tests (cores). Analysis of Further, some factors influencing the relationship between the local
results has shown a large scatter of the core concrete strength both in a strength provided by core specimens and the in situ strength of the
single member and among members extracted by the same story of a structural member as a whole, have been highlighted.
building. Lower scatters have been detected for the NDT results with Regarding the variability of concrete mechanical properties throughout
the exception of the surface ultrasonic velocity (see Figure 3). As a single structural members and among different sampling locations,
result of these findings, the role of some factors influencing the in situ investigations based on a wide series of experimental data gathered
concrete properties has been carefully evaluated, and some criteria to from surveys carried out on structures assessed for seismic vulnerabil-
suitably select locations for sampling have been provided. A procedure ity were carried out.
for the evaluation of the concrete strength based on the Sonreb method, Main results obtained are briefly outlined below:
using both core and NDT measurements, has been set up and widely no general trends have been recognized regarding the spatial vari-
validated, clearly showing its higher prediction capacity when com- ability of the key mechanical properties of concrete throughout column
pared to the relationships currently available in the technical literature. members as a possible result of the combination of the effects of the
It requires that the relationship between the in situ concrete strength load pattern and the casting process;
and the NDT measurements is experimentally derived for the specific although carried out on members already cracked and damaged, the
concrete under test. results of sonic tests are affected by scatters smaller than those deriv-
As for the possible damage on core specimens due to drilling, the ing by the compression tests on concrete samples; further, the ratio
results have shown that the strength reduction suffered by cores can be between ultrasonic velocity derived by indirect and direct measures are
significantly influenced by the original strength value of the in situ con- normally distributed around the average value of 0.75;
crete. Consequently, adopting a constant coefficient to take into account rebound tests have confirmed the substantial impossibility of recog-
drilling damage, as suggested in the technical literature, can determine nizing general trends in the spatial variability of the mechanical prop-
incorrect results. On the contrary, it appears suitable adopting coeffi- erties of concrete and led to values affected by scatters quite similar to
cient values, obtained during the research, which are inversely propor- the destructive ones.

14 RESEARCH - Seismic behavior


Regarding the treatment of uncertainties in determining the character- CF for a Limited knowledge (CF=1.35).
istics of materials and more generally of existing building parameters, a For each value of the CF the demand/capacity ratio has been evaluated
fuzzy-logic based approach for uncertainty treatment has been set up and positioned in the structural performance distribution, for the three
and a computer code for its implementation has been developed. knowledge levels, with reference to the structure at hand. The CF value
In addition, the prediction capability of some formulations provided by considered as the exact one has been defined as the value correspond-
the current technical literature was verified based on experimental ing to the demand/capacity ratio value with 5% exceedance probabili-
investigations through non-destructive and destructive tests on existing ty. In this way, the CF for material strength is evaluated on the basis of
structures. the probabilistic evaluation of structural performance taking into
Based on a a critical review of available literature, a database was pre- account all the intrinsic and epistemic uncertainties, including uncer-
pared that collects literature data on non-destructive tests on concrete tainties on the testing methods.
specimens for concrete grade assessment. Dispersion of experimental A simplified method is proposed too, based on a limited number of
results has been estimated and the influence of uncertainties coming Monte Carlo simulations, which is able to approximate the probability
from the concrete grade estimation on seismic capacity of RC existing distribution of the structural parameter. This can be a basis for the
buildings has been investigated. development of simple procedure to use in for evaluation of structural
Some results obtained during the project have been reported in papers safety.
published on journals and in proceedings of Conferences (e.g. Masi and A second objective was to develop a procedure for evaluation of mater-
Vona, 2008; Marano et al., 2008; Olivito et al., 2008). ial strength and calibration of CF based on the application of Bayesian
method, to take into account the number and the reliability of the in-
3.11 FC: Calibration of Confidence Factors situ tests carried out.
A first aim of the task has been the evaluation of the confidence level The Bayesian method allows to employ destructive and non-destructive
on structural safety of existing buildings given by seismic structural testing results to update a prior probability distribution function.
assessment carried out according to the indications of the Italian OPCM Destructive and non-destructive testing results are separately
3431, 03-05-05, Annex 2. employed, taking into account individual testing reliability (reliability
Uncertainties in reliability structural analysis are due to material prop- due to testing errors and errors in regression curve that provides the
erties, structural details and condition of the structural elements. The material resistance as a function of the testing parameter). More than
prior distribution of the considered uncertainties takes into account one test method can be employed performing consecutive up-dating of
their mechanical effects. The proposed probabilistic models are subse- the probability distribution function.
quently updated by in-situ information. The statistics reliability of the mean value is improved by applying the
A parameter describing the structural performance is defined as the confidence interval for the mean; a 95% lower confidence level is con-
demand/capacity ratio and its probability distribution is assessed by a sidered, which represents the value for the structural assessment.
Monte Carlo simulation. Each realization corresponds to an application In order to facilitate its evaluation, a simplified procedure is defined.
of the capacity spectrum method and needs the execution of a structur- The material strength value for structural assessment can be obtained
al linear static analysis. The Bayesian up-dating of the structural relia- scaling with an appropriate Confidence Factor a weighted mean of the
bility is carried out by a Markov Chain Monte Carlo algorithm. The sampling mean values obtained by, both, different testing methods and
structural performance prior probability distribution function is evalu- prior information:
m
ated in two different cases: 1) taking into account the uncertainties mD = FC m'''inf,mf (1)
about material properties and structural details; 2) updating the struc-
where:
tural assessment based on in-situ tests and inspections.
m' +n x +n x
The updating process consists of two different levels: in the first,
destructive tests and relating errors are taken into account; in the sec-
[
m = f DM DM NDM NDM
1+nDM+nNDM ] (2)
ond, non-destructive tests and relating errors are taken into account.
Structural seismic performance has been evaluated in three cases: a) by where xDM and xNDM are the sampling mean of the destructive and non
using the mean material strength values (CF=1); b) by using the mean destructive tests, respectively; nDM and nNDM are the corresponding
material strength values scaled by the CF for a Normal knowledge sampling dimension.
(CF=1.2); c) by using the mean material strength values scaled by the Generally, if Mi is the i-th testing method adopted, the material strength

RESEARCH - Seismic behavior 15


for the assessment is: ses were often found un-conservative in terms of interstorey drifts or
m' + n x chord rotations;
[
m = f i Mi Mi
1+inMi ] (3) A new pushover method that explicitly takes into account the tor-
sional behaviour of asymmetric-plan buildings was defined. The effec-
where xMi is the sampling mean of the i-th testing method and nMi its tiveness of the proposed procedure was evaluated by comparing the
dimension. seismic demand of selected case studies with that obtained through
The CF can be expressed in an explicit form as a function of the both nonlinear dynamic analyses and other pushover methods;
Bayesian coefficient of variation Vm for the median value of the mater- The nonlinear analyses on the regular and irregular buildings have
ial strength: shown the importance of damping in nonlinear dynamic analysis. More
FC = a+cVwm (4) specifically, as the hysteretic model improves, the damping should be
The parameter Vm, which estimates the reliability of the available infor- decreased. Furthermore, viscous damping is hard to assign and little or
mation, is defined as: no indications are given in the published literature to guide the user. A
n value of 2-3% damping for the elastic modes appear reasonable, but no
sm
Vm = m =
m
 i 2 Mi 2
ss,Mi+st,Mi
n x (5)
final indications were found;
In NTHA, for both natural and generated accelerograms, the applica-
i 2 Mi M2 i tion of the seismic input in the principal directions of the structure may
ss,Mi+st,Mi
underestimate the demand, the structural demand varies considerably
where s2s,Mi and s2t,Mi are the sampling variance and the variance of the as the seismic input direction changes, more so for natural accelero-
regression curve of the i-th testing method, respectively. grams. However, as for the number of input ground motions to use in
The Eq. (4) has been calibrated for concrete strength by applying the nonlinear dynamic analyses, enhancements to the EC8 requirements
least squares method. were proposed;
For irregular buildings, pushovers applied in different directions indi-
A Monte Carlo method has been used to simulate sampling with
cated demands and capacities that depend on the direction considered.
destructive and non destructive testing and the resulting equation for
Partial results on the L-shaped building are shown in Figure 4.
the CF is:
Shear collapse in existing buildings not designed according to the
FC = 0.9 + Vm (6) capacity design rules, is often dominant. If only the flexural capacity is
modelled, most frames are verified both at the Ultimate and at the Near
The Eq. (6) is effective if samples have been extracted from homoge-
Collapse Limit State. However, additional checks show that the shear
neous zones of the structure. If in the structure potential non homoge-
capacity has already been reached on several elements prior to reach-
neous zones are identified, the t-Student test can be executed on the
ing the design accelerations. This indicates the need to use models that
mean values extracted from the two zones. If the t-student test identify
consider shear failure too (this is very rare in the available software).
non homogeneous zones these must be separately evaluated consider-
Furthermore, early shear failures point to possible retrofitting actions,
ing two different median values for concrete strength with two CF.
A method has also been investigated for the evaluation of reliability of
the correlation function for the assessment of material strength by in
situ tests.

3.12 IRREG: Assessment of the Nonlinear Behavior of Buildings, with


Emphasis on Irregular Ones
The main results of the project are as follows:
Regarding the validation of the modified pushover procedure pro-
posed by Fajfar (2002), it was found that it is conservative for plan-reg-
ular framed structures with respect to NTHA, while it is unreliable for
shear wall structures. For irregular frames, multi-modal procedures
have to be used in order to improve the accuracy of static non linear
Fig. 4- Pushover analyses on L-building: Influence of seismic input angle on the demand: displacements (left) and rotation
analyses. Furthermore, predictions drawn from non linear static analy- ratios (right).

16 RESEARCH - Seismic behavior


such as shear strengthening of beams and columns; Table 1. Literature review on mixed-type buildings
Codes National International Total
Synthetic expressions were developed that express the plan-regular-
7 15 22
ity of a building with respect to stiffness and strength distributions. Damages after earthquake1 National International Total
Also, simple corrective coefficients were developed to compute the 16 63 79
structural demand increase as the structural eccentricity increases. In Mixed-type building2 Experimental Analytical studies Total
7 1 8
alternative, a simplified procedure was developed that can help opti- Confined masonry3 Experimental Analytical studies4 Total
mize the structural design by producing a plan-regular building; 27 24 51
Severe convergence problems were encountered in commercial codes 1 Reports and other papers enclosing field observations after earthquakes.
that use lumped plasticity models. These problems were related to the 2 Building in which elements of reinforced concrete and masonry are not bonded.
3 Building in which bearing masonry walls are confined by reinforced concrete elements.
software limitations rather than the modelling selection. As expected, 4 Analytical studies, including also proposal for design and for the modelling of confined

fibre section models provide a better, more physically-based prediction masonry structures.

of the section response. It is however important to extend the model to


include shear failure, in order to avoid post-processing of the results; The main conclusions inferred by the study can be summarized as fol-
Different Limit State definitions for performance assessment were lows. In many codes the confined masonry structures are taken into
considered, such as interstorey drift, plastic rotation and chord rotation. consideration, even though usually just empirical dimensional rules are
It was shown that different measures lead to different, sometimes very given, while methods for assessment and design are lacking and, when
different, capacity predictions. The limits usually assumed for inter- present, they differ substantially from a code to another. Concerning the
storey drift result in larger chord and plastic rotation than limits pro- mixed structures (unconfined masonry), only a small number of codes
posed by the European and American Codes. Furthermore, the proce- gives some indications, which are usually about the behaviour factor
dure provided by Eurocode 8 to compute the chord rotation capacity and/or the distribution of horizontal forces among different structural
yields predictions that are very different from those obtained analyti- elements. A similar trend was found also when gathering papers on field
cally, that is integrating the member curvature throughout the plastic observations, experimental tests and analytical studies: a lot of data
hinge length; were found concerning confined masonry structures while just a few
A simplified method for deriving fragility curves and evaluating the were gained about other kinds of mixed structures. In general it is pos-
probability of failure was proposed. The method is based on an incre- sible to state that confined masonry can withstand to seismic actions,
mental application of the so called N2-Method with natural response given that materials are of good quality and good constructions rules are
spectra, whose irregularity covers the record-to-record variability of the followed. On the contrary, some deficiency, such as small amount of
structural response without the need for performing non-linear dynam- transverse steel in columns, high thickness of mortar joint, lack of RC
ic analysis. element near the openings, can generate the bed behaviour of this kind
of structures. Concerning the mixed structures (other than confined
3.13 MIX: Assessment and Strengthening of Mixed-type (Masonry/RC) masonry), the main problems which arise in the assessment and design
Buildings are related to the distribution in plan of the horizontal actions, to the
The main activities of the first goal have involved both the literature determination of the behaviour factor and to the study of the connec-
review of mixed-type buildings and the modelling aspects related to tions between masonry and reinforced concrete elements. In literature,
the RC substructure of such structures. A report was produced indications about these points are few and further studies are needed.
(Decanini et al., 2006) including the main results obtained in the With reference to the second goal, the main problems concerning the
Task. non-linear analyses on RC-masonry mixed buildings have been high-
The work involved the study of 160 papers, as indicated in Table 1. lighted. Particular attention has been devoted to the seismic action dis-
Confined masonry is listed separately to underline the difference in tribution between different technology elements. The results of the
the amount of studies between those constructions and structures in analysis carried out on 3D RC-masonry mixed building (see Figure 5)
which elements of different technologies (masonry/reinforced con- with external walls and internal frames, underline the fundamental
crete) are not bonded. The latter case includes constructions in which masonry role to withstand horizontal action, while a significant transla-
masonry and reinforced concrete frames are present at the same level tion increase could be offered by introducing RC frames. This research
and construction in which reinforced concrete frames are placed in has shown the growing capacity offered by mixed building to bear the
the upper floors only. seismic action by increasing the RC elements stiffness. The increasing

RESEARCH - Seismic behavior 17


related masonry samples lead to the following main conclusions: i) cal-
carenite masonry exhibits middle compressive and shear strengths, and
ductile behaviour up to the collapse for each of the three considered
loading directions; it behaves like an orthotropic material in which the
ratio between the two elastic moduli depends on the mortar properties;
ii) clay brick masonry is about twice resisting in compression, but duc-
Fig. 5- Analyzed models (Nardone et al., 2008). tile behaviour was observed only under diagonal loading, when the
shear behaviour of the mortar joints is involved; iii) lightweight concrete
RC elements stiffness is significant for the seismic action distribution masonry has very low resistance capacities along all the loading direc-
both in linear and non-linear field. From the analyses it has emerged tions; the results obtained from this kind of masonry were rather scat-
that, by increasing the stiffness of internal RC frames, the maximum tered, because vertical mortar joints are not provided for, due to the ver-
sustainable seismic action of mixed building increases, while the rate tical profile of the resisting elements. The experimental values of elas-
of seismic action supported by masonry walls decreases (Nardone et al., tic moduli and resistances showed that the values deduced by the M.D.
2008). The important slab role in sharing the seismic action between 87 provisions are not always reliable. In particular, the shear strength
the vertical resistant elements and the importance of slab to perform values are strongly underestimated by this code for all masonry tested.
this function in order to avoid undesirable behaviour of the building has
The cyclic tests on the infilled meshes of RC frames have made it pos-
been underlined.
sible to verify that the cross-section of an equivalent strut can be deter-
In addition to that, concerning the first model, the analyses highlighted
mined by using the procedure and parameters shown in Amato et al.,
that the presence of the RC frame is generally detrimental for the
2008, where a not negligible role is exerted by the transverse strain
masonry wall. In fact, the resisting total base shear of the mixed system
ratio in the diagonal direction and the compression level transmitted to
is smaller than that of the masonry wall alone when the contribution of
the columns after the masonry infills have been constructed. The later-
the frame wall is smaller than about 15-20% of the total base shear.
al resistance of the infill can be deduced from the masonry shear
Moreover, when the top connection is non effective, hinges develop at
strength; this can be translated into a compressive strength to be con-
the top and the bottom of RC columns. Therefore, it seems appropriate
ferred to the equivalent strut, by means of a suitable criterion taking
to design the RC elements considering seismic action deriving from the
into account mainly the disconnection arising between frame and infill.
pertinent vertical loads. While, if the connection between masonry and
Finally, these tests showed that the calibration of the adopted model
frames are effective, it would be expedient to assign the whole of the
leads to sufficiently approximate results (Figure 6).
horizontal loads to the masonry (Decanini et al., 2008).
The shaking table tests on the 3D scaled building (Figure 7) have been
In general, the analyses performed on the 3D model highlighted the dif-
ficulties in modelling the masonry and the connections between differ- carried out by assuming a natural input accelerogram (Herceg-Novi,
ent elements (beam-masonry, floor-masonry) and indicated the strong
influence on the global structural behaviour of the connections effec-
tiveness. The parametric analyses showed the importance of the mason-
ry tensile strength among other mechanical parameters on the global
response and highlighted that the RC elements remain elastic and give
a negligible contribution to the overall performance. Finally, the com-
parison between the mixed structures results and those obtained for the b
a
masonry structures confirmed how the practice of replacing masonry
walls with RC frames, in the interior of old masonry buildings, can have
negative consequences on the vulnerability of the building themselves.

3.14 TAMP: Influence of Infills on Structural Response


The resisting elements that were chosen to construct commonly used
c
masonry infills were: calcarenite ashlars, hollow clay brick and hollow
lightweight concrete blocks. The results of the compressive tests on the Fig. 6- Cyclic tests on infilled RC frames: a) test set-up; b) experimental results; d) validation of model.

18 RESEARCH - Seismic behavior


sions giving the desired interstory drift.
The results of this Task, synthetically presented here, are discussed in
detail in the final Report of Task TAMP. The following remarks can be
made, concerning their use and possible improvement of provisions: -
the proposed model of equivalent diagonal strut could be utilized for
analyses of existing structures designed to bear only gravitational loads;
a
- the expressions proposed by M.D. 87 would be revised, also includ-
ing values to be assigned to the transverse strain ratios; - infills would
be considered for evaluation of the period of vibration of the structure
and definition of structural regularity.

3.15 SCALE: Behavior and Strengthening of Stairs


b
According to the literature, the existing stairs can be classified into two
main categories depending on the static behaviour of the stair steps: (i)
stairs with steps performing as cantilever beam, stair type A (see Figure
Fig. 7- Results of shaking table test (PGA = 0.3 g): a) 3D building; b) crack distributions.
8a), (ii) stairs with simply supported steps stair, type B (see Figure 8b).
Generally the stair type B are used worldwide, in Europe and USA,
Montenegro 1979), scaled to three levels of PGA. These tests showed
while the stair type A, with inclined beams, are much more adopted in
that: - the fundamental period of vibration of the structure is influenced
Europe (Tecnica y Pratica del Hormigon Armado, 1989, Reynolds and
by the presence of the infills (PGA = 0.04 g); - the crack distribution
Steedman, 2002, Guerrin and Lavaur, 1971). According to the USAs
depends on the location of the infills with opening (PGA = 0.3 g); - the
manuals a great scatter in stair structural solutions can be found (Berry,
infilled structure bears a PGA value (0.54 g) that proves to be about
1999).
three times the value that had been deduced by pseudo-dynamic tests
According to the manual design criteria (Marrullier, 1910; Rosci, 1939;
on the corresponding non-infilled structure.
Santarella, 1953, 1957; Pagano, 1963; Migliacci, 1977) stair type A
With regard to the numerical investigation, two series of four-storey and could be designed considering only gravity loads, any seismic actions
twelve-storey RC frames have been considered. Frames of the first could not be taken into consideration. The permanent and live loads on
series were designed to bear only gravitational loads; the ones of the the steps generate on the beam element (bs1-bs2-bs3) a torsional
second series were designed according to the EC8 provisions. The moment T and a distributed load producing on the beam shear force V
results lead to the following remarks: - the presence of infills implies and bending moment M. Each flight step is designed modelling it as a
greater quantity of input energy for the structure; nevertheless, the cantilever beam subjected to a distributed load.
infills can dissipate a lot of this energy so that the total balance is As it is explained in several dated manuals (Santarella, 1953, 1957;
favourable to the resisting elements of frame with respect to the case of Pagano, 1963), the design bending moment into the beam (bs1-bs2-
bare structure; - in the frames of the first series the infills can exert a bs3) is evaluated on the basis of different static schemes corresponding
decisive role towards a seismic event; - in the frame of the second series to different constraints at the extreme ends of the beam. In particular,
a non-uniform distribution of infill in elevation would be considered in two extreme constraint conditions are suggested: full constraint and
defining the structural regularity. simply supported. The torsional moment is considered of relevant
Finally, the studies regarding the calibration of a three-strut model importance, it leads to add transversal reinforcement (stirrups) along
arrived at the conclusion that this approach is useful for the compre- the length of the beam (bs1-bs2-bs3). The adopted values of the tor-
hension of the local behaviour of columns in case of degradation due to sional moment depend on the hypothesis upon the flexural stiffness of
infill-induced shear effects. For what concerns response in terms of dis- the inter-storey slab: flexible and rigid diaphragm.
placement, single strut models seem to give satisfactory results. Strut About stair type B, manuals indicate two limit structural schemes: (i)
parameters can be directly obtained from the mechanical characteris- an horizontal beam full constraint at the end, (ii) an horizontal beam
tics of infill components, i.e., mortar and blocks. Moreover, a design cri- simply supported at the extreme ends. Normally bending moment and
terion for a protecting dissipative bracing systems has been developed. shear are the internal forces taken into consideration. In the manuals of
The proposed approach allows to define the required minimum dimen- the construction time the only severe prescription is regarding the

RESEARCH - Seismic behavior 19


a b

Fig. 8- Stair typologies: (a) stairs type A with steps as cantilever beams, (b) stairs type B with simply supported steps. Fig. 9. Results of the push-over analysis in terms of base shear versus roof displacement (Cosenza et al., 2008).

design of the steel bars: a reinforcing bar should not bend to form angles ing representative of the studied sample; non linear static analyses (sta-
that favour pull-out of the concrete cover (Pagano, 1963). tic push over analysis) finalized to the evaluation of the role of the stairs,
The Italian stair design practice during the period 1954-1980 has been of their elements and modelling is performed.
analyzed in order to identify the most common typologies and the effec- The building without any stair is defined as reference. Two models have
tive adopted design criteria. As already remarked, according to the been considered to study stair type A with inclined beams and stair type
dated technical manuals and codes stairs could be designed for only B having reinforced concrete slab. For each structure, different model-
gravity load. The predominant stair type in the studied building sample ling have been adopted to evidence the influence on the global response
is type A; the flight steps are cantilever elements constraint to inclined of: biaxial bending modelling in the beams of the substructure stair;
beams having one point of discontinuity in the 53% of the cases, two bending moment-axial force (M-N) interaction into the inclined ele-
points of discontinuity in the 37% of the cases and is directly connect- ments (beam and slabs); bending moment-shear (M-V) interaction into
ed to the column without any discontinuity in the 5% of the cases. The the inclined elements and columns.
stair type B is present in the sample with incidence of 3%. The design In general, the presence of stair brings to an increase of lateral strength
practice of the most common type of stairs, composed by flight steps and to a reduction in displacement capacity with respect to the build-
constrained into a beam, is herein studied. The static design scheme ing without stair (Cosenza et al., 2007a). On the contrary, the results
has a great scatter; beams were designed considering a maximum have confirmed the need to utilize biaxial bending modelling and to
moment M+=qL2/a in the midspan with a=12 (30% cases) or a=8 account for the interaction of the different internal forces (Cosenza et
(30% cases), while the minimum moment at the extremes of the beam al., 2008) as: bending moment-axial force interaction that characterizes
is obtained with a=12 in most cases (76% cases). the inclined elements, and the bending moment-shear interaction that
Regarding the influence of stairs on the seismic capacity, this prelimi- governs the behaviour of squat columns. Shear failure becomes pre-
nary study on the structural typologies of the building sample has evi- dominant in the squat columns and in the reinforced concrete slabs and
denced the following problems related to the presence of stairs: distri- precedes the conventional ductile failure (see Figure 9).
bution of seismic forces (not considered in the design), different model-
ling design of stair structure, material strengths, element detailing. 3.16 NODI: Behavior and Strengthening of Beam-Column Joints
The structural typology of stairs generally introduces discontinuities Main results obtained in this Task are made up by the execution and
into the typical regular reinforced concrete skeleton, composed by analysis of the experimental program on beam-column joints without
beams and columns; in fact, the sub-structure stair is an assemblage strengthening, as well as by the execution and analysis of some tests on
of inclined elements as slabs or beams. All these elements contribute to strengthened and unstrengthened beam-column joints and on base
increase the stiffness of the stair due to the elastic behaviour of inclined joints. Numerical-experimental comparisons have been performed, as
elements and of squat columns. For these reasons the elements that well, regarding some highly representative test campaigns available in
constitute the stair are often characterized by a high seismic demand: the literature and relevant to some of the tests carried out. A careful lit-
the squat columns are subjected to high shear demand that can lead to erature review was carried out relevant to: (i) experimental programs
a premature brittle failure; the inclined beams, differently from the hor- carried out by other researchers, (ii) joint capacity models, and (iii) code
izontal beam, are defined by high variation in axial forces that can mod- provisions on beam-column joints.
ify the resistance and deformability of all these elements. A wide experimental program was carried out on full scale external
All these aspects are discussed with a series of analysis on a RC build- beam-column joints relevant to typical existing RC buildings having

20 RESEARCH - Seismic behavior


A parallel experimental program was devoted to the test (under imposed
increasing cyclic displacements applied to the beam end) of 4 external
reduced scale beam-column joints provided with smooth bars, later
retrofitted after a first series of tests. For 2 of them only the anchorage
of the beam bars had been restored by welding threaded bars to the
ends of the longitudinal beam bars and bolting them to steel plates
placed on the column external surface. For the other 2 joints, besides
a b restoring the anchorage, also vertical and horizontal carbon fiber fabrics
had been applied on the column, below and above the joint panel.
Fig. 10- Test results on a Z2 joint (design for medium seismic zone, low axial load): a) damage pattern at drift=7%, and b)
force-drift relationship. Moreover, other 2 real scale beam-column joints were built and tested
under cyclic loads, either with or without retrofitting, to evaluate the
different Earthquake Resistant Design (ERD) level. Quasi-static tests
increase of joint performance. As for the first 4 joints, the restoration of
have been performed with 3 loading cycles for each drift value gradu-
the anchorage of the beam bars proved to be very efficient, since it pro-
ally increased from 0.25% up to the total failure of the joint. Results
vided increases in strength up to 300% with respect to not retrofitted
showed yielding force equal to about 20 kN for the NE joints (gravity
joints. The additional carbon fiber reinforcements did not provide
loads only designed), and about 40 kN for both Z2 joints (designed for
noticeable increases in strength, because the cracking of the joint,
seismic zone 2, medium seismicity) and Z4 joints (designed for seismic
which would have required the carbon fiber contribution, did not occur.
zone 4, very low seismicity), as a consequence of the minimum require-
Regarding the other 2 joints, it was observed that the not-retrofitted one
ments on reinforcement amount prescribed by the Italian code.
attained low strength values due to the lack of steel reinforcement for
Observed failure mechanism in all the joints showed a wide and heavy
negative moment in the beam. As regards the joint retrofitted before
cracking in the beam, due to the small amount of the longitudinal rein-
testing, the carbon fiber fabrics applied on the beam significantly
forcing bars as the beams were not loaded by the floor slabs in the con-
increased both strength and displacement ductility.
sidered building model. Joint failure was generally caused by the ten-
Additional research activities were mainly focused on experimental
sile failure of the reinforcing bars in the beam. A different as well inter-
tests on full scale RC columns (base joints not-strengthened and
esting behaviour was displayed by some tests on Z2 joints (Masi and strengthened) tested under constant axial load and monotonic or cyclic
Santarsiero, 2008), where a wide cracking also in the joint panel and a flexure. The strengthening systems consisted in confinement by par-
softening mechanical behaviour (Figure 10) were observed, due to the tially wrapping unidirectional carbon or glass layers around the element
reduced amount of the applied axial force (=0.15). Drift value (coinci- at the base. On some specimens, in addition to confinement, steel
dent with chord rotation value) at failure in the NE joints is about 3.0%, angles (in some cases, anchored at the foundation) were placed in cor-
while in more ductile Z2 and Z4 joints, values equal to about 4.5% have respondence of the member corners. The specimens were designed
been detected. according to the Italian codes in effect during the 60s and 70s with the
As for the contribution of joint panel own deformations to the total drift aim of reproducing typical dimensions, rebar amount and details com-
of the sub-assemblage, experimental results showed that it is rather low, mon at that time. Main results achieved during the Project are as fol-
being always lower than 10% even in the heavily damaged specimens. lows: (i) regardless of the axial load value, the FRP confinement pro-
Further, interesting results have been found by comparing the strength duces significant increases in terms of ductility, especially if a GFRP
of the joint panel provided by the tests and that one obtained applying (glass) jacket is used; (ii) the arrangement of the longitudinal steel
the European (CEN, 2004) and the Italian Code (NTC, 2008) expres- angles unconnected to the foundation leads to higher ductility levels
sions to evaluate the capacity of beam-column joints. Results show a than those measured for members strengthened by only FRP systems;
good estimation ability of code expressions that have been able to pre- this ductility gain is lower for columns tested under n=0.40, even if in
dict which of the specimen was subjected to diagonal cracking of the these cases the unconnected angles also provide an improvement of the
joint panel. Regarding the ductile capacity, the difference between NE flexural strength; (iii) when the longitudinal angles are anchored to the
and seismic specimens was lower. In all the tests with high axial load a foundation the flexural strength of the RC columns significantly
failure mechanism with extensive cracking of the beam and evident increases, but a reduction of the available ductility is observed.
deterioration of the concrete at the beam-column interface has been A numerical-type activity focused on the analysis of the performance of
noted. RC beam-column joints through numerical simulations by using the

RESEARCH - Seismic behavior 21


F.E. software DIANA, validated by means of experimental test cam- force. Different approaches have been proposed by the different teams
paigns available in literature (e.g. by the Shiohara working group) as involved in the Task, comparative analyses have been carried out and
well as some of the tests performed. In particular, many non linear multiple applicative perspectives have been covered.
analyses have been performed on typical existing external beam-col- A specific software has been developed and made available at the pro-
umn joints as they can be found in real buildings built in the past. The ject website (www.reluis.it) for download. It is based on a Fortran lan-
definition of typical deficiencies found in real beam-column joints and guage procedure and takes advantage of a user-friendly Visual Basic
the analysis of the main parameters governing the structural behaviour interface and multiple language platforms. The computational engine of
of such joints allowed to highlight some effective strengthening solu- the software has been also used for the development of an extended
tions, especially for external joints. Further, such work allowed to vali- parametric analysis aimed at the evaluation of the influence of the biax-
date some theoretical models able to predict the behaviour of both ial actions not only on the ultimate strength but especially on the cross
external and internal joints, as well as to validate the expressions pro- section ultimate curvature. In particular, such influence was studied on
posed by some codes to predict the failure of the joint panel. square RC cross-section characterized by different values of axial load
Some results obtained during the project have been reported in papers and geometrical percentage of reinforcement. The study has been
published on journals and in proceedings of Conferences (e.g. Masi et developed in order to define simplified analytical formulation to easily
al., 2008). predict the ultimate curvature reduction in the case of biaxial bending
and axial load with respect to the case of uniaxial bending. Figure 11
3.17 BIAX: Behavior and Strengthening of Columns under Combined gives a view of the program interface (top right), an example of typical
Axial Load and Biaxial Bending and Shear results in terms of generalised relationship between the reduction of
The presentation of the results follows the research outline that has ultimate curvature compared with the reference uniaxial value
been above discussed. In particular, methodological and numerical (=u,biax/u,uni), normalised axial load  and the inclination of the
contributions are presented separately from the main experimental
findings. This choice is actually related to the nature of the results and
their impact on applicative aspects of structural seismic design, name-
ly codes and design and assessment practice.
In particular, the detailed review of technical literature that has carried
out during the early stages of the research pointed out the relevance of the
type of reinforcement used for construction of existing buildings. In fact,
it has been demonstrated that experimental response of r.c. members can
be affected by type of reinforcement, smooth or deformed, and by bond
interaction between steel and concrete especially in post-yielding phase.
As a consequence, a concrete effort has been devoted to perform a com-
parative analysis of inelastic performances of members depending on
type of reinforcement; this means that both numerical and experimental
activities have been calibrated in order to cover at local -strength and
ductility of cross sections, bond under static and cyclic loads of rebars-
and at global stiffness and rotation capacity of members-.
Results of both theoretical and experimental activities have been pub-
lished in the context of National and International conferences and
meetings. Specific attention has been paid also to continuous education
of young engineers and technical updating for practitioners. Diffusion
of software packages (free download of executables and tutorials) and
research results has been supplied on Reluis website.

The first set of results refer to the development of methods for numeri-
cal analysis of cross sections subjected to generalised bending and axial Fig. 11- Effect of biaxial bending on cross section local deformation (Di Ludovico et al. 2008a, 2008b).

22 RESEARCH - Seismic behavior


Fig. 12- Comparison between exact (fibre method) and approximate approaches for bare (left) and FRP reinforced r.c.
members (right) (Monti and Alessandri, 2008).

stress plane angle


(top and bottom left) and finally a comparison
between simplified and refined results (bottom right).
An alternative method has been proposed and implemented. It arrives
at defining closed-form equations for performing assessment of existing
RC columns with two-way steel reinforcement, under combined biaxial
bending and axial load. Starting from the load contour method, an effi-
cient procedure for estimating the strength/deformation section capaci-
Fig. 13- Fibre model of the cross section (left), force-displacement curve for a column (center), normalised equivalent stiffness
ty has been developed. In addition, simple closed-form equations for (right) (Verderame et al., 2008a, 2008b).
computing section uniaxial resisting moments and ultimate/yielding
At local scale, a number of bond tests on smooth rebars were devoted to
curvature has been defined.
the definition of a constitutive stress-slip law to be used in numerical
The method has been also extended to cross sections reinforced with
simulations.
FRP, resulting in a very effective guide for fast implementation of
Figure 13 reports sample data concerning specimens (left), typical
results in more general software packages and direct application by
cyclic experimental data (center) and finally the idealized constitutive
structural engineers using easy to manage electronic sheets.
law calibrated against the test (right).
Figure 12 shows an example of results that can be obtained according
At large scale, eight experimental tests on r.c. square or rectangular full
to this design tool, both for bare and FRP-strengthened members and
scale columns under constant axial load and uniaxial bending were per-
its ability to give reliable results with a reduced computational effort.
formed. Monotonic or cyclic action were applied on specimens. Details
Another relevant achievement is related to the response of the whole
about the experimental program are reported in Table 2 and Figure 14.
member under axial force and biaxial bending. In particular, the atten-
Each test was performed under displacement control.
tion has been focused on the development of the actual stiffness of the
member depending on the stress level and influence of biaxial bending. Table 2. Summary of experimental tests on r.c. columns
UNIAXIAL TESTS - Cross Section BxH (cmxcm)
A fibre model of the member has been implemented in order to provide 30x30 30x30 50x30 30x50
recommendations for characterization of equivalent stiffness to be used Longitudinal Reinforcement 8 12 8 12 12 12 12 12
in elastic analysis. The study carried on has been limited, at the Steel Rebars type Plain Deformed Plain Deformed
Normalized Axial Load,  0.2 0.2 0.1 0.1
moment, to the definition of the problems which must be taken into Type of action Monotonic Cyclic Cyclic Cyclic
account while passing from a cross section to the whole member, Bosco Number of tests 2 2 2 2
et al. 2008. In this context, it is worth noting the contribution of UNICH
that carried out nonlinear analyses of reference regular and irregular Primary experimental outcomes clearly indicate that the contribution of
reinforced concrete buildings and on problems related to the selection the base rotation on the global deformation mechanism is noticeably
of input ground motion, Canducci et al., 2008. different in case of columns reinforced by plain or deformed rebars,
however, the overall member global deformation and energy dissipation
In compliance with the main issues derived from the theoretical analy- capacity is not strongly affected by the bond performances of the inter-
sis of r.c. members and cross sections, experimental activity has been nal rebars. The global deformation capacity in the case of plain rebars,
carried out at different scales. is mainly due to a localized source of deformability at the column foun-

RESEARCH - Seismic behavior 23


Fig. 14- Experimental tests on columns.

dation interface (fixed end rotation), while a more diffused crack pattern
along the column end was observed in columns reinforced by using Fig. 15- FrontPage cover of two produced catalogues on precast concrete buildings.
deformed steel rebars. The significant influence of P- effect on the
global behavior of specimens has also clearly emerged by the experi- been drafted and published in a specific booklet (Figure 15). Every type
mental tests; if such effect is neglected, the ultimate rotation recorded has a short description with sketches and indications about the years of
on columns reinforced by deformed steel rebars is clearly less than that production, the regions of destination and the relative diffusion with
observed in columns reinforced by using plain rebars. However, due to respect to the global precast production. Some notes are added with
P- effect the ultimate rotations related to the two different columns indication of possible behaviour deficiencies for seismic destinations.
typologies is very close. Such result can be explained by considering Being a key point of precast construction, special attention has been
two main aspects: the higher strength of members reinforced with addressed to the seismic behaviour of connections. Some tests have
deformed rebars and the higher slope of the softening branch of the been performed to quantify their seismic capacity following a standard
shear-drift curves, if P- effect is considered. A calibration of a numer- approach.
ical model able to take account of specific aspects related to bond of Five principal categories of connections have been considered:
smooth rebars and anchoring end details has been also developed - floor-to-floor connections between adjacent floor or roof elements;
(Verderame et al., 2008c). - floor-to beam connections between floor or roof elements and the sup-
porting beam;
3.18 PREFAB: Behavior and Strengthening of Prefabricated Industrial - beam-to-column connections between the beam and the column;
Structures - column-to-foundation connections which provide the base support to
The first activity consisted of a wide survey of the existing buildings the columns;
produced from the 50s up to today and of the rendering in a reasoned - cladding-to-structure connections for the support of the wall panels.
way of the investigation results. This activity enjoyed the support of the Two level of tests have been performed:
National Association of prefabrication industries Assobeton which - particular tests: referred to the qualification of single connectors
involved ten member companies to provide the design documentation inserted between two overdimensioned blocks and subjected to the
of a number of constructions built in different times. So about 150 pro- principal action expected in the structural system (Figure 16, left);
ject documentations have been collected covering some decades of pro-
duction. Since information about far times were lacking, this survey has
been integrated with the historical memory of some experts, exploiting
their knowledge together with the old bibliography of specific journals.
From the examination of the project documentations a synthetic
description for any building has been recorded in a standard format,
summarising its features in a specific form of easy reference. This work
led to the printing of the booklet Precast structures: list of projects of
existing buildings which provides a good evidence of some decades of
precast production.
A complete catalogue of the different types of precast structures has Fig. 16- Details of testing of structural connection.

24 RESEARCH - Seismic behavior


- local tests: referred to the qualification of the connection included been examined (Felicetti et al., 2008).
between two significant portions of the elements, representing the struc- In parallel, a testing frame has been set up for push over and cyclic tests
tural arrangement and subjected to the relevant components of the on beam-to-column connections. The common type with a couple of
action (Figure 16, right); bars protruding from the column and passing through the holes of the
Previously a standard protocol for testing has been drafted. It defines beams is examined both in longitudinal and transverse direction.
the six parameters: The experimental test results in terms of force displacement allowed to
- strength: maximum value of the force which can be transferred characterise the connection roof element to beam, as global ductility,
between the parts; ultimate strength and dissipative capacity. In Figure 18, force-dis-
- ductility: ultimate plastic deformation compared to the yielding limit; placement curves are illustrated (monotonic test on rigid blocks) com-
- dissipation: specific energy dissipated through the load cycles; paring the traditional existing connection with a modified solution of the
- deformation: ultimate deformation at failure limit; connection. While a brittle behaviour is expected for the traditional
- decay: strength loss through the load cycles compared to force level; existing connection (concrete crash of the beam edges), a ductile behav-
- damage: residual deformation at unloading compared to the maximum iour is then obtained if the steel plate in the connection is opportunely
displacement; reduced. Almost similar conclusions can be obtained for quasi-static
to be measured both by: cyclic tests as reported in the research reports and papers.
- push over tests following a monotonic increase of displacement: Some other tests have been performed on dry bearing in order to mea-
- cyclic tests following an alternate history of displacements (Figure 17). sure the friction factor of neoprene pads over the concrete surface and
A number of tests on roof-to-beam connections (8 push-over and 11 their deformation parameters. Tilting tests and on inclined plane have
cyclic) have been performed. A complete report is available with the been made with different levels of normal loads (Magliulo et al., 2008).
results of the experimentation. Three types of steel connectors have The comparison between tests results and friction coefficient values
provided by PCI Handbook (1999), CNR 10018 (1999), and UNI-EN
1337:3 (2005), is shown in Figs. 19 and 20. In Figure 19 the neoprene
compressive stress (s) is reported on the horizontal axis, while the shear
one ( ) on vertical axis; in Figure 20 on this axis the friction coefficient
is presented. It is evident that PCI Handbook and CNR 10018 curves
well approximate the experimental data linear regression curve; this
does not happen in the case of UNI-EN 1337:3 curve. However CNR
10018 provides a bit larger friction strength with respect to the experi-
mental results, while PCI Handbook provides larger friction strength
only for compressive stress lower than 3 N/mm2. Furthermore, the tests
Fig. 17- Specific energy and cyclic behaviour of the connection. results confirm the light increment of friction strength, which corre-
Force

Displacement

Fig. 18- Push-over test: comparison of the two different solution (traditional vs. slightly modified).

RESEARCH - Seismic behavior 25


sponds to a light decrement of the friction coefficient, as the compres- 4. DISCUSSION
sive stress increases.
On the base of experimental results, the following relationships for neo- 4.1 MND: Non-Destructive Methods for the Knowledge of Existing
prene-concrete friction coefficient are proposed: Structures
m= 0.49 if s 0.15N/mm2 The research activities scheduled within the Task have been to a great

if 0.15 < s 5N/mm2 extent carried out without significant delays or anticipations, and the
m = 0.1+ s
 main objectives have been achieved.
where s is the compressive stress and
=0.055 N/mm2; s=5 N/mm2
is the neoprene maximum compressive strength according to CNR 4.2 FC: Calibration of Confidence Factors
10018. These formulations along with the linear regression curve of The results obtained agree with those expected, concerning the devel-
tests mean results are plotted in Figure 21: the two curves are almost opment of a Bayesian procedure for material strength evaluation and
coincident. the calibration of confidence factors.

4.3 IRREG: Assessment of the Nonlinear Behavior of Buildings, with


Emphasis on Irregular Ones
The results obtained by the Task are in line with the objectives origi-
nally established. Several papers by the different research units have
been published or are in press in international journals or conferences.
The only problems encountered by some of the units originated from
convergence issues in nonlinear codes. This is a well-known problem,
but some software failed to converge on a regular basis, thus delaying
Fig. 19- Comparison between compressive-shear stress curves provided by PCI Handbook, CNR 10018 and UNI-EN 1337:3 advances in the research. However, overall, the task followed the sched-
and tests regression
ule of work originally outlined.

4.4 MIX: Assessment and Strengthening of Mixed-type (Masonry/RC)


Buildings
The Task has pursued the proposed objectives. In particular, aspects of
modelling the behaviour of mix-type buildings through non linear analy-
sis have been investigated. These analyses have highlighted the different
steps in which the resistant elements withstand the seismic action.
The analyses performed allowed the identification of the main parame-
ters affecting the structural behaviour of mixed building and gave indi-
cations on feasible modelling and verification criteria.
Fig. 20- Comparison between compressive stress friction coefficient curves provided by PCI Handbook, CNR 10018 and
UNI-EN 1337:3 and tests regression curve.

4.5 TAMP: Influence of Infills on Structural Response


The objectives that have been pursued by this research are consistent
with the expected ones. Nevertheless, it must be observed that the
experimental calibration of the parameters defining the cyclic behav-
iour of the proposed equivalent diagonal strut model is affected by the
following main limitations: - it has been made considering only square
meshes of infilled RC frames; - the possible presence of an infill with
opening has not been considered.
These limitations, due to not sufficient time and resources, did not allow
the model to be validated by its use to reproduce the experimentally
Fig. 21- Proposed compressive stress - concreteneoprene friction coefficient relationship. detected response of the 3D infilled building subjected to the shaking

26 RESEARCH - Seismic behavior


table tests. actually exhaustive, since trial biaxial tests have been designed de-
pending on the findings of the theoretical work, but not completed. This
4.6 SCALE: Behavior and Strengthening of Stairs results in the need to extend and validate the results obtained in the
The Task has pursued the proposed objectives. A detailed investigation context of the present task and give a direct contribution to the devel-
of the main stairs typologies and of the most used design procedures opment of specific design recommendations for members under com-
have been performed; in particular, a report including the main results bined axial load and biaxial bending.
was developed (Cosenza et al., 2007b). Numerical investigations have
been performed in order to understand the seismic behaviour, and the 4.9 PREFAB: Behavior and Strengthening of Prefabricated Industrial
possible failure mechanisms of buildings having the most common stair Structures
typologies. The results have confirmed the need to utilize biaxial bend- The research activities scheduled within the Task have been to a great
ing modelling and to account for the interaction of the different internal extent carried out without significant delays or anticipations, and the
forces as: bending moment-axial force interaction that characterizes the main objectives have been achieved.
inclined elements, and the bending moment-shear interaction that gov-
erns the behaviour of squat columns. Shear failure becomes predomi- 5. VISIONS AND DEVELOPMENTS
nant in the squat columns and in the reinforced concrete slabs and pre-
cedes the conventional ductile failure. An experimental set-up has been 5.1 MND: Non-Destructive Methods for the Knowledge of Existing
designed on the basis of some simulations performed by using different Structures
modelling: dimensions of a single span frame with inclined beam, loads A large amount of RC buildings, both private and public, now placed
and resisting-wall have been defined. in seismic zones, were originally designed taking into account only
gravity loads and without explicitly provide ductile detailing. An
4.7 NODI: Behavior and Strengthening of Beam-Column Joints extraordinary rehabilitation program needs to be implemented on
The research activities scheduled within the Task have been to a great such buildings, where an accurate evaluation of the available seismic
extent carried out without significant delays or anticipations, and the capacity is important to set up cost-effective interventions.
main objectives have been achieved. Investigations have a crucial role to adequately know the structure to
be evaluated. For this reason, there is an increasing need to set up
4.8 BIAX: Behavior and Strengthening of Columns under Combined and put at disposal of technicians and other involved stakeholders
Axial Load and Biaxial Bending and Shear sufficiently reliable as well as not very expensive methods to estimate
Summary of results reported in the previous sections leads to recognize in-situ material properties. Number of tests required to suitably apply
that the development of the work basically complies with the initial these methods have to be as low as possible, thus making the total
schedule. In particular, as numerical analyses and software deliver- required budget sustainable to building owners and thus further
ables are considered, a good agreement with the program can be encouraging their use. To this purpose, the results obtained in this
addressed. Task confirm that a smart combination of NDTs and direct tests (such
In fact, different approaches and numerical procedures have been set as core extraction) gives effective solutions from both the economical
and made available to technical community. They cover at different lev- and technical point of view.
els the need of tools for checks required by modern codes in terms of Future research work should be devoted to the following:
strength and local deformation. This applies both to design and assess- Provide methods more and more capable of achieving effective
ment of existing un-strengthened concrete structures and to seismic results in terms of prediction capability of concrete properties taking
upgrading using FRP materials. Interaction between groups involved in into account both intrinsic randomness and epistemic uncertainty.
the study of irregular structures and of use of FRP for seismic upgrad- NDTs currently available on concrete do not provoke damage on
ing of structures is another positive aspect of the work. structural members but on some other building components (e.g. par-
When experimental program are concerned, it is worth noting that a rel- titions, infills, plaster, etc.) thus determining remarkable repair costs:
evant contribution to the knowledge of bond mechanisms for smooth new methods are necessary to make them really not very expensive.
bars has been given and an approach to the comparative analysis of per- As for reinforcement, taking into account the heavy damage caused
formances in terms of rotation capacity of full scale r.c. members with by the extraction of steel bars, non destructive methods to estimate its
smooth and deformed bars has been accomplished. The work is not mechanical properties need to be set up.

RESEARCH - Seismic behavior 27


5.2 FC: Calibration of Confidence Factors can provide an engineering estimate without adding complexity to the
The work performed by the Task has focused on problems related to the convergence procedure;
definition of a reliable assessment of structural seismic performance. Possible guidelines on how to consider other sources of nonlineari-
A probabilistic model for structural performance has been developed ties/failures, such as bond-slip, beam-column failure, etc.
and a method for calibration of Confidence Factors has been proposed.
Furthermore, a procedure for evaluation of material strength has been 5.4 MIX: Assessment and Strengthening of Mixed-type (Masonry/RC)
developed based on the application of the Bayesian method, taking into Buildings
account, both, amount and reliability of the in-situ tests performed. Considering the obtained results, the development of research aimed at
In addition to that, with reference to the prior probability distribution of evaluating the seismic behavior of mixed-type buildings can be envis-
structural detailing parameters, a preliminary list of possible structural aged as follows:
defects was prepared, in which, for each defect, a set of possible values Establish through non-linear analyses the influence of RC-masonry
and their relative weights are envisaged. connections on the global behavior of the building. Particular attention
To proceed further along this path, the following developments are should be paid at intersections between beams and perimeter masonry
needed: walls.
A complete procedure should be developed defining all phases of the Evaluation of q-factors. This objective should be pursued through the
knowledge process of an existing building. The procedure shall give a implementation of (static and dynamic) non-linear analyses.
CF calibrated on the basis of the distribution of the assessment results Evaluation of the seismic response (numerical and not only) of other
conditional on the acquired knowledge. structural configurations of mixed-type buildings not included in this
Further studies are needed regarding the evaluation of material study.
strength values from in-situ non-destructive tests, whose reliability
depends on the reliability of the regression curves used to transform the 5.5 TAMP: Influence of Infills on Structural Response
test parameter into a material strength value. This research gave useful results concerning the effects of masonry
A questionnaire, meant to be addressed to professional engineers as infills on the lateral response of RC frames, supported by experimental
a survey, has been prepared. The results of such survey shall be useful validations.
in creating a thorough database of structural defects and their proba- Possible developments could be aimed to pursue the following further
bility distributions. main objectives:
to revise the available empirical expressions linking the masonry
5.3 IRREG: Assessment of the Nonlinear Behavior of Buildings, with properties to those of its components, by carrying out further tests on
Emphasis on Irregular Ones masonry samples made of other kinds of resisting elements;
Several directions for future work have emerged from the project, most to generalize the proposed model of equivalent diagonal strut, by eval-
of them related to seismic code enhancements: uating experimentally the role of factors that have not been considered
Further studies for assessing the applicability of nonlinear pushover here: infilled mesh geometry, different compression levels on the frame
procedures to plan irregular buildings; columns after the masonry infills have been constructed, presence and
Further studies on different engineering demand parameters, such as size of an opening in the infilled mesh of frame.
interstory drift, chord rotation, plastic hinge rotation, as means for
assessing the structural response; 5.6 SCALE: Behavior and Strengthening of Stairs
The need for the definition of the damping to be used in NTHA, Considering the obtained results the future research developing would
depending on the model level of refinement. Using 5% damping results be oriented to the evaluation of the influence of the stair on the seismic
in an un-conservative assessment; response analysing the following aspects:
A clear definition of spectrum compatible accelerograms (according Experimental assessment of the strength capacity and deformability
to EC8) to be used in NTHA; of squat columns. In this way, the behaviour of squat columns would be
The need to establish a framework for Performance Based Earthquake better understood considering the large number of models found in lit-
Engineering, based on a fully probabilistic approach; erature that are not completely exhaustive for a reliable assessment of
Assessment of the current modelling capabilities for shear failure pre- the seismic behaviour of reinforced concrete structures.
diction, with the possible development of simplified approaches that Experimental evaluation of the seismic response of 2-D and 3-D

28 RESEARCH - Seismic behavior


frames with inclined elements and squat columns. The comparison of fruitful approach to the development of tools for the theoretical estima-
different test results would allow to evidence the influence of the intro- tion of strength and curvature ductility of members has been carried
duced specific elements such as inclined beams and squat columns. out. Flexural mechanisms are clearly identified both at local and glob-
Numerical and experimental evaluation of the influence of the stair in al scale and an encouraging capacity of simulation is demonstrated by
buildings with different location of infill walls. The numerical analysis both static and cyclic proposed models of members with smooth bars.
would be performed by using static and dynamic tools. This studies This means that numerical sensitivity analyses can give a positive con-
would allow the evaluation of the interaction between stairs and infills tribution to an optimized design of an experimental program able to
that are commonly modelled in different manner. confirm numerical forecasts, show possible points of weakness of the
theories and/or give an insight on specific aspects of the behavior under
5.7 NODI: Behavior and Strengthening of Beam-Column Joints severe cyclic loads. The large variety of existing constructions and the
The behaviour of beam-column joints can strongly affect the seismic diffusion of smooth bars in many very urbanized areas exposed to seis-
global behaviour of RC building structures. Some mechanisms (e.g. mic risk point out the need to continue the investigation on such type of
concrete cracking, slippage of longitudinal reinforcing bars, etc.) can be members and even develop customized strengthening techniques using
responsible of additional deformability, on one hand, while, on the other advanced materials.
hand, can alter the capacity design assumptions on the framing struc- Despite such positive feedbacks of the research, it is worth noting that
tural members (beams and columns) and on the joint member itself. For further work is strongly recommended to assess:
this reason, research activities need to be increasingly devoted to devel- the behavior of short columns, and
op accurate capacity models of beam-column joints to reliably evaluate the flexure-shear interaction in presence of smooth reinforcement.
the performances of RC building structures. In fact, the observed deformation mechanisms can produce effects on
Research carried out in this Task already provided important results the strength and ductility of members subjected to complex forces, like
relevant to the role of the key behavioural parameters of RC beam-col- those generated on columns of irregular constructions.
umn joints, thus giving useful suggestions on the reliability of current
code expressions and on possible improvements. However, many other 5.9 PREFAB: Behavior and Strengthening of Prefabricated Industrial
issues need to be further studied regarding both as-built and strength- Structures
ened joints (already damaged or not damaged), skilfully combining pur- The experimental campaign and the numerical investigations carried
posely designed experimental investigations, review of experimental out have highlighted that the most vulnerable buildings are those with
campaigns reported in the literature, and accurate numerical simula- disarticulated diaphragm behaviour. However, as emphasised in
tions. Palermo et al. (2008), an accurate study on the modelling of connec-
Among others, some future research developments can be recognized as tions needs to be done in order to correctly predict the overall response
follows: of precast concrete buildings.
design and execution of extensive experimental programs on joint
specimens having different characteristics (e.g. internal or external, bi- 6. MAIN REFERENCES
or tri-dimensional, shape, beam type, etc.) well targeted on the types
representative of the Italian and European built environment; 6.1 MND: Non-Destructive Methods for the Knowledge of Existing
experimental and numerical validation of different strengthening Structures
techniques based on the comparison of the relative performance and - Masi A., Vona M. (2008). La stima della resistenza del calcestruzzo
application limits, particularly dealing with tri-dimensional joints and in situ: impostazione delle indagini ed elaborazione dei risultati,
joints with embedded beams; Progettazione sismica, Anno I, No. 1, IUSS Press, ISSN 1973-7432.
codification of test protocols to make possible and promote experi- - Marano G.C., Morrone E., Mezzana M. (2008), Approccio ibrido
mental result exchange. fuzzy per lintegrazione e linterpretazione delle prove non distruttive,
Atti del convegno Valutazione e riduzione della vulnerabilit sismica di
5.8 BIAX: Behavior and Strengthening of Columns under Combined edifici esistenti in cemento armato, E. Cosenza, G. Manfredi, G. Monti
Axial Load and Biaxial Bending and Shear editors, Polimetrica International Scientific Publisher, Roma, Italy.
This Task research activity dealt with a specific, but relevant aspect of - Olivito R.S., Spadea G., Carrozzini A., Spadafora A.R. (2008). Strut-
seismic design and assessment of r.c. constructions. An effective and ture esistenti in cemento armato: controlli e verifiche mediante tecniche

RESEARCH - Seismic behavior 29


non distruttive. Atti del convegno Valutazione e riduzione della vulne- tecniche per le costruzioni.
rabilit sismica di edifici esistenti in cemento armato, E. Cosenza, G. - D.M. Lavori Pubblici, 16 Gennaio 1996. G.U. 5-2-1996, N.29, Norme
Manfredi, G. Monti editors, Polimetrica International Scientific Pu- tecniche per le costruzioni in zona sismica.
blisher, Roma, Italy. - Decanini L.D., Liberatore L., De Sortis A., Benedetti S. (2006).
Rapporto sui comportamenti, osservati in laboratorio e dopo terremoti
6.2 FC: Calibration of Confidence Factors severi, sulle prescrizioni di diverse normative, nazionali e internazio-
- Franchin P., Pinto P. E., Pathmanathan R., (2008). Assessing the nali, a proposito di edifici a struttura mista muratura-c.a.. Rapporto di
adequacy of a single confidence factor in accounting for epistemic ricerca RELUIS. Novembre 2006.
uncertainty, Convegno RELUIS Valutazione e riduzione della vulne- - Decanini L.D., Tocci C., Liberatore L., Benedetti S. (2008). Analisi
rabilit sismica di edifici esistenti in cemento armato, Roma 29-30 numeriche non lineari agli elementi finiti su modelli piani e spaziali di
maggio 2008. edifici a struttura mista muratura-c.a.. Rapporto di ricerca RELUIS.
- Jalayer F., Iervolino I., Manfredi G., (2007). Influenza dei parametri Gennaio 2008.
di modellazione e dellincertezza associata nella valutazione sismica di - NAA-80, 1980. Normas Antisismicas Argentinas.
edifici esistenti in cemento armato, Proceedings XII Convegno ANI- - Nardone F., Verderame G.M., Prota A., Manfredi G. (2008). Analisi
DIS, Giugno 2007. comparativa su edifici misti C.A.-Muratura. Valutazione e riduzione
- Jalayer F., Iervolino I., Manfredi G., (2008). Structural modelling della vulnerabilit sismica di edifici esistenti in c.a. Roma 29-30
uncertainties and their influence on seismic assessment of existing RC Maggio 2008. MIX-04.
structures, submitted to Structural Safety, 2008. - O.P.C.M. n.3431, 2005. Suppl. Ord. n.85 alla G.U. n.107 del
- Monti, G., Alessandri S., Goretti, A., (2007). Livelli di conoscenza e 10/05/2005 e s.m.i.
fattori di confidenza. Proceedings XII Convegno ANIDIS, Giugno 2007.
- Monti G., Alessandri S., (2008). Confidence factors for concrete and 6.5 TAMP: Influence of Infills on Structural Response
steel strength, Convegno RELUIS Valutazione e riduzione della vul- - Amato G., Cavaleri L., Fossetti M., Papia M. (2008). Infilled frames:
nerabilit sismica di edifici esistenti in cemento armato, Roma 29-30 influence of vertical loads on the equivalent diagonal strut model,
maggio 2008. Proc.14th WCEE, Beijing, China, CD-ROM, Paper 05-01-0479.
- Cavaleri L., Fossetti M., Papia M. (2005). Infilled frames: develop-
6.3 IRREG: Assessment of the Nonlinear Behavior of Buildings, with ments in the evaluation of the cyclic behaviour under lateral loads,
Emphasis on Irregular Ones Structural Engineering and Mechanics, Teckno Press, (21) 469-494.
- EuroCode 8 (2003a). Design of Structures for Earthquake Resistance, - Decanini L.D., Bertoldi S.H., Gavarini C. (1993). Telai tamponati
- Part 1: General Rules, Seismic Action and Rules for Buildings, Draft soggetti ad azioni sismiche, un modello semplificato, confronto speri-
no. 6. mentale e numerico, Atti 6 Convegno ANIDIS, Perugia.
- EuroCode 8 (2003b). Design of Structures for Earthquake Resistance, - Durrani A.J., Luo Y. H. (1994). Seismic retrofit of flat-slab buildings
- Part 3: Strengthening and repair of buildings, Draft no. 3 (Final with masonry infills, Proceedings from the NCEER Workshop on
Project Team Stage 34), 2003. Seismic Response of Masonry Infills-Report NCEER-94-0004, 1-3.
- Fajfar P. (2002). Capacity Spectrum Method Based on Inelastic - Mainstone R.J. (1971). On the stiffness and strengths of infilled
Demand Spectra, Proceedings of the 12th European Conference on frames, Proceedings Institution of Civil Engineers, London,
Earthquake Engineering, Paper 843, London (UK). Supplement IV, Paper 7360S, 57-90.
- FEMA 356 (2000). Prestandard and Commentary for the Seismic - Stafford-Smith B. (1966). Behavior of square infilled frames,
Rehabilitation of Buildings, Prepared by A.S.C.E., Washington, D.C. Journal of the Structural Division, ASCE, 92 (1), 381-403.
(U.S.A.). - Stafford-Smith B., Carter C. (1969). A method of analysis for infilled
- ATC 40 (1996). Seismic Evaluation and Retrofit of Concrete Build- frames, Proceedings Institution of Civil Engineers, London, (44) 31-48.
ings.
6.6 SCALE: Behavior and Strengthening of Stairs
6.4 MIX: Assessment and Strengthening of Mixed-type (Masonry/RC) - Berry R. (1999), The construction of buildings. 5th edition Vol. 2.
Buildings Blackwell Science.
- D.M. Infrastrutture, 14 Gennaio 2008. G.U. 4-2-2008, N.29, Norme - Biskinis D.E., Roupakias G.K, Fardis M.N., (2004).Degradation of

30 RESEARCH - Seismic behavior


shear Strength of Reinforced Concrete Members with Inelastic Cyclic CEN Comit Europen de Normalisation (2003). Eurocode 8 Design
Displacements. ACI Structural Journal title no. 101-S76. of structures for earthquake resistance - Part 3: Assessment and retro-
- Cosenza E., Mariniello C., Verderame G.M., Zambrano A. (2007), Il fitting of buildings (draft n. 6), prEN 1998-3, Brussels.
ruolo delle scale sulla capacit sismica degli edifici in C.A. Atti Del XII - Masi A., Santarsiero G. (2008). Sperimentazione su nodi trave-colon-
Convegno Anidis Lingegneria Sismica In Italia, 10-14 Giugno 2007, na in c.a. progettati con diversi livelli di protezione sismica: primi risul-
Pisa, Italia. (in Italian). tati, Atti del convegno Valutazione e riduzione della vulnerabilit sismi-
- Cosenza E., Verderame G.M., Zambrano A. (2007), Rapporto sul- ca di edifici esistenti in cemento armato, E. Cosenza, G. Manfredi, G.
lanalisi tipologica delle scale e criteri di progettazione statica e sismi- Monti editors, Polimetrica International Scientific Publisher, Roma,
ca. Rapporto di ricerca RELUIS. (in Italian). Italia.
- Cosenza E., Verderame G.M., Zambrano A.(2008), Capacit sismica - Masi A., Santarsiero G., Moroni C., Nigro D., Dolce M., Russo G.,
degli edifici esistenti in C.A.: Linfluenza delle scale, Proc. Reluis2rm08 Pauletta M., Realfonzo R., Faella C., Lignola G.P., Manfredi G., Prota
Valutazione e riduzione della vulnerabilit sismica di edifici esistenti A., Verderame G.M., (2008). Behaviour and strengthening of RC
In C.A. Roma 29-30 Maggio, E. Cosenza, G. Manfredi, G. Monti Eds., beam-column joints: experimental program and first results of the
Polimetrica International Scientific Publisher, Isbn 978-88-7699-125- research activity in the framework of DPC-Reluis Project (Research
5. (in Italian). Line 2), Proc. of the 14th World Conference on Earthquake Engineering,
- Marrullier E. (1910), Guida Pratica per la costruzione degli edifici con Beijing, China.
speciale riguardo al cemento armato, Editori Torinesi. (in Italian). - NTC 2008, Decreto Ministeriale 14 gennaio 2008, Norme tecniche per
- Migliacci A. (1977), Progetti di strutture, raccolta delle lezioni tenute le costruzioni, Suppl. or. n.30 alla G.U. n.29 del 4/2/2008.
presso il Politecnico di Milano negli anni accademici 1966-67, 1967-
68, Parte seconda, Edizioni Masson, Milano. (in Italian). 6.8 BIAX: Behavior and Strengthening of Columns under Combined
- Pagano M. (1963), Strutture, Liguori Editore. (in Italian). Axial Load and Biaxial Bending and Shear
- Perri E. (1966), Ingegneria Sismica, Unione tipografico- Editrice - Alessandri S., Monti G. (2008). Design equations for the assessment
Torinese. (in italian). and FRP-strengthening of reinforced rectangular concrete columns
- Reynolds C. E., Steeedman J.C., (2002). Reinforced concrete design- under combined biaxial bending and axial loads. Journal Mechanics of
ers handbook, Spon Press Taylor & Francis Group London. Composite Materials, 44 (3), pp. 197-324.
- Rosci L. (1939), Manuale pratico di volgarizzazione del Calcolo del - Di Ludovico M., Verderame G.M., Iovinella I., Cosenza E. (2008a).
Cemento Armato, G. Lavagnolo, II edizione, Torino. (in Italian). Domini di curvatura in pressoflessione deviata di sezioni in c.a.
- Santarella L. (1953), Il cemento armato, le applicazioni alle co- PARTE I: Analisi a fibre. Proc. Reluis2Rm08 Valutazione e riduzione
struzioni civili ed industriali, Vol. II, 13, Edizione Hoepli. (in Italian). della vulnerabilit sismica di edifici esistenti in c.a. Roma 29-30 mag-
- Santarella L. (1957), Il cemento armato, monografie di costruzioni ital- gio, E. Cosenza, G. Manfredi, G. Monti Eds., Polimetrica International
iane civili ed industriali, Vol. III, 13, Edizione Hoepli. (in Italian). Scientific Publisher, ISBN 978- 88-7699-125-5, pp. 629-638.
- Sezen H., Moehle J.P., (2004). Shear Strength Model for Lightly - Di Ludovico M., Verderame G.M., Iovinella I., Cosenza E. (2008b).
Reinforced Concrete Columns, Journal of Structural Engineering, Domini di curvatura in pressoflessione deviata di sezioni in c.a.
ASCE November 2004, pp.1692-1703. PARTE II: Valutazione semplificata. Proc. Reluis2Rm08 Valutazione
- Tecnica y Pratica del Hormigon Armado. (1989) Ceac ed. Barce- e riduzione della vulnerabilit sismica di edifici esistenti in c.a. Roma
lona.(in Spanish). 29-30 maggio, E. Cosenza, G. Manfredi, G. Monti Eds., Polimetrica
- Zhu L., Elwood K.J., Haukaas T., (2007). Classification and seismic International Scientific Publisher, ISBN 978- 88-7699-125-5, p. 639-
safety evaluation of existing reinforced concrete columns, Journal of 648.
Structural Engineering, ASCE. 1316-1330. - fib (2003) Seismic assessment and retrofit of reinforced concrete
buildings. Bulletin 24, ISBN 978-2-88394-064-2.
6.7 NODI: Behavior and Strengthening of Beam-Column Joints - Bosco M., Ghersi A., Leanza S. (2008). Force displacement relation-
- Braga F., De Carlo G., Corrado G.F., Gigliotti R., Laterza M., Nigro D., ship for r/c members in seismic design. Proceedings of the 14th World
(2001). Meccanismi di risposta di nodi trave-pilastro in c.a. di strut- Conference on Earthquake Engineering. October 12-17, Bejing, China.
ture non antisismiche, Atti del X Congresso Nazionale Lingegneria - Canducci G., Camata G., Spacone E. (2008). Problematiche con-
Sismica in Italia, Potenza-Matera, Italia. nesse alla descrizione del moto sismico. Proc. Reluis2Rm08 Valuta-

RESEARCH - Seismic behavior 31


zione e riduzione della vulnerabilit sismica di edifici esistenti in c.a. II - SEISMIC ASSESSMENT AND RETROFIT OF EXISTING
Roma 29-30 maggio, E. Cosenza, G. Manfredi, G. Monti Eds., Poli- BRIDGES
metrica International Scientific Publisher, ISBN 978-88-7699-125-5,
pp. 327-334. 1. INTRODUCTION
- Panagiotakos T., Fardis M.N. (2001). Deformation of R.C. members
he perception of the risk associated to the seismic vulnerability of
at yielding and ultimate. ACI Structural Journal, 98 (2), pp. 135-148.
Park R., Paulay T. (1975). Reinforced concrete structures. J. Wiley
and Son, Inc., N.York, 769 pp.
T the transportation infrastructure, and in particular to that of bridge
structures, on the part of both the relevant authorities and the profes-
- Verderame G.M., De Carlo G., Manfredi G., Fabbrocino G.(2008a). sion is a quite recent acquisition in Italy. This is possibly due to the fact
Laderenza ciclica in campo elastico delle barre lisce. Parte I: la spe- that in the last two major events that have struck the Country in the sec-
rimentazione, Proc. Reluis2Rm08 Valutazione e riduzione della vul- ond half of the 20th century (Friuli 1976 and Irpinia 1980) the trans-
nerabilit sismica di edifici esistenti in c.a. Roma 29-30 maggio, E. portation infrastructure has not suffered significant distress. In particu-
Cosenza, G. Manfredi, G. Monti Eds., Polimetrica International lar, in Friuli the construction of highways was just at the beginning. In
Scientific Publisher, ISBN 978-88-7699-125-5 the Apennine crossing of the A16 highway the bridges did undergo
- Verderame G.M., Ricci P., Manfredi G., Fabbrocino G. (2008b). some damage, mainly due to the inadequacy of the bearing devices, but
Laderenza ciclica in campo elastico delle barre lisce. Parte II: la mo- this was promptly remedied by the owner through the systematic adop-
dellazione. Proc. Reluis2Rm08 Valutazione e riduzione della vulnera- tion of the then innovative technique of seismic isolation.
bilit sismica di edifici esistenti in c.a. Roma 29-30 maggio, E. Co-
senza, G. Manfredi, G. Monti Eds., Polimetrica International Scientific On the other hand, it can be observed that this delay in the apprecia-
Publisher, ISBN 978-88-7699-125-5 tion of the risk is not exclusive to Italy. For example, it is enough to
- Verderame G.M., Ricci P., Manfredi G., Cosenza E. (2008c). La mention that it took twelve years after the spectacular failures of quite
capacit deformativa di elementi in c.a. con barre lisce: modellazione modern bridges (Figure 1, left) during the San Fernando (1971) earth-
monotona e ciclica. Proc. Reluis2Rm08 Valutazione e riduzione della quake, for the Federal Highway Administration (FHWA) to publish a
vulnerabilit sismica di edifici esistenti in c.a. Roma 29-30 maggio, E. first document titled Retrofitting guidelines for Highway Bridges
Cosenza, G. Manfredi, G. Monti Eds., Polimetrica International (FHWA-ATC, 1983). Still, in 1989, despite of the large retrofit program
Scientific Publisher, ISBN 978-88-7699-125-5, p. 617-628. set up (later proved to be fully inadequate), the Loma Prieta earthquake
exposed substantial deficiencies in bridges in California (Figure 1,
6.9 PREFAB: Behavior and Strengthening of Prefabricated Industrial right).
Structures
- Felicetti R., Lamperti M., Toniolo G., Zenti C., (2008). Analisi speri-
mentale del comportamento sismico di connessioni tegolo-trave di
strutture prefabbricate. XVII Congresso C.T.E., Roma, 5-8 Novembre,
Italy, pp. 867-874.
- Magliulo G., Capozzi V., Fabbrocino G., Manfredi G., (2008). Deter-
minazione sperimentale del coefficiente di attrito neoprene-calcestruz-
Fig. 1- Damage to bridges during the San Fernando, 1971 (left) and Loma Prieta 1989 (right) events.
zo per la valutazione della vulnerabilit sismica delle strutture prefab-
bricate esistenti. Proc. Reluis2Rm08 Valutazione e riduzione della The situation as briefly outlined above is sufficient to understand that
vulnerabilit sismica di edifici esistenti in c.a., Roma, 29-30 maggio, E. the state of the art on seismic assessment and retrofit of bridges still
Cosenza, G. Manfredi, G. Monti Eds., Polimetrica International Scien- needs to be advanced in several areas. The research undertaken within
tific Publisher, ISBN 978-88-7699-129-5, p. 717-724. this Line of the DPC-Reluis Project aimed at providing a contribution
- Palermo A., Camnasio E. and Poretti M., (2008). Role of Dissipative in this direction. The areas considered to be of prioritary interest were
Connections on the Seismic Response of One-Storey Industrial Build- assessment methods, retrofit criteria and techniques, abutments and
ings, Proc. of the 14th World Conference on Earthquake Engineering, foundations, with the final goal of producing a comprehensive docu-
Beijing, China. ment with guidelines and example applications. This result, which has
been achieved, represents the first European document on the topic and

32 RESEARCH - Seismic behavior


could be envisaged to form the basis for a future addition to the text on the seismic assessment and retrofit of existing bridges.
Eurocodes system.
3. RESEARCH STRUCTURE
2. BACKGROUND AND MOTIVATION
The program was articulated into five main tasks:
Starting from the year 1992 on funding from the FHWA, a vast research 1) Identification of bridge typologies:
program has been undertaken in the US to clarify several aspects relat- Bridge typologies characterising the Italian road- and railway systems
ed to the seismic assessment and retrofit of bridges. were identified through contacts with the main national and regional
The first product of the above research appeared in 1995 in the form of administrations, as well as with major contractors. In particular, exist-
the Seismic Retrofit Manual for Highway Bridges (FHWA, 1995): its ing contacts were exploited and new ones were established with ANAS,
further development has led to the Seismic Retrofitting Manual for Autostrade, Italferr, RFI, Ferrovie della Calabria in order to acquire
Highway Structures: Part 1 Bridges and the Seismic Retrofitting detailed documentation on a number of representative structures.
Manual for Highway Structures: Part 2 Retaining structures, Slopes, 2) Assessment methods:
Tunnels, Culverts and Roadways (FHWA- MCEER, 2005). Existing methods developed for the assessment of buildings were
In Europe the Eurocodes system includes a normative document for the extended to the deal with bridge structures, and methods specifically
seismic design of new bridges, which is at least partially based on the devised for bridges were further developed. The goal was to fine-tune
recent concepts of performance-based design: Eurocode 8 Part 2 (CEN, several methods of increasing level of accuracy and required effort to be
2005a). This document, however, is not matched by a companion one used according to the importance and size/regularity of the bridge.
covering existing bridges, differently with the situation of buildings, for These include displacement-based linear and non linear static meth-
which such a document is available in the form of Eurocode 8 Part 3 ods, as well as simplified behaviour-factor-based methods. Special
(CEN, 2005b). attention was also devoted to modelling for non-linear analysis.
In the year 2003 a firm change of direction towards the harmonisation 3) Retrofit criteria:
with Eurocode 8 has occurred in the Italian normative framework for This task focussed on the specific aspects of the application of tradi-
seismic design. In that occasion the priority was given to the drafting of tional and innovative retrofit techniques to structural elements typical
documents for the design of new structures, both buildings and bridges. of bridges. The program of activity included the execution of an exper-
A document for existing structures was also introduced, but again lim- imental campaign aimed at establishing the effectiveness of alternative
ited to buildings. These documents served later as the basis for the pro- retrofit techniques. This task also included the development of seismic
duction of the seismic chapter of the current Eurocodes-aligned nation- isolation solutions to be applied in the retrofit of bridges that were not
al design code produced by the Ministero delle Infrastrutture initially designed to be isolated, as well as bridge-specific seismic iso-
(DM2008). lation design criteria.
The need for a document dealing explicitly with the problem of assess- 4) Assessment of abutments, earth-retaining structures and founda-
ing and retrofitting bridges in seismic areas dates back actually to 2003, tions:
when the update of the seismic design code was accompanied by the Abutments and foundations are often weak elements in existing
obligation of assessing, within the time limit of five years, all the strate- bridges. The goal of this task was to advance the state of the art in the
gic structures and infrastructures in the Country. Adhering to this oblig- seismic analysis and assessment of these components, an area still
ation and with reference to bridges, with funding from the Civil characterised by the widespread use of mostly empirical or convention-
Protection Department, the National Agency for Roads and Highways al approaches.
(ANAS) has launched a program for the assessment of all its bridge 5) Model applications to bridges of different typologies:
structures. Further, the theme is of pressing interest due to the wide- It was planned that under this task at least one bridge for each of the
spread activity currently ongoing on the Italian highway network to main typologies identified under task 1 was to be subjected to a detailed
increase its traffic capacity. assessment and retrofit design, and then documented in an application
manual to complement the pre-normative document.
Within the above context the DPC-Reluis research project, with its Line The guidelines and the companion manual represent the main
3 is intended to respond to the outlined needs, and in particular to that outcome of the project.
of producing a document to be used as a proposal of a future normative The five broad tasks outlined above were split into a number of sub-

RESEARCH - Seismic behavior 33


Table 1. Subdivision of the research activity into tasks and sub-tasks. (50+9100+50 m and tall piers), Bardonecchia bridge (742 m) and the
Millaures bridge (680 m, composite steel-concrete). The second high-
way, built between 1965 and 1975 shows a greater typological variabil-
ity, which can be reduced, however, to a few homogeneous sets.
Representative bridges are: the Borgotaro viaduct (slab bridge with sev-
eral interconnections), the Narboreto bridge (430 m), the Rio Verde
viaduct (2?65+695+76 m and very tall piers, h=150 m) and the
Roccaprebalza South viaduct (1345 m and tall piers), Rio Barcalesa
(743m).

As it regards the infrastructure in the Abruzzi region typologies and


conditions were monitored along A14 and SS16. Data sheets, consist-
ing of 7 sections that provide location, type, category, geometrical and
tasks according to the table below. The table also shows the progress of environmental characteristics, condition, and photographic description,
activity over the whole project duration of three years. The progress were compiled to catalogue all 52 bridges of the latter road. Bridges
achieved during the second year is briefly summarised in the following were classified according to structural type, material and geometry. The
Table 1. bridge conditions, the piers, and the cracking and vegetative maps were
considered.
4. MAIN RESULTS
Bridge structures on the A1FiBo and A16 were scrutinised, and a
The main results of the research activity are summarised in the following selected number of bridges either representative of the most frequent
according to the research structure described in the previous section. typologies or significant for their outstanding design was identified. A
further screening of the set including these structures and those identi-
4.1 Task 1: Identification of bridge typologies fied on the TBF and PLS has led to the definition of the final case-stud-
The main bridge typologies on several seismic-prone portions of the ies for the detailed applications and the development and calibration of
Italian railway and road/highway networks have been identified during assessment methods.
the first year of activity. In summary, the data collected from various
sources (mainly national/regional administrations) pertain to the Torino- In the roadway and railway system of Calabria the most common typolo-
Bardonecchia-Frejus (TBF) and the Parma-La Spezia (PLS) highways gies are single-stem or frame piers with full, single-, or multi-cellular
(Politecnico di Torino), the Firenze-Bologna (A1FiBo) portion of the A1 hollow-core cross-section; simply supported decks, made up of rein-
Milano-Napoli and the Apennine portion of the A16 highways (Universit forced or prestressed concrete girders and a cast-in-place RC deck
di Roma La Sapienza and Universit di Roma Tre), the Adriatic por- slab. Two study cases were selected for the study of non-conventional
tions of the A14 Bologna-Canosa highway and of the SS16 Adriatica protection/retrofit techniques. The first one is the Follone viaduct on the
state-road (Universit di Chieti-Pescara), the Roma-Viterbo (RMVT) and A3 Salerno-Reggio Calabria highway, where the spans have been con-
Roma-Sulmona (RMSu) railways (Universit di Roma Tre), the regional nected with by means of longitudinal devices, while the second case is
railway and roadway networks of Calabria (Universit di Cosenza). the Val di Leto viaduct, on a provincial road, which was recently retro-
fitted using oleodynamic devices.
Structural typologies characterizing the TBF and the PLS highways are
quite different. The first highway, built in between 1983 and 1992, Finally, the data collected during the survey of the two railway lines
includes rather uniform typologies: a) about 300.000 m2 of precast seg- RMVT and RMSu, have allowed selection of four masonry arch bridges,
mental box girder bridges with pier heights up to 90m and span lengths two per line, to be used as case-studies for the calibration of analysis
between 40 and 100m, b) about 200.000 m2 of girder bridges in con- methods for masonry bridges.
crete and in composite steel-concrete with pier heights between 5 and
30m and span lengths between 20 and 80m. Representative bridges 4.2 Task 2: Assessment methods
are: the Borgone viaduct (20+2640+20 m), the Ramat viaduct In summary, Task 1 has shown that the relatively many important

34 RESEARCH - Seismic behavior


bridges crossing wide valleys in the mountain tracts of the Central and m = 0.3mpila + mpulv + mimp (1)
Southern Apennine (A1FiBo, PLS, A3) represent a negligible percent- (m + 0.3mpila)Hp + mimpHimp
H  pulv m (2)
age of the total bridge stock, made up essentially of bridges with sim-
ply-supported decks (prestressed or reinforced concrete girders plus
y = 1 yH2 / 3 (3)
slab) with single stem of frame piers. For this reason a special effort has
been devoted to devising a simplified non-linear method suitable for the
u = y + (u y)  lp(Hlp / 2) (4)
analysis of bridges with simply-supported decks (Universit di Roma
La Sapienza). As it regards statically indeterminate bridges (continuous T = 2 mi / k = 2 miy / Vy
(5)
decks) a distinction was made between those with special configuration, max= SDe(T) T  Tc o q 1
for which inelastic dynamic analysis is in most cases the method of
choice, and simpler bridges, for which two methods have been thor- q [
max= SDe(T) 1+(q1) Tc
T ] T < Tc (6)

oughly explored: the Modal Pushover Analysis (MPA) method


(Universit di Roma La Sapienza) and the Secant Mode Superposition The guidelines and application manual give a detailed description of
(SMS) method (Universit di Pavia). Finally research has also focussed the method, of the safety verifications of bearings, piers and founda-
on two more issues, namely the always debated problem of directional tions, and a complete worked-out example.
combination rules (Universit di Chieti-Pescara) and the non linear
modelling of seismic protection devices (Universit di Cosenza). 4.2.2 VERIFICATION OF APPLICABILITY OF MPA METHOD TO BRIDGE STRUC-
TURES

4.2.1 SIMPLIFIED NON LINEAR METHOD FOR BRIDGES WITH SIMPLY-SUPPORT- The MPA method by Chopra and Goel (2002) has been devised for the
ED DECKS
analysis of tall buildings. Its applicability as an alternative to inelastic
For these bridges it is possible to set up an ad hoc assessment proce- dynamic analysis and adaptive pushover methods for the assessment of
dure which represents a convenient trade-off between simplicity and
accuracy. The reference model is that of a vertical cantilever with a con-
tinuous distribution of mass, on top of which rest the pier cap and the
deck masses. As long as the pier height is not such as to make higher
mode contributions significant, in the transversal direction each pier
a
represents a single-degree of freedom oscillator (see Figure 2a). In the
longitudinal direction the entire bridge can also be represented as a
SDOF system if seismic restrainers are provided that minimise the rel-
ative movements of adjacent decks on top of the pier caps (see Figure
2d). In this case the system has mass equal to the sum of the tributary
b
masses of the piers and resisting force sum of the resisting forces of the
piers (assuming that maximum displacements are permitted by the
abutments joints).

The method consists of a simplified non linear static analysis in which


c
the force-displacement laws are constructed based on the results of
moment-curvature analysis of the pier bases (see Figure 2b). The fol-
lowing equations give the tributary mass, Eq.(1), the effective height in
the transversal direction, Eq.(2), the yield and ultimate displacement
(see Figure 2b and c), Eq.s (3) and (4), the period, Eq.(5), and the cor-
responding demand displacement, Eq.(6), as for single-mode conven-
d
tional pushover analysis. The effective height equals the piers height in
the longitudinal direction. Displacement capacity follows from ultimate
displacement with appropriate safety factors. Fig. 2- Simplified non linear method for the assessment of bridges with simply-supported decks.

RESEARCH - Seismic behavior 35


Fig. 3- Longitudinal profile of the Rio Torto viaduct (A1FiBo).

The variation of the lateral load distribution, from one based on the
modal (elastic) displacement shape to another based on the (plastic)
displacement shape at failure, does not affect appreciably the results.
The best estimate of the displacements by the MPA (i.e. the one
derived taking the optimal reference DOFs) matches reasonably well
that from TH. It is worth noting that a comparable amount of approxi-
mation on the response of the structure is obtained both in the elastic
and in the plastic response regimes. This observation, together with the
previous one, indicates that the main approximation of the method, i.e.
being based on the initial elastic modal vector, may not represent a
major limitation.
Differences between the nodal displacements estimated by the MPA
with respect to those by the TH are found to be in the order of 15%,
independently of the intensity level of the ground motion. Analogous
results are observed also for the curvatures at members end-sections,
resulting in almost coincident patterns of plastic hinge location and pre-
dictions of members failures.

Fig. 4- Two piers of the Rio Torto viaduct (A1FiBo). For the considered case, the application of the MPA method has shown
bridge structures has been investigated through its application to the Rio to lead to fully acceptable results. Such a favourable conclusion still
Torto viaduct (see Figures 3 and 4), one of the case studies selected for awaits substantiation from a larger number of applications. These
the project. The structure, built at the end of 50s, is characterized by results have led to the introduction of the method among the allowed
thirteen-span twin decks realized with two girders and top slab. The methods in the draft guidelines.
twelve supports consist of a pair of framed piers, one under each deck.
Each pier is a multi-storey reinforced concrete frame with variable 4.2.3 SMS METHOD
height, realized with two circular columns of diameter D =120160 cm. The Secant Mode Superposition method consists essentially of an iter-
The results from inelastic dynamic analysis have been taken as bench- ative multi-modal response spectrum analysis on a structural model
mark for the purpose. The response has been compared for several with secant stiffness properties and equivalent viscous damping. The
intensity levels (to assess the influence on accuracy of the level of non- procedure can be summarised in the following steps:
linearity in the response) and in terms of different response quantities, Step 0: A starting displacement profile and stiffness distribution are
both local and global (section curvatures and element displacements). assumed;
Step 1: The stiffness matrix of the equivalent linear structure is
The comparison provided the following indications: assembled;
The location of maximum modal displacement is the best choice as Step 2: Modal analysis is carried out;
the reference degree of freedom (DOF) for estimating the demand on the Step 3: Displacement in each vibration mode are obtained either from
structure. Each significant mode is therefore characterized by its own an over-damped elastic or from an inelastic displacement spectrum;
reference DOF. Step 4: Modal contributions are combined to yield displacement pro-

36 RESEARCH - Seismic behavior


file and moment distribution (different combinations rules were exam-
ined);
Step 5: Two response indices are computed, that evaluate conver-
gence on displacement profiles and force distributions, while checking
that the structural capacity is not violated.
Step 6: A final response index is obtained as an average of the first
two and checking the convergence of the proposed iterative procedure. Fig. 5- Pier section and specimens built at the University of Pavia.

The method has been thoroughly tested on the six typological bridges, Pier height of 6 m (aspect ratio equals 4);
with regular and irregular configurations, and different number of spans Longitudinal reinforcement: 8010 (rL = 1.05%) with an overlapping
and span length. Verification of the method is versus non-linear time- length equal to 20 diameters (200mm) at the base of the pier;
history analysis in terms of maximum deck displacement, and maxi- Transversal reinforcement: stirrups 6/150mm (rV = 0.38%);
mum pier shear forces has been carried out. Axial load equal to 1000kN ( = 4.3%) or 2000kN ( = 8.6%);
Concrete Rck400;
4.3 Task 3: Retrofit measures Steel FeB44K.
The experimental part of the research activity of Line 3 has been car-
ried out at the University of Pavia and of Roma Tre. The two experi- The retrofit intervention aimed to restore the tensile stress path from the
mental campaigns have focussed with different goals on the testing of pier section to the foundation, avoiding at the same time any plasticiza-
piers. The tests performed in Pavia were aimed at ascertain the effec- tion of the overlapping region. The new stress path created using longi-
tiveness of FRP retrofit measures in order to confine hollow-core piers tudinal FRP strips applied to the overlapping region is expected to
with insufficient lap-splices, while those performed in Roma Tre were cause the plastic hinge shift upwards where the longitudinal steel is
aimed, using large-scale specimens, at the characterisation of the well anchored allowing for an efficient energy dissipation.
response of frame piers built in the 60s. During the design phase different possible solutions have been consid-
ered concerning the retrofit materials (carbon, aramid or glass FRP), the
Finally, as it regards masonry bridges, a comprehensive survey of the retrofit geometry (width and length of the region to be retrofitted), the
existing retrofit techniques for this type of structures has been carried techniques to be used for the anchoring of the FRP strips to the foun-
out by the University of Genova. dation. This was possible employing a numerical FE model developed
to predict/reproduce the tests results.
4.3.1 EXPERIMENTAL ACTIVITY ON FRP STRENGTHENING FOR INSUFFICIENT Regarding the materials, the final choice was to use carbon FRP (C-
LAPSPLICE FRP): the analyses indicated that this material is the only one able to
Four 1:2 scaled bridge piers were designed with an insufficient over- sustain the acting tension forces. Too many FRP layers would have been
lapping length of the longitudinal bars across the critical zone that needed to carry the same force using glass or aramid fibres, affecting
should lead to an early loss of the lateral strength due to bar slippage. the effectiveness of the retrofit intervention.
The built specimens (see Figure 5 left and middle) have the following
characteristics: For what concern the geometry of the retrofit intervention, the final
Hollow-core rectangular cross-section (see Figure 5, right) with ex- solution was to apply longitudinally two C-FRP layers on the four sides
ternal dimensions 800x1500mm and wall thickness of 150mm; of the specimen. As far as the exploitation of the material strength is

Table 2. Considered retrofit materials


Material Description fu[MPa] e[MPa] eult[%] Layer thickness [mm]
SRP 3x2 High Density 1167 77773 1.50 1.1938
SRP 12x High Density 948 64811 1.46 1.1938
CFRP High Modulus 3000 390000 0.77 0.165
GFRP Alkali Resistant 1700 65000 2.62 0.23
AFRP High Modulus 2800 105000 2.67 0.214

RESEARCH - Seismic behavior 37


concerned this choice appears to be questionable since the fibres the performed tests: here the lateral load carrying capacity drops quite
applied to the pier sides parallel to the imposed motion will not have quickly because of the bars sliding. The red curves are the result of the
the same stress as those on the other two sides, but the adopted solution numerical model.
seemed to be the only possibility to assure the maximum stress diffu-
sion across the pier section. It is worth mentioning that even though 4.3.2 FINITE ELEMENT MODELLING CALIBRATED TO THE EXPERIMENTAL TESTS
anchoring 1500kN force to the foundation of the scaled specimen would RESULTS
have been probably feasible using a steel collar fixed to the foundation Given the large number of seismically under-designed bridges, that
with some high-strength steel bars taking advantage of the deep foun- need to be assessed and potentially retrofitted due to insufficient lap-
dation of the specimen, moving back to real structures the anchoring to splicing, the development of an efficient analytical model to simulate
the foundation of the tensile force induced in the FRP by a seismic the response of FRP-retrofitted elements was deemed critical. A quite
excitation would have been much more difficult, if not unfeasible. simple though effective finite element model was developed using
Spreading the tensile force on the four sides of the pier, the anchoring Seismostruct (Seismosoft, 2006). Figure 8 shows the adopted numeric
is clearly easier. Between different possibilities initially considered to model. The longitudinal FRP layers have been represented like an ele-
anchor such force, final choice was to use an anchoring system realised ment itself. Rigid links have been used to place the FRP at the right
with FRP too. The idea was to employ aramid connectors, normally distance from the longitudinal axis of the retrofitted member in order for
used to transfer shear stresses. If this solution will be found to be effec- them to be able to give the right contribution to the flexural strength of
tive, as it seems from its design, multiple advantages will arise both on the pier. Furthermore, each FRP element has pinned connection at both
the economic and technologic sides. ends in order to be subjected to pure axial load. On the other hand, the
FRP wrapping can be modelled in approximation without adding ele-
Due to external constraints only two piers have been tested within the ments to the model, since its main effect is the increased concrete con-
duration of the project, those without the FRP retrofit in the lap-splice finement that can be represented by the confinement factor already pre-
region. The tests confirmed that, as expected, lap-splice with an over- sent in adopted concrete stress-strain representation (Mander, 1988;
lapping length equal to 20 times the diameter of the spliced bars is Martinez-Ruenda and Elnashai, 1997). To tests the effectiveness of the
insufficient to assure the anchoring of the bars (see Figure 6a). The tests
also underlined that the effectiveness of the lap-splice decreases while
the axial load increases: that is because of the higher stresses and dam-
ages (such as partial concrete spalling) in the overlapping region.
Figure 7 shows the base shear-top displacement diagrams derived from

Fig. 8- FRP retrofitted pier model.


Fig. 6- Test Set-up (a) and open crack at the pier base during the test with 2000kN axial load (b).

adopted finite element model, the behaviour of 1:4 scaled square hol-
low section piers from previous experimental campaigns (Calvi et al.,
2005 and Pavese et al., 2004) has been reproduced through push-over
analysis.

4.3.3 EXPERIMENTAL ACTIVITY ON LARGE-SCALE SPECIMENS OF FRAME PIERS


Large-scale tests on framed piers have been undertaken at the
Universit of Roma Tre. This typology, characteristic of many old
Fig. 7- Base shear-top displacement diagrams of the two as-built pier (a) N = 1000kN (b) N = 2000kN. viaducts of the Italian highway system, has been chosen for its high

38 RESEARCH - Seismic behavior


Fig. 9- The viaduct Rio Torto.
beam, (see Figure 11a), a suitable grid of displacement transducers has
been placed on this beam, in order to measure the cracks amplitude, in
the other two specimens. In the second two mock-ups a different fail-
ure mechanism occurred on the same transverse beam. In particular,
during the second test both beam-column joints collapsed (see Figure
11b-c), while in the third one, only the left end of the beam failed in
shear (Figure 11d), with a simultaneous failure of the right beam-col-
umn joint (Figure 11e). This outcome was a nice experimental verifica-
tion of the effect of material fluctuation on determining which amongst
similarly resistant failure mechanisms actually occurs in reality.

Fig. 10- The viaduct Rio Torto.

seismic vulnerability. Among the representative bridges, a framed pier


from the Rio Torto viaduct has been chosen (see Figures 3, 4 and 9).
For the experimental program three mock-ups of pier 12 without retro-
fit have been realized and tested, with the goal of characterizing its
cyclic response and relative collapse mechanism (see Figure 10).
Subsequently, one or more reinforcing systems were meant to be
applied to the tested piers, for repeated testing to check the efficiency
and the reliability of the proposed reinforcing solutions.
A ductile flexural failure was predicted with formation of plastic hinges
for this pier while, on the contrary, all three specimens failed in shear,
either in the transverse beam or the joints, suggesting that the formula
employed for the evaluation of the shear strength tends to overestimate
the ultimate shear.

Since the first test has shown a premature shear failure of the transverse Fig. 12- Experimental force-displacement cycles.

a b c d e

Fig. 11- Failure mechanisms of the transverse beam in the three tests.

RESEARCH - Seismic behavior 39


4.4.1 A SIMPLIFIED NON LINEAR DYNAMIC MODEL FOR THE ANALYSIS OF
ABUTMENTS
A simplified model for the dynamic analysis of diaphragm walls retain-
ing dry cohesion-less soils with horizontal back-slope subjected to seis-
mic excitation has been developed (Franchin et al, 2007a). The model
is based on the well-known one-dimensional Winkler approximation
and on the non-linear shear-beam model for the ground layers on both
Fig. 13- left: Comparison between theoretical and numerical force-displacement curves; right: base column rotations.
sides of the wall (see Figure 14). The model can include anchor-ties and
can account for non-linearity in all of its elements (retained soil,
The differences between the failure mechanisms of the three piers, how-
anchors and wall). According to preliminary numerical applications,
ever, have a little influence on the global behaviour, as shown in Figure
which include validation of the proposed model results versus those of
12, which compares the global force-displacement cycles of the three
a refined plane-strain nonlinear finite-element analysis carried out with
specimens.
a commercial code, the model appears to yield quite accurate predic-
The experimental results have been compared with the results of a
tions of static and dynamic bending moment distributions and perma-
numerical model, which was set up using the non-linear code
nent wall displacements.
OpenSEES. Shear failure has been introduced using a shear force-
Next the developed model has been applied for the analysis of the
deformation relationship with a tri-linear backbone and an appropriate
response of a diaphragm abutment prior and after upgrading interven-
degradation law, included in a fiber non-linear element, using the sec-
tion with change of the support conditions and insertion of tie-backs
tion aggregator command. The yield-penetration at the base of the col-
(Franchin et al 2007b). The analysed structure is represented in Figure
umn effect is particularly relevant due to the presence of plain steel
15.
bars. This phenomenon, if neglected, can induce an overestimation of
The application of the model has shown its versatility in assessing the
the structural stiffness. This effect has been taken into account using a
system response in its existing state and in progressive states of upgrad-
zero-length element placed at the column base with a properly modify
stress-strain law of the steel bars. Finally, the buckling phenomenon
has been taken into account using a corrected constitutive law of steel.
The FE model used was able to reproduce accurately both the global as
well as the local behaviour, as shown in Figure 13.

4.4 Task 4: Abutments and foundations


The activity under this task has been carried out at University of Rome
La Sapienza, and has dealt with two distinct problems: a) the develop-
ment of an efficient non linear method for the analysis of diaphragm-
type abutments, free standing and retrofitted with tie-backs; b) the
review of the literature on soil-foundation-structure interaction with the
goal of providing detailed indications for practitioners to be included
into the assessment guidelines.
Fig. 15- Diaphragm abutment retrofitted with anchor ties.

Fig. 14- The developed model for diaphragm type. Fig. 16- Results of the abutment analysis: left, moment diagrams; right, top displacement time-histories.

40 RESEARCH - Seismic behavior


ing, in terms of both forces (Figure 16 left) and dynamic displacements soil-foundation system, consisting of the stiffness and damping func-
(Figure 16 right). To the extent that it has been validated at present, the tions (of the frequency) to be assigned at the pier base. This impedance
model represents a very efficient tool for realistic design and assess- includes the evaluation of the frequency dependent dynamic group
ment purposes. effect, i.e. the modification of the impedance obtained as a simple sum-
mation of the individual pile impedances to account for the interaction
4.4.2 CRITICAL REVIEW AND RECOMMENDATIONS ON METHODS FOR THE of the wave-fields produced by each pile.
ANALYSIS OF SOIL-FOUNDATION-STRUCTURE INTERACTION Evaluation of the response in the frequency domain. This has been
A comprehensive review of the literature on the treatment of the done both with a purpose-made code and with a commercial finite ele-
response of deep foundations has been carried out. This has led to iden- ment software that implements frequency-domain analysis (Sap2000).
tifying the available methods and their pros/cons. After the survey a The resulting power-spectrum of the displacement components can be
selection has been made of those procedures considered suitable for integrated to yield the root-mean-square (RMS) or standard deviation of
practical application and some numerical applications have been car- response, from which maxima to be used in verification are readily
ried out to assess the relevance of the phenomenon (input motion mod-
ification by kinematic interaction and foundation flexibility), in terms of TRANSLATION ROTATION
the response of the superstructure.
One example is the bridge structure shown in Figure 17. It is a simply-
supported prestressed concrete deck of span length 30.0m typical of the

STIFFNESS Kq
STIFFNESS

Italian highway construction practice of the 50s-70s with piers con-


sisting of a single-stem with hollow-core circular section. The dimen-
sions are in the figure. The foundation consists of a mat on 5 piles of
CAMPING Cq
CAMPING

Fig. 17- Simply-supported deck on single-stem hollow-core pier founded on piles. Fig. 18- Complex impedance at the pier base: stiffness (top), camping (bottom), translation (left) and rotation (right).

1.5m diameter. Pile length is 20m. The bridge has 6 spans and a pier obtained by multiplication for the peak factors.
of height 20m has been considered. Soil can be classified based on the
available information as type D. The structure has been modelled as Figure 18 shows the real (stiffness, top) and imaginary (damping, bot-
shown in Figure 17e, i.e. as a three-degree of freedom system (includ- tom) parts of the complex impedance at the base of the pier, for the
ing horizontal and rocking component of the base). translation (left) and rocking (right) displacement components. These
The analysis has been carried out in the frequency domain using the are reported for two different values of the shear wave velocity Vs, both
substructuring approach. The steps of the analysis include: compatible with the soil type D. The figure also reports the stiff-
Evaluation of the modification of the surface free-field motion (sup- ness/damping obtained by simple summation of the single pile contri-
plied as an acceleration response spectrum) due to the kinematic inter- butions. Comparing the latter with those of the group allows to appreci-
action between soil and pile group. This step provides the power spec- ate the frequency-dependent effect of the pile-to-pile interaction. This
trum of input displacement at the pier base. effect reduces, by more than 50% in this case, the total stiffness.
Evaluation of the complex frequency-dependent impedance of the Finally, Figure 19 shows the power spectral densities of the response in

RESEARCH - Seismic behavior 41


assessment. In this section only a limited overview of the applications
is given to illustrate the work done. A more detailed description can be
TOTAL DISPLACEMENT Us (m)

STRUCTURAL DEFLECTION (m)

found in the final report for the Line 3. Table 3 reports all the analysed
bridges.

4.6 Guidelines and Application manual


The activity only briefly summarised in the previous sections has rep-
resented a necessary support for undertaking the task of writing what
was the final product expected from Research Line 3: a proposal for a
Fig. 19- Results of SSI analysis on a bridge pier: power spectral densities of the response in terms of total displacement (left) guidance document on the seismic assessment of existing bridges, and
and structural deflection (right).
a companion set of example applications. The task, carried out by
terms, on the left, of total displacement (relative to input motion, i.e. University of Rome La Sapienza, has gone through several rounds of
sum of the foundation translation, the structure deflection and the trans- scrutiny by all the units. In its final version it represents the first
lation due to rigid foundation rotation), and on the right of the structur- European document on the topic and could be envisaged to form the
al deflection only. Results are reported for the two Vs values and, for basis for a future addition to the Eurocodes system. Indeed, the docu-
reference, for the fixed-base response. As it can be seen, as expected, ment is fully in line with Eurocodes and reflects to some extent the
the fundamental period of the system elongates considerably due to the experience on the seismic assessment of existing structures gained with
introduction of the foundation flexibility: it starts at T=0.83 s in the the use of Eurocode 8 Part 3 on buildings. It is also in line with the rel-
fixed base case, and reaches about 1.5s and 1.75s for Vs=200m/s and evant chapters of the DM2008, related to seismic design of bridges, and
100m/s, respectively. This increases the total displacements. The drifts, incorporates its most recent developments on the definition of seismic
however, are considerably reduced as shown in Figure 19b. action.
The above application, as well as the others carried out, allowed to
introduce in the guidelines quantitative indications on the need for The document produced consists of four chapters and two appendices:
inclusion of SSI into the modelling. Chapter 1: gives an introduction to the problem of seismic assessment
of existing bridges;
4.5 Task 5: Numerical application to case-studies Chapter 2: contains the guidelines;
All research units have contributed in producing a vast amount of case- Chapter 3: is an overview of the most common retrofit measures and
studies that have been of considerable usefulness in checking consis- criteria employed, without entering into the specifics of their design,
tency and practicality of the indications that now form the guidelines for making reference for this purpose to specialised texts on the topic;

Table 3. List of case-studies analysed according to the assessment guidelines


Unit Case-study Description Analysis
Torino Narbareto (PLS) 4 simply supp. spans, circular hollow-core piers Elastic RS analysis + q-factor
Torino Rio Barcalesa (PLS) 7 simply supp. spans, polygonal bi-cell. piers Elastic RS analysis + q-factor
Torino Borgotaro (PLS) Hollow-core slab deck, highly irregular plan, Figure 21 Elastic RS analysis + q-factor
Torino Rio Verde (PLS) 9 spans, steel deck with hollow-core RC piers Elastic RS analysis + q-factor
Torino Ramat (TBF) 5 spans, box-section, steel pier Elastic RS analysis
Chieti Vasto Marina (SS16) 15 spans, frame piers Elastic RS analysis + q-factor
Chieti Della Valle (A25) 10 spans, box-section, single-stem hollow-core piers Elastic RS analysis + q-factor
Roma Tre Rio Torto (A1FiBo) 13 spans, inelastic time-history analysis Inelastic time-history analysis
La Sapienza Rio Torto (A1FiBo) 13 spans Modal pushover analysis
La Sapienza Standard viaduct (E45) 5 spans, simply-supported, and continuous after section Simplified
Simplifiednon-linear
non-linearmethod,
method,pushover,
pushover,
widening
widening linear dynamic
linear dynamic
Cosenza Follone (A3) 4 simply supp. spans, retrofitted with link system Inelastic time-history analysis
Cosenza Val di Leto 5 simply supp. spans, retrofitted with oledynamic devices Inelastic time-history analysis

42 RESEARCH - Seismic behavior


Fig. 20- The Narbareto viaduct.

North carriageway South carriageway

Fig. 21- The Borgotaro viaduct.

Chapter 4: contains the numerical examples that illustrate the appli- composite steel-concrete deck, with and without seismic isolation;
cation of the methods presented in the guidelines. There are four appli- assessment of the Rio Torto viaduct by means of inelastic time-histo-
cations covering: ry analysis.
assessment, by means of the simplified non linear method, of a typi- Appendix A: presents the fundamentals of the response to multiple-
cal simply-supported bridge with single-stem cantilever piers in its pre- support excitations and reviews a number of methods that can be
sent state; employed to analyse bridge structures for this effect;
assessment, by means of linear and pushover analyses, of the previ- Appendix B: presents the fundamentals of the soil-foundation-struc-
ous bridge in two different configurations, with a new continuous, wider, ture interaction phenomenon and reviews a number of methods that can

RESEARCH - Seismic behavior 43


STATIC SCHEME

Fig. 22- The Della Valle viaduct.

RETROFIT VIA STEEL BARS AND NEOPRENE PADS

Fig. 23- The Follone viaduct.

44 RESEARCH - Seismic behavior


be employed to analyse bridge structures for this effect. the possibility of defining the concept of residual reference life.
Though it is admitted that in our Country it seldom occurs that the deci-
The main body of the manual is represented by chapters 2 and 4, as sion to demolish a bridge can be taken several years in advance, it may
well as by the appendices. In the following the most significant or prob- happen that, due to planned substantial modification of the traffic
lematic aspects are briefly reviewed and commented. capacity of the link, it will be economically more convenient at a future
date to replace the bridge. In this case, if seismic upgrade must be
4.6.1 CHAPTER 2, GUIDELINES: DEFINITION OF THE SEISMIC ACTION undertaken, the concept of residual reference life may be invoked to
The seismic action is defined, in line with DM2008, by means of an assign to VR a more realistic reduced value. This possibility is not cur-
elastic acceleration or displacement response spectrum characterized rently included in the guidelines, though it is regarded as being in line
by an average return period specified as a function of the limit state of with the possibility allowed for existing structures to derogate from
interest. standard safety levels dictated for new structures.

The return period TR is obtained from the probability of exceedance 4.6.2 CHAPTER 2, GUIDELINES: METHODS OF ANALYSIS
PVR over the reference life VR. The latter is given in DM2008 as the With respect to classification of methods in static and dynamic, linear
product of two factors, the nominal life VN and the use factor CU. The and non linear, now common to all modern seismic design codes and
minima for PVR for each limit state are given in DM2008. giving rise to the usual four alternatives, the guidelines restrict some-
what the field of applicability of linear analysis. This is not unexpected.
In the tentative applications of the guidelines it was raised the problem For new well-designed structures the role of analysis is a relatively
of the value to be attributed to VN and CU, especially with reference to minor one, due to the many constraints (arising mainly from global and
the first one. The uncertainty may arise in the choice between 50 and local capacity design) that guide the design. On the other hand, when
100 years for VN, when considering bridges over highways. The assessing an existing structure, the accuracy in the analysis may have
DM2008 indicates 50 years for bridges of ordinary dimensions, typolo- a major economic impact on the retrofit, possibly avoiding it altogether.
gy and importance, and 100 years for bridges of large dimensions and
strategic importance. One would then be probably directed towards The guidelines admit linear analysis of two types only: modal analysis
100 years, in consideration of the importance of the bridge (it is on a with unreduced elastic spectrum and verifications in terms of deforma-
highway). The next choice is that of CU which leads unambiguously to tion/forces (subject to stringent conditions on the response regularity),
2.0, since highways are roads of type A according to the Italian classi- and modal analysis with a spectrum reduced by a limited value of the
fication of roads (i.e. considering, again, the functional importance of behaviour factor of q=1.5.
the road on which the bridge is located). The above choices would imply The main methods put forward by the guidelines are non linear static
a reference life of 200 years and, for the life-safety limit state, a TR of and dynamic analyses. As already anticipated in 4.2.1, a simplified
about 2000 years. It is observed that this conclusion would not to be in non linear static method is proposed for the very frequent case of
line with the safety criteria contained in Eurocode 8 Part 2 (Bridges) bridges with simply supported decks. For continuous irregular bridges
which indicates for highway bridges an importance factor I=1.3 to be the use of more recent pushover variants (adaptive and/or multi-mode)
applied to the action with TR = 475 years. This multiplication leads in is introduced as an alternative to full-fledged inelastic time-history
most of Italy to an action with a return period of about a 1000 years. analysis. The allowance for more than single-mode invariant pushover
This latter in turn is consistent with a reference life of about a 100 represents a small step forward with respect to Eurocode 8 Part 2,
years, which is also the design life specified in the Eurocodes for other which builds upon the results of recent wide-ranging studies on the per-
actions (e.g. corrosion). formance of such methods in the analysis of bridges [see for ex.
(Casarotti 2005), (Kappos et al, 2005), (Isakovic and Fischinger, 2005),
An official response to the mentioned problem, whose relevance needs (Lupoi et al, 2007), (fib, 2007), as well as the draft document Inelastic
not to be underlined, cannot but come from the competent authorities, methods for seismic design and assessment of bridges by Task Group
which are in charge of choosing the safety levels. 11 of the European Association of Earthquake Engineering].

Within the framework of the definition of the reference life one aspect 4.6.3 CHAPTER 2, GUIDELINES: SAFETY VERIFICATIONS
that deserves particular consideration in the case of existing bridges is The guidelines introduce a format for bi-directional verification for both

RESEARCH - Seismic behavior 45


deformations and forces. In particular the format reads: dent motions at the supports representative of the local soil conditions,

( )( )
Dx 2 Dy 2 which can be applied using currently available commercial finite ele-
+ 1 (7)
Cx Cy ment codes (Monti and Pinto, 1998);

where Dx and Dy denote the demand quantities along the two orthogo- For what concerns soil-foundation-structure interaction the guidelines
nal axes x and y, with Cx and Cy denoting the corresponding capacities. give a classification of the approaches and present with some detail the
This format becomes, in terms of chord rotations and shear forces: substructuring method, in its application to pile (Novak 1974, Makris
 and Gazetas, 1991 and 1992) and caisson foundations (Gerolymos and
2 2
(x/u,x) + (y/u,y) 1 (8)
Gazetas, 2006a,b). In this method the structure and the soil-foundation
(Vx/Vu,x) + (Vy/Vu,y) 1
2 2

(9) system are separated and studied accordingly. The study of the soil-
foundation system consists of the solution of so-called kinematic inter-
In the above equations the demand terms are understood as the com-
bined effect of both orthogonal components of the seismic action. For action and inertial interaction problems, leading to the modified input
example, with reference to chord rotation, for the case of multi-modal motion for the structure and to (complex) impedance to be put at the
non linear static analysis one has: structure base, respectively. Then the structure is analysed, with a flex-


x=xG [(
N
2 2
ible support condition, under the previously determined modified
i=1 xEx,iG) + (xEy,iG) ] (10)
motion. All the formulas necessary to perform this procedure are pre-
where the directional combination is of the SRSS type and the summa- sent in the Appendix.
tion is over the modes.
5. DISCUSSION
4.6.4 APPENDICES
The matter covered in these two appendices, i.e. the response of bridge The main objective of the project, which was the production of the draft
structures to different motions at the piers bases and the effect of the guidelines and their application manual, has been met. In this respect
soil-foundation system deformability in modifying the input motion as the Research Line was successful, since the product has been delivered
well as the response of the structure, has been always mentioned in and its quality is believed to be high.
codes without, however, neither precise quantitative indications on the Though it wasnt explicitly included into the remit for the Line, it must
instances in which these phenomena have to be accounted for, nor of be noted that the research group initially intended to cover in the guide-
physically sensible yet practically applicable methods to do it. The rea- lines both structural concrete and masonry bridges. In spite of the
son for this resides clearly in the insufficient advancement on basic research carried out, however, this more ambitious goal could not be
research. In drafting the guidelines, however, it was considered appro- achieved.
priate to provide a presentation of selected state-of-the-art approaches Research on this front was essentially under the responsibility of the
which are susceptible of practical application. unit of Genova. This unit has produced during the three years of the
project a considerable amount of high-quality research that has been
For what concerns the effect of multiple-support excitation, the guide- regularly documented in the annual as well as the final reports, and it
lines indicate that the phenomenon should be accounted for whenever is also available in research reports from the unit uploaded on the pro-
soil conditions along the bridge belong to different soil categories. The ject website. Quoting from the final report the issues dealt with by the
guidelines also present: unit cover the following: i) statistical characterization of the Italian
A stochastic model of the motion at the supports (Der Kiureghian, bridge population; ii) mechanical models for solid clay brickwork,
1996) that can be used either to generate samples of correlated motions needed for detailed and simplified structural models; iii) in field testing
to be used in time-history response analysis or in linear random vibra- of masonry bridges, aiming at the identification of the main mechanical
tion analysis; properties of the materials and of the bridge as a whole; iv) laboratory
The multiple response spectrum method (Der Kiureghian and testing of brickwork prisms; v) reduced scale testing aiming at identify-
Neuenhofer, 1992), which provides a solution for the random vibrations ing the load carrying capacity and the collapse mechanisms of shallow
problem of a system subjected to multiple inputs based on the use of the and deep arches taking into account the fundamental collaboration of
corresponding input displacement response spectra; the so called non structural elements; vi) reduced scale testing aim-
A simplified proposal for time-history analysis employing indepen- ing at identifying the dynamic properties of shallow and deep arches

46 RESEARCH - Seismic behavior


taking into account the collaboration of the so called non structural tion of expected loss related to any given bridge. Looking now at the
elements; vii) Limit Analysis procedures for the analysis of masonry problem of bridge protection from an higher perspective, the attention
bridges taking into account the contribution of all the bridge elements; should be directed at the bridges as components of road links forming
viii) retrofitting techniques for the bridge and its components. As it a transportation infrastructure. The seismic performance of the single
may be seen all of the investigated topics are of clear scientific interest, bridge would then be put in relation with the performance of all other
though not specifically relevant to seismic assessment of bridges. This bridges to be able to estimate the overall decrease in functionality of the
is simply the unavoidable consequence of the international lack of fun- whole infrastructure. In this respect the very challenging problem of
damental knowledge on the seismic behaviour of masonry bridges. determining the loss in traffic capacity of a damaged bridge represents
an essential element.
6. VISIONS AND DEVELOPMENTS
7. MAIN REFERENCES
The research carried out within Research Line 3 has included the state-
of-the-art into a document usable for assessing the protection level of - Casarotti C. (2005), Adaptive pushover-based methods for seismic
bridges against a number of limit-states. assessment and design of bridge structures PhD thesis ROSE School,
There are certainly several areas where improvement is possible and Pavia, Italy.
desirable, and in particular these are: - CEN (2005), Eurocode 8 Part 2: Seismic design of bridges
The non-linear static analysis for bridges of complex geometry; European Committee for Standardization, Brussels, Belgium.
The ultimate strength and deformation capacity of structural members - CEN (2005) Eurocode 8 Part 3: Assessment and retrofitting of exist-
such as those encountered in bridge structures (e.g. polygonal multi- ing structures European Committee for Standardization, Brussels,
cell. hollow-core cross-sections) Belgium.
The generation of ground motions for multiple-support excitation. - Chopra A.K., Goel R.K. (2002), A modal pushover analysis proce-
While generated motions are being progressively replaced with record- dure for estimating seismic demands for buildings Earthquake
ed ones for the analysis of buildings, their use appears unavoidable for Engineering & Structural Dynamics Vol 31(3), pp. 561-582.
the analysis of bridges whenever different motions must be considered - Der Kiureghian, A. (1996), A coherency model for spatially varying
at the supports. Currently available procedures are in need of consider- ground motions Earthq. Eng. & Struct. Dyn. Vol. 25, pp. 99-111.
able improvement. Der Kiureghian, A., Neuenhofer, A. (1992), Response spectrum
The vast literature on SSI needs to be acquired and digested by struc- method for multi-support seismic excitations Earthq. Eng. & Struct.
tural engineers to become a practical tool. This a crucial aspect in view Dyn., 21: 713-740
of the displacement-based framework of the guidelines and the corre- - DM2008 (2008), Nuove norme tecniche per le costruzioni Decreto
sponding need for more accurate evaluations of deformations. Ministeriale del Ministero delle Infrastrutture 14/1/2008.
- FHWA (1995), Seismic Retrofitting Manual for Highway Bridges,
The guidelines do not cover the seismic isolation technique. The reason Publ. FHWA-RD-94 052, Federal Highway Administration.
for this choice is that the design of seismic isolation does not vary - FHWA-ATC (1983), Retrofitting guidelines for Highway Bridges
between new and existing bridges. Seismic isolation, however, will cer- Report ATC-06-2, Applied Technology Council, Redwood City,
tainly see much further diffusion in the coming years, for new as well as California.
existing bridges, while isolation device technology continues to evolve - FHWA-MCEER (2006), Seismic retrofitting manual for Highway
rapidly with the ensuing need of developing appropriate analysis and Structures. Part 1- Bridges.
design techniques. In this respect this can be regarded as an ongoing - fib (2007) Seismic bridge design and retrofit structural solutions
research topic. Bulletin 39, International Federation for Structural Concrete.
- Franchin P., Pinto P.E., Noto F (2007), A nonlinear dynamic model
To the extent that solutions to the problem of assessing the protection of for seismic analysis of earth-retaining diaphragm-walls Proc 4th Int.
a bridge against its ultimate state can be considered to be sufficiently Conf. Earthquake Geotech. Engng, Thessaloniki, Greece.
mature, the next important passage is that of being able of estimating - Franchin P., Pinto P.E. (2007), Analysis Of Diaphragm-Type Bridge
structural and monetary damage as a continuous discrete function of Abutments Before And After Seismic Upgrading Proc 1st US-Italy
seismic intensity. Achievement of this goal would allow for the estima- workshop on seismic design and assessment of bridges, Pavia, Italy.

RESEARCH - Seismic behavior 47


- Gerolymos N., Gazetas G. (2006), Winkler model for lateral response III - INNOVATIVE MATERIALS FOR THE VULNERABILITY
of rigid caisson foundations in linear soil Soil Dyn. & Earthq. Engng MITIGATION OF EXISTING STRUCTURES
Vol.26, pp. 347-361.
- Gerolymos N., Gazetas G. (2006), Development of Winkler model for 1. INTRODUCTION
static and dynamic response of caisson foundations with soil and inter-
face nonlinearities Soil Dyn. & Earthq. Engng Vol.26: 363-376. The use of fiber reinforced polymer (FRP) materials for the strengthening
- Isakovic T., Fischinger M. (2005), Higher modes in simplified inelas- of masonry and concrete structures, represents a valid alternative to tra-
tic seismic analysis of single-column bent viaducts, Structural ditional techniques. Indeed, many advantages are provided by using
Engineering International. FRPs: lightweight, good mechanical properties, corrosion-resistant, etc.
- Kappos A.S., Paraskeva T.S., Sextos A.G. (2005), Modal pushover In Italy, the use of FRP materials for reducing the seismic vulnerability
analysis as a means for the seismic assessment of bridge structures of existing structures has been allowed for the first time through O.P.C.M.
Proc. 4th European workshop on the Seismic behaviour of irregular and 3274 and more recently by the D.M. 14.01.2008, that refer to the Italian
complex structures, Thessaloniki, Greece (Paper 49). National Research Council Design Guidelines (CNR-DT 200/2004) for
- Lupoi A., Franchin P., Pinto P.E. (2007), Further probing of the suit- the external strengthening of existing structures with FRP materials.
ability of push-over analysis for the seismic assessment of bridge struc- These guidelines provide, within the framework of the Italian regula-
tures Proc. of COMPDYN07, Crete, Greece tions, a document for the design and construction of externally bonded
- Makris N., Gazetas G. (1991), Dynamic pile-soil-pile interaction. FRP systems for the strengthening of existing structures. In particular,
Part I: Analysis of Axial Vibration Earthq. Eng. & Struct. Dyn. Vol.20, several issues concerning the seismic rehabilitation of Reinforced
pp115-132. Concrete (RC) and masonry buildings have already been dealt but a fur-
- Makris N., Gazetas G. (1992), Dynamic pile-soil-pile interaction. ther investigation is still required.
Part II: lateral and seismic response Earthq. Eng. & Struct. Dyn. Within this context, the main aim of this research line has been the
Vol.21, pp 145-162. experimental validation of design indications provided by the CNR-DT
- Monti G., Pinto P.E. (1998), Effects of multi-support excitation on 200/2004 guidelines.
isolated bridges Tech. Rep. MCEER 98/0015 pp. 225-247. The main topics investigated in this research task can be summarized
- Novak M. (1974), Dynamic stiffness and damping of piles Canadian as follows:
Geotech. Jnl Vol.11: 574-598. - the mechanical behaviour of FRP materials;
- the cyclic behaviour of RC elements strengthened by means of FRP;
- the mechanical and chemical anchorage devices for FRP systems;
- the ductility increasing of RC columns confined with FRP;
- the RC joint strengthening with FRP;
- the masonry strengthening with FRP;
- the historical structure strengthening with FRP;
- the quality control and monitoring of FRP applications;
- the innovative fibers (steel fabrics, natural fibers, FRP grids, etc.) and
matrices (organic and inorganic);
- the mechanical behaviour of concrete structures reinforced with fiber
reinforced polymer bars;
- innovative strengthening techniques (near-surface mounted (NSM)
technique, FRP prestressed systems).

2. BACKGROUND AND MOTIVATION

The research activity has been performed through experimental tests


and theoretical studies mainly devoted to the development of simple
methods of analysis and design rules in order to improve the indications

48 RESEARCH - Seismic behavior


provided by CNR-DT200-2004. interaction force between FRP materials and the masonry substrate,
FRP are ideal products for structural retrofitting and seismic upgrading. experimental tests are needed. In this research line, semi-destructive
Nonetheless the small knowledge on the durability of the system is one and non-destructive techniques have been also investigated for the
of the main drawbacks to the use of FRP reinforcement in Civil quality control and monitoring of FRP applications to masonry struc-
Engineering. In particular, structural adhesives usually represent the tures, according to CNR DT 200/2004 Guidelines.
weakest point of the reinforced system and their mechanical behaviour
and durability performance need to be investigated. The first problem 3. RESEARCH STRUCTURE
in using composite materials for structural reinforcement is the deter-
mination of their mechanical properties. The bond between FRP and In order to guarantee an optimal organization of the research, the
concrete is a very important issue because the debonding is a very brit- Research Units have been grouped into the following ten Tasks, each
tle failure mechanism and must be avoided. one with a specific topic:
According to performance-based design or seismic evaluation of RC - Task 8.1: the mechanical characterization of FRP systems at fixed
buildings, it is crucial to provide a correct evaluation of the strength and environmental conditions under cyclic actions;
ductility capacity of the RC columns and beams as well as of beam-col- - Task 8.2: the delamination under cyclic actions and design of anchor-
umn joint. Experimental and direct observation of damages occurred age mechanical devices for FRP systems;
during recent earthquakes strongly highlighted this need. - Task 8.3: the confinement of RC and masonry columns subject to com-
The effectiveness of FRP systems for seismic vulnerability mitigation of bined flexure;
masonry structures is still in debate, despite it has moved a huge inter- - Task 8.4: the strengthening in flexure and in shear of RC structural
est, becoming the outstanding system in the market for this type of elements with FRP fabrics and near surface mounted (NSM) rods;
applications. - Task 8.5: the beam-column and beam-foundation joint reinforcement
Indeed CNR DT 200/2004 has been the first guideline to provide with FRP;
design criteria for the FRP seismic strengthening of masonry buildings. - Task 8.6: the design criteria for the seismic retrofit of RC and RC-
However, the retrofit design of masonry structures is still not a com- masonry composite structures with FRP;
pletely solved problem. This is due to the fact that the masonry struc- - Task 8.7: the design criteria for the seismic retrofit of masonry struc-
ture is load dependent and thus the FRP could be placed in an inactive tures with FRP;
area of the resistant mechanism. Furthermore, masonry can activate a - Task 8.8: the strengthening of masonry structural elements with FRP
large number of local mechanisms which interact with global behaviour systems;
of masonry buildings. The non linear seismic assessment of FRP rein- - Task 8.9: the strengthening of masonry vaulted elements with FRP
forced masonry structures is included also in the D.M. 2008 rule. The systems;
non linear analysis requires the knowledge of the constitutive law of the - Task 8.10: the quality control and monitoring of FRP applications to
masonry material both in the unreinforced or strengthened situations. existing masonry and RC structures.
In the recent years, the scientific research has been focused on the safe-
guard of historical buildings. Accordingly, CNR DT 200/2004 has been 4. MAIN RESULTS
published in order to provide design criteria for the use of FRP systems
for strengthening existing structures and to avoid their incorrect appli- 4.1 Task 8.1: the mechanical characterization of FRP systems at fixed
cation. The Guidelines deal with different types of FRP applications to environmental conditions under cyclic actions
masonry and reinforced concrete structures and take into account the The aim of the sub-task was to mechanically characterize FRP systems.
important phases of quality control and monitoring that should follow a The study has been focused on durability and mechanical behaviour of
strengthening application. Several aspects affect the effectiveness of structural adhesives and FRPs, the mechanical characterization of FRP
FRP systems such as the surface preparation and FRP installation. bars and strips, the values of the safety factors proposed in CNR DT
Moreover, once FRP strengthening intervention has been carried out, 200/2004 and the effects of elevated temperatures and freeze-thaw
monitoring by non-destructive or semi-destructive tests should be per- cycling on FRP.
formed to ensure the quality and effectiveness of the strengthening sys-
tem. It is worth noting that due to the increased number of composite Durability and mechanical behaviour of structural adhesives and FRPs
material applications and in order to get a better understanding of the Several tests to determine the mechanical properties of composite mate-

RESEARCH - Seismic behavior 49


Fig. 1- a) Execution of the punch-tool test and specimen after collapse, b) execution of the torsion test and specimen after
collapse.
Fig. 2- a) Anchor system for large diameter GFRP bars, b) numerical analysis.

rials and structural adhesives have been performed. Conforming to the riences for testing steel ropes and prestressing steel tendons and the
ASTM requirements, the glass transition temperature (ASTM D3418), shape of the resin head from test investigation. Numerical analyses
porosimetry and the coefficient of thermal expansion (ASTM D360) were also performed to investigate the effects of anchor parameters such
were determined. Adhesives were also tested under tensile (ASTM as cone slope angle, thickness of resin head and friction coefficient
D360), compressive (ASTM D695) and flexural loading (ASTM D790). between the anchor body and the resin head. Pull-out and beam tests
Adhesive shear strength was determined by punch tool tests (ASTM were also executed.
D732). Finally adhesive cylinder specimens were tested under pure tor-
sion load. Experimental studies and numerical analyses were developed to define
practical tests for the characterization of FRPs and adhesives mechan-
Adhesive dumb-bell specimens were prepared for tensile testing and ical properties. The main aim of this action was to provide the
then artificially aged in an environmental chamber in order to analyze Composites Kit Test - COKIT; a practical tool for professionals and
possible detrimental effects on the adhesive mechanical properties. engineers operating in the field of FRPs applications and dealing with
Exposition to deicing salts, freeze-thaw cycles and moisture may in fact FRP materials for structural retrofitting and rehabilitation. The techni-
deteriorate the mechanical properties with consequences on the dura- cal document Istruzioni per la caratterizzazione ed il controllo di
bility performances of strengthened structures. Tensile tests were per- accettazione di materiali fibrorinforzati per il rinforzo strutturale
formed conforming to the requirements of ASTM D360. In all the con- COKIT was thus published and could represent an annex of the CNR
ditioning treatments, significant losses in adhesive stiffness and tensile DT 200/2004 Recommendations.
strength were measured. The stiffness and tensile strength reductions
after exposure to salt spray fog solution may be approximated by Refinement of the safety factors proposed in Design Recommendations
straight parallel lines as described in the Arrhenius life-temperature
The environmental conversion factors provided in the guidelines of the
relationship. Fatigue tests on adhesive dumb-bell specimens were
Italian National Research Council (DT200) were analyzed on the basis
finally performed to attain the fatigue failure curves for the adhesive
of the results of artificially aged adhesive specimens tested under ten-
joint.
sion. Exposition to deicing salts, freeze-thaw cycles and moisture leads
Mechanical characterization of FRP bars and strips

Tensile and relaxation tests were performed on FRP bars with particu-
lar attention to the gripping system. Then, experimental tests and
numerical simulations were performed to develop simple, economical
and effective systems for the characterization of composite materials
and adhesives. In particular, an anchor system for tension testing of uni-
directional fiber reinforced plastic (FRP) bars of large diameter was
developed. In the system suggested each end of the bar is embedded in
a conical polymeric head that fits a conical hole inside the anchoring Fig. 3- Stiffness and tensile strength retention for the structural adhesive subject to freeze-thaw cycles of five hours each
between 18 and +4C for a total duration of about 2 months (FT), to salt spray fog for one month or three months (SF)
device. In the anchor system, the anchor body shape came from expe- and to one month humidity (HU).

50 RESEARCH - Seismic behavior


Effects of elevated temperatures and freeze-thaw cycling on FRP lami-
nates behavior

The performances at elevated temperatures and/or at freeze-thaw


cycling exposure of structural members strengthened by using exter-
nally bonded FRP laminates are mainly related to two aspects: the bond
behaviour between FRP and the member substrate; the mechanical
properties of laminates themselves. The latter aspect has been very lim-
ited experimentally investigated; only few tests have been performed to
evaluate the residual tension strength of FRP coupons after exposure to
Fig. 4- a) Steel-CFRP specimen b) Reduction in stiffness of retrofitted specimens during fatigue tests; c) S-N curve and
comparison between the fatigue resistance of the steel-CFRP bond for a stiffness reduction of 5% (blue circles) and of 15%
elevated temperatures or freeze-thaw cycling. Thus, experimental ten-
(red squares) and that of EC3 welded detail categories.
sion tests on carbon FRP (CFRP) laminates both under controlled tem-
perature and relative humidity conditions or after freeze-thaw cycles
to the deterioration of the mechanical properties of composite materials exposure have been carried out. In particular, due to reduced capacity
and in particular structural adhesives. On the basis of the experimental that commercially available resins have to transfer loads over fibres
results, the safety factors suggested in the CNR DT 200/2004 recom- around glass transition temperature, Tg, two new systems based on
mendations may be considered as appropriate, but in aggressive envi- epoxy resin have been formulated and characterized by dynamic
ronments the use of a slightly lower conversion factor seems to be more mechanical analysis (DMA). The main goal of the new formulated sys-
suitable. tems was to increase Tg, the elastic modulus in the rubber region of the
resin and to improve their performances under freeze-thaw cycles. Two
Tests were performed to refine the safety factor of FRP-steel systems: different approaches were investigated. First a new epoxy system
the fatigue behaviour of steel structures retrofitted by using FRP mate- (namely neat epoxy) was formulated and cured at 60C after an hour at
rials was investigated, SN curves were defined and the fatigue resis- room temperature. Secondly, in order to improve the mechanical prop-
tance of the steel-CFRP bond was compared to the one of welded detail erties of epoxy matrix by curing at room temperature, a nanocomposite
categories described in the Eurocode 3. system was obtained by direct dispersion of preformed nanodimen-
sioned silica particles to the neat epoxy resin.
After performing pull-pull delamination tests on FRP-concrete speci-
mens, cylinders were obtained from each concrete prism. Based on
Eurocode 2 compressive and splitting tests were carried out to deter-
mine the conditioning effects on concrete degradation. As a conse-
quence of the environmental conditioning, concrete characteristic
strength is assumed to increase by 16% for salt spray fog conditioned
specimens and to decrease by 3% for specimens subject to freeze-thaw
cycles.

Fig. 6- Specimen geometry; Temperature and relative humidity exposure profiles; test setup.

The experimental results point out that the developed formulations of


epoxy resins provide a significant increase of ultimate strength and
strain of CFRP coupons both at room and elevated temperatures with
respect to commercial systems, without significant change of the elastic
modulus. Negligible influence of a low number of freeze-thaw cycles
was observed on the mechanical properties of coupons independently of
matrices. Experimental outcomes strongly confirmed that the use of
matrices characterized by higher values of Tg and elastic modulus in the
Fig. 5- Statistical distribution of the coefficient kG for specimens subject to a) salt spray fog and b) freeze-thaw cycles. rubber region with respect to those traditionally available on the mar-

RESEARCH - Seismic behavior 51


ket, could allow to overcome one of the main limit of FRP laminates single shear push-pull test. All the tests have been performed under
related to their poor performances under elevated temperatures. displacement control of the FRP free end. In order to evaluate the vari-
ability of the results when different set-ups are adopted, the coefficient
4.2 Task 8.2: the delamination under cyclic actions and design of of variation (COV) for each set of homogeneous experimental tests has
anchorage mechanical devices for FRP systems been calculated. The scatter of the results is in general small (COV
Different experimental set-ups can be found in the scientific literature about 10%), lower than that of the tension strength of the concrete, usu-
dealing with FRP-concrete bond tests and it has been observed that dif- ally equal to 20-30%. For the plates, the scatter of the results is simi-
ferent test methodologies may give different values of the debonding lar for the different Labs, whilst for the sheets the dispersion is usually
force. This task research intended to define a standard FRP-concrete higher. The results obtained by Lab 3 are very stable in both cases and
bond test to be used to evaluate the maximum transmissible force by an close to the mean values. This study allowed to define a set of rules for
FRP anchorage, to be included in the new version of the Italian code for the standardization of bond tests to be used to evaluate the maximum
design of strengthening interventions with FRP. transmissible force by an FRP concrete anchorage.

Experimental Round Robin test on FRP concrete bonding Cyclic tests of FRP-concrete debonding under cyclic loadings

An extensive experimental campaign on FRP-concrete debonding has Many strengthened structures are subjected to fatigue loads (i.e. roads
been carried out by five different Italian Laboratories (University of and railways bridges) or to shorter but more intense cyclic actions as
Bologna, University of Naples Federico II, University of Sannio, seism: in this cases, the FRP-concrete interface is subject to cyclic
Polytechnic of Milan and University of Calabria). The tests were devot- stress regimes which can lead to premature debonding of the FRP lam-
ed to the definition of a standard test procedure for the bond strength inate from the concrete substrate and cause the FRP failure in most
evaluation. According to the Round Robin procedure, 50 concrete cases, unless appropriate local measures are taken to prevent it.
prisms (same batch) strengthened with CFRP plates and sheets have In order to develop a more economical design for FRP-strengthened
been prepared by the same operator and subject to bonds test in five structures, research line investigated the debonding phenomenon of the
different Laboratories. The sets of homogeneous specimens have then FRP reinforcements (both plates and sheets) from the concrete sub-
been subject to bond test by five laboratories of the University partners strate under cyclic actions. In particular, tests on little prismatic speci-
using different test set-ups (Figure 7). mens have been performed by applying FRP plates and sheets on con-
crete prisms and testing them under both monotonic and cyclic actions
Twelve specimens (6 strengthened with sheets, 6 strengthened with without inversion of sign.
plates), with two different bonded lengths (100 mm and 400 mm), have The experimental results of Single Shear Test (SST) performed on CFRP
been tested by each laboratory, repeating three times the same type of reinforcement applied on little prismatic concrete specimens and char-
test. As for the test set-ups (Figure 7), all the Laboratories adopted a acterized by high bond length values (400mm) showed that:
- the influence of load-unload cycles up to 70% of Pmax,M was negligi-
ble for CFRP sheets and plates;
- a low number of load-unload cycles (40) up to 90% of Pmax,M reduced
the debonding load of about 10% in the case of CFRP plates but did
not affect particularly the bonding behaviour of CFRP sheets;
- by increasing the number of load-unload cycles (up to 300) between
70% and 90% of Pmax,M, the debonding load of concrete specimens
reinforced with CFRP sheets decreased by a percentage factor equal
to 10%;
- the transfer of shear stresses at the FRP-to-concrete interface due to
the CFRP reinforcement bond length larger than effective one,
allowed to mitigate noticeably the effect of cyclic actions imposed up
LAB4 to 90% of Pmax,M;
Fig. 7- Experimental set-ups adopted by the five different Laboratories. - a degradation of interface behaviour has been recorded after the onset

52 RESEARCH - Seismic behavior


of debonding, with reduction of maximum shear stress;
- the effects of cyclic actions were more significant on plates rather than
sheets and its influence increased with number of cycles;
Moreover similar SST tests performed on CFRP reinforcements charac-
terized by lower bond length values (50-250mm) allowed to observe
that:
- design relationships provided by Teng et Al. and by main internation-
al codes for evaluating the effective bond length values are conserva-
tive for both sheet and plate reinforcements;
- referring to plates the effective bond length values, experimentally
evaluated by means of the monotonic tests, were noticeably lower than
predicted; Fig. 8- Debonding force for sheets with anchorage length (a) L=400 mm, (b) L=100 mm, and for plates with anchorage
length (c) L=400 mm, (d) L=100 mm.
- even if reinforcement bond length values were significantly low, the
cyclic tests outcomes confirmed that the influence of load-unload
cycles up to 70% of Pmax,M was negligible for CFRP sheets and plates
due to elastic behaviour characterizing FRP-concrete interface up to
such load level;
- Also a further low number of load-unload cycles (10) up to 90% of
Pmax,M did not affect particularly the bonding capacity of CFRP rein- Fig. 9- Coefficient of variation (COV) of the debonding force for (a) sheets and (b) plates.

forcement due to the transfer of shear stresses at the FRP-to-concrete


interface. Such transfer was more significant on plates rather than ratio; applying different FRP reinforcement ratios and checking the
sheets: nevertheless, bond lengths particularly conservative for plates confinement sensitivity to the number of plies.
allowed to better mitigate cyclic action effects;
- in order to better predict design bond length values, using two differ- Tested specimens represent real scale building columns designed
ent relationships for sheets and plates, respectively, could be worth- according to dated codes for gravity loads only. The design concrete
while. strength is 23.1 MPa to simulate concrete mixes used in past decades.
Concrete cylinder specimens per each casting have been prepared in
4.3 Task 8.3: the confinement of RC and masonry columns subject to order to characterize the concrete with standard procedures. The used
combined flexure steel is characterized by a yield strength of 414 MPa and a modulus of
The main goal of this research task has been to validate the design elasticity of 200 GPa. Particular care has been devoted to construction
equations provided by CNR DT200 for the confinement of RC and details, namely: hooks, longitudinal and transverse steel reinforcement
masonry members. In particular, experimental tests have been carried ratios and concrete cover specifications. Special care has been taken to
out on both real scale and scaled columns wrapped by using tradition- avoid local failure at the top and the bottom ends of the columns plac-
al FRP (CFRP and GFRP) or an innovative typology of FRP system ing steel ties with reduced spacing. Internal steel (bars and ties) rein-
made of basalt material fiber. forcement ratio is the same for each group of specimens, designed per
minimum code requirements. The minimum specimen dimensions are
Confinement of real scale RC columns subject to axial load 360 x 510 mm2. The load has been applied concentrically under a dis-
placement control rate. The load has been conducted in five cycles in
An experimental campaign has been carried out on full scale reinforced increments of one fifth of the expected capacity for each specimen.
concrete (RC) columns concentrically loaded and confined by means of Each loading-unloading cycle has been repeated once. Strain gages,
FRP (Glass FRP and Basalt FRP). Five series of tests were planned, for potentiometers and LVDTs are used for strain and displacement data
each series a reference unconfined column was tested and used as acquisition. In particular, strain gages have been applied on column
benchmark. The test matrix has been designed to assess the confine- surface and internal bars whereas LVDTs have been placed in order to
ment effectiveness: applying the same reinforcement ratio and check- obtain vertical and horizontal column displacement. Strain data acqui-
ing the effect of the shape, the side aspect ratio and the area aspect sition have been obtained by strain gages applied on FRP sheets too.

RESEARCH - Seismic behavior 53


Fig. 10- Test Matrix. (a) Column S-1-5GA (b) Column R-1-8H

Fig. 11- Columns strengthened by means of glass (a) and basalt (b) FRP.

The data have been elaborated in order to investigate on volumetric Confinement of RC cylindrical specimens strengthened by means of
strain and Poisson ratio as a function of load level. The main results of basalt fibers and inorganic matrix
the experimental campaign can be summarized as follows: a significant
increasing in the axial displacement and a little increase in the ultimate The effectiveness of such system as a confinement technique has been
load of the FRP strengthened columns compared to their benchmark. analyzed by means of an experimental campaign on concrete cylindri-
The hollow columns have shown a failure mode characterized by cal specimens. The effectiveness of the proposed technique is assessed
bulging, followed in such case by the rupture of the fibers. A more evi- by comparing different confinement schemes: 1) uniaxial Glass Fibre
dent failure mode has been shown by the remaining columns for which Reinforced Polymer laminates; 2) alkali-resistant fibreglass grid bond-
the fiber rupture has always accompanied the concrete spalling. ed with a cement based mortar; 3) bidirectional basalt laminates pre-

Fig. 12- Load versus vertical and horizontal strains.

54 RESEARCH - Seismic behavior


Fig. 13- Typical failure modes.

impregnated with epoxy resin or latex and then bonded with a cement
based mortar; 4) cement based mortar jacket. The main objectives of the Fig. 15- Failure modes.

experimental program were: a) to investigate on the effectiveness of schemes were experimentally analyzed in order to evaluate and com-
confinement based on basalt fibres pre-impregnated in epoxy resin or pare the effectiveness of the proposed strengthening techniques: 1) uni-
latex and then bonded with a cement based mortar (BRM); and b) to axial glass FRP laminates (GFRP) wrapping; 2) uniaxial carbon FRP
compare the performance (in terms of peak strength and ultimate axial (CFRP) laminates wrapping; and 3) uniaxial basalt FRP (BFRP) lami-
strain gains) of different confinement techniques using advanced mate- nates wrapping. In particular 9 tests, were performed on square tuff
rials with respect to GFRP laminates jacketing. masonry (external tuff blocks and inner core filled with tuff chips and
The investigation was carried out on 23 concrete cylindrical specimens mortar) scaled columns (mass density equal to about 1530 kg/m3): side
with a diameter of D = 150 mm and a height of H = 300 mm. average dimension equal to 220mm; and average height of about 500
mm corresponding to 8 courses of tuff bricks (height-width ratio equal
to 2.27). Masonry was made by scaled yellow Neapolitan tuff bricks
(50x50x100mm) and a pozzolan (local volcanic ash) based mortar
(thickness of 12mm). Further 9 tests were performed on square clay
brick masonry scaled columns (mass density equal to about 1700
kg/m3): side average dimension equal to 260 mm, and average height of
about 560 mm corresponding to 8 courses of clay bricks (height-width
ratio of 2.20). Masonry was made by clay bricks (55x115x255 mm) and
a pozzolan (local volcanic ash) based mortar (thickness of 13 mm).
Fig. 14- BRM wrapping installation procedure.

Experimental outcomes showed that:


BRM confining system could provide a substantial gain both in com-
pressive strength and ductility of concrete members inducing a failure
mode less brittle than that achieved in the GFRP wrapped members;
lower performance were observed by concrete confinement provided Fig. 16- Specimen details (dimensions in mm): (a) tuff masonry; (b) clay brick masonry.

by a primed glass fiber grid bonded with cement based mortar with
respect to BRM and almost no influence was generated by the jacket- Masonry columns were tested through monotonically applied axial com-
ing with mortar only. pressive loading under displacements control mode with a rate of 0.005
maximum ultimate axial strain increases were provided by GFRP mm/s.
laminates wrapping.
The experimental outcomes showed that:
Confinement of rectangular masonry columns subject to axial load GFRP and CFRP jackets led to similar compressive strength gains
on tuff masonry columns under axial loads.
An experimental campaign dealing with 18 square cross-section both GFRP and BFRP confining system led to similar compressive
listed faced tuff and clay brick masonry scaled columns subjected to strength gains of brick masonry columns under axial loads. BFRP wrap-
uniaxial compression load. In particular, three different confinement ping was more effective in terms of global ductility increase (i.e. ulti-

RESEARCH - Seismic behavior 55


a significant increase in load carrying capacity and ductility after FRP
strengthening, which identified the columns as ductile elements despite
the brittle nature of the unconfined masonry. Differences in mechanical
behavior, due to the geometry of the columns, to the nature of different
materials, to different strengthening schemes, and to the amount of rein-
forcement, have been taken into account. The calibration of design
equations recently developed by Italian National Research Council,
CNR was conducted to compare analytical prediction and experimental
results.

Fig. 17- Stress-axial strain relationships and specimens failure mode: (a),(c) and (e) tuff masonry; (b),(d) and (f) clay brick
masonry.

mate strain gain equal to 413% and 259% for BFRP and GFRP wrap-
ping, respectively) even if the mechanical external reinforcement ratio
of FRP laminates was lower then GFRP ones; such result could be
explained by the higher values of ratios efl/efu recorded on BFRP lam-
inates.
The use of high values of laminates unit height may significant reduce
the effectiveness of FRP wrapping systems since it could be detrimen- Fig. 18- Limestone and clay brick masonry specimens.
tal to the quality of confinement execution.
The presence of voids and protrusions on masonry members reduces
The results obtained from the experimental campaign confirmed that
the ultimate transverse strain on FRP reinforcement with respect to that innovative strengthening techniques, using FRP sheets and bars, are
typically achieved on concrete members. effective when confinement of masonry compressed elements is need-
ed. Two types of masonry were investigated: the first made with clay
Another experimental campaign has been carried out in order to show bricks, the second made with limestone blocks. Even if the properties
the behavior of columns built with clay or with calcareous blocks, com- of the constituent materials were different, in both cases a significant
monly found in southern Italy, especially in historical buildings. increase was measured in terms of peak load and ultimate axial defor-
Rectangular masonry columns were tested for a total of 33 specimens; mation. Two construction schemes were considered: full core and hol-
uniaxial compression tests were conducted on columns taking into low-core columns; the last type reproduces the patterns often found in
account the influence of several variables: different strengthening historical buildings. External and internal FRP confinement were test-
schemes (internal and/or external confinement), curvature radius of the ed, separately and combined. The proposed techniques are strongly
corners, amount of fiber-reinforced polymer (FRP) reinforcement, recommended when a seismic retrofit is needed, since the external con-
cross-section aspect ratio and material of masonry blocks. Materials finement introduces a plastic behavior of the compressed masonry
characterization was preliminarily carried out including a mechanical which indicates a large capacity in storing elastic energy which is taken
test on plain masonry. For all cases the experimental results evidenced by the fibers placed in the transverse direction. The presence of inter-

56 RESEARCH - Seismic behavior


nal bars used as an internal confinement system is recommended in
addition to external FRP layers if ductility constitutes a main issue,
since in columns strengthened only with bars the ultimate load was
increased but brittle behavior of unconfined masonry remained.
Columns with hollow core also showed a significant increase of
mechanical properties when confinement was applied, especially in the
cases of GFRP external sheets combined with internal bars.

Confinement of circular masonry columns subject to axial load

Fig. 20- Specimens after failure.


An extended experimental investigation has been performed in order to
show the mechanical behavior of circular masonry columns built with
Displacement capacity resulted increased in all cases; strengthened
calcareous blocks that may be commonly found in Italy and all over
columns tested showed an extended postpeak plastic branch in the load
Europe in historical buildings. Different stacking schemes were used to
versus displacement curves;
build the columns, aiming to simulate the most common situations in
existing masonry structures. Carbon FRP sheets were applied as exter- Columns confined with three 100 mm wide sheets showed higher
nal reinforcement; different amounts and different schemes of confining mechanical properties with respect to the same columns confined with
reinforcement were studied. The experimental program included a new two 150 mm wide sheets;
reinforcement technique made by using injected FRP bars through the Damage caused by overloads applied in the precracking stage before
columns cross section. The structural behavior of masonry columns strengthening did not reduce the mechanical properties of FRP-con-
damaged under different levels of load and strengthened by using FRP fined columns;
reinforcements has been also investigated. Presence of internal FRP rebars acted as an effective confining sys-
tem for cross sections composed by four blocks;
Application of design equations by Italian CNR furnished conserva-
tive results for complete FRP wrapping, whereas prediction of strength
for masonry confined with CFRP strips showed a reduced scatter with
respect to experimental results.

4.4 Task 8.4: the strengthening in flexure and in shear of RC structural


elements with FRP fabrics and near surface mounted (NSM) rods.
A new technique for the shear and flexural strengthening of RC struc-
tural elements has been investigated in this research task. In particular,
the use of Near Surface Mounted rods (NSM) for structural upgrading
has been deeply analysed by both analytical and experimental investi-
gations.
Further, the effectiveness of FRP laminates, traditionally used to
strengthen RC or masonry members, has been investigated with refer-
ence to full scale prestressed concrete (PC) girders.

Fig. 19- Geometry and dimensions (mm) of columns.


NSM bars shear contribution: a calculation procedure
Important remarks follow:
High increase in ultimate strength and strain were evident after A calculation procedure suitable for practitioners has been developed
strengthening; by simplifying a more sophisticated predictive model recently devel-
Complete FRP jacketing was much more effective than discontinuous oped (Bianco 2008; Bianco et al. 2009a-b). That procedure briefly con-
wraps; sists of: a) evaluating the average structural system composed of the

RESEARCH - Seismic behavior 57


average-available-bond-length NSM strip confined to the correspond- The maximum effective capacity for the case of shallow concrete frac-
ing concrete prism whose transversal dimensions are limited by the ture and a resulting resisting bond length whose value is equal (u = 4)
spacing between adjacent strips and the beam cross section width to the effective bond length can be evaluated by:
(Figure 20); b) determining the comprehensive constitutive law of the
A Csf
average system above (Figure 21); c) determining the maximum effec-
tive capacity that the average system can attain during the loading
Ld { sf 2
2A3f,max [
Vfi,eff = 1  A1C1 Ld f,max+ 2 2  arcsin(1A3f,maxLd)+
max

 (1)
process of the strengthened RC beam by imposing a kinematic mecha-
nism and d) determining the NSM shear strength contribution by sum-

+(1A3f,maxLd) 1(1A3f,maxLd)2
2 ]}
ming the contribution provided by each strip. The constitutive law where:
(Figure 22) and in turn the equations to determine the maximum effec- L J l3sin(q+
) l2sin(q+
)
A1= p 3 ; A2=LpJ3l; A3= ;
tive capacity assume different features depending on the main phe- 4 0J1 2 0J1
nomenon characterizing the ultimate behaviour of the average structur- sf J 0J1 (2)
C1=1 0 2 1 ; C2sf =
al system of the specific case at hand. Hereinafter, for the sake of brevi- l l2
ty, the main features of that computational procedure are shown only for 21
f,max=1= (3)
the case of shallow concrete fracture (u = 4) and a resulting resisting Ldsin(q+
)
bond length whose value is equal to the effective bond length (Figure Actual Vf and design value Vfd of the NSM shear strength contribution
23). Further details can be found elsewhere (Bianco 2008). can be obtained as follows:
The predictions obtained by that calculation procedure were also 1 1 l max
Vfd=  Vf =  (2Nf,intVfi,eff sin
) (4)
appraised on the basis of experimental results (e.g. Dias et al. 2007). Rd Rd

where Rd is the partial safety factor divisor of the capacity that can be
assumed equal to 1.1-1.2 according to the indeterminateness of the
input parameters.

Bond between NSM bars and surrounding concrete: experimental and


analytical investigation

Pull-out test were carried out to investigate both the qualitative and
quantitative influence of some of the involved parameters on the bond
performance (De Lorenzis and Galati 2006, Galati and De Lorenzis
Fig. 21- Main features of the calculation procedure: a) average-length NSM strip and concrete prism of influence, b) adopted
local bond stress slip relationship, c) NSM strip confined to the corresponding concrete prism of influence and semi-pyramidal
2006). Those parameters encompass: ratio between depth and width of
fracture surface, d) sections of the concrete prism.
the slit, kind of epoxy-based adhesive used as binding agent, distance
of the NSM bar from the edge of the concrete prism, distance between
adjacent bars and employment of external FRP strips used to confine
the joint. Tests were carried out by means of a tangential-pull device to
apply the load, LVDT transducers to measure the slip at both the loaded
and unloaded extremity and strain-gauges throughout the adhered
Fig. 22- Possible comprehensive constitutive law of an NSM CFRP strip confined within a prism of concrete: (a) concrete
that reaches the free extremity (u=1) or strip tensile rupture (u=2), (b) superficial and/or absent concrete fracture and length of the bar to measure the deformations along the joint. The mea-
ultimate resisting bond length smaller (u=3), equal (u=4) or larger (u=5) than the effective bond length and (c) deep concrete
fracture (u=6). sured quantities were processed to obtain the local bond stress-slip
relationship for the different values of the test parameters. Cyclic tests
were also carried out subjecting the joint at a limited number of cycles
whose maximum load was assumed equal to different percentages of the
peak static load. The cyclic tests were useful to evaluate the joint resid-
ual strength such as the one following a seismic action.
Fig. 23- Maximum effective capacity along the CDC for the case u=4: a) comprehensive constitutive law; b) capacity An analytical investigation has followed the experimental program
Vfi,CDC (;); and c) imposed end slip Li,CDC (;); distribution along the CDC for different values of the CDC opening angle 
and d) effective capacity as function of the angle . above (Rizzo and De Lorenzis 2007-2009b). In fact, the local bond

58 RESEARCH - Seismic behavior


stress-slip relationship obtained in the pull-out tests has been modelled contribution to a RC beam.
by suitable analytical functions whose unknowns were calibrated for the
different values of the test parameters. The local bond stress-slip rela- Experimental investigation on full-scale prestressed concrete beams
tionship obtained by the cyclic tests was also modelled by analytical strengthened by means of CFRP
functions. Then, the numerical solution of the governing differential
equation has allowed the peak pull-out load be determined as function Every year, several prestressed concrete (PC) bridge girders are acci-
of the available bond length. dentally damaged by over-height vehicles or construction equipment
The pull-out tests were also simulated by a FE model, both in the Linear impact. Although complete replacement is sometimes deemed neces-
and Non Linear range. The Linear FE model was adopted to evaluate sary, repair and rehabilitation can be far more economical, especially
the bond-induced stresses on a plane transversal to the bar, evaluating when the time and the social cost of the method are drastically reduced.
the maximum stresses for different values of the geometrical and The numerous advantages provided by the use of FRP laminates are
mechanical parameters of the joint and estimating so, local tangential leading in a sharp increase on their use for bridge construction
stress inducing the first-cracking in both resin and concrete. After that, strengthening. Experimental investigations were conducted in order to
a Non Linear model was developed by modelling: a) the several mate- validate such strengthening technique on PC damaged members and
rials according to the fracture mechanics and b) concrete/adhesive and accurately assess the upper limit of damage amount beyond which FRP
adhesive/FRP interfaces by employing interface elements. laminates are no longer adoptable as repair solution.
Starting from such purposes, an experimental campaign was conducted
Experimental and analytical investigation on the shear strengthening on five full-scale (13.0m long, 1.05m high) PC double T-beams with a
contribution provided by NSM FRP bars on RC beams reinforced concrete slab, designed according to ANAS (Italian
Transportation Institute) standard specifications. One beam was used as
Four points bending tests were carried out on RC beams strengthened control, and the other four were intentionally damaged in order to sim-
in shear by NSM FRP bars (De Lorenzis and Rizzo 2006, Rizzo and De ulate a vehicle impact by removing the concrete cover and by cutting a
Lorenzis 2006-2009a). Those beams were designed in such a way that different percentage of tendons (17% on two specimens and 33% on the
the theoretical failure mode, for both the strengthened and un-strength- remaining two). The repair, by using externally bonded carbon FRP
ened beams was due to shear-tension. Parameters investigated were: (CFRP) laminates, aimed at restoring the ultimate flexural capacity of
spacing, type and inclination of the NSM bars and the shear-span-to- the member, taking particular attention to the laminates anchoring sys-
depth ratio. Some beams strengthened by NSM strips were also tested tem. In particular, one test was performed on the control beam (refer-
in order to assess the relative effectiveness of the two techniques. The enced as S1), two tests were carried out on intentionally pre-damaged,
system of FRPs was extensively equipped to measure the deformations to simulate an over-height vehicle collision, beams (named S2 and S3,
in the bars crossing the CDC. Tests have highlighted the possibility of respectively) and the remaining two on pre-damaged specimens
a global failure modes consisting in the detachment of the strengthened upgraded by using two and three plies of CFRP laminates anchored by
cover from the underlying beam core. Such mechanisms had not been using U-wraps (named S4 and S5, respectively).
pointed out by previous investigations. In Figure 24 and Figure 25 the test setup and experimental load deflec-
Two models were developed to predict the NSM shear strength contri- tion curves are reported.
bution: a) one more simplified and b) a more sophisticated one. The for-
mer was based on the Mrsch truss and the employment of a perfectly The experimental study has shown that: 1) a loss of strands equal to
plastic local bond stress-slip relationship. The latter takes into account 17% and 33% caused a flexural capacity decrease equal to 20% and
a more realistic local bond stress-slip relationship and the interaction 26%, respectively; 2) to restore the ultimate flexural capacity of the
between existing steel stirrups and NSM bars. The different local bond undamaged PC specimen by using CFRP laminates it is necessary to
stress-slip relationships obtained in the former phase of the investiga- prevent fibers debonding; 3) U-wraps (width wf= 100mm spaced at
tion were employed to carry out some comparison. From those compar- pf=150mm) were able to significantly delay debonding but if damaged
isons it was possible to point out the great importance of the fracture existing concrete is patched by cementitious mortar, a perfect bond has
energy as opposed to the shape of the local bond stress-slip relation- to be guaranteed during the cross section restoration to prevent local-
ship. This phase of the investigation has led to the development of use- ized debonding of longitudinal reinforcement and thus fully exploit the
ful formulae for the evaluation of the NSM FRP shear strengthening potential effective FRP strain increase; 4) CFRP laminates increased

RESEARCH - Seismic behavior 59


in existing buildings. Such deficiencies were analyzed by means of
parametric analyses to evaluate the possibility to apply external
strengthening on members characterized by poor concrete quality, low
transverse reinforcement ratios, inadequate confinement due to lacking
stirrups (especially in external joints), low bond performance of smooth
and ribbed longitudinal reinforcement in columns and beams.
Numerical analyses evidenced typical failure modes, crack patterns,
Fig. 24-- Test setup. influence of mechanical and geometrical properties on the behavior of
joints. Some numerical/experimental comparisons were made based on
significative tests available in scientific literature (for instance per-
formed by Prof. Shiohara working group) or tested, during the RELUIS
Project, by UNIBAS R.U., allowing the numerical F.E.M. model to be
validated. The behavior of the joints controls the global seismic behav-
ior of an entire structure and a building in particular, so that its assess-
ment is a crucial task in the strengthening design. Based on such analy-
ses, the validity of analytical models for unstrengthened joints, avail-
able in scientific literature, was checked. Such check was supported by
local information provided by the detailed, refined, F.E.M. analyses.
Fig. 25- Experimental load deflection curves.
This numerical tool was a primary tool to understand the damage evo-
both stiffness and flexural moment capacity of PC damaged beams lution and to assess the reliability of the main assumptions, equations
(maximum moment recover equal to about 12% and 20% for specimens and procedures according to the Quadruple flexural resistance in rein-
with 17% and 33% of strands loss, respectively; 5) the strengthening forced concrete beam-column joints (Shiohara, 2001) proposed by
intervention led to weak failure mode with a global ductility loss. Prof. Shiohara working group. To provide a direct, practical tool, ori-
The experimental outcomes qualify the application of FRP technique, ented to the profession more than a nonlinear refined F.E.M. analysis,
already adopted in several cases of impacted PC bridges, as an effec- it was evaluated the opportunity to extend such consolidated model to
tive tool to restore the flexural capacity of PC girders; however the cal- the case of externally bonded FRP strengthening of beam column joints.
ibration of theoretical expressions for the computation of the design This model is based on the solution of a system of equilibrium equations
FRP strain level considering the benefits provided by anchoring sys- referred to the four rigid bodies in which the joint can be ideally divid-
tems is strictly necessary. ed, sometimes neglecting compatibility. This model is able to account
for different failure modes and for the contribution of externally bond-
4.5 Task 8.5: the beam-column and beam-foundation joint reinforcement ed FRP strengthening. The main assumptions can be recalled:
with FRP Diagonal cracks in the joint form an angle of about 45
According to performance-based design or seismic evaluation of RC Normal concrete stresses acting on the main cracks can be reduced
buildings, it is crucial to provide a reliable evaluation of the strength to an equivalent force
and ductility capacity of the beam column joints: experimental and Longitudinal reinforcement provides only axial forces, so that any
direct observation of damages occurred during recent earthquakes dowel action is neglected
highlighted this. There is a global symmetry both on the horizontal and vertical plane
This task focused on some aspects: namely, cracking of the joint panel, The joint shear can be evaluated according to the sketch in Figure 26.
longitudinal reinforcement bars slipping are deformability sources and
they could alter the capacity and interaction of beam and column mem- Vf = T + C's + C'c Vc (1)
bers and the joint itself. F.E.M. modeling was adopted as an assessment Vf = T + T' Vc (2)
tool. The finite element code TNO DIANA 9.1 was adopted to simulate
and to analyze numerically some real beam column joint sub assem- where T and T'are the tractions in the longitudinal bars at the joint sec-
blages, characterized by nonlinear mechanical properties and geomet- tion, respectively; C'c is the compression force in concrete, while C's is
rical detailing, smooth bars, structural deficiencies, as commonly found the compression force in the bars. Vj is the joint shear; and Vc is the col-

60 RESEARCH - Seismic behavior


Fig. 26- Details of the forces due to the elements converging in the joint panel. Fig. 28- The joint system and the external forces.

umn shear. In Figure 27 the joint division is shown: there are four rigid umn length; and lb is the beam length.
and interacting bodies. Each body can be associated to three equilibri-
um equations. The symmetry of the joints allows the system to be Horizontal and vertical equilibriums of forces follow:
reduced to six equations. Moreover, Figure 28 shows the relation bet-
ween the column shear, Vc, and joint shear, Vb, based on the equation Fx=0, F1F2+C1sinqC2sinq+Vc=0 (6)
Vb = m  Vc, where m = lc / lb. Fy=0, F3F4+C1sinq+C2sinqNcF7=0 (7)

Where Nc is the vertical force in the column; F7 is the traction in the


FRP reinforcement in the y direction and reduced to an equivalent
force.
To evaluate the column shear capacity of an unstrengthened joint, the
F1, F2 and F5 forces can be assumed equal to the yielding forces of the
corresponding steel reinforcement while F6 and F7 are equal to zero.
The five unknowns are Vc , F3, F4 , C1 and C2 and can be evaluated solv-
ing the system (3)-(7).
(a) Forces in reinforcement and FRP (b) Forces in concrete
To evaluate the column shear capacity of a strengthened joint, the F6
Fig. 27- Internal forces.
and F7 forces can be equal to the strength capacity of the external FRP
The independent equilibrium equations are five. Horizontal and verti- reinforcement both in the x and y direction, respectively. To account for
cal forces equilibrium and moment equilibrium based on point o (the both the tensile failure of the strengthening or for a possible FRP
middle of the joint panel) can be expressed as follows: debonding, the FRP capacity is given by the minimum between tensile
strength and debonding force evaluated, for instance, according to CNR
Fx=0, F1F2F3+C1sinq+C2sinqNbF6=0 (3) DT200.
Fy=0, F3F4+C1sinqC2sinq+mVc=0 (4)
lb 1 1 lj 4.6 Task 8.6: the design criteria for the seismic retrofit of RC and RC-
Mo=0, mV + j (F F )+ j (F F ) C =0 (5)
2 c 2b 1 2 2c 3 4 2 2 masonry composite structures with FRP.
Where Nb is the horizontal force in the beam; F1, F2 are tractions in the The applications of Carbon FRP (CFRP) and Glass FRP (GFRP) mate-
longitudinal bars of the beams at the joint section; F3, F4 are tractions rials have grown during last years; at present, seismic applications have
in the longitudinal bars of the columns at the joint section; F5 is the become comparable if not more frequent than those related to lacks due
traction in the stirrups spread along the height of the joint and reduced to gravity loads. The Italian guidelines for FRP interventions (CNR-DT
to an equivalent force; F6 is the traction in the FRP reinforcement in 200, 2004), deal with the use of composite materials to seismically
the x direction and reduced to an equivalent force; C1, C2 are compres- upgrade under-designed RC and masonry structures. From seismic
sions as shown in Figure 27b; jb, jc are the internal lever arms in the standpoint, FRP strengthening is regarded as a selective intervention
beam and column respectively; q is the inclination angle of the main technique. Based on the main deficiencies of the existing structure, the
cracks and assumed equal to 45; Vc is the column shear; lc is the col- driving principles of the intervention are based on two main strategies:

RESEARCH - Seismic behavior 61


1) preventing potential brittle failure mechanisms (i.e. shear failure, lap tively. The design of the rehabilitation was based on deficiencies under-
splice failure, buckling of longitudinal reinforcement in compression) lined by both the test on the as-built structure and the theoretical
and soft story collapse mechanism, and 2) increasing global defor- results provided by the post-test assessment. They indicate that a retro-
mation capacity of the structure, either by enhancing the ductility of fit intervention was necessary in order to increase the structural seismic
plastic hinges without their relocalization or establishing a correct hier- capacity; in particular, the theoretical results showed that the target
archy of strength by relocalizing the plastic hinges according to capac- design PGA level of 0.30g could have been sustained by the structure
ity design criteria. The retrofit strategy is obtained by combining the if its displacement capacity was increased by a factor of 48%. In order
above principles; the definition of the retrofit scheme depends on the to pursue this objective, the retrofit design strategy focused on two main
gap between actual and target performance of the specific structure, aspects. First it was decided to increase the global deformation capac-
costs, functional characteristics and importance of the structure. ity of the structure and thus its dissipating global performance; such
Experimental studies, aimed at validating the effectiveness of FRP to objective was pursued by confining column ends with two plies of
achieve the above goals, are reported in the following. GFRP laminates. Moreover, the second design key aspect was to allow
the structure to fully exploit the increased deformation capacity by
Seismic strengthening of an under-designed RC structure with FRP avoiding brittle collapse modes. To achieve this goal corner beam col-
umn joint panels were strengthened by using two plies of quadri-axial
The outlined seismic strengthening strategy effectiveness was experi- GFRP laminates as well as a wall-type column for its entire length with
mentally investigated within the European research project SPEAR two plies of the same quadri-axial GFRP laminates used for the above
(Seismic PErformance Assessment and Rehabilitation). The structure joints (see Figure 30).
under examination was designed and built with the aim of creating a
structural prototype featuring all the main problems normally affecting
existing structures: plan irregularity, dimensions of structural elements
and reinforcement designed by considering only gravity loads, smooth
reinforcement bars, poor local detailing, insufficient confinement in the
structural elements and weak beam column joints (see Figure 29). The
structure was subjected to pseudo-dynamic tests, both in its original
configuration and retrofitted by using GFRP.

Fig. 30- Column confinement and shear strength of corner joints (a); shear strength of wall-type column and retrofitted
structure overview (b).

Fig. 29- (a) Plan view and (b) 3D view of the SPEAR structure.

Fig. 31- Base shear top displacement curves for as-built and FRP retrofitted structure.
The structure in its original configuration was subjected to experimen-
tal tests with maximum peak ground acceleration (PGA) of 0.20 g. Since The assessment of structural global performance, before and after the
both theoretical and experimental results showed that the as-built strengthening intervention, was performed by nonlinear static pushover
structure was unable to withstand a larger seismic action, a retrofit analysis in longitudinal direction (positive and negative X-direction,
intervention by using FRP laminates was designed. Once the design of PX and NX, respectively) and in transverse direction (positive and neg-
the GFRP retrofit was provided, the structure was subjected to a new ative Y-direction, PY and NY).
series of two tests with the same input accelerogram selected for the as
built specimen but scaled to a PGA value of 0.20g and 0.30g, respec- In Figure 31, the theoretical base shear-top displacement curves for the

62 RESEARCH - Seismic behavior


Fig. 33- Details of the designed rehabilitation system.

has been investigated with reference to a PC bridge, named Torrente


Casale which is part of the Salerno-Reggio Calabria motorway. The
Fig. 32- Theoretical seismic performance comparison at 0.3g PGA between as-built and FRP retrofitted structure. bridge, built in the seventies, has been recently enlarged (2001) in
order to satisfy the new traffic demands. Due to the recent issuing of a
as built and FRP retrofitted structure are depicted with reference to new seismic Italian code, it was decided to assess the bridge capacity,
direction NX (where the maximum capacity-demand gap was recorded both for gravity and seismic loads, in relation with the new design pro-
for the as-built structure at the significant damage limit state LSSD). visions.
Figure 31, clearly shows that the FRP retrofit is able to greatly increase The bridge existing documentation has been investigated and both
the global deformation capacity of the structure, slightly affecting its destructive and non-destructive tests have been performed in order to
strength. The comparison between the seismic structural capacity and determine concrete and steel reinforcement mechanical properties.
both elastic and inelastic demand is reported in Figure 32 (for direction Once the bridge geometry and the material mechanical properties were
NX) by using the Capacity Spectrum Approach (CSA) (Fajfar, 2000). determined, a theoretical analysis was performed showing that the
Figure 31 clearly shows that column confinement provides the structure bridge piers (circular cross-section, diameter D=1m, and total height
with significantly enhanced ductility, allowing it to achieve the theoret- H=6m) were not adequate to sustain the seismic actions. Thus use of
ical inelastic demand by only modifying the plastic branch of the capac- SRPs spikes as columns flexural reinforcement combined with CFRP
ity curve. laminates wrapping of columns ends has been investigated to increase
both member strength and ductility; the structural upgrade was com-
After that columns and joints were wrapped with GFRP, the retrofitted pleted by increasing the shear capacity of column cap through CFRP
structure was able to withstand the higher (0.30g PGA) level of excita- U-wraps (Figure 33). The effectiveness of such technique with respect
tion without exhibiting significant damage. After tests, FRP was to a traditional one, based on RC jacketing, has been assessed. The
removed and it was shown that the RC core was neither cracked nor main construction phases of the rehabilitation intervention are reported
damaged. The comparison between the experimental results provided in Figure 33.
by the structure in the as built and GFRP retrofitted configurations
highlighted the effectiveness of the FRP technique in improving global 4.7 Task 8.7: the design criteria for the seismic retrofit of masonry struc-
performance of under-designed RC structures in terms of ductility and tures with FRP.
energy dissipation capacity. The main design problems concerning the seismic vulnerability mitiga-
tion of masonry structures are the strengthening of masonry panels for
Seismic rehabilitation of PC bridges by using FRP and SRP materials in plane and out of plane actions, the improvement of FRP action by
means of clamping and connection mechanical devices and procedures
An innovative strengthening technique based on the combined use of for the definition of the reinforcement layout in seismic analyses.
Fiber Reinforced Polymer (FRPs) and Steel Reinforced Polymer (SRPs) In order to test the design methods proposed by CNR DT200/2004, sev-

RESEARCH - Seismic behavior 63


eral seismic studies of complex monumental masonry buildings have analysis with this type of discretisation is not a viable solution. Studies
been completed by carrying out detailed linear dynamic seismic analy- on discrete representation of masonry walls by using a refined truss
ses and FRP retrofit design. In the design several combinations of structure to represent the compressive load paths inside the masonry
strengthening techniques were explored in order to point out the real panels showed that this type of discretisation is very simple, allows for
applicability of the cited guidelines. complicated constitutive laws, allows easily equilibrium checks, and
The obtained results cover the following areas: produces very reliable load displacement curves.
a) Verification of the feasibility of a FRP design through simple modifi- f) The basic choice of associate Mohr Coulomb or Drucker Prager
cations of the normal design activity; elastic plastic constitutive laws can be effective only in limited cases,
b) Definition of the feasibility bounds for this retrofit technique in case where dilatancy is not dominant. More effective analyses require a non
of monumental buildings; associate zero dilatancy rule, which is however not common in the pro-
c) Detail types and level of detailing required in real applications; fessional engineering software. More refined damage rules are actually
d) Open FRP design problems not covered by the guidelines or lacking under way, but not for normal design activity. Truss elements however,
of necessary information; allow to introduce complex behaviour by means of geometric discreti-
e) Comparison of different modelling techniques making use of plate, sation, and by this way, crack tracking can be pursued, even if in a very
shell, beam, truss elements to represent the masonry structure; rough representation.
f) Comparison of different material constitutive assumptions for the non
linear analysis of strengthened walls and buildings. 4.8 Task 8.8: the strengthening of masonry structural elements with FRP
The main results of the completed research are summarized in the fol- systems.
lowing: The research activity has been performed through experimental tests
a) The FRP reinforcement net design is easily implemented by a sim- and theoretical study. In particular, the experimental program has been
ple modifications of the normal design activity. The only needs are: inte- carried out with reference to masonry panels strengthened with FRP
gration of the stress distributions in order to obtain the stress resultants and with SRG and subjected to in plane loads, while the theoretical
and preparation of a verification sheet including the design rules for study concerns the modelling of masonry walls strengthened with FRP.
bending and shear of a FRP reinforced masonry panel. Actually many
companies have implemented DT 200/2004 rules in freeware software Masonry panels and building strengthened with FRP
which can be helpful in this activity.
b) The feasibility bounds for this retrofit technique checked for three Several experimental tests have been performed with reference to
monumental buildings analysed, are very different if stuck or mechan- masonry panels strengthened with different FRP materials (carbon and
ically fastened reinforcement is used. In fact, only slight diffuse glass) arranged according to various configurations.
increase of resistance is obtained by using externally bounded FRP In-plane shearcompression tests have been performed on full-scale
nets. Owing to increase significantly the safety of the building, mechan- tuff masonry panels consisted of two-layered walls with the inner part
ical devices are mandatory. In this last case true building retrofit is pos- filled with mortar and chips from yellow tuff blocks, considering differ-
sible. ent FRP materials and strengthening configurations. In particular, two
c) The set of the detail types needed for practical applications is very sets of panels have been strengthened with grid pattern carbon fibre
large and many connection types are not yet fully investigated although unidirectional strips (CFRP), made with three horizontal and vertical
practically employed. Fibre ropes, bars, fasteners, metallic inserts are strips on each face and two sets of panels have been strengthened with
mixed in a way almost never covered by existing experiments and the same layout, but doubling the number of plies. Similarly, four sets
guidelines. of panels have been symmetrically strengthened with a grid pattern on
d) Open FRP design problems include cement matrices, thermal both sides of the panels, but with glass bidirectional fibre strips
cycling, delamination in compression, real bond of the regularization (GFRP). A further set of panels has been made selecting a different geo-
primer to rough masonry surfaces, behaviour of the mechanically fas- metric configuration by arranging FRP laminates along the diagonals of
tened FRP elements after internal delamination. both sides of the panels and considering both a CFRP and a GFRP
e) Several modelling techniques have been employed in order to carry cross layout with either one or two FRP plies.
out the seismic studies. Plate elements in linear dynamic analysis allow Important experimental evidences have emerged from the performed
for both in plane and out of plane evaluation, but non linear static tests underlying the role of the configuration of the FRP strengthening

64 RESEARCH - Seismic behavior


system on the failure mechanism of the tested panels. In fact, while in damages; in the second phase the prototype has been repaired by
some cases it has been observed that the debonding of the FRP strips selecting GFRP strips arranged according to the Italian Code
has been the main responsible of the panels failure, in few cases the (Ordinanza 3431/2005, CNR DT 200/2004); in the third phase, dynam-
tensile rupture of the FRP strips has occurred. The different observed ic tests have been performed on the strengthened prototype.
failure mechanisms have particularly influenced the behaviour of the Experimental evidences have shown that the applied GFRP strips allow
strengthened panels both in terms of strength increase and post-peak to perform fast repair interventions in order to make operational mason-
behaviour (fracture energy). ry structures severely damaged by earthquake and at risk of after-
Further tests have concerned square masonry panels composed of clay shocks. In fact, the repaired prototype has showed reduced openings at
bricks, strengthened with different FRP configurations and subjected to the horizontal joints (about one-third of the maximum opening observed
compression diagonal load. In particular, some of the panels have been in the case of un-strengthened prototype) and an increasing of the lat-
strengthened considering both vertical and horizontal CRFP and GFRP eral strength (about +34%). A further aspect emerged from this tests
strips, while other panels have been strengthened arranging FRP strips has concerned the importance to provide effective anchorages for FRP
along diagonal directions of the panels. Different levels of the strength strengthening elements in order to avoid the delamination phenomenon
increase and different failure mechanisms have been observed. In par- which is particularly influenced by cyclic actions.
ticular, while localized cracks pattern have characterized both the case
of un-strengthened panels and the case of panels strengthened by FRP Masonry panels strengthened with SRG
only on one side, a more diffuse crack distribution has been observed
in the case of panels strengthened on the two opposite sides. In several Beside the traditional Fiber Reinforced Polymers (FRP), the research
cases the debonding of the FRP strips has been also observed. investigated the possibility of application of innovative composite mate-
From the experimental studies conducted on FRP strengthened mason- rials, called Steel Reinforced Grout (SRG), based on high strength steel
ry panels, considerations useful for improving the Document CNR- wires (Ultra High Tensile Strength Steel) forming that are assembled
DT200/2004 have been deduced. In particular, it has been observed into a fabric and embedded within a cementitious grout. This applica-
that for the evaluation of the design shear strength of FRP strengthened tion in fact could combine, to the traditional advantages proper of FRP,
masonry panels (eqn. 5.16 of the CNR-DT200/2006), the contribution the performances of this new material, reducing installation and mate-
of the masonry component can be evaluated through the eqn. 5.17 (used rial costs, and inducing an increase of ductility. Both composites, FRP
in the case of reinforced masonry elements) only if the FRP shear and SRG, can be used with perforated or solid brick to form a new
strengthening is coupled with FRP vertical elements fixed both at the strengthening system, called LATLAM ring beam, that can be used to
base and at the top of the panel. For this reason it has been suggested effectively construct the roof ring beams of a masonry structure. This
to include in the CNR-DT200 specific design indications concerning new system, can be also subjected to a pretension force. The following
this aspect. In fact, it is important to underline that, when the FRP conclusions may be drawn from the developed research:
shear strengthening system is not coupled with a flexural FRP strength- - the analytical model developed to determine the mechanical behav-
ening system with efficient anchor elements, the contribution of the iour of LATLAM ring beams has shown good agreement with experi-
masonry material in terms of shear strength is the same of the case of mental results and can be incorporated in design provisions;
un-reinforced masonry elements. - the experimental tests, performed on full scale prototypes of LAT-
A further aspect deduced with reference to tests on masonry panels LAM ring beams, demonstrated good results in terms of load carrying
strengthened by FRP has concerned the case of fibres arranged along capacity;
to the diagonal directions of the panel. From the experimental evi- - LATLAM ring beams proved to be a good substitute, either under a
dences and from global analyses it has been possible to affirm that the technical and economical perspective, of traditional reinforced con-
contribution of the FRP can be evaluated only considering the compo- crete ring beams.
nent parallel to the shear load.
A scale model of a typical tuff masonry buildings has been constructed Modelling and analysis of masonry walls strengthened with FRP
and tested on the seismic simulation shaking table at the structural lab-
oratory of CESI, Bergamo, Italy. In particular, the test procedure has A further subject of the task research activity has concerned the mod-
consisted in three phases: in the first phase dynamic tests have been elling and the analysis of masonry elements strengthened with FRP.
conducted on the un-strengthened prototype in order to induce some Indeed, recent codes extended the use of displacement-based design

RESEARCH - Seismic behavior 65


methodologies, such as the pushover analysis, to the case of masonry This aspect has revealed the important role of the FRP system to
buildings. Thus, different modelling approaches able to capture the increase the energy dissipation capacity of masonry structures. The
structural behaviour of masonry panels strengthened with FRP have increase both in terms of strength and ultimate deformation capacity of
been examined: a macro-micro modelling approach; a macro modelling strengthened elements has been confirmed by all the modelling
approach; a frame model approach. approaches evidencing the capability of the simple frame model to cap-
The first model relies on a homogenization approach combined with ture the global response of masonry elements strengthened with FRP
limit analysis suitable for the evaluation of the collapse loads and fail- and consequently the possibility of using this simple model in a design
ure mechanisms of FRP reinforced masonry panels. The application of process of the FRP strengthening system for masonry structures.
FRP strips on masonry has been treated adopting a simplified multi
step approach. In the first step the un-reinforced masonry, regarded as 4.9 Task 8.9: the strengthening of masonry vaulted with FRP systems.
a periodic heterogeneous material, has been substituted with a homo- The aim of the research task is the development of models and proce-
geneous macroscopic material using a homogenization technique. In dures for the analysis of vaulted structures reinforced with FRP. In par-
particular, an estimation of the homogenized unreinforced masonry ticular the behaviour of arches, vaults and domes subjected to seismic
strength domain has been obtained by means of a micro mechanical action is studied in order to investigate the behaviour of these typolo-
model based on the lower bound theorem of limit analysis. In the sec- gies as built and strengthened with FRP materials.
ond step, FRP strengthening has been introduced on the already The study is performed with simplified analytical procedure and with
homogenized masonry material. numerical methods. Experimental tests are further expected for the val-
The second model is based on the use of both 2D or 3D nonlinear idation of the models.
behaviour finite elements and interface elements. In particular, special The research activity has been focused on the experimental tests on the
yield criteria coupled with nonlinear constitutive laws characterized by topic available in literature for a comparison with the analytical and
softening response have been selected in order to simulate the behav- numerical results.
iour of masonry material both in tension and compression. A special On the basis of the acquired data, a simplified analytical model for the
nonlinear constitutive law has been also considered for the interface evaluation of the ultimate load of arches and portal frames reinforced at
elements in order to simulate the debonding mechanism of FRP which the intrados and/or extrados has been developed.
characterizes in several cases the failure mechanism of masonry ele- The main results are related to the definition of a methodology for the
ments strengthened with FRP. evaluation of the ultimate load of one-dimensional vaulted structures
Finally, the third modelling approach consists of a simple but effective reinforced with FRP and subjected to vertical and horizontal forces.
1D frame element able to predict the response of masonry structures, Starting point of the procedure is the assessment that the FRP presence
eventually reinforced with FRP materials. In fact, the proposed ele- does not allow the formation of the collapse mechanisms, which char-
ments presents some peculiarities both for converting the geometry of a acterise the ultimate behaviour of the un-reinforced structure. As a con-
masonry panel in the geometry of the equivalent frame and for account- sequence the cinematic approach of the limit analysis, usually adopted
ing the nonlinear behaviour of masonry and FRP materials. Special for this kind of structure cannot be applied. The proposed analytical
components have been also provided in the developed model in order model, starting from the analysis of the ultimate behaviour of the un-
to account for the shear failure modes. reinforced structure, identifies the location of the hinges, up to make the
Several numerical applications have been carried out using the pro- structure statically determined, and then to be solved with equilibrium
posed modelling approaches and considering experimental cases condition. Obviously in this case and contrarily to the limit analysis, the
deduced from literature. The obtained results, regarding both un- mechanical characteristics of the masonry and of the FRP and mainly
strengthened and FRP-strengthened simple panels and masonry the debonding phenomenon of the composite material, significantly
faades, have shown the reliability of the proposed models to reproduce influence the structural response.
the experimental behaviour of masonry elements underlying some The model has been developed and applied to assigned geometries of
peculiarities due to the presence of the FRP strengthening system. In arches and portal frames, subjected to static vertical and horizontal
fact, the micro-macro and the macro modelling approaches have par- loads, with FRP reinforcement at the intrados. The obtained results
ticularly underlined the role of the FRP strengthening system to change have been expressed in terms of interaction domains in the plane of the
the stress path and, consequently, to induce different damage states vertical and horizontal forces.
characterized by a diffuse crack path along the strengthened elements. Afterwards, the procedures for the evaluation of the ultimate behaviour

66 RESEARCH - Seismic behavior


of masonry arches reinforced with FRP have been extended to other In the Laboratory of Structures and Materials of the Department of Civil
typologies, such as the masonry portal frames reinforced at the intrados Engineering of the University of Rome Tor Vergata, has been realized
or extrados. a masonry portal frame, constituted by two columns with rectangular
In particular has been defined the formulation of a secant linear rela- section (24cm x 49cm) and height of 2m, and an arch with internal
tionship defined by the maximum load and by the related displacement radius Ri = 130cm and square section with side equal to 24 cm.
at the brittle failure, to be adopted for the estimate of the seismic behav- The portal frame was constructed with masonry block on mortar beds,
iour of the structure. This model, even if simplified, can be suitable for and then reinforced with a layer of FRP at the intrados of the arch and
simulating the response of vaulted structures subjected to vertical and on the internal surface of the columns. No anchorage of the composite
horizontal loads, strengthened with FRP sheets. In this case, indeed, if has been given at the foundation level.
the FRP is applied at the intrados, the failure is due, generally, to its A constant vertical load (v) has been applied on the arch key and two
delamination, while if the composite material is glued at the extrados, horizontal increasing loads, with equal intensity (H/2) are given at the
the failure is due to compression of shear crisis of the masonry. Anyway arch abutments, which simulate a seismic action.
the global behaviour of the structure is often characterised by an almost The experimental results on both the arches and a portal frame appear
linear response up to the maximum load, ad generally with a brittle fail- to validate the simplified models and the procedures developed in this
ure. research program for the design and check of masonry vaulted systems
Numerical FEM procedures in bidimensional field, have been further reinforced with FRP.
developed. The comparison between the results obtained with the two
methodologies has allowed the validation of the simplified models. 4.10 Task 8.10: the quality control and monitoring of FRP applications
Parallel to the analytical-numerical models, an experimental program to existing masonry and RC structures.
has been set up, on vaulted structures reinforced with FRP. Several masonry and reinforced concrete structures have been utilized
Three circular arches, subjected to a key load, have been tested. The for the application of fiber reinforced composite materials, with the aim
first one represents the reference un-reinforced, the other two have of carrying out quality control and monitoring tests, in accordance to
been strengthened with FRP at the intrados. The tests have been per- CNR DT 200/2004. For each structure, special working sheets have
formed in displacement control, with the aim of evaluating the post- been developed for a proper characterization of the building from a geo-
peak behaviour and the softening branches. metrical, mechanical and logistic point of view. On these structures
The results obtained have allowed the preliminary validation of the semi-destructive tests, non-destructive tests and delamination tests
model related to arches reinforced at the intrados. have been performed.
The last phase of the research was devoted to the validation of the
developed models, through experimental tests on masonry vaulted ele- Semi-destructive tests
ments. In particular, besides the preliminary tests in masonry arches
reinforced with FRP at the intrados, a test has been performed on a This test consisting in pull-off and shear tearing tests conducted on dif-
masonry portal frame, in full scale, reinforced with FRP and subjected ferent types of FRP materials, applied on the in-situ structures.
to vertical and horizontal loads. The obtained results allowed the vali- Moreover, pull-off tests and shear tearing tests have been conducted in
dation of the analytical models and assessed the validity of the assumed the laboratory of the Department of Structural Engineering of University
simplified hypotheses.
of Calabria on concrete elements as well, reinforced by carbon fiber
sheets and laminates. The latter tests results have been utilized for a
comparison with a number of experimental tests conducted at the
Universities of Bologna, Naples, Sannio and Milan, in the framework of
Task 8.2 Round-Robin tests, with the aim of obtaining a standard test
procedure for delamination tests and evaluating the scattering between
experimental results derived from tests conducted on specimens real-
ized by the same worker, but tested in different laboratories. The tests
have been conducted using the same device used for outdoor tests. Pull-
off tests, used to assess the properties of the strengthened substrate,
Fig. 34- Portal frame under construction. Fig. 35- Portal frame during the test. have been carried out attaching a thick circular 75 mm diameter steel

RESEARCH - Seismic behavior 67


plate to the FRP and isolating it from the surrounding FRP with a core
drill, taking particular care in avoiding heating of the FRP system while
a 1-2 mm incision of masonry substrate was achieved. The test consists
in pulling off the steel plate by means of an ad hoc device (Figure 36-
a), obtaining the ultimate pull-off strength value expressed in kN
(Figure 36-c). Whereas, shear tearing tests are used to assess the qual-
ity of bond between FRP and masonry substrate. These tests can be
conducted only when it is possible to pull a portion of the FRP system
in its plane located close to an edge detached from the masonry sub-
strate. The tests have been carried out using the same ad hoc device
used for pull-off tests. In particular, metallic elements have been set up
onto the masonry wall and through the FRP strip, with the aim of con-
necting the entire test device. Then, the FRP element has been tight-
ened until collapse (Figure 36-b), obtaining the failure tearing force,
expressed in kN (Figure 36-d). For what concerns in situ tests, 16 rein-
Fig. 36- Pull-off test (a) and shear tearing test (b); Pull-off test result (c) and shear tearing test result (d).
forced concrete structures and 17 brick and stone masonry building
have been considered, for a total number of more than 300 tests. The ments reinforced by CFRP sheets and laminates have been realized.
FRP materials have been applied in the form of strips having the From mechanical characterization tests, a value of 25 N/mm2 for con-
dimensions of 500x200 mm and 50x200 mm in the case of r.c. struc- crete compressive strength was found. The Research Unit of University
tures and the dimensions of 500x300 mm e 50x300 mm in the case of of Calabria tested 3 prisms reinforced by 3 CFRP strips and CFRP lam-
masonry structures for the execution of pull-off and shear tearing test, inates applied on their surfaces. Pull-off and shear tearing tests have
respectively. Both carbon, glass and natural fiber composites have been been conducted, with the aim of studying the failure mode and debond-
utilized. ing of FRP from the substrate.
According to CNR DT 200/2004, FRP application may be considered The results of pull-off tests conducted on both strips and laminates
acceptable if at least 80% of the tests return a pull-off stress not less respected the limit values suggested in CNR DT 200/2004, and also in
than 10% of masonry support compressive strength, or not less than these cases failure occurred in the substrate, showing the effectiveness
0.9-1.2 MPa in the case of reinforced concrete structures, provided that of the FRP application. On the other side, shear tests showed a differ-
failure occurs in the support itself. For what concerns shear tearing ent behaviour of the composite materials for strips and laminates. In
tests, FRP application may be considered acceptable if at least 80% of particular, in the case of laminates, the collapse occurred suddenly and
the tests return a peak tearing force not less than 5% of masonry sup- the whole composite debonded from concrete, with an ultimate value of
port compressive strength, whereas it has to be not less than 24 kN in the shear force higher than 24 kN (limit value suggested in CNR DT
the case of reinforced concrete structures. 200/2004). In the case of CFRP strips, a partial delamination of the
For what concerns masonry structures, taking into account the com- composite occurred and the limit value was never reached.
pressive strength of the support, both pull-off and shear tearing experi-
mental results respect the limit values suggested in CNR DT 200/2004. Non-destructive tests
In the case of concrete structures, pull-off results are in accordance to
CNR DT 200/2004, whereas in shear tearing test results the limit value Non-destructive tests, consisting in thermographic tests were conduct-
of 24 kN suggested for reinforced concrete structures, has never been ed on both brick masonry elements built in the laboratory and on the
reached. This is probably due to the small dimensions of the FRP same real masonry structures used for the above mentioned semi-
strips. In fact, CNR DT 200/2004 Guidelines dont provide any instruc- destructive tests, reinforced with different FRP materials. These tests
tion about the FRP dimensions that should be utilized to reach the are usually carried out to characterize the uniformity of FRP applica-
above mentioned limit value. tion. Both, the active and the passive thermography technique have
In the case of pull-off tests, failure has always occurred in the substrate, been adopted, in which thermal energy is applied externally onto the
for each type of FRP material applied, as expected. test object or the natural infra-red radiation emitted by the object due
The semi-destructive tests conducted in the laboratory on concrete ele- to a sufficient exposure to sun light can be utilized, respectively.

68 RESEARCH - Seismic behavior


The first Infra-Red thermographic tests have been conducted in the lab- the surface preparation. The profilometer used is the DRSC produced
oratory of the Department of Structural Engineering of University of by Miami University, which gives quality and quantity information. This
Calabria on brick masonry macro-elements, which have been rein- instrument using laser striping, highlights the rough concrete surface by
forced by carbon fiber strips placed in different directions onto the thin slits of red laser light at an angle of 45 degrees, and the surface is
specimen surfaces. From the test a mortar joint was clearly visible due observed at 90 degrees by an high-resolution (tiny) board CCD camera
to the non plane substrate surface. (Figure 38a). The video image of the laser stripes is digitized with a
The in situ masonry and reinforced concrete structures have been uti- PCMCIA frame-grabber. The projected slit of light appears as a straight
lized for thermographic tests as well, to verify the quality of bond line if the surface is flat, and as a progressively more undulating line as
between FRP and the substrate, before and after the conduction of the the roughness of the surface increases. Lasers with one to eleven stripes
described semi-destructive tests. For instance, some tests have been were used.
conducted on a reinforced concrete structure on which FRP strips have The roughness degree can be identified by several parameters obtained
been previously applied for shear tearing tests. From the thermograms, by laser profilometry and each one can give specific information on
a crack could be noticed corresponding to the strip subjected to the roughness and its particular properties. They can be classified in ampli-
semi-destructive test. Such a technique is then useful for the detection tude parameters (Ri of Figure 37) and slope parameters (ia of Figure
deterioration and damage in the structures. 37a). The first group is sensitive to roughness morphologies, where the
surface is either stepped or slotted and might be described as a discon-
Influence of roughness surface on the debonding force of FRP tinuous roughness, the second group is insensitive to roughness mor-
phologies and is more useful for characterizing continuous roughness.
In order to investigate the effect of the concrete surface preparation
Table 1 Specimens and summary of test sample
method on the roughness surface and debonding force of FRP, an exper- N Specimen Dimension Mean Compressive Formwork Type of
imental campaign has been carried out. The specimens have been pro- (mm) Strength (N/mm2) (wooden) Compaction
duced with different formworks (staves, panels) and different com- 20 ? 15 staves beating
160 x 400 x 600
20? 20 panels vibration
paction types (beating, vibration). Moreover, two different concrete
strengths have been used in realizing the specimens in order to evalu-
ate also their influence on the surface preparation method efficacy.
Thirty 15 cm-length standard cubes have been also poured and used to
evaluate the mechanical properties of concrete (according to Italian
standards). Mean compressive strength (Rcm = 15 or 20 MPa) from the
compression tests has been obtained by standard cube at an age of 50
days, corresponding to same period of the first profilometer tests. The
Fig. 37- Roughness parameters.
number of specimens considered in the present experimental campaign
and the two different casting processes are shown in Table 1. Figure 37b shows some output parameters provided by the profilome-
After curing, four different methods for surface preparation have been ter: Rmax maximum peak and valley, rough measure of the vertical dis-
applied on concrete prisms in order to study the effect of the treatment tance between the highest peak and the lowest valley; Re measures the
on the FRP-concrete bond strength: Grinding the upper surface of the vertical distance between the highest peak and the centerline of the
concrete block has been grinded with a stone wheel to remove the top profile; Rv measures the vertical distance between the lowest valleys
layer of mortar, just until the aggregate was visible; Sand Blasting the and the centerline of the profile; R measures the average of all individ-
concrete surface has been sand blasted in order to remove the whole ually measured peak to valley heights, Rp roughness profile index,
mortar over the aggregates, so obtaining a very rough concrete surface; defined as the ratio of the true length in the fracture surface trace to its
Brushing the surface has been brushed with a twisted steel cord bond- projected length in the fracture plane; iA is the micro-average inclina-
ed to a rotating disc; Scabbling impacting the substrate at variable tion angle, representing the average of the pixel to pixel angles of the
angle with a metallic tip to create a chipping and powdering action. The stripe profile.
driving mechanism is compressed air. In order to define a unique parameter for describing the surface rough-
In order to examine exhaustively the concrete surface, an extensive ness, profilometer parameters were analyzed and correlated. For each
campaign of laser profilometer analysis was carried out before and after parameter given by the profilometer, average value and covariance have

RESEARCH - Seismic behavior 69


been conducted on two half bricks placed inside a properly designed
steel frame and connected between them by means of two carbon, glass
or natural FRP strips glued on both sides of the specimens. The frame
has been designed in such a way as to avoid any hindrance to the spec-
imen collapse. The specimens have been subjected to uniaxial tensile
tests under displacement control by means of an electro-mechanic test-
Fig. 38- (a) Laser profilometer, (b) geometry of FRP-strengthened prism and (c) experimental set-up.
ing machine, with a capacity of 100kN, connected to a personal com-
puter. Some specimens have been also monitored by means of unidi-
rectional strain gauges applied both on the brick surface and on the
fibres.
From the experimental tests, the specimens failure mode has been
analysed. Collapse has occurred for delamination in almost all cases.
However, some specimens have reached failure due to fibres crack, as in
Fig. 39- Average values (a) and covariance (b) of IR parameter for different surface preparations. the case of glass fibers and natural ones due to the different stiffness of
the FRPs. More than 30 specimens have been tested. In the case of spec-
been calculated for evaluating the quantity and quality factors of rough-
imens reinforced by carbon and glass fibers, the load-displacement
ness. The covariance can provide for information on surface homogene-
curve relative to failure for delamination obtained from the strain gages
ity; both quantities are interesting especially regarding the efficiency of
applied on FRP and brick surfaces was almost linear until failure which
concrete surface preparation methods. In the following, the roughness
occurred suddenly and for debonding of the fabric from the reinforced
is described by coefficient IR = Ria, where R and ia have been
bricks, with consequent removing of part of the substrate surface. The
described before. The parameter IR is used to give information on the
presence of relevant traction stresses is observed at the attachment of the
absolute value of the roughness and on its specific shape. Profilometer
FRP strip with the bricks where the delamination phenomenon begins.
analysis allows to correlate the casting methods with various degrees of
For what concerns the tests conducted on bricks reinforced by natural
roughness. In fact, the specimens casted with staves are more rough
fibers, the results, obviously, cannot be compared to the ones derived
because the disconnection of the staves increase the surface irregulari-
from brick reinforced by CFRP or GFRP; however, the tests have shown
ty. The specimens compacted by means of the vibration are more rough
interesting properties of natural fibers for not bearing applications such
due to different positions of aggregates and the presence of vacuum pro-
as on ancient masonry structures where FRP mechanical behaviour
duced by air bubbles.
sensibly affects the global behaviour of the structure.
The roughness of the concrete surface has been investigated before and
Finally, for each tested specimen, an accurate study of the substrate
after the preparation; its difference is a way to evaluate the efficiency of
after failure has been conducted for the exact definition of the delami-
each surface preparation method. In Figure 39a,b are shown the aver-
nated surface dimensions; the crack begins at a depth of about 10 mm
age values and the covariance of the IR parameter for all the different
surface preparation methods. Figure 39a shows that all the surfaces
prior to the treatments have very similar roughness while after them the
mean value of the IR parameter is particularly high in the case of scab-
bing and sand-blasting. On the contrary, all the surface preparation
methods strongly reduce the statistical dispersion of IR parameter
(Figure 39b); the scabbing and sand-blasting provide for the higher
level of homogeneity.

Delamination tests

Delamination tests have been carried out on bricks reinforced with FRP
materials, with the aim of studying both the collapse load value and the
failure modes that can occur if FRP materials applied during strength-
ening interventions collapse. In particular, delamination tests have Fig. 40- Failure mode of bricks reinforced with CFRP strips subjected to delamination tests.

70 RESEARCH - Seismic behavior


and this distance obviously depends on the experimental equipment, been focused on both masonry and reinforced concrete members. In
whereas the depth and the form of delaminated substrate appear almost particular, advanced FRP materials made of basalt fibers (bonded
constant and dependent on the testing modality (Figure 40). with traditional or inorganic cement matrix) have been used for the
From both in-situ and laboratory tests, important information were confinement of RC and masonry columns. In addition, the mechanical
obtained regarding composite materials behaviour and experimental properties of several FRP systems at fixed environmental conditions
procedures. In particular, it was noticed that shear tearing tests are and/or under cyclic actions, the behaviour of RC members confined or
extremely affected by instrumental errors that can take place during the strengthened in flexure and shear as well as of beam-column and
test conduction. In fact, while FRP laminates allow a perfect alignment beam- foundations joints have been deeply analyzed. Design criteria
of the composite with respect to the hydraulic pull machine, in the case for the seismic retrofit of masonry members and structures have been
of FRP strips the applied force is not perfectly aligned with the FRP also provided. The outcomes of experimental activities have, in most
concrete interface, and peeling stresses are generated, so causing a sig- cases, confirmed the reliability of CNR-DT200/2004 provisions, espe-
nificant reduction of the debonding force. cially for RC members. Regarding to masonry members, test results
Then, on the basis of the above described experimental results and in indicate that some coefficients of theoretical expressions provided by
accordance to workshops organized during the Research Activity, some CNR-DT200/2004 have to be refined and some other experimental
variations to the CNR DT 200/2004 guidelines, with particular regard results could be necessary to derive theoretical relationships specifi-
to shear tearing limit value and experimental procedure, were proposed. cally targeted at different masonry substrates and failure modes.
In particular, shear tearing test should be carried out following two dif-
ferent procedures, namely direct and indirect procedures. The direct 6. VISIONS AND DEVELOPMENTS
procedure is preferred if the load can be directly applied on the FRP
glued onto the specimen surface, especially in the case of laminates. If Innovative materials for the vulnerability mitigation of existing struc-
the FRP system is made of strips to be prepared in-situ, the indirect test tures have been largely analysed in the research activity. Further, new
procedure is preferred, in which load is applied by means of steel plates typologies of FRP systems, also made of several fibers and matrices
glued onto the FRP surface. types are spreading; on these new typologies further investigations are
strongly necessary in the future. The study of the feasibility of using
5. DISCUSSION composite systems made of inorganic matrix strengthened with natural
fibers fabrics is needful. In fact, the idea of using natural fibers is due
In this research task a large number of experimental tests has been to economic and environmental sustainability suggested by the use of
performed in order to validate the design equations provided by CNR- such materials. These studies could allow new FRP systems to be
DT200/2004 and to calibrate some coefficients included in these inserted in the actual guidelines in order to increase their use in the
equations to better fit theoretical predictions and experimental civil engineering applications.
results. The indications provided by CNR-DT 200/2004 for the vulnerability
Experimental tests have been carried out on both real scale or scaled mitigation of existing structures have been accepted by recent Italian
members by using traditional FRP systems or innovative typologies of guidelines, but further experimental tests will need, especially to miti-
FRP systems and advanced materials. The experimental tests have gate the vulnerability of existing masonry structures.

RESEARCH - Seismic behavior 71


7. MAIN REFERENCES mechanical model for simulating the NSM FRP strips shear strength
contribution to RC beams, Engineering Structures, Vol. 31 n. 4, Else-
- Aprile A., Benedetti A., Cosentino N., (2006), Seismic Reliability of vier.
Masonry Structures Strengthened with FRP Materials, 100th - Bruno D., Greco F., and Lonetti P. (2005), A 3D delamination mod-
Anniversary Earthquake Conference, San Francisco, paper n 1677. elling technique based on plate and interface theories for laminated
- Aprile A., Benedetti A., Steli E., Mangoni E., (2007), Seismic Risk structures, European Journal of Mechanics A/Solids, 24, 127-149.
Mitigation of Masonry Structures by Using FRP Reinforcement, - CNR DT 200/2004, Guide for the Design and Construction of
FRPRCS-8 University of Patras, Patras, Greece, paper n 424. Externally Bonded FRP Systems for Strengthening Existing
- Anselmi V., Aprile A., Benedetti A., (2005), Safety and reliability of Structures, Italian National Research Council, Rome (2004).
structures including ductile and brittle elements, ICOSSAR 2005, - De Lorenzis L., Rizzo A., (2006), Behaviour and capacity of RC
Augusti, Schuller, Ciampoli (eds), pp. 2183-2188, Rotterdam, ISBN beams strengthened in shear with NSM FRP reinforcement, 2nd Int. fib
90 5966 040 4. Congress, Napoli-Italy, June 5-8, Paper ID 10-9 in CD.
- Ascione F., Feo L., Olivito R.S. and Poggi C., La qualificazione del- - De Lorenzis L., Galati D. (2006), Effect of construction details on the
lesecuzione degli interventi di rinforzo strutturale con FRP a margine bond performance of NSM FRP bars in concrete, Proceedings fib
delle recenti Istruzioni per la progettazione, lesecuzione ed il control- Congress, Napoli, Giugno 2006.
lo di interventi di consolidamento statico mediante lutilizzo di com- - Dias S.J.E., Bianco V., Barros J.A.O. (2007), Low strength concrete
positi FRP, Proceedings of National Italian Conference Ambiente e T cross section RC beams strengthened in shear by NSM technique,
Processi Tecnologici La certificazione di Qualit dei materiali e dei Workshop Materiali ed Approcci Innovativi per il Progetto in Zona
prodotti da costruzione (in Italian), Naples (2005). Sismica e la Mitigazione della Vulnerabilit delle Strutture, University
- Bastianini F., Olivito R.S., Pascale G. and Prota A. (2005), Controllo of Salerno, Italy, 12-13 February.
di qualit e monitoraggio dei rinforzi in FRP, LEdilizia (in Italian), - Galati D., De Lorenzis L. (2006), Experimental study on the local
139, 66-71. bond behavior of NSM FRP bars to concrete, Proceedings CICE 2006,
- Benedetti A. and Steli E. (2007), Analytical Solution of the Shear Miami, USA, December 2006.
Displacement Curve for Reinforced Masonry Panels, The Tenth North - Galati D., De Lorenzis L. (2008), Effect of construction details on the
American Masonry Conference, St. Louis, Missouri, ISBN 1-929081-28- bond performance of NSM FRP bars in concrete, Advances in
6. Structural Engineering, Multi-science, in stampa.
- Benedetti A., Camata G., Mangoni E., and Pugi F., (2007), Out of - Milani G., Rotunno T. , Sacco E. and Tralli A. (2006), Failure Load
Plane Seismic Resistance of Walls: Collapse Mechanisms and Retrofit of FRP strengthened masonry walls: experimental results and numeri-
Techniques, The Tenth North American Masonry Conference, St. Louis, cal models, Structural Durability & Health Monitoring, 2 (1), 29-50.
Missouri, ISBN 1-929081-28-6. - Olivito R.S. and Zuccarello F.A. (2006), Indagine sperimentale per il
- Benedetti A., Mangoni E., Montesi M, Steli E. (2007), Verifiche di controllo dellapplicazione di materiali FRP a strutture murarie medi-
Sicurezza ed Interventi di Consolidamento Della Chiesa di S. Martino ante prove semi-distruttive e non distruttive, Proceedings of National
in Casola, INARCOS, 680, pp. 411-423. Italian Conference on Materials and Structures Experimentation (in
- Benedetti A., Steli E., (2008), Analytical models for sheardisplace- Italian), Venice.
ment curves of unreinforced and FRP reinforced masonry panels, - Rizzo A., De Lorenzis L. (2006), Analytical Prediction of Debonding
Construction and Building Materials, 22, pp. 175-185, Failures in RC Beams Strengthened in Shear with NSM FRP
doi:10.1016/j.conbuildmat.2006.09.005. Reinforcement, Proceedings CICE 2006, Miami, USA, December
- Bianco V. (2008), Shear strengthening of RC concrete beams by 2006.
means of NSM CFRP strips: experimental evidence and analytical mod- - Rizzo A., De Lorenzis L. (2007), Modelling of debonding failure for
eling, PhD Thesis, Dept. of Struct. Engrg. And Geotechnincs, Sapienza RC beams strengthened in shear with NSM FRP reinforcement,
University of Rome, Italy, submitted on December 2008. Proceedings FRPRCS8, Patras, Luglio 2007.
- Bianco V., Barros J.A.O., Monti G. (2009a), Bond Model of NSM FRP - Rizzo A., De Lorenzis L. (2009a), Behavior and capacity of RC
strips in the context of the Shear Strengthening of RC beams, ASCE beams strengthened in shear with NSM FRP reinforcement,
Journal of Structural Engineering, in press. Construction and Building Materials, Elsevier, Vol. 23, No. 4, pp. 1555-
- Bianco V., Barros J.A.O., Monti G. (2009b), Three dimensional 1567.

72 RESEARCH - Seismic behavior


- Rizzo A., De Lorenzis L. (2009b), Modeling of debonding failure for - Savoia M., Ferracuti B. and Mazzotti C. (2003), Delamination of FRP
RC beams strengthened in shear with NSM FRP reinforcement, plate/sheets used for strengthening of r/c elements, Proceedings of
Construction and Building Materials, Elsevier, Vol. 23, No. 4, pp. 1568- Second International Structural Engineering and Construction (ISEC-
1577. 02), Rome, Italy.

RESEARCH - Seismic behavior 73


Experimental Research on Seismic
Behavior of Precast Structures

ver the last two decades an extensive experimental and theore- bad seismic behaviour. Industrial one storey precast buildings were
O tical research activity aimed to investigate the seismic beha-
viour of precast structures has been carried out at European scale.
defined as inverted pendulum systems to which a very low value of the
behaviour factor was recognised. This was in conflict with the natio-
The results of this activity allowed to consolidate a good knowledge of nal codes that did not make any difference between cast-in-situ and
the seismic behaviour of precast structures and contributed to the precast frames. This was in conflict also with the experience of past
achievement of prefabrication in Europe with outstanding realizations earthquakes, where precast structures, except for the dry supports,
in terms of both quality and reliability. showed a very good ductile behaviour despite their non seismic desi-
gn. The lack of experimental data was adduced to justify the heavy
The research was developed within six research programmes. The penalisation of precast structures in this code.
first stage developed between 1992 and 1996 during the drafting of In order to verify the validity of this penalisation the Italian
the first ENV version of Eurocode 8 (EC8). The initial draft of the spe- Association of prefabrication industry promoted a campaign of analy-
cific rules for precast structures gave them the presumption of a very tical investigations that have been developed at Politecnico di Milano.
The investigation have been started with a set of cyclic tests on pre-
cast columns. Figure 1 shows one of the prototypes of precast columns
tested at ELSA Laboratory of the Joint Research Center of the
European Commission at Ispra, Italy. Cyclic and pseudodynamic tests
were performed for different reinforcement amount and axial actions.
An example of force-displacement diagrams obtained from cyclic
tests is shown in Figure 2.a. The energy dissipated over the half-
cycles was compared with the maximum value of dissipated energy
associated to a perfect elastic-plastic cycle, as shown in Figure 2.b.
The results of these tests confirmed:
- a good ductile behaviour with specific dissipation around 0.4, as
typical for cast-in-situ columns;
- a more reliable behaviour due to the absence of bar splices and to
the stable position of stirrups during the casting of concrete (precast
columns are cast in an horizontal position);
Fig. 1- Prototype of precast column.
Dimensionless values

Half-cycles
(a) (b)

Fig. 2- (a) Force-displacement diagrams obtained from cyclic tests. (b) Energy dissipated over the half-cycles compared with the maximum value of dissipated energy associated to a perfect elastic-plastic cycle.

74 RESEARCH - Seismic behavior


Fabio Biondini, Giandomenico Toniolo
Department of Structural Engineering, Politecnico di Milano, Milan, Italy

Figure 3.a, with four critical cross-sections dimensioned for a moment


mFh/2, may dissipate the same amount of energy which the hinged
arrangement shown in Figure 3.b dissipates in its two critical cross-
sections, dimensioned as they are for a double moment M=Fh2m. In
fact it is the global volume involved in dissipation, and not the num-
ber of plastic hinges, that gives the total amount of energy dissipated
by the structure.
Fig. 3- Energy dissipation in one-storey frames: (a) monolithic and cast-in-situ; (b) hinged and precast. To demonstrate this assumption, the two prototypes shown in Figure 4
were considered, the first cast-in-situ with monolithic connections,
- the fundamental importance of a narrow spacing of the stirrups the second precast with hinged connections. They have the same ove-
against the early buckling of longitudinal bars; rall dimensions, with the size and reinforcement of the columns cho-
- displacement ductility ratios between 3.5 and 4.5 consistent with the sen to achieve the same vibration periods and the same design sei-
code provisions for cast-in-situ frames. smic capacity in terms of base shear strength. With the combination
of the different heights and cross-sections (Figure 5) a number of
The ENV EC8 was therefore published with precast structures no types of frame were selected for both the frames.
more treated as inverted pendulum, but still penalised with a lower
behaviour factor with respect to cast-in-situ systems. The seismic response of the prototype was investigated in probabili-
stic terms for lognormally distributed material strengths and under
The seismic behaviour of cast-in-situ and precast structures has been artificial accelerograms, randomly generated so to comply with the
also investigated by means of proper numerical models on probabili- design response spectrum. A Monte Carlo simulation based on a large
stic bases. This investigation was carried out by means of non linear
dynamic analyses reproducing the real vibratory behaviour of the
structures under earthquake conditions. The aim was to demonstrate
that, under the same seismic action, the monolithic frame shown in

(a) (b)

Fig. 4- Prototypes of one-storey frames: (a) monolithic and cast-in-situ; (b) hinged and precast.
(a)

(b)

Fig. 5- Cross-sectional details of the columns. Fig. 6- Statistical distribution of overstrength ratio k: (a) monolithic and cast-in-situ; (b) hinged and precast.

RESEARCH - Seismic behavior 75


(a) (b)

Fig- 7- View of the structural prototypes with (a) monolithic and (b) hinged beam-column connections.

sample of incremental nonlinear dynamic analyses taken up to colla- 1000 accelerograms in one of the cases studied. These results prove
pse was therefore carried out for each prototype to compute the stati- that precast structures have the same seismic capacity of the corre-
stical parameters of the overstrength , ratio of the computed value sponding cast-in-situ structures, and confirm the correctness of the
over design value of the seismic capacity. Figure 6 shows the distri- values given by the code to the behaviour factor of concrete cast-in-
bution of overstrength computed for the two prototypes for a set of situ frames (q=4.5).

Fig. 8- Displacement time-histories for one of the pseudodynamic tests: numerical (thick lines) versus experimental results (thin lines).

76 RESEARCH - Seismic behavior


(a) (b)

Fig. 9- View of the structural prototypes with roof elements with axis (a) parallel to the direction of the seismic action (Prototype 1), and (b) orthogonal to the direction of the seismic action (Prototype 2).

The third stage of research developed during the revision of EC8 for of the seismic capacities of cast-in-situ and precast structures, and
its conversion to the final EN version. The preceding analytical at the same time the validation of the analytical model used in the
demonstration was effective, but an experimental confirmation was previous numerical investigation. Figure 7 shows a view of the full
still necessary. Therefore, taking advantage of the Ecoleader pro- scale prototypes. The comparison of the results obtained for these
gramme for the free use of the large European testing facilities, two prototypes highlights the expected large seismic capacities (about
pseudodynamic tests on full scale prototypes have been performed at ag=1g) of this type of structures and confirms the overall equivalen-
ELSA Laboratory. The aim was the direct experimental comparison ce of the seismic behaviour of precast and cast-in-situ structures (see

(a) (b)

Fig. 10- Experimental tests carried out on a connection. (a) Test set up. (b) Failure mechanism.

RESEARCH - Seismic behavior 77


(a) (b)

Fig. 11- Tree-storey full-scale prototype. (a) Transversal and (b) longitudinal section.

Figure 8). Growth research projects showed the good seismic performance of
precast structures under condition that the connections are properly
The fourth stage of research was developed within the Growth pro- over-dimensioned. The last aspect to be still clarified is therefore the
gramme. Two further prototypes consisting of six columns and a mesh actual behaviour of connections under seismic excitation. Based on
of beams and roof elements were designed to investigate the seismic these needs, the European research program Safecast has been recen-
behaviour of precast structures with roof elements placed side by tly launched to investigate the seismic performance of connections in
side. Figure 9 shows a views of these prototypes and of the testing precast systems. This project will involve a large campaign of experi-
plants. The prototypes differ only for the orientation of the beams and mental static tests carried out on single specimens, such as the con-
roof elements with respect to the seismic action. Common hinged con- nection shown in Figure 10, as well as pseudo-dynamic tests on a
nections are used between roof elements, beams and columns. The three-storey full-scale prototype shown in Figure 11. The results of
control of the pseudodynamic test is based on two degree of freedoms, the Safecast project are expected to complete the large research pro-
associated with the top horizontal displacements of the lateral frames, gram developed in Europe over the last two decades which provided
and of the central frame. Also the effects of cladding panels on the significant advances in the understanding of the seismic behaviour of
structural response has been investigated. precast systems and in the definition of reliable design criteria for this
type of structures.
The measured top displacements of lateral and central columns
during the pseudodynamic tests resulted practically coincident. This ACKNOWLEDGEMENTS
result proves that double connection between beams and roof ele-
ments gives a rotational restraint in the roof plane which enables the A number of partners participated to the research, coming from the
activation of an effective diaphragm action, even if the roof elements principal European countries subjected to seismic hazard. The natio-
are not connected among them. After the pseudodynamic tests both nal associations of prefabrication industry were involved (ANIPB for
prototypes have been subjected to a cyclic test under imposed displa- Portugal, ANDECE for Spain, ASSOIBETON for Italy, SEVIPS for
cements up to collapse. With a ultimate displacement du360 mm and Greece, TPCA for Turkey). Also the Italian associations of cement
a yielding displacement dy80 mm, a global displacement ductility industry AITEC and ready-mix concrete ATECAP participated to
equal 4.5 is deduced, as assumed by the new final version of EC8 for some stages. The research providers were: JRC Joint Research
the behaviour factor of precast frame systems. Centre of Ispra (of the European Commission), LNEC Laboratorio
Nacional de Engenharia Civil of Lisbon, Politecnico di Milano,
The results of the investigations carried out under the Ecoleader and University of Ljubljiana, NTUA National Technical University of

78 RESEARCH - Seismic behavior


Athens, ITU Istanbul Technical University, and the private labora- researches coordinated by prof. Giandomenico Toniolo of Politecnico
tories LABOR and LUGEA (I). The users were some producers of di Milano. Many other theoretical and/or experimental researches
precast structures (Magnetti Building I, Gecofin I, Civibral P, have been performed on precast structures by initiative of single com-
Prelosar E, Proet GR), and some auxiliary companies (Halfen D, panies for their specific interests. And this wide activity qualifies the
DLC I). prefabrication industry as one of the most advanced sectors in growth
What described in the present report refers to the series of European and innovation.

RESEARCH - Seismic behavior 79


State-of-the-art on the Research on
Structural Concrete in Italy

1.INTRODUCTION lary tension caused by the formation of water menisci developed in cap-
illary pores and responsible for the shrinkage of the cement paste
he research activity carried out in Italy on the structural concretes (Figure 1).
T is very busy in both the academic compartments and in the indus-
trial operations. The present report summarizes some of the most impor-
tant works in this area including three aspects of this activity:
- shrinkage compensating concrete in the absence of wet curing;
- properties of concretes with recycled aggregates;
- use of bottom ash from municipal solid wastes incinerators.

2. SHRINKAGE COMPENSATING CONCRETE IN THE


ABSENCE OF WET CURING

Shrinkage-compensating concretes have been extensively used in the


last forty years to minimize cracking caused by drying shrinkage in
reinforced concrete structures.
The first and most diffused system to produce shrinkage-compensating
concretes involves the use of expansive cements, according to ACI 223- Fig. 1- Water menisci interact with C-S-H fibers determining the shrinkage on cement paste,The New Concrete.
98, instead of ordinary portland cement. All these special binders are
based on a controlled production of ettringite. Recently [3], the combined addition of a shrinkage-reducing admixture
Another effective method to produce shrinkage-compensating con- with a CaO-based expansive agent has been found to be very success-
cretes, not covered by ACI 223-98 but commonly used in some coun- ful in producing restrained expansion of laboratory specimens protect-
tries, like Italy or Japan, lies in the use of a CaO and/or MgO based ed from water evaporation for just 1 day by using a plastic sheet and
expansive agent. This technology seems to be more advantageous with then exposed to air (60% R.H).
respect to that based on the ettringite formation from an economical as The influence of the SRA on the length change behaviour of a shrink-
well as from a practical point of view. age-compensating concrete includes two different aspects:
Recently, the addition of a shrinkage-reducing admixture (SRA) has - the  effect in Fig. 2 due to a reduction in shrinkage when the con-
been found to improve the behavior of CaO based shrinkage-compen- crete is exposed to drying, as expected for the presence of a shrinkage-
sating concretes especially in the absence of an adequate wet curing [1]. reducing admixture;
Although the actual cause of this synergistic effect has not been com- - the unexpected  effect, which is an increase in the restrained expan-
pletely explained, the use of this technology in construction industry sion when the concrete is protected from drying with respect to that
has been increased, in the last five years, particularly in Italy, with very obtained without SRA, all the other parameters being the same.
interesting results. By using a combination of CaO and SRA, then, it is possible to reduce
In the present report three remarkable examples of special reinforced the amount of expansive agent needed to obtain a fixed restrained
concrete structures are presented in which the use of CaO-SRA based expansion. This reduces the risk of residual un-reacted lime in the con-
shrinkage-compensating concretes was successfully carried out in order crete.
to prevent shrinkage related cracks and/or joints excessive opening in Furthermore, the performance in terms of initial restrained expansion
the presence of adverse curing conditions which are normally not suit- and final restrained shrinkage (or residual expansion), of SRA + CaO-
able for the use of this technique. based shrinkage-compensating concretes is less dependant on the cur-
SRAs (Shrinkage-Reducing Admixtures), are generally based on propy- ing efficiency so that the practical use of this technique is easier and
lene-glycol ether, neo-pentyl glycol or other similar organic substances, the results are more reliable.
that are able to reduce the drying shrinkage of concrete up to 50% if The synergistic effect in Figure 2 has been confirmed by Maltese et al
used in 1-2% by mass of cement. [4] who have found that the use of a CaO-based expansive agent with a
According to Berke et al. [2] the effectiveness of SRA must be ascribed shrinkage reducing admixture allows to obtain mortars less sensitive to
to the decrease in the surface tension of water .This reduces the capil- drying. These authors hypothesize that the synergistic effect of the

80 RESEARCH - Concrete
Mario Collepardi*
* ENCO, collepardi@encosrl.it

SRA-CaO combination must be ascribed to the massive formation of London, U.K.) had proposed the construction of several architectural
CaO elongated crystals during the first hours of curing. concrete walls (20 meters high and 60 meters long) having a sinuous
The same authors in [5] propose another mechanism of action: since the shape and no contraction joints (Figure 3).
SRA is an organic hydrophobic molecule, it could reduce the water sol- A special CaO-SRA based shrinkage-compensating self-compacting
ubility of CaO, retarding its reaction and, then, increasing the concrete (SCC) was studied in order to assure a marble-like look, as
restrained expansion according to Chatterji [6]. required by the designers, even in the presence of a very congested
Otherwise, Tittarelli et al. [7] have found that SRA doesnt affect the reinforcement (Figure 4) and, in the same time, to avoid the formation
speed of CaO reaction with water. of shrinkage related cracks along the surface.

Fig. 2- Schematic view of the influence of SRA on the length change behavior of a shrinkage-compensating concrete.
Fig. 3- View of bent and joint-less walls of the MAXXI, Rome, Italy.
Although this synergistic effect has been confirmed by several authors,
the actual mechanism of action needs further investigations in order to
be completely understood.
Notwithstanding this lack of knowledge, the use of this technology, in
the construction industry, has been growing in the last 5 years with
many successful and very interesting results.
In the second part of this paper, three remarkable case histories of spe-
cial reinforced concrete structures are presented in which the use of
CaO + SRA-based shrinkage-compensating concretes was successfully
carried out in order to prevent shrinkage-related cracks and/or joints
excessive opening in the presence of adverse curing and thermal con-
ditions.
The difficulties encountered in using this technique, in each case, will
then highlight describing the countermeasures which have been taken
to overtake them.

2.1 Case History of shrinkage compensating concrete in the absence of


wet curing: MAXXI of Rome

The Museum of Arts of XXI century (MAXXI) in Rome was the first rel-
evant Italian construction in which a SRA + CaO-based shrinkage-
compensating concrete has been used (2004-2006).
For this very prestigious building, the designers (Zaha Hadid Limited, Fig. 4- Example of steel congestion in a typical wall of MAXXI, Rome, Italy.
RESEARCH - Concrete 81
In order to demonstrate the effectiveness of this type of concrete in off- three days (to assure a correct hydration of the concrete cover) shrink-
set the formation of shrinkage cracks, its performances were compared age compensating concrete was designed in order to warranty a resid-
to those of an ordinary CaO-based shrinkage-compensating concrete ual restrained expansion of about 200 mm/m even in case of deficient
(without SRA) and of a plain SCC mixture without expansive compo- curing consisting in just 24 hours of protection by the formwork.
nent and SRA. Figure 6 shows the length change of the reinforced prismatic specimens
Table 1 shows the composition of these three SCCs having the same w/c manufactured with the three different SCCs according to ASTM C 878.
(0.48) and approximately the same cement dosage (350 kg/m3). Specimens were not put under water for 7 days as specified in ASTM C
878 test method but were protected with a plastic film for just 24 hours
Table 1 Composition of three different SCC (to simulate the protection offered by the formwork) an then exposed to
Mix CaO+SRA CaO Plain unsaturated air (60% R.H.) at 20C.
Cement CEM II A/L 42.5R (kg/m3)* 350 348 347 This curing condition was later introduced as curing method B in the
Limestone filler (kg/m3) 150 149 183 last version of the Italian standard UNI 8147 in addition to the curing
Gravel 4-16 mm (kg/m3) 847 884 871 method A previously specified, consisting in a total immersion in
Sand 0-4 mm (kg/m3) 908 916 903 water for 7 days as in ASTM C878. Actually, the curing method B
Water (kg/m3) 167 167 166 appears to be more realistic and similar to jobsite conditions.
Acrylic superplasticizer (kg/m3) 6.3 6.2 6.3 Even under these un-favourable conditions of curing, the CaO-SRA
CaO-based Expansive Agent 35 35 \ shrinkage-compensating concrete performed very well since the
Viscosity modifier (kg/m3) 4.2 4.1 4.3 restrained expansion after 24 hours of protection with a plastic film was
SRA 4.0 \ \ as high as 560 mm/m and, even after 140 days of exposure to unsatu-
(*) Blended Portland-limestone cement according to EN 197/1 rated air, a residual restrained expansion of about 250 mm/m was
recorded. On the contrary, the conventional CaO-based shrinkage com-
Figure 5 shows the strength development with time of the three com- pensating concrete showed a lower initial expansion (at lest 320 mm/m)
pared SCCs (CaO-SRA, only CaO and Plain). The strength of the expan- which completely disappeared after a week of exposure to air after
sive concretes was higher than that of the plain mix. This is probably which, the concrete started to shrink.
due to the consumption of a small part of mixing water caused by the Obviously, the plain concrete showed the worst performance reaching a
transformation of CaO into Ca(OH)2 which happens when the concrete restrained shrinkage of about 550 mm/m after 60 days when some
is still in the plastic state and to the consequent reduction of the actual cracks appeared on the specimen surface.
w/c. Comparing the behaviour of the CaO+SRA-based mix to that of the con-
On the other hand, a slight decrease in the compressive strength of the ventional shrinkage-compensating concrete, both the  and  effect of
SRA + CaO mix was recorded if compared to that of the CaO mix due Figure 6 can be detected.
to the presence of SRA as experienced in [8]. On the basis of the above results, the customer and the contractor
Although it was specified to protect the concrete surface for at least decided to adopt the SRA+CaO-based shrinkage-compensating SCC

Fig. 5- Strength development of three different SCCs. Fig. 6- Length change with time of the three different SCCs.
82 RESEARCH - Concrete
for the manufacturing of all the architectural concrete walls of MAXXI. age, of the various specimens manufactured.
Since it was the first time the contractor used an SCC, it was decided to As expected, the specimens manufactured and cured at 20C per-
carry out several field tests, before starting with the manufacturing of formed well showing a residual restrained expansion in the range of
the actual walls, in order to optimize all the casting procedures and test 210280 mm/m after 28 days of exposure.
the suitability of formwork. It was, then, a good chance to test on a real The prolonged mixing (at the same temperature of 20C) caused a
scale the effectiveness of the expansive technique. decrease of the initial as well as in the residual expansion as reported
Two field tests were successfully carried out in March and April of 2004 in [2].
with no cracks formation in two long minor walls of the basement. A little higher decrease was recorded in the expansion of the specimens
A third test carried out in June in order to verify the behaviour of the manufactured and kept at 30C and cast after 5 minutes of mixing.
expansive concrete in the presence of high temperature failed since Anyway the behaviour of these specimens can be considered accept-
after two weeks, some cracks appeared on the wall surface. The maxi- able.
mum temperature during the casting operation was as high as 35C and On the contrary, the combination of a high temperature of manufactur-
checking the transport documents of the trucks mixer it was verified ing and curing and a prolonged mixing cause a strong reduction in the
that, because of the congested traffic of Rome, the time elapsed between initial restrained expansion which was completely cancelled after just
the starting of mixing, in the batching plant, and the casting of concrete one week after which the concrete started to shrink.
into the forms had been in the range of 60-90 minutes, notwithstanding The problem was not eliminated by increasing the amount of expansive
the batching plant were located near the jobsite. agent up to 45 kg/m3 so that, being impossible to assure a transporta-
For this reason the cause of the failure was ascribed to a combined tion time lower than 60 minutes, the contractor decided to delay the
effect of the high temperature and of a too prolonged mixing time. This begin of the main wall construction to the autumn and to stop it during
hypothesis was confirmed by laboratory tests in which some ASTM the whole next summer.
C878 prismatic specimens were manufactured at 20C (with raw mate-
rials kept at 20C for 24 hours before the use) whereas other similar 3. PROPERTIES OF RECYCLED CONCRETES
specimens were manufactured at 30C (with raw materials kept at 30C
for 24 hours before the use). In both cases, some specimens were put The problem of recycling industrial wastes is of vital importance for a
into the forms after 5 minutes of mixing whereas the others were kept in sustainable progress in order to avoid disposals in the environments and
the mixer (in movement) for 60 minutes before casting at the same tem- possibly to save resources for the next generations. Such a problem has
perature of manufacturing (20 or 30C). After setting time (about 6 already been faced in using fly ash, silica fume and blast furnace slag,
hours) the specimens were demoulded and protected with a plastic film all wastes coming from industries other than cement and concrete.
till 24 hours, at the same temperature of manufacturing (20 or 30C). During the last decade a similar problem has been found for wastes
Successively, the specimens were exposed to unsaturated air (60% coming from the construction industry and from the concrete in partic-
R.H.) at the temperature of manufacturing (20 or 30C). ular [9] [10].
Figure 7 shows the behavior, in terms of restrained expansion or shrink- These wastes can be recycled as aggregates for concretes with two
advantages:
first, to save the environment specially in countries, like Nederland
and Belgium, where the available area to build is very limited;
second, to recycle this waste as aggregate specially in areas where
natural or artificial aggregates are scarce.
Therefore, concrete recycling, by using the readily available concrete as
an aggregate source for new concrete or pavement subbase layers, is
gaining importance because it protects natural resources and eliminates
the need for disposals.
Concrete recycling is a relatively simple process. It involves breaking,
removing, and crushing the existing concrete into a material with a
specified size and quality.
Fig. 7- Restrained expansion or shrinkage in different manufacturing and curing condition. The quality of concrete with recycled aggregate depends on the quality

RESEARCH - Concrete 83
of the recycled material used, the most important aspect being the ori- lower modulus of elasticity of the concrete with respect to the concrete
gin of the recycled material such as concrete or demolition, the latter with virgin aggregates. For the same reason, drying shrinkage and creep
including waste from brick walls and other type of rubbles. The crushed of concretes with coarse recycled aggregates are much higher (25-50%)
and sieved material, must be deprived by contaminating products such with respect to the virgin aggregates. The difference can be still higher
as wood, paper, plastic, and bitumen. This recycled material can be if also fine recycled aggregate is used.
used for pavement subbase layers. The permeability of a concrete with recycled aggregate is higher than
The process of recycling demolished concrete is based on four steps: that of the corresponding concrete at a given water-cement ratio. Again,
selection of wastes; the cement paste of the recycled aggregate is responsible for this draw-
crushing concrete blocks; back because the cement paste is more porous and permeable of the
removing of contaminating products; virgin stone.
mixing with virgin aggregates. The frost-resistance of the concrete with recycled aggregate is strongly
Reinforcing steel and other embedded items, if any, must be removed, reduced by the amount of fine fraction of the recycled aggregate.
and care must be taken to prevent contamination by other materials, such Therefore, in concrete exposed to freezing and thawing cycles the frac-
as: asphalt, soil and clay balls, chlorides, glass, gypsum board, sealants, tion of recycled aggregate smaller than 4 mm should be removed.
paper, plaster, wood, and roofing materials which can be troublesome. The fine fraction of recycled concrete smaller than 4 mm is responsible
The mechanical plants where to recycle concrete structures are not very for all the above mentioned limits of the concrete with respect to the
different from those adopted to treat crushed virgin aggregates. corresponding concrete manufactured with virgin aggregates. However,
If the material is devoted to concrete production, further crushing and if the fine fraction is ground very finely (smaller than 0.1mm) it can be
sieving are needed before mixing it with virgin aggregate [11]. used advantageously in manufacturing self-compacting concrete as
The crushing characteristics of hardened concrete are similar to those filler to improve the cohesiveness of the concrete [13, 14].
of virgin rock and are not significantly affected by the quality of the
original concrete. Recycled aggregates can be expected to pass the 4. USE OF BOTTOM ASH FROM MUNICIPAL SOLID
same tests required for conventional aggregates. The recycled concrete WASTES INCINERATORS
can be batched, mixed, transported, placed and compacted in the same
way as conventional concrete. Special care is necessary when using Mineral solids in form of fly ash and bottom ash are produced by burn-
recycled fine aggregate. Only up to 10% to 20% recycled fine aggregate ing municipal solid wastes in incinerators (MSWI). Fly ash is negligi-
is beneficial. The aggregate should be tested at several substitution ble and it is so chloride-rich that it cannot be used as mineral addition
rates to determine the optimal rate. in cement-based mixtures for reinforced concrete structures.
On the other hand, bottom ash is about 25% with respect to MSWI and
3.1 Properties of recycled fresh concretes its chloride content is negligible, so that it could be potentially used as
mineral addition for manufacturing concrete mixtures. However, ground
The amount of mixing water of the coarse recycled aggregate is about bottom ash (GBA) from MSWI does not perform as well as other miner-
5% more with respect to that of virgin aggregate at given size. This al additions (silica fume or fly ash produced by coal burning) due to the
value becomes as high as 15% when the recycled aggregate contains presence of aluminium metal particles which react with the lime formed
also the fine fraction. This effect is due to the rough texture of the aggre- by the hydration of Portland cement and produce significant volume of
gate and the cement paste surrounding the recycled aggregate. hydrogen in form of gas bubbles which strongly increase the porosity of
However, the use of superplasticizers and mineral additions can com- concrete and reduce its strength.
pletely overcome this drawback [12]. Due to this drawback, a new process was developed to completely sep-
arate the aluminium metal particles through a mechanical removal of
3.2 Properties of recycled hardened concretes metals and a special wet grinding of bottom ashes. At the end of the
process GBA was used as an aqueous slurry to replace Portland
Due to the higher porosity, related to the lower density, the recycled cement.
aggregates are responsible for the lower strength of the concrete with Some researches have actually shown the pozzolanic activity of ground
respect to the concrete manufactured with virgin aggregates. MSWI bottom ashes showing their reactivity with lime or portland
Due to the lower rigidity, recycled aggregates are responsible for the cement clinker [15,16]. Nevertheless, no successful use of MSWI bot-

84 RESEARCH - Concrete
tom ashes as mineral addition in concrete has been reported, because tom ashes from MSWI in comparison with concretes containing 20% of
of the side effects of this addition. According to Bertolini et al [17], the coal fly ash or 10% of silica fume.
main side effect is related to the evolution of hydrogen gas after mixing The performances of GBA with mean sizes of 3 and 5 mm were higher
due to the presence of metallic aluminium. In the alkaline environment than that of the coal fly ash particularly at 1-60 days. The finest ground
produced by the hydration of portland cement (pH around 13), corro- bottom ash (with a mean size of 1.7 mm) performs as well as silica fume
sion of some metals (mainly aluminium) produces a great amount of in terms of compressive strength, water permeability, chloride diffusion
gaseous hydrogen. After placing and compaction of concrete, this gas is and CO2 penetration.
entrapped in the fresh material, producing a network of bubbles that These results appear in particular to be very interesting from a practi-
leads to significant reduction in the strength and increase in the per- cal point of view since it will be possible to manufacture big amounts of
meability of the hardened concrete. a pozzolanic material as effective as silica fume (which is not available
The present report summarizes the results of a research [18] aimed at and very expensive) in producing high-performance concrete in agree-
developing suitable treatments to allow the use of MSWI bottom ashes ment with a sustainable progress for the re-use of waste materials
as mineral additions for the production of structural concrete without instead of a their disappointing disposal.
the evolution of hydrogen gas due to the presence of metallic alumini-
um particles. 5. CONCLUSIONS
Ground bottom ashes (GBA) from municipal solid waste incinerators
(MSWI) were manufactured according to a new technology based on a The research on the progress of structural concrete in Italy is very
high degree of separation of metals including the heavy ones, the wet active. Three sections of this activity have been illustrated in the pre-
grinding process, and other specific technical solutions to completely sent report:
remove the aluminium metallic particles. At the end of the process, a shrinkage-compensating concrete for crack-free structures even in
fluid slurry was obtained with particle size in the range of 1-5 mm. By the absence of a wet curing;
changing the wet grinding time three GBA were produced with a mean recycled concrete with special application of a fine powder material
particle size of 5 mm, 3 mm and 1.7 mm. devoted to self-compacting concrete;
Compressive strength and durability measurements were carried out in bottom ash from municipal waste incinerators as pozzolanic materials
concretes where Portland cement was replaced by 20% of ground bot- for high-performance concrete.

RESEARCH - Concrete 85
REFERENCES MET/ACI International Conference on Superplasticizer and Other
Chemical Admixtures, Sorrento, 2006, pp. 103-115.
[9] N. Kashino and Y. Ohama, RILEM International Symposium
[1] M. Collepardi, The New Concrete, pp. 347-362, Second Edition, Environmental Conscious Materials and Systems for Sustainable
Ed. Tintoretto, Villorba, Italy, 2010. Development, 2004, pp. 179-186.
[2] N.S. Berke, L. Li, M.C. Hicks and Bal, Improving concrete per- [10] J.Zh. Xiao, J.B.Li, Ch. Zhang, On relationship between the
formance with Shrinkage-Reducing Admixtures, 7th CAN- mechanical properties of recycled aggregate concrete: An
MET/ACI International Conference on Superplasticizer and Other overview, Materials and Structures, 39, pp. 655-664, 2006.
Chemical Admixtures in Concrete, Berlin, Germany, Ed. V.M. [11] M. Bassan, M. Menegotto, G. Moriconi, Precast Structural Con-
Malhotra, 2003, pp. 37-50. crete with Recycled Aggregates, Proceedings, the 2nd fib Con-
[3] M. Collepardi, A. Borsoi, S. Collepardi, J.J. Ogoumah Olagot, R. gress, Naples, ID 6-19, 2006.
Troli, Effects of Shrinkage-Reducing Admixture in Shrinkage [12] V. Corinaldesi, G. Moriconi, Role of Chemical and Mineral
Compensating Concrete Under Non-Wet Curing Conditions, Admixtures on Performance and Economics of Recycled-
Cement and Concrete Composities, 6, 2005, pp. 704-708. Aggregate Concrete, Proceedings of the Seventh CANMET/ACI
[4] C. Maltese, C. Pistolesi, A. Lolli, A. Bravo, T. Cerulli and D. International Conference On Fly Ash, Silica Fume, Slag and
Salvioni, Combined Effect of Expansive and Shrinkage Natural Pozzolans in Concrete, Editor V.M. Malhotra, ACI SP
Reducing Admixtures to Obtain Stable and Durable Mortars, 199, Vol. 2, pp. 869-884, 2004.
Cement and Concrete Research 12, 2005, pp. 2244-2251. [13] V. Corinaldesi, G. Orlandi, G. Moriconi, Self-compacting con-
[5] C. Maltese, A. Lolli, C. Pistolesi, A. Bravo, and T. Cerulli, crete incorporating recycled aggregate, Innovations and De-
Combined effect of expansive and shrinkage reducing admixture velopments in Concrete materials and Construction, Proceeding
on microstructure of mortars and concretes, Proceeding of the of the International Conference, Editor Ravindra Dhir, Dundee,
International Conference on Durability of Concrete, Editor: V.M. UK, pp.9-11, 2002.
Malhotra, Montreal, 2006, pp. 781-796). [14] V. Corinaldesi, F. Tittarelli, L. Coppola G. Moriconi, Feasibility
[6] S. Chatterji, Mechanism of expansion of concrete due to the pres- and Performance of Recycled-Aggregate Concrete Containing
ence of dead burnt CaO and MgO, Cement and Concrete Fly Ash for Sustainable Building Development, Proc. of the
Research 25, 1995, pp. 51-56. Three-Day CANMET/ACI Int. Symposium On Sustainable
[7] F. Tittarelli, S. Monosi, G. Moriconi, R. Troli, Effects of a shrink- Development of Cement and Concrete, SP 202-11, pp. 161-180,
age compensating Admixture in Shrinkage-Compensating 2001.
Concretes, presented in 9th ACI International Conference on [15] K.A. Paine, R. K. Dhir, V.P.A. Doran, Unprocessed and
Superplasticizer and Other Chemical Admixtures, Seville, 2009. processed incinerator bottom ash as a cement bound material in
[8] M. Collepardi, R. Troli, M. Bressan, F. Liberatore, G. Sforza, R.K.Dhir, T.D.Dyer, K.A. Paine (Eds.) Use of incinerator ash,
Crack-Free Concrete for Outside Industrial Floors in the Absence Proceedings of the International Symposium organized by the
of Wet Curing and Contraction Joints, Suppl. of 8th CAN- Concrete Technology Unit, University of Dundee (UK) 20-

86 RESEARCH - Concrete
21/3/2000. lepardi, MSWI ashes as mineral additions in concrete, Cement
[16] A. Macias, E. Fernandez, S. Goi, A. Guerrero, Valorizacion de and Concrete Research, 34, 1899-1906, 2004.
las cenizas de inceneracion de residuos solidos urbanos en los [18] M. Collepardi, S. Collepardi, D. Ongaro, A. Quadrio Curzio, M.
materiales de construccion, Papel de los sectores cementero y Sammartino, Concrete with bottom ash from municipal solid
de la construccin en la gestin y reciclado de residuos, CSIC, wastes incinerators, paper accepted for publication on Second
Madrid, 2001. International Conference on Sustainable Construction Materials
[17] L. Bertolini, M. Carsana, D. Cassago, A. Quadrio Curzio, M. Col- and Technologies (SCMT), Ancona, Italy, June 2010.

RESEARCH - Concrete 87
Damages of LAquila Earthquake

1. INTRODUCTION construction may undergo severe damage but must avoid collapse in
order to protect the lives of its occupants. The earthquake on April 6 of
his chapter summarizes the main findings of a chapter of the spe- last year was of an intensity comparable to that found in the new
T cial issue of Progettazione Sismica devoted to LAquila earthquake
the author edited and which is about to appear. In particular, in the fol-
Technical Regulations of Construction, and many buildings in LAquila
historic center sustained damage compatible to those norms (the acce-
lowing the papers of Carocci and Lagomarsino (2010) regarding lerograms recorded by instruments present in the area of the seismic
masonry buildings, Cosenza et al. (2010) for reinforced concrete buil- crater revealed a spectrum only slightly higher than that in the buil-
dings, Di Ludovico et al. (2010) about school buildings, Casarotti et al. ding plans).
(2010) for hospitals, Menegotto (2010), Faggiano et al. (2010) for indu- Therefore, one can attest that well-constructed masonry buildings, in
strial structures, Dolce et al. (2010) for lifelines, are reported. other words those built with connections that allow the structure to act
as a unified organism (clamping between orthogonal walls, junctures
2. MASONRY BUILDINGS between walls and floors) are able to sustain damage without manife-
sting fragility collapse.
Are masonry buildings able to withstand a strong earthquake such as This article will not deal with monumental structures, but with minor
the one that struck LAquila on the 6th of April, 2009? Is it possible to masonry ones, common in the many historic centers in and around
repair damaged buildings, guaranteeing an adequate safety level to LAquila, where severe damage and collapse was diffuse. Wishing to
their inhabitants in an area with such a high level of seismic risk? take into consideration the number of victims as a parameter of vulne-
These are the questions asked by researchers and government techni- rability, outside of LAquila there were 97 victims in masonry buildings,
cians, but especially by those people who lived through this tragic event compared to 14 victims in r.c. buildings (it should be said that in these
and long to see the restoration of historic centers and return to their towns the percentage of r.c. buildings is much lower than in LAquila
homes but are also afraid. and the buildings were also lower).
The answer to these questions must be complex and detailed because What are the factors that justify the different behavior between the
there were so many factors that influenced the seismic behavior of apartment buildings in LAquila and the buildings in the surrounding
Aquilan constructions. historic centers? Effectively, there are two reasons: construction quality
In this earthquake, more than others, the effects of localized seismic and the presence of incongruous subsequent restoration.
amplification played an important role. If one analyzes the macro-sei- In the masonry buildings of the historic center of LAquila, most of
smic consequences of many of the historic centers in the Aterno Valley which were rebuilt after the tragic earthquake in 1703, a series of struc-
(south-east of Aquila), one will immediately observe that the villages tural modifications characteristic of the LAquila rules of thumb were
that were struck the hardest (Onna and Villa SantAngelo, I=9-10) were easily recognizable: wooden beams (elements built into the wall thick-
near other villages where the damage was limited: Onna (I=9-10) which ness and connected externally by way of small tie rods), to improve con-
is only 1500m as the crow flies from Monticchio (I=6). Near the town of nections between walls; the connection of roofs to the tops of the walls,
Villa SantAngelo, the distance between the chief town (I=9) and the vil- utilizing external wooden posts. These rules were adopted even in the
lage of Tussillo (I=8) is less than 700 m. Even within the town of smaller surrounding historic villages, but often with lesser construction
LAquila, there were zones where the damage was clearly concentrated. know-how and lower quality building materials.
Especially in LAquila, it was noted that most victims were found in Even after the earthquake in 1703, the historic centers of the Aterno
reinforced concrete (r.c.) buildings (135 versus 52 in masonry buil- Valley were struck by significant earthquakes, in particular, the one in
dings). In the historic center, damage to churches was severe, in some Avezzano in 1915. In fact, the repairs and seismic reinforcements are
cases with extended collapse, but the increased vulnerability of these easy to discern (scarp walls, buttresses, and tie rods adjacent to
structures is well-documented by history (in the past, churches were masonry walls) and in many cases, the partial reconstruction of colla-
often found to have the highest concentration of victims). Some severe psed portions. These interventions often functioned well, but in others
damage was rather diffuse in apartment buildings, but did not lead to the vulnerability remained, proving once again how difficult it is to
collapse except for localized cases in loggias or stairwells; almost perform truly efficient seismic restorations and improvements.
always found in abandoned or poorly maintained buildings. As far as recent restorations were concerned, (only in certain cases done
Modern criteria for structural safety, based on tests of differing limit sta- for consolidation purposes) it should be said that while LAquila
tes (performance-based design), state that during a rare earthquake the appears to be better preserved (due to both the large number of protec-

88 RESEARCH - LAquila Earthquake


Gaetano Manfredi
Department of Structural Engineering, Universit degli Studi Federico II, Naples, Italy

ted buildings as well as perhaps the number of uninhabited buildings street front and the contextual absence of building compactness in the
in certain areas), the smaller historic centers were subject to frequent depth of the fabricated bodies (for example: San Demetrio Colle).
changes: enlargements, transformations, added floors, restored floors Such generalized characteristics are largely identifiable in centers
and roofs. These interventions were often done with building materials analyzed herein, even if in our conscience it is not possible to establish
and techniques incompatible with the original structures: alterations in a correlation between the type of building aggregate and the damage
weight displacement, differing rigidity between elements, and dange- sustained. In any case, the observations relative to the scale of the
rous increases in mass. aggregate are important not only for identifying where the main vulne-
The final important question is if it is possible to repair severely dama- rability lies and differentiating it from that of single buildings, but espe-
ged constructions and rebuild collapsed parts of aggregates with tradi- cially for establishing the project phases for restoration and reconstruc-
tional construction techniques (i.e. stone), eventually modified on the tion.
basis of the experience of this earthquake. To this end, an examination The diffusion of arches placed between facing walls along the streets is
of construction damage to masonry is proposed based not only on the a repeated tendency in these villages subject to earthquakes. However,
identification of vulnerable areas, but also on resistance (anti-seismic it is not possible to prove that the volume of overpasses (built onto the
regulations) as a means of truly learning something from the test of an inhabitable surface area) though widely present in many of these cen-
earthquake. ters (for example: Santo Stefano di Sessanio), is the result of the evolu-
tion of the arches, even if it is obvious that both tend to create a rather
2.1 Typological aspects of masonry buildings in the historic centers of good connection between the facing blocks and constrict wall move-
LAquila ment where they are located. Street names such as Contrada dellArco,
Chiassetto dellArco, Via sotto gli Archi, present in nearly all of the cen-
The historic centers of the LAquila territory are principally made up of ters we visited, indicate that this construction technique dates way
simply organized masonry buildings. Even the grouping within connec- back.
ted buildings seems to be to simply follow the rules imposed by the oro- Buttresses, built-on scarp walls and tie rods should be counted among
graphy of the terrain. the pre-modern anti-seismic defenses that systematically characterize
Due to the areas extended (including seismic) history, construction the building range of these centers. Each technique can be identified
peculiarities and specific characteristics for the arrangement of aggre- by materials analysis, production, function and use after the many
gates for each of these centers have been adopted, and deserve to be earthquakes that have struck this area throughout history. In fact, many
taken into consideration, but will not be dealt with in this article. of these elements have been utilized in restoration phases while in other
Nevertheless, with reference to aggregates, one can affirm that the cases, they seem to have become the rule of thumb for the partial or total
urban centers are characterized by smaller dimensioned blocks of buil- reconstruction of buildings.
dings that develop along the morphology of the terrain where they were Evidence of the damage done by historic earthquakes is still often visi-
built. ble, such as the often seen out of axis external walls of buildings, and
In the centers located on mountainsides (for example: Fossa, Casentino, the presence of stone elements, often hewn after being taken from older
Tussillo, Castelnuovo) or those built on plateaus (for example: Poggio collapsed buildings and reutilized in post quake reconstruction..
Picenze, SantEusanio Forconese, Villa SantAngelo), both cases are Given the close relationship that links these centers to the territory
characterized by sloping terrain. One can observe the two types of typi- (agricultural/pastoral economy), the inhabited buildings still show tra-
cal aggregates: parallel blocks and blocks built orthogonally to the ces of their rural past with ground floors usually for storage or stalls and
incline as well as a wide range of variants in between. the living spaces in the upper floors accessed by means of external
Buildings were arranged according to the conformation of the terrain staircases along a balcony.
and the existing buildings, and the relation to seismic vulnerability can The structure of these stone stairs is usually surrounded by pillared
be listed schematically: height differences between walls on opposite masonry that tends to create volume towards the street front; in cases
sides of the street; scaling of contiguous building fronts; number and where the building has another floor, the additional volume is used as a
height of external exposed walls, and the placement of the building in loggia while access between the two inhabited upper levels is solved by
the aggregate. way of an internal wooden stair.
In the oldest centers where the building aggregates developed along In some of the centers visited, the houses showed more urban cha-
side streets, the aggregates are characterized by the continuity of the racteristics in how they were set up on the street front while at times

RESEARCH - LAquila Earthquake 89


maintaining the external stone stairway to access the inhabited upper one pitch towards the street or two slopes with a central ridge (depen-
floor. In any case, a common peculiarity seemed to be the add-on of ding on the configuration of the whole aggregate). The wooden beams
advancing volume on the front of the house for more living space. are always positioned without overhang; even in the case of buildings
Along with the mono-celled buildings on the street front, which we have placed on corners or at the head of a block where the configuration of
referred to up till this point, larger and more complex structures exist the pitch is a triangular pavilion shape (therefore with the presence of
without being large enough to be considered real apartment buildings two directions of slope). The attention paid to not creating overhanging
(with the exception of sporadic cases in the smallest villages). These structures and nevertheless the wooden elements were almost always
were derived mostly from the fusion of pre-existing smaller buildings connected to the walls. The attention to the way rooms were construc-
with the addition of an upper floor. ted for reducing seismic vulnerability is also observed in many small
The stall-hayloft is another very common type of building which com- details, such as the light eaves made from wooden and bamboo struc-
pletes the overview of the minor architecture in these centers. Its func- tures and the balconies entirely made of wood. Such construction mea-
tion determines its appearance, and though it is completely distinct sures were evidently realized in order to limit the lethal effects of dama-
from the living area, it is often placed contiguously to the house and the- ge on the lives of the inhabitants.
refore inserted into the building mesh. The most important difference To conclude this chapter which focuses on the characteristics of the
from a construction and structural point of view is that these construc- buildings in minor centers of the LAquila area, it is necessary to men-
tions represent the largest part of the wall structure (where some walls tion two aspects that should be considered in the observation and eva-
reach up to 10 m in length) with the presence of dilated wall light which luation of damage: the techniques adopted in recent restorations and
corresponds to the systematic use of roof coverings connected to exter- the conservation state of the buildings.
nal walls. In relation to the first of the two aspects mentioned above, it should be
From our observations, it was possible to conclude that up until now, the observed that as long as masonry walls were the dominant technique
wall thickness of the cell walls were adequate for the dimensions of the used for building, transformations on existing buildings took into consi-
buildings (3 levels above ground and usual light openings); in fact these deration this vulnerability and often brought about seismic improve-
prove to be between 50 and 70 cm except at ground level or in rare ments within the limits of the techniques available, eliminating weak-
cases where stone vaults are present where up to 1 m of thickness can nesses and introducing protective measures.
be found. Congruous tapering was also observed in upper floors; scant Unfortunately, recent transformations appear mostly linked to damage
wall thickness was observed only on rare occasions, always in the pre- after the earthquake. The most common intervention is the substitution
sence of more recent interventions. of the original roof covering with a new one which generally followed the
As far as horizontal diaphragms are concerned on the first floor (the same configuration of the preceding one, but was at times made from an
division of lower floors used for storage or stalls and inhabited upper r.c. structure with hollow or infill panels or with metal beams instead of
floors), there were always barrel vaults with the generatrix perpendicu- wooden ones. The result of these changes with regards to seismic vul-
lar to the faade. The construction technique calls for a mesh of rough nerability was quite often negative.
hewn stone elements and the presence of compact buttresses. As far as the second aspect is concerned, it seems clear that the con-
The upper floors usually have simple wooden flooring. However, brick servation state of the building played an important role in whether or
covered metal beams were frequently noted, certainly from substitu- not it was damaged, but also with reference to the buildings near it in
tions in the last century (for example after the Avezzano earthquake). the aggregate. In fact, in cases where the adjacent buildings were poorly
Metallic beams, hollow flat blocks covered by flooring were more constructed due to many decades of abandon, there was less damage
recent. In some cases, the beams were connected to bars or welded pla- than expected due to the stabilizing contribution of the contiguous cells.
tes and anchored to the masonry. In some of the centers visited, the presence of many restoration work
Another type of horizontal diaphragm is made of thinly layered brick zones was observed (for example at Villa SantAngelo).
vaults (two thin layers of interwoven bricks) which have proven to be As far as we could discern, in most cases the techniques these workers
very vulnerable; these vaults are also relatively recent (from the end of were utilizing were far from those that should have been adopted for the
the 19th century to the first decades of the 20th century). knowledgeable recovery of historic heritage. On a positive note, also in
The roof coverings are made of wood with the constant presence of view of the reflections in the next chapter, we also encountered that the
overhangs made of dripstone or wooden elements. In buildings in aggre- restoration and reuse of abandoned masonry structures had been going
gates, the ridge of the roof is parallel to the street front and may have on for some time, and was probably linked to a search for available

90 RESEARCH - LAquila Earthquake


living space close to the city center but also due to a slightly higher tou- floors. At times, the new building eliminated certain walls orthogonal to
rism interest (also foreign). the faade yet present in the lower floors when raising the height, thus
altering the passage through transverse walls (and generally rotating the
2.2 Damage observations: vulnerability and resistance orientation of the floors).
Figure 2 shows the case of a building where one of the walls orthogonal
The description of the damage observed will be discussed by topic to the faade was absent on the top floor, yet traces can be observed on
according to the most important building aspects (at times re-examining lower floors. The excessive distance between the transversal walls ren-
the topics mentioned above) and by referring to the principal damage dered overturning nearly inevitable in the case of seism.
and collapse mechanisms. Such construction defects, observed on many occasions, can have easily
Subdivision is necessary in order to shed light on the probable causes been revealed in the buildings in the historic centers and the vulnera-
that facilitated or more aptly limited damage. In this way, characteri- bility derived from such configurations found and resolved preventati-
stics of vulnerability will be highlighted as well as the strengths which vely.
should lead to further reflection for future reconstruction.

2.2.1 Structural organization of buildings and position in the aggregate


As far as the placement of the building in the aggregate is concerned, it
has been revealed that the configuration of the corner or the head of the
block has proven to be the most disadvantageous as widely noted befo-
re. In Figure 1, we can observe three buildings each placed at the head
of a block. All of them sustained the collapse of the front wall and a
great portion of the two side walls. Good organization and regularity of
the vertical load elements of the building represents a point of strength
with regards to seismic action. The orientation of orthogonal walls to the
faade, the placement and quantity of openings and the connection with
the horizontal elements determines the greater or lesser flexional trim
(vertical and horizontal) of the walls. Irregularity or changes introduced Fig. 2- LAquila: the absence of transverse walls on the upper floor of the building.
in the configuration of the whole often proves fatal during an earth-
quake. The vulnerability derived from the absence of clamping between walls
One often noted case is that of the buildings created by the fusion of of buildings built at different times was confirmed (growth phases and
pre-existing contiguous buildings and the later addition of one or more evolution of the buildings in an aggregate). Figure 3 shows the collapse
of the portion of a wall due to construction discontinuity by the simple
placement of a wall next to the pre-existing corner.

Fig. 1- Tempera: collapse of buildings placed at the head of a linear aggregate. Fig. 3- Poggio Picenze: local collapse due to discontinuity.

RESEARCH - LAquila Earthquake 91


Beyond typical growth or urban congestion, many violations of building the house, and consequently, the use of spaced trusses are necessary
codes were found in the transformations of aggregates in historic cen- (Figure 5-6). From a construction point of view, it is clear that the con-
ters. In other words, demolition and reconstruction of weight-bearing nections between roof and walls were put into place at the time of ori-
walls in different positions in order to obtain small extensions or recon- ginal construction as demonstrated by the position of lintels which
figurations of the block produced negative alterations of cells. In these necessarily implies their placement before mounting the boards and the
cases, the original clamping was lost and nearly impossible to restore subsequent roof covering. Finally, it should be noted that the use of con-
on the newly placed walls, and thereby eliminated connecting elements. necting trusses to the masonry walls by means of wooden lintels was
In the case shown in Figure 4, a cut off wooden tie rod is visible right also consistently found in churches in the LAquila region, and there-
inside the internal wall, probably during restoration of the building or fore one can assume that this technique was generally associated with
its apartments. This alteration, which eliminated an anti-seismic con- construction configurations with great light.
struction norm put in place by previous craftsmen certainly contributed Connections by means of wooden lintels were also observed in larger
to the collapse that was verified in that building unit. and more important buildings like that in Figure 7 which gives an
example of the positive contribution offered by the roof covering. The
corner building shows systematic tie rods between floors and the top of

Fig. 4- Villa SantAngelo: violations: the external wooden connection to the metallic tie rod was cut during fusion when the
wall direction was altered.

Fig. 5- Villa SantAngelo: a wooden truss leans on a wooden lintel.


2.2.2 Roof Coverings
As already mentioned, wooden roof structures generally do not
overhang, demonstrate well-placed orientation, and are usually well-
connected to the masonry walls evidenced by metallic or wooden con-
nections visible from the external walls. Such connections are associa-
ted to the very common presence of pavilion type roof coverings or the
triangular portion of the pavilion roof, where the presence of ridge raf-
ters or slope necessitates the organization of the mesh of the wood to eli-
minate thrust. In roofs whose main direction is limited by masonry
walls they are posed upon, it was not rare to find that an effective con-
nection to the wall rims contributed to limiting the damage and protru-
sion of the external walls.
The habit of connecting, often by way of wooden lintels, roofs to
masonry walls seems to be linked to the dilated dimensions of the cells;
in fact this modality is constantly found in uninhabited spaces such as
Fig. 6- Casentino: masonry cyma of a hayloft and the truss touching the wooden lintel (note how the internal connection is
stalls-haylofts. In the latter, the floor area is usually larger than that of thanks to a board nailed to the actual lintel).

92 RESEARCH - LAquila Earthquake


the wall put into place when the masonry walls were elevated. One can
also observe the presence of wooden tie rods which are connected to the
trusses which create the pavilion roof structure (the presence of a tie rod
may indicate the absence of an orthogonal wall to the faade). In any
case, it is interesting to note how the building has cracked in the floor
with walls (second mode mechanism) while initial cracking is entirely
absent on external walls.

Fig. 9. Villa SantAngelo: diffused collapse of the masonry cyma without involving the window openings.

the external walls without involving the architrave structure of the ope-
nings on the highest floor (Figure 9). In both cases, damage may have
been caused by movement relative to the placement of wooden elements
Fig. 7- Poggio Picenze: a small corner building of a block built on a slope. and the effect of the addition of wall openings which weakened wall
strength.
Below, the Figures demonstrate how damage can be attributed to a lack
Localized damage in the tympanum walls (which is one of the most
or modification of the roof structure. Damage caused by the poor qua-
commonly found types of limited damage observed in the LAquila cen-
lity or advanced state of decay of the wooden elements of the roof was
ters) may be more or less extended depending on the surrounding con-
very common. This was due to the lack of connection to the masonry
ditions, where the tympanum wall was placed in the aggregate (differing
(along with the scarce conditions of placement), the lack of connection
height to the contiguous buildings), or in relation to the quality of the
between main and secondary elements of the roof covering, or the bad
masonry pattern at the top of the wall.
conditions of the wooden boards (which if in good condition guarantee
Extensive damage involving the top of the external walls may be attri-
a very limited bracing effect).
buted to where the wooden mesh elements (primary and secondary)
In these cases, local collapse was recorded in the tympanum of the cell
were not interconnected which led to greater displacement at the top
walls (Figure 8), in other words extended collapse to the fascia above
part of the wall during the quake.
Localized collapse to a portion of the masonry cyma illustrated in
Figure 10 is presumably due to thrust in the rafters of the slope of the
pavilion roof. Localized cracks are also visible near the corners, per-
haps caused by horizontal displacement transferred to the walls by the
angular rafters of the pitch. One can also observe the collapse of the
obstruction of an arched opening.
It was dramatic how many cases of damage were unmistakably caused
by the substitution of traditional wooden rafters with structures which
although appeared to imitate the original had vast differences in weight
and rigidity.
Both in the cases of heavy stringcourses as in those where the weight
is due to infill roof panels, the effect produced by the seism was fatal.
Fig. 8- Poggio Picenze: local collapse of tympanum walls due to holes punched in the rafters. The external walls below were seriously damaged and the roof cove-

RESEARCH - LAquila Earthquake 93


ring structure remained on top of what remained of the building
(Figure 11-12).

2.2.3 Horizontal diaphragms and vaults


In the case of floors, it was generally observed that modifications played
a negative role in seismic behavior; recent substitution interventions of
original horizontal structures were quite diffused, in particular in the
largest centers, and they seem to have led to the type of damage discus-
sed below.
The use of floors with metallic beams and brick (small brick arches or
thin hollow flat blocks) was common for nearly all of the last century
and may improve the overall behavior in some cases when well-made.
However, such interventions require localized clamping to the masonry
(of less impact than that necessary for the insertion of r.c. stringcourses)
and in a certain increase in the vertical load on the masonry. These
Fig. 10- Vallecupa: localized collapse of a portion of the masonry cyma.
loads inevitably influence one of the masonry parameters favoring
damage to the transverse connections in the wall and phenomena of
localized instability.
In Figure 13, observe how the portion of the wall collapsed from the
ground floor up to the height of the lintel of the window on the top floor.
The collapsed portion was limited to the vertical alignment of two
columns of puncturing and the nesting seems to have been favored by
the presence of an r.c. floor positioned over the original thin brick vault
and by the presence of the addition of a staircase leading to the attic.

Fig. 11- Tempera: the heavy substituted roof caused the crumbling of the walls of the floor beneath it.

Fig. 13- LAquila: a second floor and a staircase in r.c. were placed on top of the original one.

An analogous effect seems to have resulted from the presence of an r.c.


floor posed in order to consolidate and increase the rigidity of the origi-
nal floor in metallic beams with thin layered brick vaults (Figure 14).
In this case, the wall collapsed from the base to the summit with the
exclusion of the r.c. stringcourse and the roof structure which remained
in position.
Fig. 12- Villa SantAngelo: the r.c. roof with infill panels induced the collapse of the masonry walls beneath it. The head of the building had a rounded corner and the collapse of the

94 RESEARCH - LAquila Earthquake


mon horizontal structure for most of the buildings. It is interesting to
observe that in the range of current seismic damage, the vaults reacted
well. In fact, even in the cases involved in extensive collapse, the vaults
of the ground floors appeared intact (Figure 16). This is certainly due to
the correct dimensions of the masonry piers as well as the good place-
ment of the vault within the complete configuration of the building and
the whole aggregate.

Fig. 14- Paganica: a thick cement layer for rigidity was posed on the horizontal diaphragm.

walls may have been determined by the rigidity of the floor panels
which formed a diagonal point which tended to expel at the corner.
The case documented in Figure 15, merits some attention due to the
fact of its common reccurrence within the complete range of damage.
The deformation localized in correspondence to the size of the horizon-
tal diaphragms, absent at the top of the external wall, may indicate
excessive floor weight and/or the lack of any connection between the Fig. 16- LAquila: the vaults remained intact notwithstanding the collapse of the facade and the upper horizontal diaphragm.

horizontal diaphragms such as a barrel vault or flooring built on beams


parallel to the faade. Both in the role of horizontal diaphragms as well as in false ceilings for
inhabited attic spaces, the brick vaults were not always spared from
damage (Figure 17). This fact is easily proven both in smaller con-
structions as well as in important and monumental buildings (the dama-
ge is visible to these types of structures in the upper floors of many buil-
dings in the historic center of LAquila). To the contrary of stone vaults,
the reduced thickness of the brick vaults rendered them more sensitive
to even the smallest movements of the imposing walls. Obviously, for an
accurate evaluation of vulnerability of this type, the many cases where

Fig. 15- Poggio Picenze: evident deformation of the masonry walls of the facade with the worst damage at the level of the
horizontal diaphragm.

As previously mentioned, the use of vaulted structures for horizontal


diaphragms is rather common in luxury buildings as well in minor ones
in the Aquilan centers.
In the latter, the vaults are generally in stone with thickness varying
between 20 and 30 cm, while in the city of LAquila, there were predo-
Fig. 17- Paganica: collapse of two thinly laid brick vaults (horizontal diaphragm on the first floor and a false ceiling on the
minantly brick head vaults (12 cm). Barrel vaults create the most com- second floor).

RESEARCH - LAquila Earthquake 95


these vaults collapsed due to the fall of the roof covering structures or
single elements placed above them should be excluded.

2.2.4 Out of plane and in plane response in masonry walls


The topic of masonry quality in the Aquilan territory is the center of
much debate, and not only technical, due to its relevance for interpre-
ting damage but also and most importantly in the choices that must be
made for reconstruction. It is the authors opinion that judgment on the
quality of masonry cannot be expressed in a univocal manner as of yet.
However, one can certainly attest that most of the building patrimony
was built based on earthquake knowledge, in particular for the syste-
matic use of clamping and tie rods between walls (besides the previou-
sly mentioned use in the connections of non-thrusting roof coverings). Fig. 19- LAquila: an elegant 17th century tie rod connects to a wooden beam placed within the masonry thickness (note the
The organization of construction systems and the realization of masonry adjacent strengthening tie rod, inserted after construction on the internal side of the wall).

techniques in many cases calls for the presence of tie rods put in place
while the wall is raised. The construction of such tie rods is easily visi-
ble today by the mass of partial collapse dating back to the reconstruc-
tion after the earthquake in 1703, when the technique now used for over
two centuries was first put into use. The clamping consists in the pla-
cement of a wooden element built inside the wall and connected at the
extremities by an iron plate and nails (Figure 18) and then anchored
externally to the corner by way of a small tie rod (Figure 19). It is very
efficient as long as the wooden element is not placed under too much
tension causing weakness nears the nails.
Obviously, the necessity of balancing costs and the availability of mate-
rials called for the elaboration of variations which were not always effi-
cient (Figure 20).
More often than not, the progressive deterioration of these elements was
in correspondence to connections between the wood and the iron plate, Fig. 20- Villa SantAngelo: a poor example of wooden beams placed in the middle of the wall thickness without external
clamping.
especially in abandoned or poorly maintained buildings.
Generally speaking, the limited presence of out of plane displacement
within the vast array of damage found can be attributed to the systema-
tic use of tie rods in building walls.
Nevertheless, it was possible to encounter certain types of out of plane
displacement after this earthquake. Figures 21 and 22 show cracks
which caused detachment from the faade. Both were caused by poor
clamping at the corners and involved of a portion of the orthogonal
walls. In both cases, the portion of the isolated faade was limited to the
height of the horizontal diaphragm.
Figure 23 illustrates a case of the start of overturning of a wall portion
relative to the two floors laid over the ground floor; here the out of plane
displacement seems to have been caused by a lack of clamping with the
masonry adjacent to the contiguous building and a lack of efficient con-
nections at the level of the second floor.
Also damage ascribed to localized displacement of portions of the
Fig. 18- San Demetrio: metallic tie rod with a nail connecting the rod to the wooden beam. masonry cyma were common, both in the form of fractures (Figure 24)

96 RESEARCH - LAquila Earthquake


Fig. 21- Castelnuovo: detachment of a wall caused by defective clamping to the lateral wall. Fig. 24- San Demetrio: localized fractures above the lintel of an opening.

Fig. 22- Paganica: overturning which involved a portion of the orthogonal wall. Fig. 25- Paganica: collapse of a portion of a wall over the lintel.

as well as the collapse of limited portions due to cracking near openings


(Figure 25).
In plane damage to walls is present in all of its characteristic variants.
In fact the form and placement of the shearing varied according to the
efficiency of the connections between the elements it was built with and
the quality of the masonry pattern.
Figures 26 and 27 illustrate two cases; in the first, the inclining fractu-
res cross the external wall, traveling indifferently to the horizontal
diaphragms; while in the second, the fracturing is concentrated in the
masonry piers between the openings of the first floor.
These situations are found wherever there was a systematic presence of
clamping or for more recent buildings the presence of stringcourses.
Regarding the validity of anti-seismic regulations, which had been
adopted in most of these buildings, one can only say that in cases where
Fig. 23- Tempera: overturning of the top portion of an external wall. the workmen had the technical capacity to follow the rules of good con-

RESEARCH - LAquila Earthquake 97


the necessary clamping to the pre-existing masonry; a defect which in
fact greatly limited its seismic resistance.
As previously mentioned, the traditional connecting systems between
converging masonry walls are the most effective: both those put into
place at the time of construction (clamping at the corners, anchored
wooden beams) as well as later strengthening interventions or repara-
tions (tie rods placed in walls). The limited presence of collapse due to
overturning of the faade (compared to damage caused by major in
plane wall displacement) leads one to believe that these systematic con-
necting measures played an essential role in conservation (Figure 29).
In this sense, the diffuse presence of shear fractures in masonry walls
should not be interpreted as a demonstration of poor masonry quality,
Fig. 26- Villa SantAngelo: shearing which crosses the entire external wall.

Fig. 27- SantEusanio Forconese: shear cracks limited to the masonry piers. Fig. 29- LAquila: a building where an overturning mechanism initiated but was effectively contrasted by the connections at
the summit thereby causing in plane resistance behavior of the walls.

but as a positive consequence of the anti-seismic measures which pro-


duced a box-like behavior, impeding out of plane mechanisms. In a
well-built stone wall, the formation of even deep shearing following
such a violent earthquake is inevitable and allows the activation of
significant dissipative capacities, limiting the risk of collapse.
Over all, much of the damage observed can certainly be attributed to
insufficient quality of the masonry pattern, whether in ruinous collapse
or the loss of only the external faade. In general, the root of the pro-
blem is a lack of building quality: bad equipment for the faade, poor
quality cement (and its excessive proportion to stone elements), absen-
ce of transverse connections between walls. Such cases are commonly
caused minor damage to adjacent buildings.
Fig. 28- Castelnuovo: masonry buttress built without clamping to the pre-existing wall. Figure 30 illustrates the case of a building where the entire upper floor
collapsed. One can observe how the seismic activity disarranged the
struction and utilized these norms, such measures functioned to limit or walls to such a point that it crumbled to the ground in rubble notwith-
modify damage mechanisms within the limits of their efficacy. Instead, standing the light roof covering (and therefore not able to induce the
Figure 28 shows the frequent case of a buttress added to a wall without shearing particularly prevalent in upper floors). On the contrary, it

98 RESEARCH - LAquila Earthquake


2.2.5 Localized damage
Always on the subject of masonry walls, localized damage was found to
structural and nonstructural elements, justifiable in a situation where
the force of the seism is extremely violent. In fact, even in cases where

Fig. 30- Castelnuovo: total collapse of the top floor due to defective masonry quality.

should be noted that the external wall of the building to the left was left
whole.
In Figure 31, the very common case of the collapse of only the external Fig. 32- Casentino: detachment of hewn stone elements around an opening.
wall is shown. In these cases, the modality of damage declared its
cause: the lack of transverse compactness of the masonry panels. Small
or badly placed stone elements, lack of regularity in the pattern and
clamping between the facing sections rendered the walls vulnerable to
horizontal displacement and set off autonomous behavior in both walls
that then very often led to collapse of this kind.
The collapse of a single wall occurred more frequently where r.c. string-

Fig. 33- LAquila: localized collapse around the borders of the openings.

Fig. 31- Villa SantAngelo: collapse caused by the lack of transverse clamping in the masonry mesh.

courses were found at the top of a building, in particular when the enti-
re roof covering was heavy and rigid. The detachment of the wall at the
level of the stringcourse is linked to the local increase of compression
which originates in out of plane flexion from the condition and blocking
vertical displacement (caused by the flexional rigidity of the roof cove-
ring). Fig. 34- Vallecupa: a chimney exposed by the seism.

RESEARCH - LAquila Earthquake 99


building stock in Italy, indicate in LAquila city a 24% of reinforced
concrete structures, 68% of masonry structures and an 8% of structu-
res whose typology is not identified (Figure 36a). Data that identify the
age of construction of the buildings (Figure 36b) indicate that 55% of
the whole patrimony was realized after 1945. Considering low inciden-
ce of RC structures on the total it is possible to infer that after 1945 new
masonry structures were still realized and that RC structures number
increased gradually over years.
From distribution of storeys number (Figure 36c) it can be observed that
only 5% of the buildings have more than three storeys. Assuming that
all buildings with more than three storeys are RC structures, it is still
possible to infer that the great majority of LAquila RC buildings have
Fig. 35- San Pio alle Camere: localized collapse of the top portion due to scarce thickness. no more than three stories.
the masonry buildings behaved well, it was natural that there was some Assuming an interstorey height that can vary between 3.0 and 3.5
small specific vulnerability. meters, period approximate formulation given by the Italian Code,
A frequent case observed was the constructive vulnerability of hewn referred to RC frame structures, gives for three storeys buildings a fun-
stone above doorways and windows. The hewn stone elements used to damental period equal to 0.4 seconds.
border openings were rarely clamped to the masonry walls within the In the following section, when comparing different Italian Code spectral
wall thickness which inevitably led to their detachment (Figures 32 and shapes, it will be possible to focus on period value ranges minor or
33). This glaring omission to the attention given to seismic vulnerabi- equal to 0.4 seconds, thus leading to a comparison between constant
lity in local anti-seismic construction deserves to be examined more acceleration parts of the spectra.
closely. The problem was probably overlooked considering that such
elements do not determine the stability of the building as a whole. 3.2 Structural and non structural damages
Further damage was found in consequence to local irregularities of
various types, such as continuity solutions which weaken masonry walls In this section main structural and nonstructural damages to RC struc-
locally, or the presence of wide parameter walls with no structural func- tures after LAquila earthquake are presented. Figuregraphic documen-
tion for the building. tation was produced in the following few days after the 6th April 2009
In the case shown in Figure 34, one can observe the collapse of the clo- mainshock. Generally speaking, damages to structural elements are not
sure of a chimney along the external wall. Figure 35 shows the collapse so frequent and they seldom involve the whole structural system, on the
of a portion of the top of a wall characterized by scarce thickness, since other hand damages to nonstructural elements such as internal and
it did not bear the load of the roof and was conceived as a simple light external infill panels involved the main part of RC structures in
closure. LAquila.

3. REINFORCED CONCRETE BUILDINGS 3.2.1 Columns and walls


Main structural damages involving RC columns can be easily interpre-
3.1 LAquila reinforced concrete building stock ted as failures caused by mechanisms which capacity design principles
can avoid or at least limit.
ISTAT 2001 data, representing the official source for information about It is worth to observe that during an earthquake columns put up with
a b c

Fig. 36- 2001 census ISTAT data for LAquila town: (a) building typology, (b) age
of construction, (c) storey number

100 RESEARCH - LAquila Earthquake


high flexural and shear demand. Maximum flexural demand joined with Figure 37(b) reports a circular column belonging to a building in the
axial force produced by gravity loads and seismic loads are located at residential zone of Pettino, realized during 80s, which shows a typical
the end of the element; in these zones (critical regions) rotational duc- damage due to axial force and bending moment; concrete cover was
tility demand can concentrate. Therefore, it is necessary to provide an spalled due to high compression strain and longitudinal bar buckling.
adequate rotational capacity and to avoid buckling phenomena on com- In this case too hoop spacing is not thick enough but probably designed
pressed longitudinal reinforcements. in perfect accordance with code prescriptions at the age of construction.
Actual Italian seismic code provides prescriptions aimed at increasing In analogy, shear demand can produce brittle failures with an outstan-
section rotational capacity: upper limit on longitudinal reinforcement ding dissipative capacity reduction of the column. Minimizing shear
percentage, fixed the flexural resistance of the section, leads to a higher resistance mechanism to transversal reinforcement spacing only can be
ultimate section curvature; a proper spacing between hoops and cross- not exhaustive. In modern design rules shear design has to involve
tie presence give, by a more efficient confinement action on concrete, capacity design principles such as fixing a proper hierarchy between
an additional increase in section deformation capacity. Additionally, a brittle and ductile failure mechanisms, that is shear and bending beha-
proper hoop spacing avoids buckling phenomena in longitudinal vior of the element.
reinforcement or at least fixes an acceptable upper bound limit for In order to prevent brittle failures during post-elastic behavior, shear
which this phenomena occur. demand has to be evaluated based on maximum flexural demand of the
On the other hand, prescriptions and structural detailing presented element. When this criterion is applied to column, it consists in a rota-
above are typical of modern design concepts that in Italy appeared for tional equilibrium, obtaining shear demand by the ratio between the
the first time in 1997 with explicative document to the 1996 code but sum of the bending moments at the end sections and the total height of
were adequately ruled only in 2003 with OPCM 3274, and finally offi- the column. It is possible to prevent a brittle failure occurrence
cially adopted in 2008 by DM 14/01/2008 and its subsequent explica- applying an amplifying coefficient to the shear demand. When shear
tive document. demand is evaluated, possible local interaction with adjacent infills is
So, according to codes that were in force before 1997, it is possible to considered; in fact when masonry panels do not completely fill the RC
find RC columns with longitudinal reinforcement percentage that bay, evaluation of shear demand is based on the height of the column
exceeds 4% limit, section dimensions not conforming to actual limita- subtracting infill panel height.
tions, hoops with a not sufficiently thick space (15-20 cm) and closed Adopting proper capacity models it is possible to take into account con-
with 90 hooks. crete degrading resistance mechanisms due to ductility cyclic demand.
Figure 37(a) presents a corner column of a RC building in LAquila These prescriptions are provided by the Italian code since 2003, befo-
historical centre, probably realized between 1950 and 1960, where re OPCM came into force shear design of columns used to be made
damages occurred at the bottom end section of the element. Presence of assuming shear demand from linear analyses; this procedure can easily
smooth bars and small diameter of the hoops (6 mm) closed with 90 lead to shear capacity underdesigned respect to flexural capacity; so no
hooks can be observed, but most significantly the absence of any tran- control of the failure mechanism can be applied and a priori the failu-
sversal reinforcement in the first 30.40 cm of the elements immediately re mechanism can be either brittle or ductile.
adjacent to the beam-column joint region. All above considerations can be confirmed by column damages repor-
a b a b

Fig. 37- Column with smooth bars and poor transversal reinforcement (a); damage to a column due to axial - bending Fig. 38- Shear failures of rectangular (a) and circular (b) columns.
interaction (b).

RESEARCH - LAquila Earthquake 101


ted in Figure 38. squat column is different respect to the typical behavior of a slender
Considering the rectangular column in Figure 38(a), whose section is element. This is why, if local interaction between the column and the
probably (30x100) cm2, belonging to a building realized in 80s, a shear concrete wall is not taken into account, a premature brittle failure due
failure involving top end section is evident. Transversal reinforcement to concrete excessive compression can more often occur.
has a hoop spacing approximately equal to 15-20 cm, that is completely Even columns belonging to staircases can show brittle failures. Most
underdesigned respect to column section dimension and consequently common staircase typologies, generally, are characterized by disconti-
respect to the section inertia, leading to a premature shear failure of the nuity elements in the regular RC frame scheme composed by beams
element. The brittle failure mechanism is to be noted, underlined by the and columns. In fact, on a side staircase is composed by inclined axis
crushing of the concrete within the reinforcements. Third and fourth elements (beam or slab), on the other side squat columns are created by
hoops from the top end of the element that are completely opened. the intersection of inclined axis element with the column.
Figure 38(b) shows shear failure of a circular column characterized by Staircase elements considerably contribute to lateral stiffness of the
a 30 cm diameter; in this case too it is possible to detect the hoop spa- whole structural system due to axial stiffness of inclined axis elements
cing not thick enough, that leads to diagonal cracking typical of shear and to higher lateral stiffness of squat column. This contribution is sim-
mechanisms and to instability of longitudinal bars in the column. ply estimable via linear analyses. On the other hand, staircase elements
In order to highlight the non secondary role played by column infill are characterized by higher shear demand that can lead to brittle failu-
interaction in determining shear failures in the elements, Figure 39 pre- re mechanisms.
sents some damages to columns. In particular in Figure 39(a) it is pos- Figure 40 shows a staircase composed by inclined axis beams, in par-
sible to recognize the brittle failure in the column due to the local inte- ticular the squat column in the corner, due to intersection with the
raction with the concrete infill that partially covers the bay frame get- beam, is characterized by a typical shear failure, due to an unfavorable
ting to 1/3 of the total height of the column. Partial infilling effectively shear demand capacity ratio. Poor transversal reinforcement in terms of
interacting with the column reduces the effective height of the element, spacing and hoop diameter is to be noted.
producing a higher shear demand that exceeds column shear capacity. Shear failures characterized reinforced concrete wall performances too.
This kind of phenomenon involves all of the columns that interact with
the concrete partial infilling.
Figure 39(b) shows an example of buildings with a partly below ground
level. This basement level is characterized by walls, often realized with
concrete, aimed at a retaining function of the adjacent embankment;
concrete wall height is limited respect to column height to allow the rea-
lization of windows. This structural solution leads to a strict reduction
of column effective height with a consequent increase in shear demand.
Moreover shear span decreasing of the element can modify shear span
ratio up to a modification to a squat column behavior. This situation is
Fig. 40- Shear failure in squat columns of the staircase.
not of secondary importance, because shear resistant mechanism of a
a b

Fig. 39- Shear failures of column adjacent to partial infilling panels (a), shear failures of squat column adjacent to basement Fig. 41- Failure in reinforced concrete walls.
level concrete walls (b).

102 RESEARCH - LAquila Earthquake


As an example, Figure 41 reports damages detected in two reinforced terized by beam-column joints without transversal reinforcement. In
concrete walls, respectively characterized by two different shape fac- this kind of situation, when cracking occurs no truss mechanism can
tors; damages consist of a spread diagonal cracking. Low longitudinal develop, leading to a strict reduction in strength capacity of the joint.
and transversal reinforcement percentage can be observed, especially if Considerations presented above are confirmed by damages observed in
compared with minimum values prescribed by actual design code. RC structures after 6th April 2009 earthquake. Figure 42(a) shows an
external beam-column joint, characterized by an extensive cracking in
3.2.2 Beam - column joints concrete belonging to the joint panel. The absence of transversal
Beam-column joints can completely modify structural complex beha- reinforcement in the joint, probably because of the vertical component
vior and their failure should be necessarily avoided in a proper seismic of the seismic action, led to local buckling of the longitudinal bars that
design, considering that these element are characterized by brittle fai- consequently brings to concrete cover spalling. It is worth to observe
lure mechanisms. In these zones, geometrically very small, demand that the absence of a proper transversal reinforcement in the joint con-
from beams and columns is concentrated and concrete panel with lon- ducted to a loss of anchorage in beam longitudinal reinforcement.
gitudinal bar is subjected to high gradients of shear and flexural Figure 42(b) shows a typical diagonal cracking failure in concrete panel
demand. Beam-column joints influence structural behavior in terms of belonging to an external joint. Crack begins at the intersection between
both ductility when concrete cracking and bar sliding occur and resi- joint and upper column and ends at the intersection between joint and
stance when brittle failures occur. lower column producing the loss of monolithic connection. Hoops
Failure mechanisms of joints are principally governed by shear and absence, in this situation too, leads to a buckling in the external longi-
bond mechanisms. In fact force distribution which allows shear and tudinal bar and it involves lower column not provided of transversal
moment transfer produces diagonal cracking and consequently joint fai- reinforcement in the first 30-40 cm.
lure due to diagonal compression in the concrete can occur, producing Other remarkable aspect in RC damage observation after LAquila
a reduction in strength and stiffness in the connection. Cyclic degrading earthquake is a peculiar loss of connection at joint-lower column inter-
of bond mechanism, on one hand, produces a reduction in bending resi- face, probably due to a not prepared cast surface that can lead to a shear
stance and in the ductility of the elements meeting in the joint; on the friction failure. Generally, there are no code prescriptions, nor Italian or
other hand an increase in lateral deformation of the level is detected. international, providing a shear friction verification at joint-column
Therefore, capacity design rules essentially aimed at avoiding brittle interface, because some execution details, such us preparation of the
failure mechanisms point to shear failure prevention in joints by means cast surface and a proper diffusion of longitudinal reinforcement along
of design rules and proper reinforcement detailing. In fact if the joint the perimeter of the column section ensure that this failure condition
collapses, a strict limitation results in resistance and deformation capa- does not limit or influence design procedure. In fact, main resistant
city of the adjacent structural elements. mechanisms in post-cracking condition are referred to (i) concrete to
Generally speaking, joint design is defined by the condition of a diago- concrete interface shear, (ii) dowel action in column longitudinal
nal stress induced by the elements meeting in the joint not exceeding reinforcement and (iii) clamping action produced by a local yielding in
allowable concrete compressive stress. Furthermore, in order to keep longitudinal bar that contributes to transfer main part of the shear
structural continuity when concrete cracking occurs, a proper transver- strength.
sal reinforcement along the whole element should be provided. a b

Transversal reinforcement allows to transfer stresses even if concrete


cracking phase has been overtaken, by means of a strut and tie mecha-
nism that can develop if longitudinal, transversal reinforcement and
concrete struts contribute to a truss formation. By these design pre-
scription in the joint a ductile mechanism in beam elements can deve-
lops, avoiding a brittle failure in the joint.
Joint design rules appeared in the Italian design prescriptions only in
2003, in fact in 1997 explicative document to 1996 code a transversal
reinforcement in joints at least equal to the hoop spacing in the adja-
cent columns was simply necessary.
Fig. 42- Joint failure with evident longitudinal bar buckling (a), diagonal cracking failure in concrete joint panel (b).
It can be gathered that all structures realized before 1996 are charac-

RESEARCH - LAquila Earthquake 103


surface. Reinforcement probably composed by only four longitudinal
bars and the absence of hoops in the zone strictly reduce dowel action.
In fact, this limitation is substantially due to a low longitudinal reinfor-
cement percentage not properly spread on the section perimeter, joined
with the spalling of the concrete cover that is the only transversal
restraint to horizontal bar sliding. A strong reduction in the axial force
essentially due to a non ordinary strong vertical component of LAquila
event has further reduced shear friction capacity. Figure 43(b) shows,
as well as Figure 43(a), a clear separation in concrete at joint-column
interface. Absence of hoops in the joint and axial force reduction due to
vertical seismic action strictly reduce, respectively, dowel action
Fig. 43- Failure mechanisms at joint column interface surfaces.
mechanism and friction mechanism at joint-column interface.
Shear friction mechanism is evidently influenced by axial force amount
and roughness (friction coefficient) of the casting surface; a not proper 3.2.3 Infills
preparation and control of the casting surface can reduce shear friction In the previous section it was emphasized how damage limitation limit
resistance. state prescriptions and verifications are essentially aimed at avoiding or
Clamping action a friction mechanism too, integrating above contri- reducing infill damage and most remarkably that this kind of prescrip-
butions is proportional to longitudinal reinforcement amount. tion was firstly introduced in the Italian code only in 1996 and better
Dowel action, not negligible in post cracking phase, is strictly connec- detailed and completed in 2003 with OPCM 3274.
ted to longitudinal reinforcement percentage but mainly, due to their Therefore, it is reasonable to assume that the main part of RC buildings
peripheral position, to effectiveness of transversal reinforcement in the in LAquila was realized without any deformability control and verifica-
zone adjacent to the sliding surface. Hoops, in fact, work as a transla- tion. On the other hand it should be emphasized that, even if a design
tional restraint to longitudinal bars involved in the mechanism. procedure according to 1996 or better according to 2008 code had been
Figure 43(a) reports a failure mechanism due to the loss of continuity at employed, involving damage limitation verification, the strong PGA
the intersection between joint and lower column. A poor treatment of characterizing LAquila event would have equally produced a spread
the casting surface can be easily detected and it can lead to a lower con- diffusion of damages to nonstructural elements such as external infills.
crete to concrete friction coefficient in correspondence with this weaker As a general rule, infill failure mechanisms can be classified in: (i) hori-

Fig. 44- Infill panel failures: diagonal cracking (a), (b) and corner crushing (c)
mechanisms.

104 RESEARCH - LAquila Earthquake


zontal sliding in the central zone of the infill panel, (ii) diagonal a b

cracking due to tensile stress in the central zone of the infill panel, (iii)
corner crushing in the direct contact application zone.
Figure 44 reports two building facades in which infill panels are cha-
racterized by a diagonal cracking mechanism. In the first case, see
Figure 44(a), it is worth to note how cracking diffusion involves infilling
adjacent to window openings and damage is concentrated at the first
levels of the building; in the second case, Figure 44(b), diagonal
Fig. 45- External infill panel failures and discrete (a) or lined up (b) connections between layers.
cracking is more emphasized by plaster layer because the external layer
a b
of the infill is composed by solid clay bricks. Figure 44(c) shows a typi-
cal corner crushing mechanism. Out of plane failure of the infilling
external layer gives the possibility to detect corner crushing mechanism
of the internal layer; other evidence is the crack, that visibly involves
the plaster but probably is deeper, localized at the top of the column
adjacent to the infill panel as a consequence of column-infill local inte-
raction.
The great majority of external infill panels are composed by double
layer infill panel, internal layers are generally realized with clay bricks;
connections between the two layers are realized by the interposition of
brick elements discretely, Figure 45(a), or lined up, Figure 45(b). Not Fig. 46- External infill panel failures without connection between layers (a) and with ineffective connection (b).
reliable efficacy of this system should be stressed.
Furthermore, in most of the observed cases, internal infilling layers are is applied by the panel to the RC element.
restrained at the four corners of the RC frame while external ones are As a global phenomenon, infill-structure interaction increases global
constrained only by the upper and lower beam by means of a little pawl. stiffness of the complex system and consequently spectral acceleration
This executive solution leads to a reduction in the interaction mechanism demand, besides it can represent a source of irregularity in plan or in
between RC frame and external infill panel in both plane and out of elevation when a non uniform distribution is present.
plane seismic forces. In fact, the low efficacy of the restraint applied to
the external panel, coupled with ineffective connections or complete 3.2.4 Criteria for regularity in plan and in elevation
absence of connections between the two layers, produces a damage Regularity criteria in plan or in elevation were introduced in the Italian
restricted to the external infill panel, which shows an out of plane fai- code in 1997 by the explicative document to the 1996 code as a quali-
lure due to seismic action in both directions as it can be detected in tative definition. Only in 2003 by means of OPCM 3274 introduction a
Figure 46. a b

Windows or door openings represent a discontinuity in the infill panel,


modifying its performance and capacity by a reduction in terms of stiff-
ness contribution and by a modification in the failure mechanism.
Figure 47 shows damages detected in LAquila RC buildings characte-
rized by a different opening position in the panel or a different opening
percentage respect to the total area of the bay.
c
Both local and global interaction effect between infill and RC structure
are not negligible. As it was previously emphasized, local interaction
between infill panel and adjacent column can bring to (i) a reduction in
the effective length of the column, an increase in shear demand and a
consequent brittle failure of the column when the panel partially fills
the frame bay; (ii) to a concentration of shear demand at the end of the
column and to a consequent brittle failure when diagonal compression Fig. 47- Infill failure mechanisms differing form opening position and percentage.

RESEARCH - LAquila Earthquake 105


quantitative definition of regularity in plan and in elevation was furni- Some peculiar cases of structural failure after LAquila event, mainly
shed. caused by irregularities in plan or in elevation, are reported in Figure
A compact shape in plan that implies geometrical limitation in the 48. Figure 48 shows a view of the structures before the earthquake, so
appendixes, a uniform distribution of resistant systems in plan, slabs without any damages, on the left, and after the event where collapse is
that can be considered as rigid respect to vertical resistant elements can evident on the right.
ensure a regular structure in plan. The first structure, Figure 48(a) and 48(b), was placed in the centre of
Conversely, a verification in mass and stiffness distribution, verifying LAquila city (Porta Napoli street); it was characterized by elevation
that storey lateral strength is not characterized by unexpected reduc- irregularity due to the non continuity of the seismic resistant scheme
tion, leads to a regular structure in elevation. over height, additionally, second level was characterized by an evident
In spite of a quantitative definition of regularity criteria, not necessarily discontinuity in terms of infill distribution, in fact on the left wing of the
these criteria are enough to avoid other irregularity sources not consi- building there is a sort of porch. Observing building collapse, a dama-
dered in the design procedure of the RC structure. It is the case of infills ge concentration at the second level is to be noted, that consequently
that by means of interaction with the structure can strictly modify stiff- produced complete failure of the upper levels.
ness and strength distribution in plan and in elevation. The two other structures proposed in Figure 48 are both placed in the
It is necessary to avoid non uniform distribution of infills or conversely residential zone of Pettino (Dante Alighieri street), close to LAquila,
it is necessary to explicitly take into account infill contribution in the and present the same shape in plan similar to a T, so they both can be
analysis procedure. For example, Italian seismic code provides some considered as irregular in plan. In addition to plan peculiarities, a dif-
prescriptions aimed at considering infill irregularity contribution in ferent distribution of infills at the first level respect to the others, due to
plan by an increase of accidental eccentricity value or in elevation by the presence of garages entrances, can be observed. Both buildings pla-
an increase of the shear demand at the storey characterized by an irre- ced in this street showed a soft-storey mechanism at the first level that
gular distribution of the infills. can be explained by infills irregularities in elevation and presumably by
a local interaction between infills and adjacent columns, leading to a
brittle failure of some columns at the first level.

4. SCHOOL BUILDINGS

One of the main objectives of the Civil Protection in the immediate


post-earthquake of LAquila of April 6th, 2009 has been scholastic buil-
a b dings damages assessment as well as the fast repair of the ones with
non structural damages only. These activities have been developed by
a joined work of Function 1 of Emergency Management and Quarter of
Department of Civil Protection (DPC), Consortium ReLUIS, Seismic
Risk Competence Centre of DPC and Public-Works Office of Lazio
Sardegna and Abruzzo.
The structural safety assessment of LAquila scholastic buildings star-
ted on April 8th, 2009. The in-site inspections have been coordinated
c d
by ReLUIS under the supervision of Function 1 of DPC at Reiss
Romuli; the inspections have been related to both LAquila and its pro-
vinces scholastic buildings.
The activity involved 62 scholastic buildings of LAquila: 53 under the
administration of LAquila municipality (6300 students on the total of
about 7000) and 9 of province (4000 students on the total of about
5000). A total of 156 structures have been investigated. The results of
e f
structural safety assessment is summarized in Figure 49.
Fig. 48- Soft storey mechanisms examples in LAquila: Porta Napoli street (a), (b), Dante Alighieri street (Pettino) (c), (d), (e)
and (f) before and after collapse occurrence. The in-site inspections on scholastic buildings in the LAquila provin-

106 RESEARCH - LAquila Earthquake


and about 21% masonry structures (see Figure 51)
In terms of damages the results of in-situ inspections showed that about
31% of RC framed structures were assessed as A (i.e. no significant
damages), about 43% as B (i.e. no significant damages on structural
members) and about 26% as E (i.e. significant damages on both struc-
tural and non-structural members). The structures with a lower level of
damage (i.e. A and B) have been mainly built between 60s and 90s
years while the structures recorded as E were mainly built between 20s
and 70s years. The RC framed structures mainly showed damages on
non-structural members (i.e. partitions and ceiling); the elementary
school of Paganica is a typical example of such kind of structures (see
Figure 52-53). An example of RC framed structure with significant
damages on structural members is the school Celestino V, see Figure
Fig. 49- Structural safety assessment on LAquila scholastic buildings. 54-55. The RC shear wall-type structures were assessed only as A
(27%) or B (73%); no significant damages on structural members were
found at all. The masonry structures, mainly built before 60s years,
were assessed as: 30% A, 24% B, and 46% E. Figure 56-57 show some
significant damages on masonry members of the elementary school
S.Elia.

Fig. 50. Structural safety assessment on scholastic buildings of LAquila provinces.

ces were performed in 64 different municipalities on 224 buildings for


a total of 309 structures; the results of such activity is summarized in
Figure 50.
The scholastic buildings of LAquila are mainly reinforced concrete
(RC) or masonry structures; in particular, about 66% are RC structures
(56% RC framed structures and 10% RC shear wall-type structures), Fig. 52- Damages on partitions at elementary school of Paganica.

a b c

Fig. 51- Structural safety assessment results on scholastic buildings of LAquila: a) A; b) B; c) E.

RESEARCH - LAquila Earthquake 107


Fig. 53- Cracks and plaster spalling on partitions of elementary school of Paganica. Fig. 56- Roof collapse of elementary school S. Elia.

Fig. 54- Column crack at elementary school Celestino V. Fig. 57- Diagonal cracks on masonry panels of elementary school S. Elia.

In the immediate post-earthquake several teams with members from


ReLUIS, Seismic Risk Competence Centre of DPC and Public-Works
Office of Lazio Sardegna and Abruzzo were involved in a further stage
of in-situ inspections on schools in order to investigate the repair pos-
sibility before the new scholastic year official opening (foreseen in
September). In some cases several destructive and non-destructive tests
were performed in order to investigate on the materials mechanical pro-
perties. As a result of this second round of in-situ inspections, a cost
estimate to fully repair these schools was also performed. The repair
interventions were planned on schools assessed as A and B; the repair
strategy and the interventions design were provided by engineers of
municipality or province under the supervision of ReLUIS and Public-
Works Office. The Public-Works Office also managed the bids for works
Fig. 55- Damages on ceiling at elementary school Celestino V. execution (see Figure 58).

108 RESEARCH - LAquila Earthquake


Fig. 58- Table of works at elementary school of Torrione. Fig. 61. Intervention on partition to avoid the overturning.

Fig. 59- Strengthening of joints on elementary school of Paganica. Fig. 62- Investigation on masonry corner of elementary school Villa Grande of Tornimparte.

According to Ordinances 3789 and 3790 and commentaries the works


involved not only the repair of non-structural members but also local
strengthening interventions on structural members (i.e. strengthening of
front and corner joints of RC structures (Figure 59-60), insertion on
masonry members of chains and braces) and non structural members
(interventions on partitions in order to avoid their overturning, to con-
nect their internal and external faces, and application of zinc coated
steel grids on partitions), Figure 61. The execution of repair and
strengthening works has been developed together with the execution of
materials non-destructive tests (rebound and sonic tests, tests on con-
crete cores and steels specimens, (Figure 62) as well as tests by using
flat jacks on masonry structures). Finally, load tests were performed on
slabs and flights.
The analysis of several cases of study allowed to optimize a series of
Fig. 60- Strengthening of joints on elementary school of Torrione. repair or local strengthening intervention on existing RC or masonry

RESEARCH - LAquila Earthquake 109


structures; a detailed description of standard repair interventions to a b

restore damaged buildings due to seismic actions has been reported in


a document developed in collaboration between DPC and ReLUIS,
Guidelines for Repair and Local Strengthening of Structural and Non
Structural Members.
The synergy between DPC, ReLUIS, Public-Works Office as well as
municipality and provinces allowed scholastic buildings assessed as A
to be opened with urgency in order to regularly perform the exams at the
end of the scholastic year. Further, the whole stock of scholastic buil- Fig. 63- Evidence of the reinforced concrete frame structure through the damaged partitions: Building 9 (a); Building 10 (b).

dings assessed as B has been opened in the period between September


Salvatore complex suffered considerable structural damage. This was
21st and October 5th. limited to small areas and primarily was due to evident issues that will
be described in detail later in this work. There was slight and relatively
5. THE CASE OF SAN SALVATORE HOSPITAL OF COPPITO limited non-structural damage and significant non-structural damage in
only a few buildings. In these latter inner partitions significantly helped
The recent earthquake of 6th April 2009 significantly hit the city of the lateral resistance by dissipating the earthquakes energy and suffe-
LAquila and its surroundings both for the serious number of casualties ring critical damages (Figure 63).
and for the damage suffered by residential and important structures. The basement and semi-basement part of the complex, mostly made up
Among the latter, one of the most important is surely the San Salvatore of reinforced concrete walls, showed a stiff box response without signi-
Hospital of Coppito, the crucial point of the hospital system in the area ficant damage. Finally, no evident damage was observed in the founda-
of LAquila, which was completely evacuated during the emergency due tion structures.
the damage at various floors of the buildings. From the point of view of the human safety, the most widespread and
A thorough analysis of the hospital facilities has shown besides the relevant non-structural damage was that of exterior faade bricks, cove-
essentially non-structural damage also some aspects of the construction ring the entire surface of all the buildings. Such a coating, not linked to
of the hospital complex, which might be crucial for the seismic respon- the interior walls, in many cases was partially or totally detached. No
se to future shocks and therefore open more critical damage scenarios. significant damage was observed on the equipment and internal mecha-
Other smaller hospitals located in neighbouring urban areas showed nical devices inside.
minor damage, allowing partial absorption of emergencies, thus redu-
cing the enormous overload on the field hospital set up close to the San 5.1 Description of the structures of the San Salvatore Hospital
Salvatore Hospital. Recent valuations, in fact, have shown that more
than 1500 injured received assistance during the days after the emer- In terms of typology, the San Salvatore Hospital complex consists of a
gency, which confirms the tremendous impact on the hospital system. series of reinforced concrete frame structures, with interior and exterior
The hospital structure has been designed in 1966 and took about 30 masonry walls, built from the mid 70s on, and put into service in the
years to build entirely. Such an information is particularly important if second half of the 90s. Some of the buildings of the complex are not
related to developments in domestic seismic regulation. In fact, the first hospital property.
Italian seismic law, considered a forerunner of the most modern ones, The buildings differ in typology, materials and heterogeneous construc-
was n. 64 of 2/2/1974, issued after the 1974 Ancona earthquake. tion details depending on the different age of construction. A covered
Previously, references to seismic law came from Royal Decrees (e.g. walkway connects the various blocks on four floors, two above ground
Regio Decreto Legge n. 2105 of 22/11/1937), while during the Repu- and two underground.
blic era there was Act n. 1684 of 1962 that followed the Campania There are several building typologies: L-shaped of 2 or 3 storeys, tower
earthquake of 21/08/1962, and was later completed by Act. n.1224 of blocks of 3 or 4 floors, in-line buildings of 2 or 3 storeys and some
5/11/1964, and by the Act n. 6090, 1969 issued after the Belice earth- ground-level buildings. Most of them have one or two basement floors.
quake. However, such references were oriented towards defining heights, The approximate date of construction and the main function of the
thicknesses, executive methods and quality of materials rather than cal- various blocks, identified on the basis of the numbering given in Figure
culation methods and design criteria. 64, was provided by the Technical Department of the hospital and is
After the 6th April earthquake, only 3 buildings out of 15 in the San shown in Table 1.

110 RESEARCH - LAquila Earthquake


Table 1 - Buildings of the San Salvatore Hospital complex, Coppito, In all the buildings of the complex there are also several structural joints
LAquila of sizes and characteristics not appropriate to shock induced move-
Denomination Function Construction age ments. The resulting local pounding between adjacent bodies caused
Building 1 Thermal Power station 1977/78 localised damage, somewhere particularly evident. Figure 66, for exam-
and refectory ple, shows the damage to a structural joint ending on the top of a column,
Building 2 Analysis laboratories 1976/77 which have caused an abnormal concentration of pressure on the joint.
Building 3 Diagnostics and radiotherapy 1976/77 Figure 67 shows the cracking continuing in the ceiling from the dama-
Building 9 Emergency room 1978 ged joint in the wall, found in a connecting walkway (Building 2).
Building 10 Pharmacy and operatory rooms 1978/79 Some local damage have been caused because of improper construction
Building L1 Direction 1983/84
Building L2 Obstetrics and gynaecology 1983/84
Building 6 Wards 1987
Building L3 Oncology 1979/80
Building L4 Infectious Diseases 1979/80
Building L5 Neurology 1983/84
Delta 6 Wards 1987
Delta 7 Medical Delta 1985 a b c

Delta 8 Surgical Delta 1980

5.2 Usability surveys and structural response Fig. 65- Damage to the coating of Building 2 (a), Building 9 (b) and Building L3 (c).

The buildings of the hospital complex have been repeatedly checked


because of a succession of significant after-shocks related to the so-cal-
led seismic swarm, and were grouped into categories depending on the
assessed damage.
All the buildings have an exposed brick wall covering, which is not
enough or not at all connected with the infill panels and the reinforced
concrete frame structure. These are therefore critical for human safety
since they can be dangerous and might fall onto the walkways (Figure
65). Another type of frequent and potentially dangerous non-structural
damage is the complete detachment of the coating of many of the
ground floor tiled walls. Fig. 66- Improper structural joint ending on the top of a column.

Fig. 64- Site plan with different blocks (the groups of buildings identify the
classification of usability as described in section 4).
(Gruppo/Group; Edificio/Building)

RESEARCH - LAquila Earthquake 111


shortly, and therefore can be usable by means of simple short term
countermeasures, among which: removing falling cladding, plaster and
detached coatings, making safe damaged false ceilings, repairing light
damage to claddings and partitions, checking all the fixings hanging
from cladding and false ceilings, in order to evaluate the possible deta-
chment risk, creating localized barriers to protect walkways, and remo-
ving unsafe portions of the outer coating. Buildings L1, L2 and L5
belong to this category, which show widespread light non-structural
Fig. 67- Damaged structural joint on wall, extending in ceiling. damage, particularly at the ground level. These three buildings will be
the first to be reopened at the end of May 2009, less than two months
after the main shock.

Group 3: partially usable buildings


Each of the Buildings 2 and 3 is composed of two different blocks sepa-
rated by a structural joint, which show different levels of damage. In
particular Buildings 2A and 3A, which face Building Delta 7 (Figure
69), are not usable, whereas light non-structural damage is common to
the rest of Buildings 2B and 3B (Figure 69).
Unlike group 1, in this group of buildings the unusable part features
high structural risk, whereas the less damaged part requires modest
intervention before to be used again.

Fig. 68- Examples of improper transversal reinforcement and insufficient concrete cover in Building 2A and in the external
connecting walkways.

detailing, at least according to current seismic design criteria, such as


the not always appropriate confinement of the structural elements, the
insufficient concrete cover (Figure 68), the presence of columns made
squat by the infill masonry walls.
Fig. 69- Buildings 2A/3A and 2B/3B (Edificio/Building).

Group 1: usable buildings


Regarding the basement, 6 metres underground, where the thermal
power station is located, Building 1 shows moderate non-structural
damage, particularly in the offices at the lowest floor. The portion hou-
sing the power plant shows a moderate damage in some beams, mainly
as a result of the relative movement of the Gerber half joint, and it is
therefore usable.
The rest of the building, however, shows greater damage in the upper a b
Fig. 70- Damage of the stairs of Building 2 (a) and Building Delta 8 (b).
storeys, even if it does not constitute a danger to the floor below, and it
is therefore to be considered unusable. Buildings 2B and 3B can be made usable with short term countermea-
sures, like those described in the previous paragraph. At the last flight
Group 2: buildings that can be made usable with short term countermeasures of the inner stairs between portions 2A and 2B, mild localized structu-
Some buildings with slight non-structural damage can be reopened ral damage was detected, probably due to concentrated rotation of the

112 RESEARCH - LAquila Earthquake


Fig. 71- Structural damage (plastic hinges and shear failures) at the top of co-
lumns (ground floor of Building 2A) caused by insufficient transverse rein-
forcement.

Fig. 72- X cracks on the external coating of Building 2A.

Fig. 73-. Important damage to partitions in Building 9.

RESEARCH - LAquila Earthquake 113


Fig. 74- Structural damage caused by the presence of belt windows, which make
the columns squat in Building 10.

knee beams (Figure 70). In this case limited restoration intervention dings, or because they feature a high percentage of severely damaged
can be provided, such as repairing cracks with resin or anti-shrinkage partitions compared to the volume of the building.
mortar, and later if necessary a seismic retrofit can be designed using Building 10 has severe structural damage at the ground floor, because
fiber-reinforced materials. of the shear failure of all the columns on the side in front of the church.
Blocks 2A and 3A are unusable because of the considerable non-struc- In this part the columns have been made squat by the infill panels inter-
tural damage and the structural risk caused by the failure of some rupted by the belt windows along the north-east side (Figure 74). It is
columns of the covered walkway over the driveway to the Emergency clear that for most of these columns the bearing capacity is seriously
Room, facing the Building Delta 7. This is the first area of significant compromised. The building also has widespread moderate to severe
structural damage. As shown in Figure 71, at the top of the columns it non-structural damage, particularly at the ground level.
is possible to recognize clear plastic hinges, shear failures, spalling of To ensure this building will not collapse, short-term countermeasures
the concrete cover with the consequent instability of the longitudinal are arranged, by means of a series of cement block infill panels between
reinforcement. All these effects are due to the scarcity or absence of the columns of the external porch and by propping the beam supported
transversal reinforcement, i.e. of confinement. This has led to a near by the damaged columns.
state of collapse, requiring urgent propping intervention. In the same In Buildings Delta 7 and Delta 8 there is widespread structural dama-
area there is widespread damage to infill panels, with the formation of ge more important at the lower levels. The damage to partitions is non-
X-cracks (Figure 72) typical of shear failure. Although such failures structural but widespread and of moderate to severe intensity. The only
may affect only infill panels, potential shear failure to the structural ele- light structural damage is found in one of the stairs.
ments under the damaged coating cannot be excluded, and therefore The two buildings are considered unusable, due to the significant exten-
they must be checked. sion of damage, together with the structural irregularity and the proximity
of buildings with structural problems (Building Delta 8 is next to Building
Group 4: buildings requiring more extensive non-structural measures 10 and Building Delta 7 is adjacent to Buildings 2A and 3A).
A number of buildings can be made usable only after significant local
demolition and reconstruction of the most damaged partitions and remo- 5.3 Usability of walkways and connections in basements
val and restoration of all unsafe or detached parts. The restoration of con-
The basements of all the inspected buildings, where accessible, have no
ditions of temporary usability in this case requires more time, and the-
refore cannot be included among short term countermeasures.
Buildings 9, L3 and L4, in this group, have suffered moderate to severe
non-structural damage to many of the ground floor partitions (Figure 73)
and a widespread light non-structural damage at the upper storeys.
On the more damaged part, the usability restoration requires demolishing
and rebuilding some infill walls, and propping a stair in Building 9.

Group 5: unusable buildings


The last group of buildings is classified as unusable, due to the signifi-
cant structural damage or because they are close to dangerous buil- Fig. 75- Connecting walkways: hollow clay block collapse of the intrados of the basement at 3 metres below ground.

114 RESEARCH - LAquila Earthquake


significant damage. There are rare and limited examples of hollow clay with a column-beam structure, are the usual type in the industrial areas
block collapse in the ceiling of basements at 6 and 3 metres below in Pile, Bazzano, Monticchio and Ocre (AQ). Buildings have generally
ground (Figure 75). one storey; two storeys are seldom observed, moreover only on a more
At the level of 6 metres below the ground there are unused spaces, limited area than the first storey. Beam supports on the columns may be
accessible only by technical personnel, without significant damage, of the saddle or bracket type. Deck beams are usually transversally ali-
except for localized seepage and settlement/shrinkage cracks. gned (with respect to the longer side of the structure plan) and they are
At the level of 3 metres below the ground, in the underground passage I-shaped, with variable depth; longitudinal alignment and inversed T-
connecting the buildings, the damage is predominantly non-structural shapes are less commonly observed. In the latter case, the shape is used
and not very extensive. It mainly consists of the damage to existing to support the tiles. The roof is generally built with tiles -shaped; less
joints and the formation of natural joints after the event, with limited often U-shapes are observed. Skylights are sometimes present. -tiles
plaster and coatings detachments, and some hollow clay blocks col- are also used for the intermediate deck, when there is one. External par-
lapse. titions are either made with bricks, or with precast reinforced concrete
In such basement levels the damage is more significant towards Build- shells, with no stiffeners.
ing 1, where some localized modest damage is present. Structural nodes are those typical of the Italian constructions: beaker
At the first floor level (3 metres above ground) there is damage to exi- footing for the foundation to column nodes, simple support for the beam
sting joints, formation of cracks, primarily to the completion of incom- to column nodes (with neoprene bearings and steel pins). Tiles are
plete joints, and the subsequent detachment, sometimes only partial, of generally directly resting on the beams, with neither horizontal restraint
plaster, false ceilings and coatings. nor neoprene bearings. Pin connections are seldom present. Tiles may
At the ground level the most critical situation is found since the con- be connected each other with the upper reinforced concrete layer, or
nection walkways between the various buildings are not protected, simply linked via steel restrainers (partly poured within the tiles con-
sometimes there is just a covered porch. The risk of falling of loose parts crete, partly welded with the next tile restrainer). Partition panels are
of partially detached coating or new portions of coating falling off, even either supported by the eaves beam or by the column, via links of many
after minor shocks, makes the need to remove partially detached or types. They are also sometimes supported by the deck tiles. A typical
already collapsed coating urgent, to protect all the paths adjacent to the shell-eaves beam connection is via a steel plate partially put within the
buildings (when there is no covered porch) from falling objects and to concrete shell during pouring. A bolt connects the plate and a steel
close off the riskiest paths. At this level there is some structural dama- angle, which is restrained to the eaves beam edge. The shell to column
ge: a column behind Building 9 stressed by a structural joint, and connection is often built with a steel plate within the column and a
another column on the corner of Building 2 with a plastic hinge at the bayonet with bushing linked to the shell, via a long bolt. This techno-
top. logy is used also to connect the partitions to the deck tiles; the steel
angles and bolts connection is less frequent. It is worth to note that
6. INDUSTRIAL STRUCTURES structural shell buildings are much less common than beam-column
buildings. A few precast industrial buildings were under completion on
6.1 Building-like industrial structures April 6th, 2009, when the earthquake struck; so it was possible to verify
the seismic behaviour of these structures under variable degrees of
Industrial buildings were built for many years as an assemblage of pre- completion.
cast reinforced concrete elements. The April 2009 LAquila earthquake
has struck, for the first time in Italy, industrial structures on a large 6.2 Damage and seismic performance analysis
scale. In fact, the Irpinia 1980 earthquake hit an area with few indu-
strial sites; and similarly happened in the Umbria and Molise earth- The response of the structural elements of the industrial buildings to the
quakes, which moreover were felt within a limited area. On the other April the 6th 2009 earthquake was generally in accordance with their
hand, the Friuli 1976 earthquake damaged industrial structures, but design level: no column collapsed, even though, in many cases a plastic
they were designed with no regard with respect to the seismic action; if hinge was observed, due to the high intensity of the seismic action
any design rule was used, this however belonged to inadequate seismic (Figure 76). In some cases such plastic hinge was not observed at the
codes. LAquila and its surroundings are instead undergoing a rather column base, i.e. at the column-foundation joint, but even one meter
strong industrial development. Precast reinforced concrete buildings, above, where the longitudinal reinforcement decreases. Furthermore no

RESEARCH - LAquila Earthquake 115


a b

Fig. 76- Plastic hinge in columns of industrial buildings: (a) FIAT garage at Pile; (b) building used for bovine-breeding at Fossa. Fig. 78- Collapse of beams due to loss support of a building at Fossa used for bovine-breeding.

ve the collapse of the beam due to support loss at the side without joint
bar, caused by too large displacements, and the pounding between the
beam and the column top fork (Figure 78).
The phenomenon of the joint bar cover splitting can be also noted at the
intermediate level of some two-storey precast buildings, where, as
already written, the beam-column joint is on corbel. The same pheno-
menon has also characterised the collapse of some tiles. In this case,
indeed, even where the joint had been fastened by a steel bar, the little
thickness of the bar cover of the beam, also characterised by the lack of
stirrups, collapsed, causing the tile support loss (Figure 79a).
Obviously, such support loss easily happened where the tile-beam con-
nection was not fixed and/or there was no connection between tiles; par-
ticularly unlucky situations were characterised by buildings in phase of
assembly, where the floor slab, joining the tiles, was not made yet.
Fig. 77- Effects of pounding between the cover tiles and the beam.
a b

plastic hinge was observed in beams or tiles due to the increment of the
vertical action. However, the damage of the precast industrial buildings
should be well analysed; indeed, it was characterised by collapse of
parts of the buildings, which, if the mainshock had happened during the
working time instead of at 3 a.m., it would have caused victims.
The static scheme of such structures is characterised by large deforma-
bility; consequently, the most of the observed damages of structural ele- Fig. 79- (a) Tiles collapse due to cover splitting and support loss of a FIAT garage at Pile (b).

ments (made by reinforced concrete) depend on the relative displace- Collapse of perimeter panels due to the breaking of the angle stirrup or
ments between the elements. Indeed, many cases of pounding between to the bolt head going out from the profile happened in a building used
elements of the same structure were observed. Furthermore, pounding as material and machine deposit at Bazzano.
between adjacent buildings was frequent, in the case of both precast However, the most important and spread damages of precast industrial
and cast in situ structures, due to the insufficiency of separation joints. buildings caused by the April 6th earthquake are those concerning the
In Figure 77 the pounding between the tiles and the beam of an indu- elements on the perimeter; indeed, the large damage of such elements,
strial precast building placed at Bazzano is shown. even though the structural typologies are different, associates precast
Confirming the numerical studies performed in the last years, the con- buildings to the in situ cast ones. The top connection of the vertical
nections represented the weak parts in terms of seismic performance of panels to the side of the gutter beam, made by a profile drowned in the
both old and new precast buildings. Some buildings have shown dama- panel, bolt with nut and angle stirrup, in some cases gave way due to
ges at the beam-column connection: the only observed case of precast the angle stirrup breaking and/or to the bolt head going out from the
beams collapse was due to the damage of such connection and to the profile (Figure 79b). This last phenomenon also caused the collapse of
following support loss. Indeed, as shown by numerical analyses, the panels connected to columns by a profile drowned in the column and
splitting of the joint bar cover happened where the thickness was mini- bayonet with bushing joined to the panel by bolt; some of these last con-
mum. In other frames of the same structure, it is also possible to obser- nections, instead, collapsed due to the bayonet breaking at bushing

116 RESEARCH - LAquila Earthquake


Fig. 80. Local failure of the beam support.

position. Some other failures where due to the going out of the whole
profile from the panel where it was drowned. A better seismic response
was shown by panels joined to the structure by angle stirrups and bolts.
In the case of perimeter elements made by bricks, the seismic action
determined their out of plane deformation, in many cases up to the
expulsion of bricks and the consequent partial or total collapse of the Fig. 82- Pictures of the sili after the seism (Figure by G. Verderame, ISPRA, F.M. Mazzolani).
perimeter element. Finally, among the carrying out mistakes, it is
noteworthy, for precast structures, the local failure of the beam support. which have induced strong deformations of the shells of the sili. Such
In Figure 80, a near collapse condition is shown, caused by the large type of damages are a clear effect of the earthquake vertical component,
column cover due to the fire protection provisions; indeed, due to such whose importance they highlighted (Figure 82).
provisions, a volume of concrete without reinforcement works as beam The Vibac sili have a metallic structure. Generally for their conception
support. sili have a very low structural weight, normally significantly lower than
the weight of the contained material. Such a characteristic implies a
6.3 Non-building-like structures: the case study of the Sili Vibac at Baz- very slender structure. It is evident that such structures are sensitive to
zano both local and global buckling phenomena. In fact the most common
failure mode is the instability of the wall panels due to the effects of the
The sili of the Vibac multinational (a chemical company which produ- axial force in compression. Such actions are due to the friction between
ces plastic films), located at Bazzano, close to Onna (Figure 81) repre- the silage material and the walls. The horizontal radial pressure, acting
sent an exceptional case of damage to steel constructions. They also on the cylinder surface from the silage material, has a stabilizing effect
represent an emblematic case of damage induced by the earthquake of against the buckling of the silos walls, giving rise to a tension stress
April 6th. The sili are used for the storage of polypropylene pearls, and field of membrane type. The distribution and intensity of the internal
they were full when earthquake struck. Some sili collapsed, some other forces in every constituting part of the silos, the cylinder and the hop-
remained standing even though strongly deformed, both locally at some per, are strongly influenced by the material extraction behaviour, which
rings and diffusely (Figure 82). in turn depends on the shape of the silo.
A more close visual inspection indicated that the collapses occurred for The Vibac sili have an elongated shape typically used for the storage of
overturning due to the crushing of the base rings and the hopper. plastic material. Therefore the predominant extraction mode is of the so
Moreover, along the sili height, deformations induced by buckling phe- called mass type, having the characteristic that the first material
nomena of the wall panels are apparent. In some cases an effect of coming out is the one inserted as first in the silos, all the material mass
pounding on the adjacent precast reinforced concrete constructions is in movement at the leakage. In case of sili with a stocky shape, the
took place, the latter have achieved the partial failure of the infills, extraction behaviour of funnel type prevails, it has the characteristic
a b that a central tube forms in the material mass, which is sucked by the
hopper. Such a tube is fed by the silage material all along the height,
the part of material external to the tube rests during the leakage. In par-
ticular, in elongated sili, when completely full, along the height of the
Fig. 81. (a) Localization of the plant VIBAC in the Bazzano municipality (AQ). (b) Sili before the earthquake. cylinder, from the higher ring bands the radial pressure grows towards

RESEARCH - LAquila Earthquake 117


the base up to assume a constant value, then at the section variation, it a b

reduces from the ring where the hopper is installed, and high stress
levels arise. Obviously in case of empty or partially full silos the beha-
viour is different, the stabilizing effect of the radial pressure is lost in
the empty part with a consequent abrupt variation of the critical stress.
Pressure variations inside the silos depend also by the leakage of the
material through the hopper, which causes a backwash effect and thus
Fig. 83- Impact of the earthquake on the road network: (a) SS80 Gran Sasso dItalia road affected by rock falls, but featuring
depressions. In order to control and regulate such an effect, sili are pro- rock-proof tunnels. (b) Distribution of traffic management solutions (updated to 01/05/09) for the 61 road tracts affected
by the earthquake (red = road closed; dark green = passable with limitations; yellow = alternating one way; light green =
vided with pressure valves. Given that it is plausible that on one side lane and velocity restictions).

the effect of the seismic vertical component provoked a sharp and


national territory, with an average daily commitment of 80 men and 70
important increment of the actions in compression in the sili walls, cau- vehicles.
sing buckling, the contemporary seismic action in all the components Rockfalls (Figure 83a) and landslides triggered by the earthquake and
accentuated the effect of possible asymmetrical distributions of pressu- aggravated by the heavy rain that hit the area in the days following the
re, due either to structural eccentricities, or to the silos filling method, event, were identified as the most problematic situations affecting the
or to the anisotropy of the silage material, causing a reduction of the sta- network mobility. However, the rock falls and landslides occurred
bilizing effect of the radial pressures themselves. Furthermore, buck- mainly in mountainous areas around LAquila, while the main road
ling could also occur due to constructional imperfections at the joints network in the city was not affected by the aforementioned phenomena.
between the coating ring bands of the silos, where joints in any case In the urban area, mobility limitations were caused by debris following
represent a discontinuity in the flow of longitudinal stress in compres- damaged and/or unsafe residential and monumental buildings adjacent
sion, with high concentrated stress. The above mentioned considera- to the roads.
tions fully justify the collapse behaviour observed during the LAquila Immediate activities for the restoration of normal mobility conditions
earthquake. included: 1) removal of rocks and soil from the roads; 2) rock slope con-
solidations; 3) enhancement of soil slope stability. These activities were
7. EMERGENCY MANAGEMENT FOR LIFELINES AND conducted employing, where possible, internal resources or activating,
RAPID RESPONSE AFTER LAQUILA EARTHQUAKE alternatively, emergency contracting procedures with external organisa-
tions. Securing of unsafe buildings adjacent to roads was carried out by
7.1 Road Network firemen.
Temporary traffic management measures were extensively implemented
ANAS S.p.A. is the agency that manages in the Abruzzo Region, as well in order to minimize road closures; these measures included traffic flow
as in the rest of the national territory, the state road network. The resi- restrictions; alternating one-way; lane and velocity restrictions (Figure
dual functionality and safety investigation of the road network were the 83b).
first priorities identified by ANAS for the management of the first phase The only significant damages occurred to the road network components
of the emergency. Physical and human resources were deployed to were the structural failure of the viaduct Corfinio on the national
achieve the following goals: 1) rapid survey of the road network to ensu- roadway SS5 and the collapse of a bridge on the main road SP36
re, at the largest possible extent, the regional mobility; 2) activation of Forconese. No further significant damages were reported to the com-
emergency contracting procedures (somma urgenza agreements) to ponents of the road networks including the numerous tunnels present in
immediately begin, where possible, activities for the restoration of nor- the Region that performed well.
mal mobility conditions; 3) damage survey of the road-network compo- The urgent need for a standardized and structured survey form to report
nents; 4) short term planning for the repair of damaged components. damages and disruptions in the road networks was highlighted while
At the same time, physical and human resources were deployed in sup- performing safety investigation and damage survey operations. A rapid
port of the Civil Defence for a first partial debris removal and for the survey form and an ad-hoc procedure were therefore identified and for-
excavations works necessary for the installation of relief campsites. It is malised while the survey work was in progress.
worth mentioning that, further to the local resources, additional ones The timely information on the mobility conditions was a key component
were used to manage the emergency. These resources were available of the effective emergency management. The Civil Defence issued daily
from few ANAS Regional compartments differently located on the a report summarising road closures, mobility restrictions and repair

118 RESEARCH - LAquila Earthquake


works carried out in the road network. Using a Geographic Information morning of April 6, a significant and sudden change in the water flow
System, GIS, the technical compartment of the Direction of Command for a main pipeline in Paganica. The immediate closure of the relative
and Control, Di.Coma.C represented this information in a cartographic shutters for that pipe was operated directly from the GSA headquarters,
format. Road closures and other temporary traffic management measu- before the technician team reached the affected site. The cause of the
res were overlaid to aerial Figuregraphs, technical regional maps, etc. rupture was identified in the fault crossing the Paganica pipe. Because
providing maps that had a fundamental role in supporting many emer- of that, the steel joint of the pipeline (diameter = 600cm; pressure 25-
gency management operations. 30atm) slip-off, causing a violent escape of water (Figure 84a).
As for the public information, emergency bulletins were regularly A connection portion at the joint, however, was still grasped for a length
issued to update in real-time the end-users about the mobility situation of 6cm. In order to quickly respond to the emergency, the repair was
in the Abruzzo Region. Communications and timely news were, as well, limited to the welding of the pipes at the joint.
posted on the ANAS website. Exception made for the aforementioned joint slip-off, no significant
Once the firth phase of the emergency was managed, efforts and resour- damage was observed to the main distribution and storage system.
ces were concentrated, on one hand, to handle the modified traffic con- Following the repair of the damaged joint it was, therefore, possible to
ditions in LAquila city due to the closure of the main road that ran restart the provision of potable water for all municipalities administered
through the city and, on the other hand, to respond to the new mobility by the G.S.A. SpA since the evening of April 6. As a lot of ruptures were
requirements created by the relief camps, and by the construction of the expected in the minor water distribution system, in order to prevent
provisional accommodation: Temporary Housing Modules M. A. P, and flooding and deterioration in the buildings already damaged, the deci-
C.A.S.E project. sion was made, not to restore the water distribution in LAquila histori-
cal centre and in the most affected villages. For these areas, the resto-
7.2 Water distribution network ration of the water provision was gradually operated starting from the
less affected zones and/or the zones with a strong need for reactivation;
Gran Sasso Acqua G.S.A. SpA is the water provider for LAquila city priority was given to the strategic services, secondly to the commercial
and for 37 municipalities in the earthquake area. The organisation and industrial activities, including the hotels to be reopened for the G8
offers an integrated water service including potable water supply, sewe- meeting, and finally to the residential buildings classified safe, after the
rage and wastewater treatment. specific AeDES survey. The partial restoration of the water distribution
The G.S.A. has 3 major supply systems (Chiarino, Gran Sasso, Water was possible because of secondary networks and of a shutter system
Oria) in addition to some secondary ones. The water supplied is tran- that allowed the exclusion of areas where the water supply was not
sported by a network consisting of approximately 900km of large dia- urgently needed. A few days after the earthquake (19 April), due to a
meter pipes and is stored in a huge number of tanks (about 200) that further slip of the fault, the welded joint of Paganica pipe broke, requi-
require continuous functional and hygienic monitoring and maintenan- ring a further repair intervention.
ce. The water is distributed from the tanks to approximately 100000 The priorities identified in the second phase of the emergency manage-
customers through a 1100 km distribution network made of quite old ment were, on one hand, the provision of the water service to the relief
cast iron and steel pipes. The pressure inside the main pipeline network campsites and, on the other hand, the management of all the activities
is quite high, reaching 30-50 atm., as well as in the distribution for restoring the water provision in LAquila City. To carry out the works
networks where it can reach 6-8 atm. for the water network connection in the relief campsites, the technical
Thanks to a remote control service and guided valves connected, through staff of the company (fully operative since the third day after the earth-
cables or wireless connection, to the main reservoirs and supply quake) was supported by the Genio Civile staff. On the other hand,
systems, it is possible to check the water flow inside the pipeline the works for repairing damages and restoring the functionality of the
network and to manage partial or total opening/closing operations direc- water service in LAquila were operated, where possible, by the G.S.A.
tly from the Gran Sasso Acqua headquarters. In particular, electroma- SpA technicians, or activating emergency outsourcing procedures for
gnetic sensors, measuring input low pressure, and electromagnetic gau- the most demanding operations. Relationships with external organiza-
ges (or Clamp on), measuring output differential pressures, are instal- tions have been unfortunately, nowadays, interrupted because of the
led in the tanks. The remote control service allows furthermore the financial difficulties that the company is undertaking due to the lack of
assessment of the water level in the tanks. income.
The equipment connected to the remote control system revealed, on the Most commonly observed damages in the minor distribution system

RESEARCH - LAquila Earthquake 119


a b
and chlorinator system) and with those for management purposes
(buildings used as offices, rooms for technical equipment and laborato-
ries).
The facility in Ponte Rosarolo is located near the historical center of
LAquila (4220'18.10''N - 1323'39.09''E). Structures were built the
60s-70s. The reinforced concrete digestion tank suffered partial colla-
pse of a longitudinal wall (Figure 85a), several vertical cracks on a tran-
Fig. 84- Impact of the earthquake on the water distribution network: (a) Joint slip-off in a main water network pipeline in
Paganica. (b) Repair on a cast iron pipe in a Paganica at the moment when some of the evacuated people were returning
sversal wall and the separation of orthogonal walls at the edges (Figure
home.
85b). The partial collapse of the wall also involved the steel pipe adduc-
ting wastewater that was connected to it. In buildings used as offices,
were the slippage/breakage of the joints and the breaking of cast iron
local technological and laboratory equipment (RC framed structure)
pipes (Figure 84b). It is important to emphasize, however, that in large
were also found cracks of both internal partitions and external walls.
part of the red zones (damaged zones with prohibited access) the
However, there were no evidences of damage to structural elements: the
water network is still closed. Because of that, it has not yet been possi-
cracks detected on non-structural elements did not represent signifi-
ble to completely estimate the extent and the spread of the damage suf-
cant damages and did not prevent the use of building. The inspected
fered by the network1.
facilities were therefore useable at the time of inspection, except the
Finally, it is worth mentioning that the drinking water purity and qua-
digestion tank that was useless. Due to this damage the tank has lost
lity has been officially tested and certified daily since the early days
water and the plant were partially closed by reducing the disposal capa-
after the seismic event. Because the G.S.A. official testing laboratory
city of about 60%. The remaining functionality was still sufficient to
was severely damaged after the earthquake, this service was guaranteed
face the demand, which was significantly reduced due to the large num-
via mobile laboratories of a neighboring water organization, C.A.M..
ber of evacuated people (approximately 30,000), housed outside the
The third phase of the emergency management focused on the con-
city.
struction of the water distribution network and connections for the sites
identified for the construction of the provisional accommodation:
Temporary Housing Modules M. A. P, and C.A.S.E project. Both the
design and the new construction of the reservoirs and of the distribution
network for these areas were committed to external organizations and
contractors. The costs for both the design and the construction of the
new reservoirs and networks for the temporary accommodation were
covered by the Civil Defence. The G.S.A. SpA will continue to be in
charge of the management of the water provision for the temporary a b c
accommodation areas. Fig. 85- Ponte Rosarolo Plant. Digestion Tank: (a) partial collapse of a longitudinal wall and of the pipe connected to it. (b)
Detail of the detachment of the orthogonal walls at the edges. (c) Displacement of the pump in the control room.

7.3 Wastewater treatment plant The structures of facility in Pile (4221'3.25''N - 1322'13.41''E),
which is situated between the town and the industrial area of LAquila
The technical visits at the wastewater treatment plants serving LAquila being the second plant serving the city, were realized in two different
(AQ), in the resorts of Ponte Rosarolo, Pile and Arischia, and at that periods (80s and 2000) with RC walls and slabs. Structural damages
located in the City of Corfinio (AQ ) have shown that examined systems were not detected, only some damages to the partitions of local offices
have similar technical characteristics, as they have the same practical occurred. With regard to the older settling tanks, characterized by a cir-
functions. Each plant was equipped both with the structures necessary cular cross section, a deterioration of the curbing RC beam was detec-
for the treatment of wastewater (primary clarifier tank, aeration tank, ted due to significant corrosion of the steel reinforcements.
digestion tank, settling tank, thickener, sludge dewatering band press The inspected structures, therefore, were viable and fully functional
despite the damages (of non-seismic origin), due to degradation of mate-
rials descending from a insufficient maintenance of the settling tanks.
1The water consumption was reduced by 30% as a result of water shut off into the red zones.
However, in the control room, a tube connected to the pump (not ancho-
Mobile water tankers were used to serve the relief camps in the first days after the quake. red) was damaged due to a displacement of 15 cm, Figure 85c. Finally

120 RESEARCH - LAquila Earthquake


it should be noted that this plant has been out of energy for three days it has shown some structural damages;
after the earthquake, so it worked through its own backup generator. - The Corfinio plant was not damaged because distant from the epicen-
The plant located in Arischia (4224'49.02''N - 13 20'25.48''E) pre- ter (approximately 50 km): the maximum acceleration recorded by the
sents reinforced concrete structures with the exception of the circular accelerometers of Sulmona station (Sul) located near the plant, is in-
tanks for leaching, consisting of circular walls of artificial masonry deed equal to 0.34 m/s2, approximately one-twentieth of the maximum
blocks connected with a RC curb at the top of the tank, and a gravity recorded at AQV Station of LAquila.
retaining wall. The structures date back to the 70s with the exception
of RC curb which was more recently constructed. Cracks on the walls 7.4 Gas distribution network
of a distribution trap and damages to the retaining stone wall, which led
to the partial obstruction of the hydraulic groove drain at the base of the Enel Rete Gas S.p.A. is the gas provider for LAquila city and for other
tank, were observed. With regard to the circular tanks, one of the two 5 municipalities in the earthquake affected area, namely Lucoli,
rotating distributors was put out of service for damage to its support; the Tornimparte, Ocre, Rocca di Cambio, Rocca di Mezzo.
cracks found on some blocks of the structure were dated before the The gas is distributed via a 621 km pipeline network, 234 Km of that
earthquake. Therefore, the inspected facilities were functional, although with gas flowing at average pressure (2.5-3 bar) and the remaining 387
the restoration of the full functionality of the hydraulic facility required Km with gas flowing at low pressure (0.025-0.035 bar).
some minor rehabilitation and repair of the tank distributor. In any case, The medium pressure network is connected to the high pressure national
the age of the plant suggests a constant monitoring even after the reme- one (namely SNAM network) through 3 reduction cabins while, about 300
dial action. reduction groups allow for the transformation of the gas transport pressu-
The treatment facility in Corfinio (AQ) situated not far from the center re (2.5-3 bar) into the gas distribution pressure (0.025-0.035 bar).
of the same town (427'25.74''N - 1350'31.78''E) is a RC construction The gas network is mainly made of steel pipes, with an average internal
built in the 90s. The central part of the longitudinal walls of the aera- diameter of intenal =125cm (external diameter external =139.7cm) and
tion tank, separated from lateral walls, shows a rotation very probably the joints are mainly welded.
occurred in large part before the seismic event, as witnessed by the The first priority identified for the management of the gas network, in
comparison of the positions of monitoring slides before and after the the first phase of the emergency immediately after the earthquake, was
earthquake; such slides were applied two years before the event: the the timely securing of the network in order to avoid explosions, gas
displacements due to the earthquake did not compromise the hydraulic leaks and fires and to allow the emergency vehicles and the USAR
seal of the joint, nor the functionality of the structure. teams to act in the safest possible way.
A comprehensive analysis of the observed damages was carried out in To ensure this priority, the entire network managed by Enel Rete Gas
relation to the position of each facility with respect to the epicenter of S.p.A. in the affected area was shut off via the closure of the 3 reduc-
the earthquake of April 6th, 2009 (UTC 01.32 hours) and to the records tion cabins. Thanks to this decision, and to the rupture of a pipeline
provided by the National Network accelerometric (RAN) available. It near Onna (Figure 86a), it was possible to timely and significantly redu-
can be observed that: ce the gas pressure and to avoid the occurrence of secondary effects.
- Ponte Rosarolo facility is located near the epicenter and close to the The subsequent closure of the 300 reduction groups ensured the full
AQK accelerometric station, which recorded ground accelerations securing of the network in less than two hours after the earthquake. In
equal to 3.7 m/s2 equal to about 50% of the maximum value recorded the days following the event, the gas valves external to each residential
for the same seismic event (station AGV - 6.6 m/s2); after the earth- building were as well closed. The pipeline damaged in Onna was repla-
quake, the plant has shown damages to the tanks with rectangular walls ced with a new one that was too rigidly connected to a reinforced-con-
larger than those found in circular tanks of the Pile plant, despite the crete support. It is worth highlighting that, as a result of the earthquake,
geographical proximity. The structural behavior of the circular tanks the Enel Rete Gas headquarters in LAquila resulted unusable. Because
was essentially better than that of the rectangular ones, mainly becau- of that the chief executive and the staff had to manage the emergency
se of the lack of structural details ensuring effective connection bet- without the support of their data, software and maps. Luckily, the natio-
ween the orthogonal walls; nal society Enel Rete Gas has, at a national level, an integrated infor-
- Arischia plant lies about 5 km from the LAquila accelerometric sta- mation system, including a data base and a geographical information
tions AQV, AQG and AQA, which recorded maximum ground accele- system GIS. Making reference to the closest Enel Rete Gas headquar-
ration values; even if distant from the epicenter (approximately 10 km), ters in Teramo and Pescara, it was possible to reprint the maps and all

RESEARCH - LAquila Earthquake 121


the documentation necessary to operate. ted to the action of the earthquake. As for the priorities, those identified
The second phase of the emergency response was focused on the acti- by the Civil Defence were followed; namely, the service was provided
vation of the physical and human resources in support to the Civil first of all to the strategic buildings, secondly to the manufacturing and
Defence. The timely provision of gas to the strategic structures was the industrial plants, and finally to the residential buildings identified as
first priority identified and was operated via mobile reduction cabins safe after the AeDES ispection. As for the testing procedures, in accor-
and gas wagons. H24 shift were organized for the local technical and dance with the procedures used by Enel Rete Gas for routine checks,
administrative teams, as well as for the teams coming from other areas an ad hoc protocol was defined in collaboration with the Civil Defence
of the national territory including the Enel Rete Gas national head- and the Firefighter Department. It was decided to reconnect each sin-
quarters in Milan. In the first month after the earthquake, the daily gle user following the fulfillment of four conditions: 1) safe dwelling
commitment of physical and human resources resulted on average (classified as A following the AeDES survey); 2) leak-tightness
approximately equal to 70 men and 35 vehicles, including equipped checking; 3) operative test of the equipment; 4) smoke test. It is worth
trucks, gas wagons and gas-leak detectors. mentioning that the Civil Defence fully covered the cost of the whole
On the same time, activities for the reactivation of the gas provision were procedure to reconnect the individual users to the gas service and that
started. The reactivation of the shut gas network required to operate gra- a dedicated phone line (Line Amica Abruzzo) was specifically set up to
dually restoring, first of all, the gas flow into the medium pressure facilitate and support the end-users in this operation.
network, secondly the gas flow in the low pressure network, up to each As a final note it is worth remembering that no damages were detected
external valve pertinent to each residential building previously closed. to the gas storage facilities.
Reactivation of the service was managed according to the following four
steps: 1) seal verification; 2) nitrogen check; 3) repair of damaged pipes 7.5 Electric power distribution network and telecommunications
and/or valves; 4) reopening. In the seal verification phase, the detection
of broken pipes and/or the possible joint slip-off was made, acting in the It was reported that two substations serving the greater LAquila had
first instance, from node to node, and further segmenting the network damaged connections between a rigid bus and insulator, Figure 87a.
when necessary. That was due to shifting of the un-anchored transformers during the
The material and equipment needed for the repair was immediately earthquake. Also due to sloshing of the cooling oil within the transfor-
available from the integrated logistics system which Enel Rete Gas mer, cooling oil pressure increased, and actuated the safety shut off fea-
uses; actually, the material normally in storage in the Battipaglia inter- ture to avoid costly damage. One of the transformers moved about 14
harbour to perform ordinary repairs and maintenance works, was sim- cm. In the distribution system, 30 posts were damaged causing severed
ply diverted to LAquila. The adopted strategy ensured the remediation links that resulted in service disruption. More than 180 pedestal type
and testing of more than 90% of the gas network in three month time connection boxes were dislocated and severed cable connections at the
after the earthquake. The diagram in Figure 86b shows how, three termination lugs that resulted in localized power failure (Figure 87b).
months after the quake, it was possible to restart the gas distribution for The Electric Power Control Center at LAquila sustained severe dama-
all the end-users with a safe home, exception made for LAquila city. ge, both building and equipment, and it had to be moved to a temporary
It is worth mentioning that the reconnection of the individual user sup- building in the yard of the building premise. It took three days to com-
plies required, on one hand, the definition of the priorities to be fol- plete the move, while the essential part of the system was functional by
lowed and, on the other hand, the definition of the testing procedures to 9 AM the day after the earthquake (Figure 86). Transformers in substa-
be carried out to certify the safety of the gas systems that were subjec- tions were not anchored. We noted that steel angles were welded on the

a b Fig. 86- Impact of the earthquake on the gas distribution network: (a) Onna (AQ),
damaged pipeline. (b) End-user gas connections activeted on June 8, 2009
(Green = end-users that can be potentially reconnected; Bleu = end-user
reconnected with respect to total that can be potentially reconnected).

122 RESEARCH - LAquila Earthquake


a b

Fig. 87- (a) Damage to rigid connection of a transformer. (b) Typical damage to pedestal box.

a b

Fig. 89- One of the relief campite in LAquila set up by the Civil Defence.

Fig. 88- (a) Steel anchors installed after the earthquake to avoid sliding of transformers. (b) Unanchored batteriesracks in
substation.
our investigation. Since tenants were not allowed back to their houses
tracks that the transformers were supported to stop sliding, Figure 88a. or apartments, most landlines were not used. Hence the demand on this
This was done after the earthquake. However the steel angles seemed circuit became much lighter.
to be under sized. In the control house of substations, the batteries were
not anchored or tied to the racks, Figure 88b. There was no batteries 7.6 Temporary housing
damage reported at these substations. Some locations were without
power for three days, e.g. wastewater treatment plant. The Italian government organizations and NGOs (Non-Government
Telecommunication service performed reasonably well. It went off air Organization) were to be commended on a great effort providing the vic-
for a couple of hours right after the earthquake. Cellular phones seemed tims with relief services and care. The military and fire brigade set up
to be the main means of telecommunication in this small community. service camps to provided needed services to the victims. Some of the
Although there was no reported damage to the physical equipment and relief campsites provided the victims with Internet services in addition
equipment building, we saw a number of temporary cellular sites to daily necessities such as medication, food, and water. In general the
deployed within the earthquake impacted areas. The increase of cell victims were very satisfied with the relief service. Many residents were
sites might have reduced the circuit overload that commonly occurs afraid to get back to their houses even when their houses (marked as
after an earthquake. Both Fire Fighters and Police used their own radio class A or B) were not condemned, due to their fear of future earth-
system as the primary communication tool. Cellular phones were also quakes and the potential for damage to their homes. Temporary housing
used to compliment the radio system. With a good backup power gene- is scheduled to be completed by September 2009 (before winter arrives)
ration plant, their communication was not interrupted. The Fire depart- for the victims, Figure 89. These houses will be on a base isolation
ment had three repeater stations, which were not damaged. A number system to protect residents from future earthquakes. There were more
of landlines were damaged or severed, as repairs were evident during than 30,000 victims settling in more than 160 campsites.

REFERENCES Sismica, 2010. IUSS Press, Pavia.


C. Casarotti, A. Pavese, S. Peloso, Seismic Response of the San Salvatore
C.F. Carocci, S. Lagomarsino, Masonry Buildings in the historic centers of the Hospital of Coppito (LAquila) during the 6th April 2009 earthquake,
LAquila area, Progettazione Sismica, 2010. IUSS Press, Pavia. Progettazione Sismica, 2010. IUSS Press, Pavia.
E. Cosenza, G. Manfredi, G.M. Verderame, Reinforced concrete buildings, B. Faggiano, I. Iervolino, G. Magliulo, G. Manfredi, I. Vanzi, Post-event analy-
Progettazione Sismica, 2010. IUSS Press, Pavia. sis of industrial structures behavior during LAquila earthquake, Progettazione
M. Menegotto, Observations on precast concrete structures of industrial buil- Sismica, 2010. IUSS Press, Pavia.
dings and warehouses. M. Dolce, S. Giovinazzi, I. Iervolino, E. Nigro, A. Tang, Emergency Mana-
M. Di Ludovico, G. Di Pasquale, M. Dolce, G. Manfredi, C. Moroni, A. Prota, gement for lifelines and rapid response after LAquila earthquake, Progettazione
Behavior of scholastic buildings after LAquila earthquake, Progettazione Sismica, 2010. IUSS Press, Pavia.

RESEARCH - LAquila Earthquake 123


Reconstruction between temporary and
definitive: the CASE project

THE IDEA should actually be considered permanent, since they had a lifetime lon-
ger than 10 years (ignoring the fact that apparently between 10 and 50
hat is the time difference that distinguishes a temporary or pro- years works can neither be called provisional, nor permanent).
W visionally home from a permanent or final? It is not easy to
respond to this question, if you consider the seemingly enduring eter-
If then the provisional does not exist from a durational point of view, it
would be useful to wonder whether it makes sense that it would exist
nity of what in Italy is built with the objective to last for months, or for looking at energy consumption, sustainable environment or pollution. It
a maximum of few years. would also be useful to wonder whether buildings could be constructed
With reference to Italy, it is enough to consider what happened after the with environmental characteristics and safety level similar to that requi-
earthquakes of Belice and Irpinia (or even in Friuli), there is therefore red for permanent ones on a temporary basis and with cost per unit
no need to further elaborate the concept. similar to provisional ones. If this should be the case, it would be logi-
On the other hand, we could refer to the technical code of 2008 [1], in cal to propose to build provisional houses with characteristics of the
which the nominal lifetime of a structure is defined as the number of permanent ones.
years in which the structure normally maintained can be used for the These ideas and others were discussed in the days directly following the
purpose it was built for, it is indicated in a table and it needs to be spe- Aquila earthquake with Guido Bertolaso for the political, administrati-
cified in the design documents. ve and economical aspects, with Mauro Dolce, Edoardo Cosenza and
It is interesting to note that the code only indicates a maximum for pro- Gaetano Manfredi for the technical and scientific aspects.
visional works (10 years) and two minima for ordinary and important A first complete conceptual proposal, with 3D-rendering and prelimi-
works (50 and 100 years respectively). If one sticks to these data, it nary calculations was submitted on April 16th, together with several
should be concluded that all the provisional works that were construc- comments. It was hypothesized to deliver the buildings for 3,000 inha-
ted in the aftermath of the earthquakes that took place after WWII bitants within 5 months, guaranteeing seismic safety by means of an

Fig. 1- One of the first sketches of the project illustrating the logic of the buildings constructed on isolated plates.

Fig. 2- A plan sketch made by architects Ragazzi e Hoffer as to illustrate the logic of the infrastructure in a court
open to pedestrians.

2 Fig. 3- One of the many 3D renderings used to illustrate design hypotheses.

124 RESEARCH - LAquila Earthquake


Gian Michele Calvi1 and Vincenzo Spaziante2
1 Eucentre Foundation - Centro Europeo di Formazione e Ricerca in Ingegneria Sismica, Pavia.
www.eucentre.it
2 Department of Protezione Civile, Rome. www.protezionecivile.it

isolation system at the level of an urban block, and proposing elevate ring the production capacity of the market. The time programme was
standards of living, technology and environment protection. The pursuit defining in four weeks the date to open the construction sites, i.e. to
of these objectives, apparently impossible, was based on the construc- start construction by mid-May, to deliver houses to 3,000 inhabitants by
tion of large isolated plates and the subsequent assembling of pre-fabri- September.
cated three-storey living units. The need for the project to be as much The economic analyses indicated an estimated cost of 120 million euro,
as possible independent from local soil conditions and from the un- VAT excluded, for 3,000 inhabitants, with a 20% uncertainty rate and
known construction technology (many different ones would have been without considering furniture, purchase of the terrain and photovoltaic
necessary to meet the deadlines) became immediately clear. To this end installations.
it was stated the need of urgently identifying the possible building tech- In preliminary calculations it was assumed to use friction pendulum
nologies compatible with the timing programme and the technical con- devices [2-8], with a radius of curvature of 4m, a vibration period of 4s,
straints, of selecting technical and commercial partners and of explo- a displacement capacity of about 300 mm, a friction coefficient between

Fig. 4- Images used in the preliminary phase to illustrate possible technologies for the assemblage of the buildings.

3 and 5% and an equivalent viscous damping between 20 and 25%.


The alternative of using rubber isolators was also taken into considera-
tion, but appeared in this specific case to be less competitive, conside-
ring the relatively low axial forces and the large horizontal displacement
demands.
In the days immediately following, several aspects that would have per-
manently defined the project were discussed and clarified:
- The reduction of each one of the isolated plates to about 20 by 60 m,
suitable to sustain three-storey buildings with each floor surface of
about 600 m2, with a capacity of about 80 inhabitants in 25 to 30 apart-
ments. Plates of this dimension should allow an adequate flexibility in
relation to the plane altitude conditions of the areas to use (at that

RESEARCH - LAquila Earthquake 125


moment unknown) and the construction technologies, also unknown; re, of personnel roles, activities and their interaction, time programme
- The definition of 150 as the approximate number of plates to construct and milestones required several days of intensive work and was com-
and therefore of about 12,000 inhabitants to settle in; pleted and formalised by May 8. The way the project is managed is very
- The division of the intervention in numerous small villages, consisting innovative with respect to the schemes that are normally adopted, and
of 4 to 20 plates and hence a number of inhabitants between 300 and not only in Italy. In fact a singlepurpose consortium was created
1600; (named ForCASE), formed by Eucentre (a non-profit foundation, centre
- The definition of a serial timesheet in which a group of 30 plates of competence for seismic risk of the department of civil protection,
should be finished about 15 days after the previous group, which founded by four public institutions and with a nature of public com-
implied a forecast of delivery of the apartments in 5 tranches for 2,400 pany in Europe) and two construction companies, (ICOP and Damiani).
inhabitants a time, with deadlines spread out between 30 September The two companies agreed to operate in this context as non-profit enti-
and 30 November; ties and not to participate in any other reconstruction activity in
- The decision to manage the entire project directly, without interven- Abruzzo. Their role would have been that of a technical office, and the-
tion of a general contractor, setting up a non-profit technical structure refore to facilitate the consortium to act on behalf of the CPD as a gene-
that responds directly to the Civil Protection Department (DPC). It was ral contractor, with the capacity to manage directly supplies purchasing,
thought that this way it would be possible to save substantial economi- to coordinate activities on the construction site, to arrange and verify all
cal resources, mainly on general additional costs and to have a more accounting matters.
accurate control on deadlines and quality of the project. Obviously, the consortium had as well the main task of carrying out all
designing and construction management activities, under the responsi-
THE ORGANISATION bility and coordination of the authors of this article. Coherently with
what has been briefly described, the operational organogram demon-
The definition of an operational, management and outsourcing structu- strates five main areas of actions: two being related to design activities,

Fig. 5- The personnel and work organization plan set up in the preliminary phase.

126 RESEARCH - LAquila Earthquake


pletion of the project;
- Finalisation and publication of call for bids;
- Stipulation of contracts;
- External checking and control;
- Relations with institutions and obtainment of permissions;
- Identification of the intervention construction sites, expropriations of
lands and related activities.

INFRASTRUCTURES AND ARCHITECTURAL DESIGN

The architectural project of a building unit, as briefly presented, favou-


red the development of different types of apartments, as a function of
family compositions, which resulted in 109 different shapes after the
Fig. 6- A simplified version of the extremely detailed and complex time schedule that allowed daily overviews on each selection of 16 contracting bids, as discussed later.
aspect of the project and the construction.
Regarding the choices on infrastructure, it needs to be highlighted that
two to management and accountancy activities and one to project coor- a first guiding concept was that of placing the settlements in the neigh-
dination. bourhood of existing villages that had suffered severe damages becau-
In order to obtain maximum efficiency, in terms of time and costs, and se of the earthquake, to be able to relocate the people within their own
to ensure quality control, three different operational modalities for con- territory, to preserve the close ties that people have with land and neigh-
tracting and execution of the work were identified: bours.
- For the activities of preparation of the construction site and infra- This general principle was confronted with technical difficulties deri-
structure works, it was decided to mainly use local contraction compa- ving from non-ideal geomorphic, hydrological and geotechnical cir-
nies; cumstances of the areas, to finalise the best possible selection of the
- For the foundation and isolation systems, it was opted to act directly areas of intervention.
as a general contractor acquiring materials and supplies, such as con- Once the settlements had been defined and sized as a function of quan-
crete, welded wire meshes, steel columns, isolation devices, formwork titative needs and land capacity, considering the dimensional and
positioning, concrete casting, etc.; morphological characteristics of the location, the problem of existing
- For the construction of the housing structures it was decided to launch infrastructures (roads, pipelines, sewing system, etc.) and of their
a call for bids that included final design and global construction, improvement and integration had to be faced.
allowing the use of any building technologies compatible with the needs Finally, the population indices could be defined, starting from figures
and available time, and selecting the proposals with the highest quality between 100 and 150 inhabitants per hectare, for location in more rural
and the lowest cost. or more densely populated areas. Such figures imply a rather sparse set-
The economic quantification of the costs for the management of all acti- tlement typology, marked by large green areas.
vities was estimated based on the pure cost of the staff assigned to this A final infrastructure index had been identified by assigning 30% of the
temporary job (in months), on a monthly cost, in general between 3,000 land surface to services and facilities, such as leisure, sport, shopping
and 12,000 euro (these are costs for the consortium, not net salaries), centres or education and religious structures.
and on a sum to cover cost of accommodation and travel, that could in Based on these premises the final urban design of the areas was com-
any case never exceed 3,000 euro per person per month. pleted, obviously combining the building units previously described
As all the activities would be executed within the non-profit framework (essentially consisting of three inhabited floors above a covered park-
that characterizes the Eucentre Foundation and the ForCASE consor- ing), also considering exposure to sunlight, valley and mountain views,
tium, the estimations were considered as a maximum not to be excee- steepness of terrain.
ded, while the real costs would be subjected to accountability checks. Driveways and walking paths were kept separated to the maximum pos-
The Department of Civil Protection would directly execute, in coopera- sible extent, generally locating vehicles roads on the outer skirts of each
tion with the consortium, all the activities related to: area, with access limited to parking lots and ground floors of the buil-
- Definition of Civil protection ordinances, possibly needed for the com- dings, also used as parking. The walking paths were designed elimina-

RESEARCH - LAquila Earthquake 127


Fig. 7- An example of the plans that were
designed for the bids of housing con-
struction, with an underground parking bet-
ween the two plates.

128 RESEARCH - LAquila Earthquake


Fig. 8- Location of the construction sites, all within the municipality of LAquila.

Preturo Sassa
Fig. 9- Examples of the urban plans for some site.

Bazzano Coppito 2

Fig. 9- Examples of the urban plans for some sites.

RESEARCH - LAquila Earthquake 129


ting all architectural barriers, connecting green areas and inhabited in which H is the height of the building and Cs is 0.05 for wall structu-
levels to road systems and parking lots with external elevators when res, 0.075 for reinforced concrete frames, 0.085 for steel frames. It is
needed. however well known that equations of this sort tend to underestimate the
The final character of the new settlements tried to combine in an opti- real vibration period resulted from a secant stiffness to yield, that for the
mal way people needs, environmental and landscape requirements, use examined buildings could arrive at values between 0.8 and 0.9 s [9, 10].
of existing infrastructures and construction of new ones, in an integra- Based on these considerations, the design period of vibration of the iso-
ted vision. lation system was selected in the range of 4 s.
Later, another problem had to be faced, i.e. how to combine each one of It was also preliminarily observed that even an extreme temperature
the specific building units (at this stage 150, in 20 different locations) variation of 30 C, leads to variations in length of about 8,5 mm on
to each one of the 16 different typologies proposed by the companies each side of the axe of symmetry, that would not induce excessive hori-
who won the call for bids. Choices had to be made in relation to con- zontal loads into the columns.
struction technology and material, external aspect, number of buildings
awarded to each company, construction plan and schedule proposed by Seismic action
each company. Seismic action and in particular spectral demands in acceleration and
Finally, considering the high environmental value of the landscape, the displacement are discussed in detail elsewhere in this volume [11].
design and realisation of the green areas was again the subject of a Here it is however important to note that the fundamental parameter to
public, international call for bids, where again cost, time and quality of be assessed for a proper design of the isolation system is the maximum
the proposal were considered to select the winning bids. displacement demand at a period of about 4 s. The spectra derived from
the registrations of April 6th show generally displacement demand of
STRUCTURAL DESIGN less than 120 mm, with one exception, the AQK registration, in which
spectra values are close to 250 mm. The code spectra for events with
Preliminary considerations return periods of 1000 years, to be used for the design of the isolation
The structural design of the buildings constitutes the fundamental ele- system, have values of about 300 mm for soil type B and 400 mm for
ment that allowed the development of the entire project and is extre- soil type E. These values can be significantly reduced in presence of
mely simple in its basic logic: two reinforced concrete plates, separated energy dissipation, as a function of an appropriate equivalent damping,
by columns and isolators, the lower one being in contact with soil and according to the  factor:
the upper one with the building. The plates were designed without
knowing the local soil properties, nor the weight and plan distribution
and structure of the buildings. Therefore for both aspects conservative
assumptions were used, to be verified later. In a few cases, some poten-
=
 10
5+
where  is the equivalent viscous damping value, that could be in the
tially selected construction location had to be discarded because the order of 10-15% for rubber bearings and of 20% for friction sliders. The
soil properties appeared to be unsuitable. values obtained from the reduction coefficients are between 0.6 and
It should be noted that the two plates are characterized by similar flexu- 0.7, with consequent estimations of displacement demands of about 250
ral actions induced by gravity, if it is assumed a uniform distribution of mm for soil type E.
the building load and of the soil reaction. Preliminary evaluations, based For the non-linear analyses the code spectrum for vertical actions has
on a column span of 6 m in both directions (convenient for parking arran- also been considered, while for the building phases it was defined as a
gements), lead to a required thickness of both plates of 500 mm. construction event consistent with what indicated in addendum A of
The weight of each building, with three floors of about 600 m2 each, was Eurocode 8, part 2 [13]. Such an event appeared to be consistent with
estimated in a maximum of 21 MN, with a consequent total maximum registrations corresponding to a magnitude of 4.0, and was thus consi-
weight of slab, dead loads and building details of between 30 and 40 dered reasonable. While the demand in terms of acceleration was signi-
MN (or an average weight per column of about 1 MN). ficant (in the order of ag = 0.10 on stiff soil), the displacement demand
The first vibration period of the building can be estimated between was negligible.
Ts = 0.25 and Ts = 0.45 s, using the equation: Eight sets of spectrum compatible accelerograms have been used for
non-linear analyses, derived from registrations made in LAquila (3
Ts  CsH0,75 records), and during the events of Imperial Valley in 1979, Loma Prieta

130 RESEARCH - LAquila Earthquake


Fig. 10- Acceleration and displacement spectra of an event with a 1000 year return-period in LAquila, according to the Italian code [1], soil category B and E, damping 5%.

in 1989, Northridge in 1994, Kobe in 1995 and Taiwan in 1999 (one for soil properties.
each of these events). Obviously other combinations may be possible, also related to the
various displacement demands for isolators placed in different positions
Isolation system (because of the eccentricity of the loaded mass, even only accidental,
The design and the verification of the isolation system was carried out the demand at the perimeter is larger than that closer to the slab cen-
considering the possibility of adopting two different configurations, cha- tral area). It was therefore allowed to bidders to propose different solu-
racterised by different devices, one based on the use of 12 elastomeric tions, provided that they were respectful of design performances and
isolators, together with 28 multi-directional sliding pot-bearings and input. The result of the call for bids, in which FPS systems were pre-
the other on the use of 40 isolators sliding on spherical surfaces, uni- ferred, should not be considered as a general demonstration of superio-
versally known as friction pendulum [FPS, 2]. rity with respect to elastomeric devices, but rather as a consequence of
Both choices are compatible with the project requirements, in different the specific conditions of this project, characterised by relatively large
ways. Actually, the smaller dissipation capacity of the system with ela- horizontal displacement demands, low vertical forces on the devices
stomeric isolators (estimated to be equivalent to 12% damping) with and relatively low horizontal stiffness (as discussed, vibration periods of
respect to the one with FPS isolators (estimated damping 20%,) requi- the order of 4 seconds were assumed). This was the reason why elasto-
res a larger displacement capacity; in the order of 300 to 360 mm for meric isolators had to be coupled with pot bearings: the use of rubber
the elastomeric isolators, versus 260 mm for the FPS, depending on the bearings alone would have resulted in stiffness values incompatible
with the requirements of the project.
In the case of the FPS devices, the force corresponding to a displaced
position is defined by the following equation:

F = Mg + Mg d ( )
R
In which Mg is the axial action (M is the mass and g the acceleration of
gravity), R = 4 m the radius of the spherical surface,  = 3% is the fric-
tion coefficient and d the displacement of the isolator.
The least favourable conditions for the verification of displacement
capacity of the isolation system versus the corresponding demand are
likely to be those of a rigid and heavy superstructure, i.e. those of a
large participating mass and deformations concentrated in the isolation
system. With a configuration of this sort, the system global characteri-
Fig. 11- Comparison of several spectra recorded on April 6th on soil type B, code spectra for an event with a 1000 year
return-period according code [1] and results of a recent research project (DPC-INGV-S5 [2]). stics (40 pieces) resulted to be as follows.

RESEARCH - LAquila Earthquake 131


Fig. 12- Force displacement response of a system of 40 isolators and heavy superstructure. Fig. 13- Force displacement response of the system considering axial force variation due to vertical acceleration and global
interaction response [2, 4].

Effective stiffness, secant to the design displacement: Verification of slabs and columns
The foundation and isolation plates have been subjected to numerous
Keff = 14615 kN/m finite element analyses, that allowed to calculate the maximum bending
and shear demand levels for several load combinations, to design the
Corresponding period of vibration of the isolation system (note that in reinforcement, generally made by welded wire meshes to favour a fast
general heavier structures are also stiffer, therefore characterised by positioning, and to verify the resulting action combinations with appro-
lower vibration periods): priate strength domains.
Local verifications for loads concentration at the column ends were also
T = 2
 M =3.29s
Keff
performed on both slabs, considering the consequences of the substitu-
tion of a bearing as well. This operation was needed in hundreds of case
Corresponding equivalent damping: during construction, when the isolators were not yet available at the
time of casting the upper slabs.
FPS = 2Mg = 0.201=20.1% The columns have been designed and verified considering either the
Keffd
case of reinforced concrete and of steel, again to allow the use of various

Fig. 14- Examples of displacement histories for an elastrometic isolator (left) and for a FPS isolator (right), subjected to events with a 1000 year return period derived from 3 registrations in LAquila, compared with capacity circles of 360 mm (left) and
260 mm (right).

132 RESEARCH - LAquila Earthquake


Fig. 15- Examples of reinforcement distribution in a section of the isolated plate.

technologies and thus reducing the operation time. For the same reason tions on the characteristics of the buildings needed to be defined, in
steel columns were in general preferred, even if more expensive, using order to keep them within the parameters assumed for the analyses and
concrete only when no steel elements were ready to be mounted. therefore avoiding unexpected responses and jeopardizing the verifica-
tion of plates, foundations and isolation system. A summary of these
Prescription for building design prescriptions follows:
As already mentioned, the final design of the home buildings was left to 1. The load resulting from the building structures shall not induce in
the bidders, to allow the use of any building technology. However, the any element of the slab foundation system local actions larger than
specifications to which the projects would have anyway to conform nee- those resulting from a uniformly distributed live load equal to 50 kN/m2
ded to be defined as to assure an appropriate safety level to the global (i.e. excluding the slab self weight).
structural system. 2. The load distribution on the plates shall exclude concentrations
The seismic demand was defined in terms of design acceleration of the potentially resulting in local collapses.
building masses, calculated with reference to the maximum value of the 3. The maximum vertical action on a single bearing should be less than
ratio between base shear and weight of the building, obtained in the 2800 kN, either for the seismic load combination and for the gravity
worst loading conditions, corresponding to those of a stiff building (T = combination at the ultimate limit state, including the weight of the
0.19 sec) with the lowest mass (1500 t). plate.
For analyses performed with accelerograms compatible with the design 4. Bearings shall not be subjected to tensile forces in any load case.
spectrum at a collapse limit state (SLC, return period  1000 years), the 5. The main period of vibration of the building (considered fixed at the
base shear always resulted less than 0.11 times the weight of the buil- base) shall not exceed 0.5 seconds.
ding. It was therefore prescribed to assume a design acceleration equal 6. The eccentricity between centre of mass of the building and centre of
to 0.1 g to verify the buildings at a life safety limit state (SLV, return mass of the plate shall be less than 5% of the total length of the plate
period  500 years). (57 m) in the longitudinal direction, and less than 10% of the length of
For the same conditions, the average inter storey drift resulted on the the plate (21 m) in the transversal direction.
range of 0.1%, a value that certainly allows a full use of the buildings 7. The maximum seismic mass of the building alone (i.e. without consi-
even after a high intensity event. dering the weight and loads of the slab), calculated including self
Together with these extremely simple design data, a series of prescrip- weight, dead load and the fraction of live load to be considered for seis-

RESEARCH - LAquila Earthquake 133


Fig. 16- Example of bending stresses in the foundation plate, for gravity
loads (1st row, moments around the two axes of symmetry in kNm/m)
and for seismic loads (2nd row: maximum values, and 3rd row: minimum
values).

18

Fig. 17- Example of local reinforcement of the foundation plate at column bases.

Fig. 18- Example of bending moment axial action strength domain, for a section of plate (at
columns centres).

Fig. 19- Example of bending action on the isolated plate during bearing substitution at different
locations.

17
19

134 RESEARCH - LAquila Earthquake


Fig. 20- Reinforcement in a foundation plate with concrete columns and steel columns with isolators on a casted plate.

mic verification shall be less than 2100 t. for bids for excavations, supply of concrete (initially about 200,000 m3,
8. The buildings shall be designed in accordance with the technical with peaks in delivery of more than 5,000 m3 per day, self compacting
code of 14/01/2008. It is accepted to represent the horizontal load equi- and aerated), supply of welded wire meshes (initially about 260,000 kN,
valent to the seismic action by means of a static force vector, to be in general with diameter 14 mm at 100 mm), supply of steel columns
applied to the building floors, according to equations given in the code, (initially 180,000 ton, diameter 800 mm), supply of isolators (initially
assuming a design acceleration Sd (T1) of 0.1 g. 6,000 pieces, including assistance to positioning) and supply of casting
forms (initially for about 336,000 m2) and on-site assistance for reinfor-
CONSTRUCTION OF THE PLATES cement positioning and pouring of concrete. All quantities were later
significantly increased, since the number of buildings passed from 150
As previously discussed, for the production of the plates, the ForCASE to 184. The prices per unit obtained through bidding have been the fol-
consortium has directly taken the role of general contractor, with calls lowing:
Contractors for the production of the plates with initial price and offers

RESEARCH - LAquila Earthquake 135


Fig. 21- Rendering and floor plans of some buildings, proposed by the bidders.

136 RESEARCH - LAquila Earthquake


Contractors, structure material, number of buildings offered and price per building

Structure

Walter

di

Self-compacting concrete 82,55 /m3 the existing norms (that already represent a high standard). The maxi-
Welded wire mesh 0,49 /kg mum time allowed for completing each building from the availability
Steel columns 2,09 /kg of the upper plate was fixed at 80 days, a proposed reduction was also
Isolators 1,427 /piece considered in the evaluation, together with a reduction of the proposed
Forms and on-site assistance 91,7 /m2 price.
Following the presentation of the 58 proposals and an accurate review,
CONSTRUCTION OF THE BUILDINGS 16 contractors were selected, with a total average amount per lot of
about 10,500,000 euro, which means an offered price reduction of
A public call for bids was launched for the construction of the buil- about 5%.
dings; including final design. The 150 buildings to be built were grou- On a total of 150 buildings, timber structures were proposed for 75
ped in 30 lots, each one of 5 buildings, allowing a bidder to present a (50%), concrete structures for 45 (30%) and steel structures for 30
proposal for a maximum of 10 lots. (20%).
Depending on the final ranking of the offers, it might have been possi-
ble to have from a minimum of 3 contractors (in case the first 3 would INFRASTRUCTURES, FURNITURE, ELEVATORS, MECHA-
each propose 10 lots) to a maximum number of 30 contractors (in case NICAL AND ELECTRICAL INSTALLATIONS, GREEN AREAS
each one would have proposed 1 lot). The basic price for any lot of 5
buildings (about 160 covered parking spots, 3,000 m2 of outside pave- To complete the project it was necessary to prepare and launch other
ment and 9,000 m2 internal living area) was fixed at 11 million euro. 5 groups of bids, in order to satisfy various needs:
The evaluation of the proposals was essentially based on the proposed - The upgrading and integration of the external infrastructure
improvement of the minimum performance characteristics foreseen by (networks of any type) with the difficult problem of interacting with the

RESEARCH - LAquila Earthquake 137


Fig. 22- Same complete buildings.

construction sites. Twenty bids were released, one for each area of gement and maintenance of a photovoltaic system, capable of produ-
intervention, inviting companies located in Abruzzo and preferably in cing about 4,500 kW. The call for bid assumed that there should have
the province of LAquila. Five companies randomly sorted out were been no cost to the administration, and was based on technical merit
invited to bid for each site. and on one fundamental economical parameter, i.e. a yearly fee to be
- The furniture and supplies necessary to immediately use the apart- paid, as a percentage of the public incentives provided to favour the
ments. In this case a public competition was set up, on four lots of use of alternative, renewable energy sources. The winner and therefo-
about 1,000 apartments each. The foreseen time to assemble the fur- re contractor was Enerpoint spa, Ener Point Energy Srl and Troiani &
niture on site was 6 days from the moment an apartment would be fini- Ciarocchi Srl., who offered to refund 9,01% of the incentives.
shed. 18 companies presented an offer, with the following four resul- - Finally, two last calls for bids were launched to complete the green
ting winners: Deltongo Industrie spa, Mobilificio Florida srl, RTI areas, simply grouping the eastern and western construction sites. The
Europea spa P.M. International Furnishings srl Martex spa and offers should obviously include land preparation, grass, bushes and
Estel Office spa. The average price reduction offered was about 34%, trees, walking and cycling paths, but also irrigation and drainage
which corresponds to an average cost for the interior furnishing of an systems, external furniture, sport and leisure fields. 19 companies par-
apartment of 9,500 euro. It has to be underlined that the specifics of ticipated in the bid and contractors were selected on cost and on eva-
the bid requested the highest possible standards also for the electrical luation of landscape beauty, environmental sustainability and mainte-
and mechanical equipment included in the offer, such as dish washer, nance and management characteristics of the offers. The selected con-
washing machine, tv set, etc. tractors were 3A Progetti S.p.a, which lowered the estimated price of
- 309 elevators to connect the various floors of the buildings and 129 39% and the Sestante Consortium, that managed to offer a 35,16%
elevators to connect the buildings to the parking ground floor. This reduction.
need derived from the specific choice of completely eliminating all
potential architectonical barriers case, in excess of what compulsory WORK MANAGEMENT, QUALITY CONTROL, SAFETY MEA-
for legal requirement. A call for bids was released for three lots of 146 SURES
elevators each; 12 companies participated in the competition.
Marrocco elevators srl, ATI S.A.S. srl Grivan Group srl, Schindler The extremely limited time available for the completion of the project
spa resulted winners with an average reduction of price of about 16%. required an extreme level of control and programming, with a conti-
- The opportunity of producing electric energy on site, collocating pho- nuous flow of information between engineering work management and
tovoltaic panels on the roofs. An estimate of about 45,000 m2 of roof construction companies and daily reports and checks.
surface was considered adequately exposed to sunshine and conse- A coordination and management system was therefore set up, focusing
quently another bid was released for the design, construction, mana- on the definition of priorities and of the main activities consequently

138 RESEARCH - LAquila Earthquake


Fig. 23- Work in progress and completed
works.

RESEARCH - LAquila Earthquake 139


required and on the identification of potentially critical elements and
work phases that could threaten the fulfilment of the programmed time
schedule.
As previously discussed, the work activities and their management and
control were organised in five sectors, corresponding to production of
the plates, buildings (including interior design and furniture), mecha-
nical and electrical installations, infrastructures roads and green areas.
For each technical sector a technical coordination structure was defi-
ned with a responsible for programming, coordination with the con-
struction companies and management of works. The general timeline of
the works was accordingly subdivided into the same five sectors. Daily
updates on the work progress and comparison with the time planning
guaranteed that each sector was closely monitored in terms of work pro-
gress, as well providing all technical personnel with an overview of the
general picture, fundamental to manage the coordination between dif-
ferent sectors. The graphical visualization of the daily progress of the
works resulted to be particularly useful for a rapid interpretation of the
complexity of the data that needed to be managed.
An idea of the large quantities of materials and labour that needed to be Fig. 24- Example of a global daily overview form.

managed, can be obtained by considering the example of the foundation


and isolation system, that on a daily basis needed an average concrete

Fig. 25- Example of a production summary overview daily form (25 October, general overview on the left, production and delivery of plates on the right).

140 RESEARCH - LAquila Earthquake


Fig. 26- Example of a daily report on the general development of the project
works (19 October).

Fig. 27- Example of a daily report on the general developments of the works in
a specific area (19 October).

RESEARCH - LAquila Earthquake 141


supply of the order of 5,000 m3, welded meshes and reinforcing bars for different category of work and giving the average cost per building, per
about 10,000 kN, about 200 steel columns (with a diameter of 800 mm), apartment and per square metre of living space.
to be provided in general in 20 different construction sites. The total cost of 655 million euro refers to a total of 164 buildings,
The efficiency of the team that was set up to program and coordinate while 150 were foreseen at the starting of the project and 184 were
the activities, allowed such a proper and precise forecasting of the actually built at the end. Applying a criterion of linear proportion it can
work progress that all construction sites proceeded always on time and be inferred that the original 150 buildings would have cost 599 million
actually all works were completed ahead of time, despite of the diffi- euro, which is in line with the 700 million that were initially estima-
culties inherent in the number of workers (more than 8,000 in some ted, since the sums indicated do not include the cost for land expro-
phases) and in the complex interaction between different work activi- priation and V.A.T. The cost of the double slab foundation system is
ties. An example of the time schedule programmed for the construction compensated to significant extent by the value of the covered parking
of the two slabs systems, with foundation and isolation, can be sum- spots, each one of them have the size of a large garage box (6 by 3
marised in Fig. 23. metre). The number boxes exceed that of the apartments. It is thus rea-
Finally, the great efficiency of the team in charge of controlling all sonable to assume that the real cost of the foundation system is actual-
aspects of safety in the work process should be noted. The extremely ly a fraction of that indicated in about 30,000 per apartment. If a
detailed and continuous checks allowed the completion of hundreds of fraction of 30% would be assigned to the foundation itself, the cost per
millions of euro worth of work in just a few months without any nota- square metre of the living space should result to be less than 1,400 ,
ble accident and with the appreciation of all external controlling insti- including foundations: a reasonable value, especially considering the
tutions, from those aimed to assure workers health to the unions. compressed time of construction, that was made possible by a conti-
nuous work over the 24 hours, with three turns of 8 hours each, day and
THE COSTS night and considering as well the very high quality of the buildings for
aspects related to energy consumption, environment and detailing qua-
The total cost of the project is split in the table below, considering the lity.

Parametric costs of the whole intervention, based on 164,29 equivalent buildings, v.a.t. not included

142 RESEARCH - LAquila Earthquake


It is interesting to note that the seismic isolation cost only about 1,5% will be delivered within February 2010, with the possibility of notable
of the total, or rather just a bit over 2% when only the building cost is anticipation if the meteorological circumstances should be favourable.
considered. The delivery of the houses started Tuesday 29 September with about
The modest cost of general and technical activities has to be noted, 500 apartments and will continue at a pace of approximately 300 apart-
made possible by the way the project was managed, extensively ments a week. The property of the buildings will be eventually assigned
discussed in the previous sections. The real share of the technical to the city of LAquila who will be responsible for the management and
costs of the ForCASE consortium (design, management, security, etc) maintenance based on pre-defined procedures, specified in detail in the
has been around 8 million euro, i.e. not much more than 1% of the total project documents. Political and economical choices, with relation to
cost. The costs of the furniture includes everything, from TV sets to the progress in the repair, strengthening and reconstruction of the buil-
bed sheets. dings damaged in the historical centre and in the city outskirt, will drive
the decision on rent costs and use of the new villages.
THE FUTURE The users of the houses are being carefully selected jointly by the city of
LAquila and the DPC, taking into account the preferences expressed by
At the moment this article is being completed (September) the last 20 the homeless people, parameters connected to the family situation (num-
buildings are being constructed, with a significantly lower cost becau- ber of components, age, economical capacity, etc.) and the localisation of
se they are built in already inhabited areas. The decision of the Civil the original place of living. A prerequisite to be considered is that a
Protection Department to build additional houses was motivated by family previous home should have been classified in the category non
upgraded in the population census. It is foreseen that these buildings easily to be repaired (type E and F in the classification of damage).

Forecast of the number of apartments and beds available in function of the foreseen completion
period (forecast of September 22nd, including the twenty buildings that had just been added, of
which forecasts are cautious)

RESEARCH - LAquila Earthquake 143


The houses will anyway become part of the citys heritage and in the the 25,000 students come from other regions.
future it will be therefore possible to reuse them to host vulnerable cate- In the near future the availability of student housing at a controlled
gories of population (such as the elderly people) or to host students, a price could become a relevant peculiarity of the university, modifying
need particularly relevant in LAquila, where a significant fraction of its attraction capacity in a positive way.

Those who have contributed to the success of the C.A.S.E. project: bani, Marco Vecchietti, Paolo Verri, Stefano Vitalini (site responsible for plates), Ro-
berta Viviani
Department of Civil Protection
Giacomo Aiello (legal responsible),Emilia Aloise, Giovanna Andreozzi, Enrico Checker team, administrative aspects
Ardito, Vincenzo Ardito, Arianna Bertelli, Mariano Bonvegna, Angelo Borrelli Giovanna Andreozzi, Maria Laura Conti, Alessandra Conti, Michele Dadamo, Gio-
(administration responsible), Fabrizio Bramerini, Cristina Capriotti, Maria Teresa vanni Di Mambro, Salvatore Fiengo, Giorgio Grossi, Emilia Aloise, Mariano Bon-
Cartolari, Mario Cera, Claudia Ciccone, Pietro Colicchio, Alessandra Conti, Marco venga, Carlo Bordini, Cristina Capriotti, Maria Teresa Cartolari, Carluccio Code-
Conti, Maria Laura Conti, Chiara Dangelo, Antonella De Felice, Giulio De Marco, ghini, Fabio Compagnoni, Dario Compagnoni, Massimo Criscuolo, Antonella De
Giovanni Di Achille, Giovanni Di Mambro, Mauro Dolce (technical responsible), Felice, Giuseppe Fasiol, Maria Cristina Ferroni, Arturo Furlan, Achille Gentile,
Riccardo Fabiani, Maria Cristina Ferroni, Salvatore Fiengo, Claudia Fiore, Alessandro Greco, Gerarda Iannarone, Giuseppe Ianniello, Giovanni Infante, Ettore
Mariasilvia Gianneramo, Beatrice Guerra, Gerarda Iannarone, Federica La Chioma, Iorio, Paolo Marchesi, Luca Pagani, Lucia Palermo, Roberto Pesolillo, Salvatore
Luisa Marinaro, Lucia Palermo, Francesca Paneforte, Ada Paolucci, Roberto Provenzano, Rosario Romano, Gianni Strazzullo, Fabiola Toni, Daniela Ursino,
Pesolillo, Giancarlo Piccione, Patrizia Picuti, Immacolata Postiglione, Giuseppina Michele Villani
Sementilli, Vincenzo Spaziante (general coordinator), Tiziana Tarduini, Vergilio
Tidei, Fabiola Toni, Angelo Vici.
Checker team, structures
Edoardo Cosenza, Gaetano Manfredi, Claudio Moroni, Paolo Pinto (President),
CASE Consortium
Paolo Zanon (assistenti: Massimo Acanfora, Claudio DAmbra, Antimo Fiorillo)
Fabio Aldrovandi, Francesco Ambrosi, Francesco Amici, Maurizio Ardingo (respon-
sible for construction safety measures), Luciano Baglione, Giovanni Bastianini, Paolo
Battegazzore, Giuliano Bellini, Maria Teresa Dolores Bertelegni, Federica Bianchi, Companies and organisations
Saverio Bisoni, Gaia Boggioni, Filippo Bonali, Barbara Borzi, Maria Benedetta Excavations: CO.GE.FER. s.p.a.; Midal s.r.l.; P.R.S. Produzione e Servizi s.r.l.
Bossi, Matteo Bottari, Vittorio Bozzetto, Roberto Brandimarte, Piero Burba, Concrete: Colabeton s.p.a.; Societ Meridionale Inerti SMI s.r.l. Steel reinforcement:
Maurizio Calderari, Andrea Caligari, Maura Castellani, Gian Michele Calvi (project La Veneta Reti s.p.a. Steel columns: A.T.I. Edimo Metallo s.p.a. /Taddei s.p.a.;
leader, designer and construction director), Salvatore Caroli, Christian Caroli, Paolo Cordioli & C. s.p.a.; Formwork and assistance: Consorzio Edile C.M. Gruppo Bison;
Caroli, Francesco Ceribelli, Antonio Coccia, Andrea Colcuc, Oliviero Comand, Sacaim s.p.a.; Zoppoli & Pulcher s.p.a. Isolators: Alga s.p.a.; FIP Industriale s.p.a.
Massimiliano Cordeschi, Filippo Dacarro, Michele DAdamo, Alberto Damiani Buildings: A.T.I. Consorzio Stabile CONSTA s.c.p.a./Sicap s.p.a.; A.T.I. Donati
(responsible for building construction), Pietro Damiani, Edi Danielis, Simonetta Di s.p.a./Tirrena Lavori s.r.l./Dema Costruzioni s.r.l./Q5 s.r.l.; A.T.I. Eschilo Uno
Nicola, Maurizio De Santis, Pasquale Di Marcantonio, Dante Di Marco, Stefano s.r.l./COGEIM s.p.a./Alfa Costruzioni 2008 s.r.l.; A.T.I. Ille prefabbricati
DOttavio, Ettore Fag, Mario Fanutti, Carlo Florio, Pierluigi Fontana, Fabrizio s.p.a./Belwood s.r.l.; A.T.I. Impresa Costruzioni Giuseppe Maltauro s.p.a./Taddei
Frau, Renato Fuchs (organisation coordinator), Nicola Gallina, Marco Gasperi, s.p.a.; A.T.I. Iter Gestione e Appalti s.p.a./Sled s.p.a./Vitale Costruzioni s.p.a.;
Fabio Germagnoli, Federico Gianoli, Daniele Gimnetti, Sergio Giordano, Stefano A.T.I. COGE Costruzioni Generali s.p.a. /Consorzio Esi; Consorzio Etruria s.c.a.r.l.;
Grasso, Carlo Lai, Massimo Lardera (responsible for infrastructure), Ignazio Locci, Consorzio Stabile Arcale; Cosbau s.p.a.; DAgostino Angelo Antonio Costruzioni
Giuseppe Lombardi, Mauro Maganetti, Giovanni Magenes, Claudio Maggi, Carlo Generali s.r.l.; Impresa di Costruzioni Ing. Raffaello Pellegrini s.r.l.; Meraviglia
Magni, Fabrizio Magni, Michele Magnotti, Gabriele Mantini, Antonio Marcotullio, s.p.a.; Orceana Costruzioni s.p.a.; R.T.I. Ing. Armido Frezza s.r.l./Walter Frezza
Paola Marotta, Sara Martini, Emanuele Meago, Paola Migliazza, Enrico Misale, Costruzioni s.r.l./ Archilegno s.r.l.; Wood Beton s.p.a. Furniture: Del Tongo
Marta Molinari, Federico Monutti, Matteo Moratti (site responsible for structures), Industrie s.p.a.; Estel Office s.p.a.; Mobilificio Florida s.r.l.; R.T.I. Europeo s.p.a./
Vincenzo Pane, Vincenzo Paolillo, Alessandro Papale, Carmine Pascale, Pierluigi PM.International Furnishing s.r.l. Infrastructure: CO.M.AB. Appalti Pubblici e
Pascale, Moreno Pavan, Fausto Pedetta Peccia, Gianfranco Peressutti, Edoardo Privati s.n.c.; Codimar s.r.l..; Codisab s.r.l.; Conglomerati Bituminosi s.r.l.;
Peronace, Michele Pescina, Paolo Petrucco (site responsible for plates and infra- Facciolini s.r.l.; G.C.G. s.r.l.; I Platani s.r.l.; Impresa Edile Di Cola Michele; Ing.
structure), Piero Petrucco, Nereo Petten, Dario Pietra, Roberto Pitolini, Federica Armido Frezza s.r.l.; Molisana Inerti Conglomerati s.r.l.; Produzione e Servizi s.r.l.;
Polidoro, Alessandro Pollini, Stefano Pozzi, Salvatore Provenzano, Bruno Quadrio, Ridolfi Idio e Figli s.r.l.; San Giovanni Inerti di Pietro Mascitti s.r.l.; Valentini
Nadia Rizzardi, Enzo Rizzi, Fabio Roiatti, Cristiana Ruggeri (responsible for mecha- Costruzioni s.a.s.; Elevators: Marrocco elevators s.r.l., ATI S.A.S. s.r.l./Grivan Group
nical and electrical installations), Gaetano Ruggeri, Mario Rusconi, Daniele s.r.l., Schindler s.p.a.; Photovoltaic panels: R.T.I. Ener Point s.p.a./Ener Point
Sambrizzi, Valentina Scenna, Matteo Schena, Michele Schiabel, Paolo Scienza, Energy s.r.l./Troiani & Ciarrocchi s.r.l.; Green areas: R.T.I. 3a Progetti/Gsa
Fabiola Sciore, Roberto Scotti, Domenico Sgr, Martino Signorile, Danilo Marco s.r.l./O.Ci.Ma. s.r.l./Bellomia-Sebastianini-Euroengineering s.r.l., Consorzio
Siviero, Luigi Spadaro, Davide Tagliaferri, Piergiuseppe Tamburri, Alessandro To- Sestante. Demolition: CODISAB SRL, A.S.M. s.p.a.; Connections to external pipe-
sello, Stefan Trenkwalder, Roberto Turino (site responsible for buildings), Diego Ur- line networks: ENEL Rete Gas, ENEL Energia, GranSasso Acqua.1

144 RESEARCH - LAquila Earthquake


REFERENCES lation bearings, International Jour. for Num. Methods in Engineering, Vol. 40,
29-49.
1. NTC (2008) - Norme Tecniche per le Costruzioni, D.M. 14/01/2008, Gazzetta Uffi- 8. Wang Y., Chung L.L., Liao W.H. (1998) - Seismic response analysis of bridges
ciale 04/02/2008, Italia. isolated with friction pendulum bearings, Earthquake engineering and structural
2. Zayas V., Low S. (1990) - A Simple Pendulum Technique for Achieving Seismic dynamics, 27, 1069-1093.
Isolation, Earthquake Spectra, Vol. 6, No. 2. 9. Priestley M.J.N., Calvi G.M., Kowalsky M.J. (2007) - Displacement based design
3. Almazan J.L., De la Llera J.C. (2002) - Analytical model of structures with fric- of structures, IUSS Press, Pavia.
tional pendulum isolators, Earthquake engineering and structural dynamics, Vol. 10. Crowley, H. and Pinho R. (2004) - Period-height relationship ofr existing
31, 305-332. European reinforced concrete buildings, Journal of Earthquake Engineering, Vol. 8
4. Calvi G.M., Ceresa P., Casarotti C., Bolognini D., Auricchio F. (2004) - Effects of (SP1), 93-120.
axial force variation on the seismic response of bridges isolated with friction pen- 11. Crowley H., Stucchi M., Meletti C., Calvi G.M., Pacor F. (2009) - Uno sguardo
dulum systems, Journal of Earthquake Engineering, Vol. 8, SI1, 187-224. agli spettri delle NTC08 in relazione al terremoto de LAquila, capitolo 1.7 in que-
5. Christopoulos C., Filiatrault A. (2006) - Principles of Passive Supplemental sto volume.
Damping and Seismic Isolation, IUSS Press, Pavia. 12. AA.VV. (2007) - Definizione dellinput sismico sulla base degli spostamenti,
6. Priestley M.J.N., Calvi G.M. (2002) - Strategies for repair and seismic upgrading progetto S5 INGVDPC, http://progettos5.stru.polimi.it.
of Bolu Viaduct 1, Turkey, Journal of Earthquake Engineering, Vol. 6, SI1, 157-184. 13. Comit Europen de Normalisation, Eurocode 8 part 2 (2006) - prEN1998-2,
7. Tsai C.S. (1997) - Finite element formulations for friction Pendulum seismic iso- CEN, Brussels.

Here and in the next pages, some pictures of the completed buildings areas.

RESEARCH - LAquila Earthquake 145


146 RESEARCH - LAquila Earthquake
RESEARCH - LAquila Earthquake 147
148 RESEARCH - LAquila Earthquake
RESEARCH - LAquila Earthquake 149
150 RESEARCH - LAquila Earthquake
RESEARCH - LAquila Earthquake 151
CONSTRUCTION
Italian High-Speed Network
A special focus on concrete structures

he new Italian high-speed Network, due to the needs of increasing the Strategic choices
T capacity of the actual railway operating lines, nearly doubling the
actual number of trains running daily, and decreasing travel time, has involv- Traffic analysis carried out on existing Italian railway network at the end of
ed in the last fifteen years, large engineering resources, construction skills the 1980 remarked the following needs: quadrupling the main passenger
and strict planning, managing organization, railway engineers supervision, transport routes, upgrading and increasing freight transport, reducing time
apart from the necessary economical huge investments. of travel for passengers trains, integrating Italian network with European
The main characteristics of the new lines and the strategic choices from the network.
infrastructural and structural points of view are presented in the next chapter. So, the first infrastructural choice was the realization of a new mixed pas-
The new lines are designed according to the national code for the design senger/freight high-speed network with a close integration with existing
and construction of railways bridges [1], and to the most advanced techno- lines and with interchange centres (interports, ports, airports).
logical standards in order to achieve the safest conditions of service, speed The close integration with the existing conventional network will produce an
and interoperability with operating railways and with the European high- increase of freight transport capacity on the historical lines, clearing the
speed lines of railway transportation of passengers and freight. existing network from the long distance passengers traffic, and an increase
More than 90 % of bridges and viaducts of the new high-speed lines are of freight traffic using new lines during specific time bands (usually at night).
realised with simply supported spans of prestressed concrete (PC) decks, The choice of a mixed traffic meant low ruling gradients (less than 12 )
so attention will be paid on their typical structural solutions. and heavy design loads (SW0/SW2) adopted in the new Italian standard
Simply supported composite steel and concrete spans, few continuous brid- for railway bridges [1]. For this reason, the standard was rewritten in 1995,
ges and some special structures (arch bridges and the cable-stayed bridge then revised in 1997.The old standards had been written 50 years before
over Po River [2]) compose the remaining part. and were related only to conventional lines.The Italian standard for railway
Main concepts of durability of concrete structures required in the design of bridges [1] has introduced LM71 (passengers traffic as showed in leaflets
railway bridges are then focused. UIC 702 and 776-1), SW/0 and SW/2 (heavy traffic) models of loads
Construction processes are then described, taking examples from the late- according to ENV 1991-3: Actions on structures, Part 3:Traffic loads on brid-
st viaducts with PC decks. ges (Ed. 1996).
Finally, main concepts in the design of continuous PC bridges are outlined. These new standards are in perfect agreement with the European Technical
Standards for Interoperability of the trains in the European High-speed
Network.
One of the main prescriptions asks for a structural design respectful to all
prescriptions for seismic areas (at least III category the minimum consi-
dered in the 1996 Italian seismic code), even in no-seismic areas, apart
from Sardinia. A proper standard was written and recently revised for the
design of railway bridges to be built in seismic area [4]. It deserves to be
pointed out that many no-seismic areas became recently seismic, in the
latest proposals of codes, giving interesting confirmation to the conservati-
ve railways code approach.
According to general seismic design principles, the adoption of special rules
and technical details is requested to guarantee a minimum ductility of the
structure, and it has direct consequences on the care for the details of
reinforcement design of piers and foundations.
Two main characteristics of the applied national code [1] are the concepts
of train-track-structure dynamic interaction and train-rail-structure static
interaction.
The dynamic interaction analysis is evaluated to check the safety of the
train and the comfort of the passengers, with an analysis of the following
parameters: all decks for high-speed railway must respect the limit value of
2.5 as maximum dynamic amplification of static deflection (impact factor
j real=j dyn /d stat) and the value of the vertical acceleration at deck mid-
span, induced by real trains running at different speed (from 10 km/h up
to 1.2 maximum speed of the line), must be lower than 3.5 m/s2.
For standard simply supported beam or continuous bridges with total
Fig. 1- Italian New High-speed Network: lines under design, construction and the operating line Roma-Firenze, built in the
70s-80s for a design speed of 250 km/h [3]. length shorter than 130 m, a simplified analysis can be adopted according

154 CONSTRUCTION - Civil engineering Works


to Annex A of [1], with a preliminary check of flexural - frequencies of vibra- with long welded rail, to avoid rail expansion devices, to ease maintenance
tion modes: the simply supported prestressed concrete decks always operations and minimize maintenance costs. Besides, this solution is usual-
respect the limit values specified in [1], being first flexural mode frequency ly preferred in those cases when bridges have to be designed in areas with
between 4 and 8 Hz. compressible soils or in river channels.
For non-conventional structures, as arch bridges, cable-stayed bridges, etc., Figures and statistics of this paper are based on more than 600 km long
a Runnability analysis is required.The analysis has to consider all dynamic high-speed lines. Attention will be paid to double-track decks, with a distan-
characteristics of the system: railway structures, suspension system of the ce between tracks of 5.0 m, designed for a train-speed of 300 km/h, and
vehicles, rail fasten system etc., track and wheel irregularities. for both heavy and passengers traffic load models [1], with rails on pre-
The static interaction analysis studies the effects on rail and bridge struc- stressed concrete sleepers on ballast.They count nearly 2300 spans of sim-
ture due to variation of thermal conditions in the structures, to the longitu- ply supported prestressed concrete bridges, and they are composed of
dinal forces associated to braking and traction, and to the longitudinal nearly 6400 precast beams or monolithic decks.
displacements due to vertical loads. For simply supported prestressed con- In order to reach this frame of prestressed concrete decks, analysis will con-
crete decks, which respect a maximum length of 65 m and small variations cern the structural characteristics such as use of pre-casting, tensioning
of the longitudinal stiffness of piers and foundations, a simplified method systems, bearings, expansion joints, all durability issues as multi-layer pro-
can be adopted as indicated in Annex B of [1]. tection systems, monitoring, methods of construction and costs.
In any other case, it is necessary to analyse advanced Finite Element models Deck The pre-stressed concrete elements are realised with both pre-ten-
to evaluate these effects. The analysis has to check rail stress limits with sioning and post-tensioning systems.The post-tensioning systems are always
maximum value of compressive stress 60 MPa and maximum value of ten- designed with bonded internal cables, even if Italian standard for railway
sile stress 70 MPa, the relative displacement between deck bridge and the bridges [1], generally speaking and under severe controls, allows also exter-
rail, and the forces acting on the bearings. nal post-tensioning. According to [1], post-tensioning cables composed by
Also reliability under service conditions is required: comfort limit state has bars should be preferred for viaducts along railways with electric traction of
to be verified for a maximum midspan deflection with the load of one direct current, and both solutions with cables composed by strands or bars
LM71 load model, increased with dynamic factor. This deflection must not can be used in structures for railways with electric traction of alternating
exceed l/2400 for design length of the span l<30m, l/2800 for current as the high-speed network.
30m<l<60m and l/3000 for l >60m. Maximum deformability of structu- In [1], special attention for durability and limiting or avoiding cracking of
res under train load is checked to keep the contact rail-wheel safe and sta- concrete is introduced: undoubtedly most limiting verifications deal with limi-
ble: deck torsion, rotation at supports and horizontal deflection have to be tation of maximum compressive stresses and, especially, strong limitation of
evaluated. The limits of deformation of the structures are similar to those tensile stresses during construction and final conditions. In particular, no lon-
pointed out in the same Eurocode, and are widely respected by common gitudinal tensile stress in PC structures is admitted, with maximum design
simply supported spans. loads and both Allowable Stress or Limit States methods of verification.
Special attention has been put in the concepts of durability of structures for Besides, cracking of concrete must be verified towards no decompression
railway bridges, introduced in [1]. As the subject deserves wide illustration limit state for verification under track equipment, where inspection is not
and details, a full paragraph has been devoted to the scope. possible.
As high-speed network is designed for the use of long welded rail, the struc- The experience of existing railway lines with concrete structures with pos-
tural system for the viaducts must avoid rail expansion devices. In all high- sible beginning of corrosion of the reinforcement and spalling of the con-
speed network, only along the Milano-Bologna line, for the crossing over Po crete, which leads to easier access to the pre-stressing tendons for aggres-
River, composed by two continuous bridges and the cable stayed bridge [2], sive agents, has been translated into design prescriptions.The required con-
two joints in the rails have been necessary to keep the expansion length crete cover to reinforcement, tendons and pre-tensioned strands has been
within allowable limits. increased, compared to Italian standard for design of structures. Minimum
Other rules and prescriptions for design and construction are taken into concrete cover thickness is required to be 3 cm for PC decks, increased to
consideration in [1] and are illustrated in the following paragraphs. All these 3.5 cm under track equipment, at least one external diameter of duct in
are finalised to have low costs of maintenance of the infrastructure, to mini- case of post-tensioning, and 3 strand diameters in case of pre-tensioning.
mise the irregularities of the track and to reach a high performance level Mix design of concrete for PC deck has to respect a 0.45 water to concrete
in the field of the durability and reliability of the system. ratio, a S4S5 concrete consistency class of at least 45 MPa characteristic
cubic strength. A quality assurance system and testing before and during
Simply supported prestressed concrete bridges every casting operation reveals the quality of mix design, which is recogni-
sed as an important factor for life and durability of PC structures.
Simply supported spans of prestressed concrete deck realised more than Bearings and expansion joints Under simply supported railway brid-
90% of the new lines. It is undoubtedly a traditional choice of Italian Rail- ges, only one kind of bearing is generally present: spherical bearings with
way Company (Ferrovie dello Stato FS) for ordinary viaducts: to better fit polished stainless steel and PTFE plate.

CONSTRUCTION - Civil engineering Works 155


Italian High-Speed Network

With this kind of bearing, rotations can occur till 0.0167 rad in all direc-
Type a
tions, in order to place the bearing without inserting packings. In order to Weight of one precast box: 455 ton (33.1 m)
avoid parasite forces arising with only one train on a double line bridge 19% of total length of viaducts
915 ton one deck weight (34.5 m)
deck, a new kind of fixed bearings has been studied: it has a special device 343.196 ton is total weight

which controls horizontal stiffness.


The cover of expansion joints are realised with dielectrical elastomeric
cushion joints, composed by neoprene reinforced with vulcanized steel pla-
tes.They allow fast bearings changing with a maximum differential lifting of
50 mm between decks, operated by hydraulic jacks between deck and pier Type b
Weight of single precast V beam: 88 ton
cap (all simply supported or continuous decks have to pass this design veri- 40.5% of total length of viaducts
650 ton one deck weight (25 m)
fication) without any operation under rails. Actually, the use of mechanical 698.476 ton is total weight

devices instead of resins avoids disease to daily train in case this operation
becomes necessary.
On pier caps and abutments of every bridge span, reinforced concrete or
steel devices (stroke end device) are required in order to avoid deck slip-
ping out of pier cap and falling, because of accidental breaking of fixed bea- Fig. 2- Two box girders deck (a) and four precast V beams and cast in situ slab (b).
rings e.g. in case of devastating earthquakes.There are pillows of reinforced
neoprene where decks may hurt against these provisions and their main- beams with post-tensioning cables. In Roma-Napoli line it was also realised
tenance or changing operations has to be assured by proper design. with two V-beams and cast in situ slab. Transversal beams are usually pre-
Italian standard for railway bridge bearings and cover of expansion joints stressed with straight cables of strands or bars.
requires these devices underpass preliminary homologation tests led by F.S. The number of transversal beams is prescribed in [1]: for a deck with two
technicians, through prototypes testing, in order to assure quality of every or more girders, at least two prestressed concrete transversal beams have
single component and of the final assembled products. to be designed out from supports and more in case of decks longer than
Piers and foundation Piers have usually circular or rectangular, full or 25 m.
empty, cross-sections, while foundations are usually realised with plinths with Strands getting out from the heads of the box girder are cut, isolated and
large diameter reinforced concrete piles. protected with the use of dielectric resin. Decks' deformability is largely veri-
In case of piers in riverbed, even if empty structural sections are adopted, fied for comfort limit state: maximum deflection at midspan for Type a is
low class of concrete is always poured inside till the river maximum level, in less than l/5600.
order to avoid unexpected water inside. Type b is composed by four precast V-beams and cast in situ slab: actual
As previously mentioned, in order to increase structural safety, all bridges are maximum length is 33.6 m.
designed considering at least low seismic condition: it focuses the designers' Beams are steam cured, pre-tensioned with longitudinal steel strands and
attention especially on reinforcement details, very important for piers and
piles. Good number of stirrups and loops for longitudinal bars and concre-
te confinement, use of hooks for good stirrup behaviour, limitation of maxi-
mum compression stress in pier concrete, no junction or superposition of
longitudinal bars in the length of 3 m from foundation, etc. are consequen-
ces of above-mentioned prescriptions.
The minimum reinforcement areas for both piles and piers is fixed to the
0.6 % area of concrete section, and spirals are admitted as stirrups in
reinforced concrete piles only if welded to longitudinal bars in every inter-
section.

Typical cross sections

The most common cross-sections of prestressed concrete decks are showed


in Fig. 2, 6 and 11; in the following, a brief description of main features is
presented for each typical cross-section.
Type a is a box girders deck, spanning till 34.5 m, generally composed by
two precast box girders, prestressed with longitudinal steel strands and con-
nected with small second step casting in the slab and with transversal Fig. 3- Prestressed concrete V beam (type b) in stocking area, Torino-Milano line.

156 CONSTRUCTION - Civil engineering Works


transversally connected with cables in transversal concrete beams; it is the
most common deck: it has been chosen for 40.5 % of total length of sim-
ply supported prestressed concrete deck.
The maximum deflection at midspan is largely verified: for type b span-
ning 25 m (22.3 % of length of all prestressed concrete decks) it is less
than l/6000.
Because of the prescription of complete absence of tensile stress during
construction and life of the bridge, and of the large amount of pre-tensio- Fig. 5- The first precast 25m long single box girder (Torino-Milano), during launching operations.
ning strands in V-beams, the technique of strand passivation for few metres
along beam ends, over supports, has been introduced for a portion of sversal beams or second step castings of concrete in head anchorages of
strands, in order to reduce even minimal cracking on heads of the beam. tendons in the required transversal beams, which always become visible
From an aesthetic point of view, shortest spans of box girders (both V-beam with time.
or cellular deck) can be put at a disadvantage, because in [1] a free hei- Anyway, in some case of grillage deck, when perspective had to be impro-
ght of at least 1.6-1.8 m inside box girder is prescribed to be left to ease ved, concrete noise barriers have been usefully adopted, covering second
inspection, leading to relevant height of the deck even for short span brid- step casting or empty spaces on pier-cap between decks for inspection.
ge. Anyway, these spans can be agreeably inserted in case of viaducts with Type e is composed by four precast I beams and cast in situ slab.
short piers. Beams are longitudinal post-tensioned with cables composed by strands
Type c is a single box girder deck, realised in two different ways and with straight and parabolic profiles.
Some are tensioned in precasting plant, then, after completing the bottom
slab and tensioning of transversal cables, the second part of longitudinal
cables is tensioned over the piers and slab is casted. Longest span of type
e is also the longest span for simply supported prestressed concrete
decks: 46.2 m.
Type e has the advantage to manage precasting and launching of one
beam at time instead of full deck, so requiring simpler technology, but, at
the same time, the operation of assembling formworks and casting con-
nection of lower slab and transversal beams, and the huge transversal post-
tensioning (no. 47 4-strands cables for 46.2 m long deck) may put it to a
disadvantage.
Type f is the original Modena viaduct: the first case of lower way U deck
for high-speed lines; the double track is realized by two independent decks,
piers and common foundation. It has two single track decks spanning 31.5
m, and a total width of 18.4 m; each deck is pre-stressed with 20 longitu-
dinal post-tensioning tendons of 12 strands. 566 km of corrugated plastic

Type c
Weight of single precast deck: 567 ton (25 m)
Fig. 4- Box girder (type a) on carriers towards launching operations, Milano-Bologna line. Weight of cast in situ box deck: 1043 ton
(43.2 m)
11% of total length of viaducts
lengths: 25 m long precast box girder with longitudinal pre-tensioned steel 173.856 ton is total weight

strands on Torino-Milano line (3.78 km long Santhi and 1 km long Carisio


viaducts) and cast in situ post-tensioned deck spanning 43.2 m (2.8 km
long Padulicella viaduct) on Roma-Napoli line.
Type d is adopted in 5.1 km long Piacenza viaduct: 150 precast spans Type d
with two cells and curved transversal profiles. It is a single monolithic box Weight of single precast deck: 970 ton
7.2% of total length of viaducts
girder of 970 ton, with a maximum length of 33.1 m. 145.500 ton is total weight

Piacenza viaduct has been provided with 119 km of corrugated plastic


ducts and it represents the first application of electrically isolated disposals
for the anchorages of 12 and 19 strands for longitudinal post-tensioning
cables. Compared to the grillage decks, the monolithic decks have the
aesthetic advantage of clean prospects and even deck sides, without tran- Fig. 6- Box girder with single cell (c) and Box girder with two cells (d).

CONSTRUCTION - Civil engineering Works 157


Italian High-Speed Network

Through periodical inspection of old railway lines and particularly of the


Roma-Firenze railway line, an analysis of the most common defects on brid-
ge deck have been made in order to work out new guidelines of design and
construction of the high-speed lines. Inadequate access for inspection,
Fig. 7- Piacenza viaduct precast deck, 33.1 m, in stocking area.
leaking of waterproofing system of the expansion joints between decks,
Fig. 8- Perspective of curved profiles of Piacenza viaduct. insufficient cover, not efficient bearings towards maintenance operations as
lifting of decks in order to change bearings, or reduced space to insert
hydraulic jacks on pier-caps, etc. have been found as most frequent defects.
These circumstances led to the introduction of new or more detailed pre-
scriptions in the Italian standard for railway bridges [1], in order to gua-
rantee better behaviour of structures with time, care for durability of con-
crete structures, tensioning cables, system of waterproofing, devices for good
Fig. 9- Four precast I beams in stocking area, from Roma-Napoli high-speed line.
Fig. 10- Four precast I beams launched over pier caps, from Milano-Bologna high-speed line. drainage and anything else whose purpose is to ensure the overall long-
term integrity of bridge structure. But, first of all, great care is put to inspec-
tion ways to check the conditions of the structures and medium-life ele-
Type e
Weight of single precast I beam: ments.
270 ton (46.2 m)
Weight of four I beams deck:
Access for inspection All bridges are designed assuring access for
1.400 ton (46.2 m)
7.6% of total length of viaducts
inspection, testing, maintenance and possible replacement of medium life
137.274 ton is total weight elements. It must be always possible to walk over bridge decks because a
width of min. 50 cm on both sides is left for maintenance people. Inside cel-
lular deck or closed box girders, as previously mentioned, a minimum hei-
Type f ght of 1.6-1.8 m must be always guaranteed; fixed stairs from deck to pier
Weight of single precast
deck: 689 ton cap must be provided every 3 spans or 100 m and from piers to the
15% of total length of
viaducts ground every 500 m, for viaducts longer than 1000 m. Over pier cap it
446.472 ton is total
weight must be always possible to pass from one deck to the following and stairs
or landings are fixed to ease the movements of maintenance people.
Finally, it must be also possible to inspect bearings and stroke end devices
Fig. 11- Four precast I beams and cast in situ slab (e) and Lower way U deck (f).
or to operate in case of replacement of bearings or neoprene pillows, so a
free height of 40 cm is left between lower side deck and top of pier-cap.
Drainage It is a key issue about durability; deck slabs are provided with
provisions for good drainage and great attention is put to design, testing and
layout of all devices. Drainage of expansion joints is assured by a flashing
tray of elastomeric material stuck with resins to slabs ends, in order to
avoid leaking over piers and to drain water out of deck sides. Over bridge
Fig. 12- The first one of 750 Modena precast decks in stocking area, Milano-Bologna line. deck, in order to protect from atmospheric agents, a thick layer of water-
Fig. 13- Beam head of Modena precast deck in stocking area, Milano-Bologna line.
proofing is extended, also beneath the footways. In case of sensible pre-
stressing system, pre-stressing strands or post-tensioning cable system just
ducts is used. In Milano-Bologna line this structural solution for prestressed beneath deck slab or critical drainage system (types a, c, d and f),
concrete deck is used for 10.7 km long double track viaducts. It is also used a sprayed polyurethane waterproofing of 3 to 5 mm is extended. Checks
in Junctions, for another 3.33 km of single-track, for a total number of 767 are been carried out on site for adhesion and thickness by F.S. technicians.
precast spans. In particular, Modena viaduct is the longest viaduct of all This surface treatment has proven to be very long life cycle performant.
high-speed lines, with its length of 7.1 km. In post-tensioned concrete structures, deck anchorages are to be avoided
The structural solution of lower way deck has been usefully adopted to on deck slab; anyway, for all anchorages, design has to avoid leakage to get
minimize structural height plus noise barrier, because the noise barriers access to anchorages, providing protection against leaking expansion joints,
became a part of the structure. Modena system of viaducts is one of the as water drips.
few cases, where aesthetics and environment impact have so strongly led Post-tensioning tendonsThe grouting of the sheaths of prestressed con-
the process of design and the solutions for structural and construction crete bridges is always done with vacuum technique with a depression of
needs, in order to have quite original deck, unique in its kind. 0.2 bar during injection. It is standard for prestressed concrete deck with
Durability post-tensioning tendons because grouting has been recognised as a key

158 CONSTRUCTION - Civil engineering Works


operation to ensure durability of prestressing. Usually, it is difficult to ensu- number of strain-gages, inclinometers, thermocouples, instrumented bea-
re complete filling of the ducts: pathology teaches us that the prestressing rings, load cells, foundation settlement meters, piezometers etc. Seldom
tendons are more vulnerable in the case of post-tensio