Seismic Behaviour and Design of Irregular and Complex Civil Structures II

Geotechnical, Geological and Earthquake Engineering
Zbigniew Zembaty
Mario De Stefano Editors
Seismic Behaviour
and Design of
Irregular and
Complex Civil
Structures II
Geotechnical, Geological and Earthquake
Engineering
Volume 40
Series Editor
Atilla Ansal, School of Engineering, Özyegin University, Istanbul, Turkey
Editorial Advisory Board
Julian Bommer, Imperial College London, U.K.
Jonathan D. Bray, University of California, Berkeley, U.S.A.
Kyriazis Pitilakis, Aristotle University of Thessaloniki, Greece
Susumu Yasuda, Tokyo Denki University, Japan
More information about this series at http://www.springer.com/series/6011
Zbigniew Zembaty • Mario De Stefano
Editors
Seismic Behaviour and

Design of Irregular and
Complex Civil Structures II
Editors
Zbigniew Zembaty Mario De Stefano
Faculty of Civil Engineering Department of Architecture (DiDA)
Opole University of Technology University of Florence
Opole, Poland Florence, Italy
ISSN 1573-6059 ISSN 1872-4671 (electronic)

Geotechnical, Geological and Earthquake Engineering
ISBN 978-3-319-14245-6 ISBN 978-3-319-14246-3 (eBook)
DOI 10.1007/978-3-319-14246-3
Library of Congress Control Number: 2013930359
Springer Cham Heidelberg New York Dordrecht London

© Springer International Publishing Switzerland 2016
This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of
the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
or information storage and retrieval, electronic adaptation, computer software, or by similar or
dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt
from the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, express or implied, with respect to the material contained
herein or for any errors or omissions that may have been made.
Printed on acid-free paper
Springer International Publishing AG Switzerland is part of Springer ScienceþBusiness Media

(www.springer.com)
Preface
Structural regularity was a logical preference for traditional civil engineering

design when the hand, static calculations motivated early researchers to simplify
structural models. What is more, application of the simplified separate, planar
models of the actual three-dimensional structures in most cases resulted in a
conservative design. However, when it comes to modern structural dynamics and
earthquake engineering, the results of similar simplifications are quite the opposite.
The actual three-dimensional seismic response of irregular structures generates
substantial additional design burdens compared to the results of direct applications
of simple planar, code formulas. The same is true when it comes to complications in
the seismic load models resulting from the wave propagation effects, i.e. spatial,
non-uniform excitations or rotational ground motion. To address such formulated
research aims, contributions coming from modern seismology are also required. In
most cases, accounting for these effects will increase the seismic design load.
Nonlinear effects during strong motion excitations, as well as modern passive and
active control of irregular structures, require yet more advanced research method-
ology. Furthermore, major seismic codes are still in the need of improvement
regarding both regularity criteria and analysis methods. That is why these effects
are being intensively studied.
This book presents a second collection of the state-of-the-art papers devoted to
these subjects and published within the framework of the Springer Geotechnical,
Geological and Earthquake Engineering series. The first book1 appeared in 2013
and had a positive response, particularly from the earthquake engineering commu-
nity. The second volume contains 30 reviewed and edited contributions presented
during the 7th European Workshop on the Seismic Behaviour of Irregular and
Complex Structures, which took place in Opole, Poland, on October 17–18, 2014.
1
Lavan O., De Stefano M. (editors), Seismic Behaviour and Design of Irregular and Complex
Structures, Springer, Geotechnical, Geological and Earthquake Engineering, Dordrecht, 2013
(http://www.springer.com/gp/book/9789400753761).
v
vi Preface
Forty participants from 15 countries took part in this successful event organized as
in the earlier workshops, by Working Group (WG) 8 of the European Association
for Earthquake Engineering: Seismic Behaviour of Irregular and Complex Struc-
tures, as well as by the Polish Group for Seismic and Paraseismic Engineering
affiliated to the Polish Academy of Sciences and by the Faculty of Civil Engineer-
ing, Opole University of Technology. Previous conferences which took place in
Capri, Italy (1996); Istanbul, Turkey (1999); Florence, Italy (2002); Thessaloniki,
Greece (2005); Catania, Italy (2008); and Haifa, Israel (2011) also resulted in
published proceedings. This conference cycle has already established itself as a
valuable venue for exchanging of ideas in the area of irregularity issues in earth-
quake engineering.
The book consists of 30 chapters and is divided into 3 parts:
• The first part, devoted to the complexity issues of ground motion modelled as
wave propagation effect, consists of seven papers referring to seismic engineer-
ing and modern seismology issues.
• The second part, consisting of 20 chapters, covers the key area of structural
irregularity effects on seismic response, both the horizontal ones, leading to
structural torsion and vertical setbacks often leading to the amplification of the
overall seismic response.
• The third part consists of three papers devoted to active and passive control of
the irregular structures.
Opole, Poland Zbigniew Zembaty

Florence, Italy Mario De Stefano
Contents
Part I Ground Motion, Rotations and Wave

Propagation Effects
1 A General Procedure for Selecting and Scaling Ground
Motion Records for Nonlinear Analysis
of Asymmetric-Plan Buildings . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Juan C. Reyes, Erol Kalkan, and Andrea C. Ria~no
2 Prediction of the “Average” Peak Nonlinear Seismic
Response of Asymmetric Buildings Under Bi-directional
Ground Motion Acting at an Arbitrary Angle of Incidence . . . . . . 13
Kenji Fujii
3 Evaluation of Torsional Component of Ground Motion
by Different Methods Using Dense Array Data . . . . . . . . . . . . . . . . 25
G.R. Nouri, M.R. Ghayamghamian, and M. Hashemifard
4 Estimation of Rotational Ground Motion Effects
on the Bell Tower of Parma Cathedral . . . . . . . . . . . . . . . . . . . . . . 35
Zbigniew Zembaty, Andrea Rossi, and Andrea Spagnoli
5 FOSREM: Fibre-Optic System for Rotational Events
and Phenomena Monitoring: Construction, Investigation
and Area of Application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
Leszek R. Jaroszewicz, Anna Kurzych, Zbigniew Krajewski,
Jerzy K. Kowalski, and Krzysztof P. Teisseyre
vii
viii Contents
6 Application of Rotation Rate Sensors in Measuring Beam

Flexure and Structural Health Monitoring . . . . . . . . . . . . . . . . . . . 65
Zbigniew Zembaty, Seweryn Kokot, and Piotr Bobra
7 Asymmetric Continuum with Shear and Rotation Strains
Including Quantum Synchronous Processes . . . . . . . . . . . . . . . . . . 77
Roman Teisseyre
Part II Seismic Analysis and Design of Irregular

Structures
8 Seismic Assessment of RC Frame Buildings . . . . . . . . . . . . . . . . . . 89
Klemen Sinkovič, Iztok Peruš, and Peter Fajfar
9 Torsional Index of an Asymmetric Building Based
on Mode Shape . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
Kenji Fujii
10 An Approximate Method for Assessing the Seismic
Response of Irregular in Elevation Asymmetric Buildings . . . . . . . 111
George K. Georgoussis
11 Application of Nonlinear Static Procedures for the Seismic
Assessment of a 9-Storey Asymmetric Plan Building . . . . . . . . . . . 123
André Belejo and Rita Bento
12 Seismic Assessment of an Existing Irregular RC Building
According to Eurocode 8 Methods . . . . . . . . . . . . . . . . . . . . . . . . . 135
Alessandra La Brusco, Valentine Mariani, Marco Tanganelli,
Stefania Viti, and Mario De Stefano
13 The Concrete Strength Variability as Source of Irregularity
for RC Existing Buildings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
Stefania Viti, Marco Tanganelli, and Marco De Stefano
14 The Influence of Axial Load Variation on the Seismic
Performance of RC Buildings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
Valentine Mariani, Marco Tanganelli, Stefania Viti,
and Mario De Stefano
15 Parametric Study of Inelastic Seismic Response
of Reinforced Concrete Frame Buildings . . . . . . . . . . . . . . . . . . . . 171
Asimina M. Athanatopoulou, Grigorios E. Manoukas, and Amfilohios
Throumoulopoulos
16 Seismic Upgrading of Vertically Irregular Existing
r.c. Frames by BRBs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181
Francesca Barbagallo, Melina Bosco, Edoardo M. Marino,
Pier Paolo Rossi, and Paola R. Stramondo
Contents ix
17 Application of Nonlinear Static Method with Corrective

Eccentricities to Steel Multi-storey Braced Buildings . . . . . . . . . . . 193
Melina Bosco, Giovanna A.F. Ferrara, Aurelio Ghersi,
Edoardo M. Marino, and Pier Paolo Rossi
18 Influence of the Interaction Yield Domain on Lateral-Torsional
Coupling of Asymmetric Single-Storey Systems . . . . . . . . . . . . . . . 205
Melina Bosco, Aurelio Ghersi, Edoardo M. Marino,
and Pier Paolo Rossi
19 Improved Nonlinear Static Methods for Prediction
of the Seismic Response of Asymmetric Single-Storey Systems . . . . 215
Melina Bosco, Aurelio Ghersi, Edoardo M. Marino,
20 Influence of the Rotational Mass Inertia on the Torsional
Seismic Response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225
Dietlinde K€
ober and Dan Zamfirescu
21 Seismic Response Trends of SDOF Plan Irregular
Structures. Simplified Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . 233
Dietlinde K€
22 Maximum Corner Displacement Amplifications
for Inelastic One-Storey In-Plan Asymmetric Systems
Under Seismic Excitation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243
Michele Palermo, Stefano Silvestri, Giada Gasparini,
and Tomaso Trombetti
23 Earthquake-Induced Pounding Between Asymmetric
Steel Buildings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255
Barbara Sołtysik and Robert Jankowski
24 Dynamic Analysis of Irregular Multistorey Shear
Wall Buildings Using Continuous-Discrete Approach . . . . . . . . . . . 263
Jacek Wdowicki, Elżbieta Wdowicka, and Zdzisław Pawlak
25 Analysis of the Dynamic Response of Masonry Buildings
with Irregularities of Localization of Bearing Elements
Due to Mining Shocks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275
Tadeusz Tatara and Filip Pachla
26 Numerical and Experimental Prediction Methods
for Assessment of Induced Vibrations in Irregular Buildings . . . . . 289
Jan Benčat
27 Stability Analysis of Żelazny Most Tailings Dam Loaded
by Mining-Induced Earthquakes . . . . . . . . . . . . . . . . . . . . . . . . . . . 303
Waldemar Świdziński, Aleksandra Korzec, and Kinga Woźniczko
x Contents
Part III Seismic Control and Monitoring of Irregular Structures

28 Optimal Drift and Acceleration Control of 3D Irregular
Buildings by Means of Multiple Tuned Mass Dampers . . . . . . . . . . 315
Yael Daniel and Oren Lavan
29 Improved Seismic Performance of RCC Building Irregular
in Plan with Water Tank as Passive TMD . . . . . . . . . . . . . . . . . . . 323
Suraj N. Khante and Rutuja S. Meshram
30 Behaviour of Asymmetric Structure with Base Isolation
Made of Polymeric Bearings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 333
Tomasz Falborski and Robert Jankowski
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 343
Contributors
Asimina M. Athanatopoulou Department of Civil Engineering, Aristotle

University, Thessaloniki, Macedonia, Greece
Francesca Barbagallo Department of Civil Engineering and Architecture,
University of Catania, Catania, Italy
André Belejo Oregon State University, Corvallis, OR, USA
Jan Benčat University Research Centre, University of Žilina, Žilina, Slovakia
Rita Bento ICIST, Instituto Superior Técnico, Universidade de Lisboa, Lisbon,
Portugal
Piotr Bobra Faculty of Civil Engineering, Opole University of Technology,
Opole, Poland
Melina Bosco Department of Civil Engineering and Architecture, University of
Catania, Catania, Italy
Yael Daniel AJS Consulting Engineers, Haifa, Israel
Mario De Stefano Department of Architecture (DiDA), University of Florence,
Florence, Italy
Peter Fajfar Faculty of Civil and Geodetic Engineering, University of Ljubljana,
Ljubljana, Slovenia
Tomasz Falborski Faculty of Civil and Environmental Engineering, Gdańsk
University of Technology, Gdańsk, Poland
Giovanna A.F. Ferrara Department of Civil Engineering and Architecture,
Kenji Fujii Department of Architecture and Civil Engineering, Chiba Institute of
Technology, Narashino, Chiba, Japan
xi
xii Contributors
Giada Gasparini Department DICAM, University of Bologna, Bologna, Italy

George K. Georgoussis Department of Civil Engineering Educators, School of
Pedagogical and Technological Education (ASPETE), Attica, Greece
M.R. Ghayamghamian International Institute of Earthquake Engineering and
Seismology, Tehran, Iran
Aurelio Ghersi Department of Civil Engineering and Architecture, University of
Catania, Catania, Italy
M. Hashemifard International Institute of Earthquake Engineering and Seismology,
Tehran, Iran
Robert Jankowski Faculty of Civil and Environmental Engineering, Gdańsk
Leszek R. Jaroszewicz Institute of Applied Physics, Military University of
Technology, Warsaw, Poland
Erol Kalkan Earthquake Science Center, United States Geological Survey, Menlo
Park, CA, USA
Suraj N. Khante Applied Mechanics Department, Government College of
Engineering, Amravati, Maharashtra, India
Dietlinde K€ober Reinforced Concrete Department, Technical University of
Bucharest, Bucharest, Romania
Seweryn Kokot Faculty of Civil Engineering, Opole University of Technology,
Opole, Poland
Aleksandra Korzec Institute of Hydro-Engineering Polish Academy of Sciences,
Gdańsk-Oliwa, Poland
Jerzy K. Kowalski m-Soft Ltd., Warsaw, Poland
Zbigniew Krajewski Institute of Applied Physics, Military University of
Technology, Warsaw, Poland
Anna Kurzych Institute of Applied Physics, Military University of Technology,
Warsaw, Poland
Alessandra La Brusco Department of Architecture (DiDA), University of
Florence, Florence, Italy
Oren Lavan Faculty of Civil and Environmental Engineering, Technion – Israel
Institute of Technology, Haifa, Israel
Grigorios E. Manoukas Department of Civil Engineering, Aristotle University,
Thessaloniki, Macedonia, Greece
Contributors xiii
Valentine Mariani Department of Architecture (DiDA), University of Florence,

Florence, Italy
Edoardo M. Marino Department of Civil Engineering and Architecture, University
of Catania, Catania, Italy
Rutuja S. Meshram Applied Mechanics Department, Government College of
Engineering, Amravati, Maharashtra, India
G.R. Nouri Department of Civil Engineering, Faculty of Engineering, Kharazmi
University, Tehran, Iran
Filip Pachla Institute of Structural Mechanics, Cracow University of Technology,
Cracow, Poland
Michele Palermo Department DICAM, University of Bologna, Bologna, Italy
Zdzisław Pawlak Institute of Structural Engineering, Poznań University of
Technology, Poznań, Poland
Iztok Peruš Faculty of Civil and Geodetic Engineering, University of Ljubljana,
Ljubljana, Slovenia
Juan C. Reyes Department of Civil and Environmental Engineering, Universidad
de los Andes, Bogota, Colombia
Andrea C. Ria~ no Department of Civil and Environmental Engineering,
Universidad de los Andes, Bogota, Colombia
Andrea Rossi Rose School, EUCENTRE, Pavia, Italy
Pier Paolo Rossi Department of Civil Engineering and Architecture, University
of Catania, Catania, Italy
Stefano Silvestri Department DICAM, University of Bologna, Bologna, Italy
Klemen Sinkovič Faculty of Civil and Geodetic Engineering, University of
Ljubljana, Ljubljana, Slovenia
Barbara Sołtysik Faculty of Civil and Environmental Engineering, Gdańsk
Andrea Spagnoli Department of Civil, Environmental, Land Management,
Engineering and Architecture, Parma University, Parma, Italy
Paola R. Stramondo Department of Civil Engineering and Architecture,
Waldemar Świdziński Institute of Hydro-Engineering Polish Academy of Sciences,
Gdańsk-Oliwa, Poland
xiv Contributors
Marco Tanganelli Department of Architecture (DiDA), University of Florence,

Florence, Italy
Tadeusz Tatara Institute of Structural Mechanics, Cracow University of
Technology, Cracow, Poland
Krzysztof P. Teisseyre Institute of Geophysics, Polish Academy of Sciences,
Warsaw, Poland
Roman Teisseyre Institute of Geophysics, Polish Academy of Sciences, Warsaw,
Poland
Amfilohios Throumoulopoulos Department of Civil Engineering, Aristotle
University, Thessaloniki, Macedonia, Greece
Tomaso Trombetti Department DICAM, University of Bologna, Bologna, Italy
Stefania Viti Department of Architecture (DiDA), University of Florence, Flor-
ence, Italy
Elżbieta Wdowicka Polytechnic Faculty, Higher Vocational State School
(PWSZ), Kalisz, Poland
Jacek Wdowicki Polytechnic Faculty, Higher Vocational State School (PWSZ),
Kalisz, Poland
Kinga Woźniczko DHV HYDROPROJEKT Sp. z o.o., Warszawa, Poland
Dan Zamfirescu Reinforced Concrete Department, Technical University of
Bucharest, Bucharest, Romania
Zbigniew Zembaty Faculty of Civil Engineering, Opole University of Technol-
ogy, Opole, Poland
Part I
Ground Motion, Rotations and Wave
Propagation Effects
Chapter 1
A General Procedure for Selecting
and Scaling Ground Motion Records
for Nonlinear Analysis of Asymmetric-Plan
Buildings
Juan C. Reyes, Erol Kalkan, and Andrea C. Ria~

no
Abstract In performance assessment and design verification of complex structural

systems including base-isolated buildings, high-rise structures and structures uti-
lizing advanced lateral force resisting components (e.g., viscous dampers),
nonlinear response history analysis (RHA) is now a common engineering tool to
estimate seismic demands. Today, majority of ground motion selection and scaling
methods are suitable for symmetric plan buildings with first-mode dominant
response. There is, therefore, a need for a robust method to select and scale records
for nonlinear RHAs of asymmetric-plan buildings with significant torsional
response. Presented here is a generalized ground motion selection and scaling
procedure called modal pushover-based (MPS) procedure. The proposed procedure
explicitly considers structural strength, determined from pushover curves, and
determines a scaling factor for each record to match a target value of roof displace-
ment. The accuracy and efficiency of the procedure is evaluated by using computer
models of symmetric- and asymmetric-plan buildings subjected to one or two
horizontal components of ground motions. Analyses for one component of ground
motions were conducted for five existing symmetric-plan buildings of 4, 6, 13, 19
and 52 stories; for two components of ground motion, 48 single-story systems and
10 multi-story buildings were analysed. Also examined here is the ASCE/SEI
7 scaling procedure for comparison purposes. This study clearly shows that the
MPS procedure provides much superior results in terms of accuracy [true estimates
J.C. Reyes (*) • A.C. Ria~

no
Department of Civil and Environmental Engineering, Universidad de los Andes,
Carrera 1 Este No. 19A-40, Bogota, Colombia
e-mail: [email protected]
E. Kalkan
Earthquake Science Center, United States Geological Survey, Menlo Park, CA, USA
© Springer International Publishing Switzerland 2016 3

Z. Zembaty, M. De Stefano (eds.), Seismic Behaviour and Design of Irregular and
Complex Civil Structures II, Geotechnical, Geological and Earthquake Engineering 40,
DOI 10.1007/978-3-319-14246-3_1
4 J.C. Reyes et al.
of expected median engineering demand parameters (EDPs)] and efficiency

(reduced record-to-record variability of EDPs) than the ASCE/SEI 7 scaling
procedure.
Keywords Scaling seismic records • Modal pushover based scaling • Nonlinear

analysis • Asymmetric-plan buildings • Selecting seismic records
1.1 Introduction
Performance-based procedures for evaluating existing buildings and proposed

designs of new buildings in the U.S. require response history analyses (RHAs) for
an ensemble of earthquake records to determine EDPs for validation of a targeted
performance criterion. Earthquake records selected for RHAs often need to be
scaled to a seismic hazard level considered.
Kalkan and Chopra (2009) developed the modal pushover-based scaling (MPS)
procedure for selecting and scaling earthquake ground motion records in a form
convenient for evaluating existing structures and proposed designs of new struc-
tures. This procedure explicitly considers structural strength, obtained from the
first-“mode” pushover curve, and determines a scaling factor for each record to
match a target value of the deformation of the first-“mode” inelastic SDF system.
The MPS procedure has been proven to be accurate and efficient for low-, medium-
and high-rise buildings with symmetric plan (Kalkan and Chopra 2010, 2011, 2012)
subjected to one component of ground motion. Recently, Reyes and Chopra (2011a,
b, 2012) extended the MPS procedure for one component (mentioned above) to two
horizontal components of ground motion.
Reyes and Quintero (2014) proposed a new version of the MPS procedure for
single-story asymmetric-plan buildings. Reyes et al. (2014) extended this procedure
to multi-story asymmetric-plan buildings. In this investigation, the developed
procedure is compared against the ASCE/SEI 7-10 (ASCE7 henceforth) ground
motion scaling procedure for 3D analysis. Based on results from nine multi-story
asymmetric-plan buildings with various plan shapes, it is shown that the MPS
procedure provides much superior results in terms of accuracy and efficiency than
the ASCE/SEI 7-10 ground motion scaling procedure.
1.2 Modal Pushover-Based Scaling (MPS) Procedure
The MPS procedure is implemented here in three phases: (1) computation of target
roof displacement and pushover analyses, (2) scaling phase, and (3) selection phase.
The step-by-step procedure is presented here in a general form (Reyes et al. 2014).
1 Procedure for Selecting and Scaling Records 5
1.2.1 Target Roof Displacement and Pushover Analyses
1. For a given site, define the target spectra Âx and Ây, in this study taken as the
median of the 5-percent damped pseudo-acceleration response spectra of two
components of the ground motions.
2. Compute the natural frequencies ωn (periods Tn) and modes ϕn of the first few
modes of linear-elastic vibration of the building. For each ground motion
component direction (x or y), identify the first, second and third modes as the
three modes with the largest effective modal mass.
3. Develop the base shear-roof displacement, V bn urn , relationship or pushover
curve by nonlinear static analysis of the building subjected to the nth-“mode”
invariant force distribution:
2 3
mϕxn
s∗ 4 mϕ yn 5
n ¼
Io ϕθn
where m is a diagonal matrix of order N with mjj ¼ m j , the mass lumped at the
jth floor level; Io is a diagonal matrix of order N with I ojj ¼ I o j , the moment of
inertia of the jth floor diaphragm about a vertical axis through the center of mass
(C.M.); and subvectors ϕxn, ϕyn, and ϕθn of the nth mode ϕn represent x, y and θ
components of ground motion, respectively. This step should be implemented
only for the first three “modes” in the direction under consideration; this step
could be omitted for the higher-“modes” if they are treated as linear-elastic
(Chopra 2007).
4. Idealize the V bn urn pushover curve as a bilinear or trilinear curve, as
appropriate, and convert it into the force-deformation, ðFsn =Ln Þ Dn , relation-
ship for the nth-“mode” inelastic SDF system using well-known formulations
[10]:
Fsn V bn urn
¼ * Dn ¼
Ln Mn Γ n ϕrn
where Fsn is a nonlinear hysteretic function of the nth modal coordinate and Mn
is the effective modal mass for the nth-“mode” (Chopra 2007).
2 3 2 3 2 3
m 0 0 1 0
Ln ϕ T Mι
Γn ¼ ¼ Tn M¼4 0 m 0 5 ιx ¼ 4 0 5 ιy ¼ 4 1 5
Mn ϕn Mϕn
0 0 Io 0 0
1 and 0 are vectors of dimension N with all elements equal to one and zero,
respectively; and ϕrn is the value of ϕn at the roof.
5. Establish the target roof displacement u^r. For a system with known Tn, damping
ratio ξn, and force-deformation curve (Step 3), determine the peak deformation
6 J.C. Reyes et al.
Dn for the nth-“mode” inelastic SDF system due to each of the unscaled ground
motions u€g(t) by solving:
€ n ðtÞ þ 2ξn ωn D_ n ðtÞ þ Fsn ¼ €ug ðtÞ ! Dn

D
Ln
Determine D ^ n as the median of the Dn values. Calculate roof displacement in

the direction under consideration of the nth-“mode” as u^ rn ¼ Γ n ϕrn D ^ n , and
compute the roof displacement in the direction under consideration u^r from
values of u^rn using a suitable modal combination method (e.g., complete qua-
dratic combination). In practical applications, target deformation D ^ n can be
computed as D ^ n ¼ CRn D ^ no , where CRn is the inelastic deformation ratio, esti-
mated from empirical equations (Chopra and Chintanapakdee 2004), D ^ no ¼
2^
ðT n =2π Þ A n and Ân is the target pseudo-spectral acceleration at period Tn.
1.2.2 Scaling Phase
6. Compute the scale factor SF for each record in the direction under consideration
by solving the following nonlinear equation: ur u^ r ¼ 0, where ur is the peak
roof displacement in the direction under consideration from the scaled records.
Because this equation is nonlinear, SF cannot be determined a priori, but
requires an iterative procedure as shown below:
(a) Select an initial value of the scale factor SF, and compute deformation Dn(t)
for the nth-“mode” inelastic SDF due to the scaled record by solving:
€ n ðtÞ þ 2ξn ωn D_ n ðtÞ þ Fsn =Ln ¼ SF €ug ðtÞ ! Dn ðtÞ
D
(b) Compute roof displacement of the nth-“mode” in the direction under
consideration: urn ðtÞ ¼ Γ n ϕrn Dn ðtÞ
(c) Compute roof displacement in the direction under consideration:
!
X

ur ¼ max urn ðtÞ
n
(d) Estimate error: ε ¼ ur u^ r

(e) Adjust the value of the scale factor SF, and repeat steps “a” to “d” until ε is
less than a tolerance value.
In this study, step 6 was implemented by a numerical algorithm. By develop-
ing steps “a” to “e”, separately for the x and y components of the record, scale
factors SFx and SFy are determined. Note that pushover curves (step 4), and
target roof displacement (step 5) will be different for the two horizontal com-
ponents of the ground motion.
1.2.3 Selection Phase
7. Select the first k records with the lower values of:
X
6
Error ¼ SFx Ax ðT i Þ A^ x ðT i Þ þ SF y A y ðT i Þ A^ y ðT i Þ
i¼4
where Âx and Ây are vectors of spectral values Âi at different periods Ti (Ti¼ T4,
T5, T6); Ax and Ay are vectors of spectral values for the unscaled records over the
same periods.
1.3 Point of Comparison (Benchmark)
The benchmark or comparison point of an EDP is defined in this study as the

median value of EDPs obtained from nonlinear RHAs of the structure subjected to a
large set of unscaled records. The accuracy and efficiency of the MPS and ASCE7
procedures are examined by comparing their median EDP estimates from subset of
records against the benchmark values, and by comparing record-to-record variabil-
ity of the EDPs.
1.4 Study Cases for One Component RHAs
For a single horizontal component of ground motion, the MPS procedure scales
each record by a factor such that the deformation of the first-“mode” inelastic SDF
system matches a target value of the inelastic deformation. When steps 1 thru 6 of
the procedure shown in Sect. 1.2 are implemented for one-component of ground
motion, the following simplifications may be made: (1) Only the fundamental mode
in the direction of analysis is used in steps 3 thru 6; (2) Target deformation D^ 1 may
be used instead of a target roof displacement u^r; (3) In order to compute the scale
factor SF in step 6, the nonlinear equation may be written as D1 D ^ 1 ¼ 0; (4) The
selection phase (step 7) may consider only the unscaled spectral values at the
second mode in the direction of analysis instead of the spectral values at T4, T5
and T6. These simplifications may reduce significantly the computational time.
A large group of representative buildings in California were selected to study the
one-component MPS and the ASCE7 procedure. This group consists of three existing
low- and mid-rise steel special moment resisting frame (SRMF) buildings with 4, 6
and 13 stories, and two existing tall steel SRMF buildings having 19 and 52 stories. A
description of these structures and complete details of their analytical models are
reported in Kalkan and Chopra (2011, 2012). For these studies a total of 21 near-fault
8 J.C. Reyes et al.
19-story building 52-story building

Bench ASCE MPS 50 Bench ASCE MPS
15 40
Story
30
10
20
5
10
0 0
0 5 0 5 0 5 0 5 0 5 0 5
Story drift ratio Story drift ratio
Benchmark ASCE7 MPS
Fig. 1.1 Inter-story drift ratios in percentage for the 19- and 52-story buildings. In each case the
marker and the horizontal line represent the median value of the EDP σ, assuming a log-normal
distribution
strong ground motions were selected. These motions were recorded during seismic
events with moment magnitude MW 6.5 at fault distances RRUP 12 km. Selected
records are listed in Table 1 of Kalkan and Chopra (2011, 2012)
For those five buildings, the median values of EDPs attributable to sets of seven
ground motions scaled by the two methods—MPS and ASCE7—were computed by
nonlinear RHAs of the buildings and compared against the benchmark EDP values.
Representative results for two tall buildings are shown in Fig. 1.1. For the 19-story
building, the first three panels of this figure show the benchmark (abbreviated as
Bench) EDPs, and EDPs for the ASCE7 and MPS procedures, respectively. The
next three panels display similar results for the 52-story building. The markers and
horizontal lines represent the median EDP value one standard deviation σ
assuming a lognormal distribution. For comparison purposes, the median bench-
mark values are kept in all sub-plots as a dashed line. The ASCE7 scaling method
grossly overestimates the inter-story drift ratios (IDRs) at almost all floors; for
example, IDRs are overestimated by as much as 80 % for the 19-story building, and
almost 170 % for the 52-story building. In contrast, IDRs obtained from MPS differ
from the benchmark results by less than 10 % in most cases. Furthermore, the
dispersion in EDPs as a result of the seismic records scaled according to the ASCE7
procedure is much larger than those from the MPS procedure.
Similar results were observed for the other three buildings analysed (low- and
mid-rise buildings); these results are not shown due to space limitations. Additional
findings can be found in Kalkan and Chopra (2011, 2012).
1.5 Study Cases for Two Components RHAs
The accuracy and efficiency of the MPS procedure for multi-story symmetric-plan
buildings subjected to two components of ground motions was examined using a
computer model of an existing 9-story steel moment frame building (Reyes and
Chopra 2011b). For asymmetric-plan buildings, two categories of hypothetical
y y y
R05 R10 R15
x x x
y y y
L05 L10 L15
x x x
T05 T10 T15
y y y
x x x
corner c1 x-girder y-girder center of mass center of stiffness
Fig. 1.2 Plan views of the nine multi-story asymmetric-plan buildings
buildings were considered: single-story and multi-story buildings. The single-story

buildings modelled are 48 structures with fundamental vibration periods Tn equal to
0.2, 0.5, 1 and 2 s, and with yield strength reduction factors R equal to 2, 3, 5, and a
value that leads to linear elastic design. Their lateral resisting system consists of
buckling restrained braced frames with non-moment-resisting beam–column con-
nections. The structures considered in the multi-story buildings category are nine
steel SMRF buildings with 5, 10 and 15 stories. Their plan shapes are shown in
Fig. 1.2, where the moment resisting frames are highlighted. The buildings are
identified by the letters R, L and T followed by the number of stories: plan R is
approximately rectangular, plan T is symmetric about the x axis, and plan L is
un-symmetric about both x and y axis. Further details of their structural systems
including the fundamental periods, mode shapes, torsional irregularity factors etc.
can be found in Reyes and Quintero (2014) and Reyes et al. (2015).
For the single-story buildings, the 30 near-fault records selected were recorded
from nine shallow crustal earthquakes with moment magnitude MW ¼ 6.7 0.2
10 J.C. Reyes et al.
R=2 Linear 2
1
0
Normalized EDP
2
1
0
2
R=3
1
0
2
R=5
1
0
T =0.2 sec T =0.5 sec T =1 sec T =2 sec
1 1 1 1
Benchmark _ MPS-Rand _ MPS-Best_ ASCE7 -Rand _ ASCE7 -Best_
Fig. 1.3 Normalized roof displacement (EDP) at point c2 along x-direction of the L-shaped plan
single-story buildings (see Fig. 1.2). For each set the marker and the vertical line represent the
median value of the EDP one standard deviation, assuming a log-normal distribution
x-direction y-direction
15 Bench ASCE MPS Bench ASCE MPS
12 L15
Story
9
6
3
0
0 1 2 0 1 2 0 1 2 0 1 2 0 1 2 0 1 2
Story drifts Story drifts
Benchmark ASCE7 MPS
Fig. 1.4 Inter-story drift ratios in percentage along x- and y-direction at corner c1 of the L-shaped
plan 15-story building (see Fig. 1.2). In each case the marker and the horizontal line represent the
median value of the EDP σ, assuming a log-normal distribution
recorded at distances (RRUP) ranging from 0.1 to 15 km. The seismic scenario for
multi-story buildings consists of 30 far-field ground motion records populated from
seven shallow crustal earthquakes with moment magnitude MW ¼ 6.7 0.2 at dis-
tances ranging from 20 to 30 km. The selected seismic records are listed in Table 1
in Reyes and Quintero (2014) and Reyes et al. (2015).
The procedure developed in Sect. 1.2 is compared against the ASCE7 scaling
procedure. Representative results from single-story buildings and 15-story building
are shown in Figs. 1.3 and 1.4, respectively. These structures have L-shaped plan
with significant plan irregularity. Figure 1.3 includes three sets of seven records
randomly selected (called “MPS-Rand” and “ASCE7-Rand”) and one set of seven
records selected by implementing an improved selection procedure (called “MPS-
Best” and “ASCE7-Best”). Roof displacements (normalized by the corresponding
benchmark results) obtained from sets “MPS-Rand” (Fig. 1.3) are accurate, and
show a low “record-to-record” and “set-to-set” variability. Only displacements are
unsuccessfully estimated for short-period structures designed for high values of
yield strength reduction factors. A considerable improvement in accuracy and
“record-to-record” variability is obtained when sets “MPS-Best” (Fig. 1.3) are used
for estimating the roof displacements; in this case, the errors are not greater than
20 %. Roof displacements obtained from sets “ASCE7-Rand” (Fig. 1.3) are in
general less accurate and show a large “record-to-record” and “set-to-set” variabil-
ity; overestimation of roof displacements is as high as 40 %.
Figure 1.4 shows story drifts at a selected corner of the building (c1 in Fig. 1.2)
from records scaled and selected according to MPS and ASCE7 procedures together
with benchmark values. This figure confirms that for the multi-story case (L15
building), the records scaled according to the MPS procedure lead to more accurate
estimates of median values of EDPs than ASCE7 scaling procedure. The maximum
discrepancies encountered by scaling records according to the ASCE7 procedure
are reduced when these records are scaled by the MPS procedure; for example, the
error in story drifts decreases from 28 % to 8 %. Due to limited space, we only show
here a representative set of results; additional findings can be found in Reyes and
Quintero (2014) and Reyes et al. (2015).
1.6 Conclusions
This paper presents summary of a general procedure for selecting and scaling
ground motion records for nonlinear response history analyses of asymmetric-
plan buildings with significant plan irregularity. Based on results from multiple
study cases, it is clearly shown that the modal-pushover-based (MPS) procedure
provides much superior computation of EDPs in terms of accuracy and efficiency as
compared to the ASCE7 scaling method. This superiority is evident in two respects.
First, the ground motions scaled according to the extended MPS procedure provide
median values of EDPs that are much closer to the benchmark values than is
achieved by the ASCE7. Second, the dispersion (or record to-record variability)
in the EDPs due to seven scaled records around the median is much smaller when
records are scaled by the MPS procedure compared to the ASCE7 procedure.
References
Chopra AK (2007) Dynamics of structures: theory and applications to earthquake engineering, 4th
edn. Prentice-Hall, Englewood Cliffs, pp 403–445
Chopra AK, Chintanapakdee C (2004) Inelastic deformation ratios for design and evaluation of
structures: single-degree-of-freedom bilinear systems. ASCE J Struct Eng 130(9):1309–1319
Kalkan E, Chopra AK (2009) Modal pushover-based ground motion scaling procedure for
nonlinear response history analysis of structures. In: Proceedings of the Structural Engineers
Association of California Convention, San Diego, CA, 2009
Kalkan E, Chopra AK (2010) Practical guidelines to select and scale earthquake records for
nonlinear response history analysis of structures, U.S. Geological Survey Open-File Report
2010-1068. Available at http://pubs.usgs.gov/of/2010/1068/. Accessed 22 Jan 2015
12 J.C. Reyes et al.
Kalkan E, Chopra AK (2011) Modal-pushover-based ground-motion scaling procedure. J Struct

Eng 137(3):298–310
Kalkan E, Chopra AK (2012) Evaluation of modal pushover-based scaling of one component of
ground motion: tall buildings. Earthq Spectra 28(4):1469–1493
Reyes JC, Chopra AK (2011a) Three-dimensional modal pushover analysis of buildings subjected
to two components of ground motion, including its evaluation for tall buildings. Earthq Eng
Struct Dyn 40:789–806
Reyes JC, Chopra AK (2011b) Evaluation of three-dimensional modal pushover analysis for
asymmetric-plan buildings subjected to two components of ground motion. Earthq Eng Struct
Dyn 40:1475–1494
Reyes JC, Chopra AK (2012) Modal pushover-based scaling of two components of ground motion
records for nonlinear RHA of buildings. Earthq Spectra 28(3):1243–1267
Reyes JC, Quintero O (2014) Modal pushover-based scaling of earthquake records for nonlinear
analysis of single-story asymmetric-plan buildings. Earthq Eng Struct Dyn 43:1005–1021
Reyes JC, Ria~no AC, Kalkan E, Arango CM (2015) Extending modal pushover-based scaling
(MPS) procedure for nonlinear response history analysis of multi-story unsymmetric-plan
buildings. Eng Struct 88:125–137
Chapter 2
Prediction of the “Average” Peak Nonlinear
Seismic Response of Asymmetric Buildings
Under Bi-directional Ground Motion Acting
at an Arbitrary Angle of Incidence
Kenji Fujii
Abstract In this study the “reduced response spectrum” is introduced for the
prediction of the “average” response of all possible angle of incidence, and the
“average” peak seismic response of an asymmetric building under horizontal
bi-directional ground motion, acting at an arbitrary angle of incidence, is predicted
based on the procedure proposed in Fujii (Bull Earthq Eng 12:909–938, 2014). In
this study, the following two definitions of the “reduced response spectrum” are
examined; (A) the geometric mean spectrum of the horizontal major and minor
components, and (B) the square root of the average of squares spectrum from all
possible angle of incidence. In the numerical example, nonlinear time-history
analyses of a four-storey torsionally stiff asymmetric building are carried out
considering the various directions of seismic inputs, and these results are compared
with the predicted results. The results show that the predicted “average” peak
response displacement for the flexible-side frame satisfactorily agrees with the
average of the time-history analyses results, either definition of the “reduced
response spectrum”.
Keywords Asymmetric building • Bi-directional ground motion • Geometric mean

spectrum • Pushover analysis • Equivalent SDOF model
2.1 Introduction
In designing new buildings for earthquake resistance or when conducting seismic

evaluation of existing buildings, a horizontal ground motion is applied to each of
the main orthogonal axes of the buildings. However, for the seismic assessment of
K. Fujii (*)
Department of Architecture and Civil Engineering, Chiba Institute of Technology,
2-17-1 Tsudanuma, Narashino, Chiba 275-0016, Japan

DOI 10.1007/978-3-319-14246-3_2
14 K. Fujii
asymmetric buildings this procedure may be inadequate because the most

critical direction of incidence of the seismic input, which would produce the
largest response, may be different from the direction of the building’s main
orthogonal axes, and the major component of ground motion may act in any
direction.
For this purpose the author has proposed a simplified procedure to predict the
largest peak seismic response of an asymmetric building subjected to horizontal
bi-directional ground motion acting at an arbitrary angle of incidence
(Fujii 2014). In this procedure, the response spectra of the two horizontal ground
motion components are assumed to be identical. The response spectrum of the
horizontal minor component is assumed to be the same as that of the major
component. These assumptions may be valid for a conservative prediction
of the “largest” peak responses. Even though this idealization is simple, it may
be too conservative for use in the seismic assessment of structures. Besides,
as Beyer and Boomer has noted in their study (Beyer and Boomer 2007),
that the “average”, or the median, structural response of all possible angle of
incidence should be considered from the aspect of the probabilistic seismic hazard
assessment (PSHA) lather than the “largest”, or the maximum, structural
response.
One of the very simple approaches to predict the “average” peak response is
using the response spectrum, which is somehow reduced considering the variation
of response spectrum of all orientation, and apply the procedure proposed in
previous study (Fujii 2014). However how to define the “reduced response
spectrum” for the better prediction of the “average” responses remains problem.
One of the candidate of the reduced response spectrum is the geometric mean
spectrum of orthogonal components (Beyer and Boomer 2006, 2007), because it is
commonly used as horizontal component definition in a ground motion prediction
equation. The another candidate of the reduced response spectrum is, from the
aspect of structural response, the square root of the average of squares spectrum
from all possible angle of incidence, because the sum of squares of velocity
spectrum of two orthogonal components is strongly related to the seismic input
energy of structure.
In this study the “reduced response spectrum” is introduced for the prediction of
the “average” response of all possible angle of incidence, and then the “average”
peak seismic response of an asymmetric building under horizontal bi-directional
ground motion, acting at an arbitrary angle of incidence, is predicted based on the
procedure proposed in Fujii (2014). In this study, two definitions of the “reduced
response spectrum” are examined.
In the numerical example, nonlinear time-history analyses of a four-storey
torsionally stiff (TS) asymmetric building are carried out considering various
directions of seismic inputs, and these results are compared with the predicted
results.
2 Prediction of the average peak response 15
2.2 Description of the Procedure
2.2.1 Definition of the “Reduced Response Spectrum”
In this study, the “reduced response spectrum” is introduced for the prediction of
the “average” peak response. The definition of the “reduced response spectrum” is
illustrated in Fig. 2.1.
A set of orthogonal axes U-V in the X-Y plane is considered, with the U-axis
being the principal axis of the first modal response (Fujii 2010, 2011). The asym-
metric buildings are N-storey buildings, with 3N degrees of freedom (3N-DOFs)
oriented for the multi-storey model presented here. Another set of orthogonal axes
ξ-ζ in the X-Y plane is considered, with the ξ and ζ axes being the axes of the
horizontal major and minor components of the seismic input, respectively.
It is assumed that the pseudo acceleration spectra of the ξ and ζ components,
p Aξ(T ) and pSAζ (T ), respectively, are similar (Fig. 2.1b).
S
The ratio of pSAζ(T ) to pSAξ(T ), γ, is constant, as shown in Eq. 2.1.
γ ¼ p SAζ ðT Þ=p SAξ ðT Þ ¼ const: ð2:1Þ
The thick dotted circles in Fig. 2.1a show the simplified seismic inputs of each
direction for the prediction of the “average” peak response, while the thin dotted
circle in this figure shows the simplified seismic input for the prediction of the
“largest” peak response. Note that in the previous study (Fujii 2014), the spectrum
of the major component pSAξ(T ) is used to predict the “largest” peak response. In
this study, pSAξ(T ) is replaced by the reduced response spectrum p SAU ðT Þ as shown
in Fig. 2.1b.
In this study, the following two definitions of the “reduced response spectrum”
are examined; (A) the geometric mean spectrum of the horizontal major and minor
components, and (B) the square root of the average of squares spectrum from all
possible angle of incidence.
Fig. 2.1 Definition of the “reduced response spectrum” for the prediction of the “average” peak
response. (a) Intensity of seismic input to each direction. (b) Reduced response spectrum
16 K. Fujii
If the reduced response spectrum p SAU ðT Þ is defined as the geometric mean

spectrum of the horizontal major and minor components, p SAU ðT Þ is calculated
from Eq. 2.2.
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
p SAU ðTÞ ¼ p SAξ ðTÞ p SAζ ðTÞ: ð2:2Þ
Substituting Eq. 2.1 into Eq. 2.2, the ratio of reduced response spectrum to major
component, γ 0 , is obtained as in Eq. 2.3.
0 pffiffiffi
γ ¼ p SAU ðT Þ=p SAξ ðT Þ ¼ γ: ð2:3Þ
Note that in general, the geometric mean spectrum of orthogonal components

depends on the angle of incidence considered, as discussed in Boore et al. (2006)
and Beyer and Boomer (2007). However in this study, the geometric mean spec-
trum of the horizontal major and minor components, defined as Eq. 2.2, is used as
the one definition of the reduced response spectrum for the simplicity.
Another definition of the reduced response spectrum considered in this paper is
expressed as Eq. 2.4.
vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
u ðπ
u1 n o2
u
p SAU ðT Þ ¼ t p SAU ðT; Δψ Þ dΔψ : ð2:4Þ
π
0
n o2 n o2 n o2
p SAU ðT; Δψ Þ ¼ p SAξ ðT Þ cos 2 ðΔψ Þ þ p SAζ ðT Þ sin 2 ðΔψ Þ
2 n o2 ð2:5Þ
¼ cos 2 ðΔψ Þ þ γ 2 sin 2 ðΔψ Þ p SAξ ðT Þ :
In Eq. 2.4, pSAU(T, Δψ) is the pseudo acceleration spectrum of the U-component,
and Δψ is the angle between the U and ξ axes as shown in Fig. 2.1.
Substituting Eq. 2.5 into Eq. 2.4, the ratio γ0 is obtained as in Eq. 2.6.
0 pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
γ ¼ ð1 þ γ 2 Þ=2: ð2:6Þ
Note that from the comparisons of Eqs. 2.3 and 2.6, the reduced spectrum defined as
(B) the square root of the average of squares spectrum from all possible angle of
incidence (Eq. 2.6) is larger than that defined as (A) the geometric mean spectrum
of the horizontal major and minor components (Eq. 2.3). Therefore, the prediction
of the “average” response by using Eq. 2.6 will result more conservative than that
by using Eq. 2.3.
2.2.2 Outline of Simplified Procedure
An outline of the proposed procedure to predict the “average” peak response is

summarized in Fig. 2.2. The fundamental assumptions of the proposed procedure
are as follows:
1. The spectra of the two horizontal ground motion components are assumed to be
identical to the reduced response spectrum p SAU ðT Þ;
2. The building oscillates predominantly in a single mode in each set of orthogonal
directions;
3. The principal directions of the first and second modal responses are almost
orthogonal.
Note that these assumptions are the same as the procedure proposed previously
in Fujii (2014) except for the use of the reduced response spectrum. Therefore, the
details of the procedure are the same as shown in previous study (Fujii 2014),
except for the seismic input.
2.3 Numerical Examples
2.3.1 Building Model
The building investigated in this study was a four-storey asymmetric building

model as shown in Fig. 2.3. This building model, with bi-directional eccentricity,
is the Model-B1 in previous study (Fujii 2014). Each frame structure was designed
according to the weak-beam strong-column concept. The longitudinal reinforce-
ments of the concrete sections were determined so that the potential hinges were
located at all the beam-ends and bottoms of the columns, and the structural wall in
the first storey. Sufficient shear reinforcement was assumed to be provided to
prevent premature shear failure. The base shear coefficients obtained from the
planar pushover analysis in both the X and Y-directions, which were the values
when the roof displacement reached 1 % of the total height HN, were 0.522 and
0.455, respectively. Further details of this model can be found in previous study
(Fujii 2014).
Figure 2.4 shows the natural modes of the building model in the elastic range.
Here, Tke is the kth natural period in the elastic range, ψ ke is the angle of incidence
of the principal direction of the kth modal response in the elastic range with its
tangent given by Eq. 2.7, and Rρke is the torsional index of the kth mode (Fujii 2014)
in the elastic range, as defined by Eq. 2.8.
18 K. Fujii
Fig. 2.2 Outline of the proposed procedure for the prediction of the “average” peak response
Fig. 2.3 Four-storey building model considered for the numerical examples. (a) Plan. (b)
Elevation (frame Y1). (c) Elevation (frame X1)
Fig. 2.4 Shape of the first three natural modes of the building model in the elastic range
X X
tan ψ ke ¼ m j ϕYjke = m j ϕXjke ; ð2:7Þ
j j
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
X X
Rρke ¼ I j ϕΘjke 2 = m j ϕXjke 2 þ ϕYjke 2 ; ð2:8Þ
j j
φke ¼ f ϕX1ke ϕXNke ϕY1ke ϕYNke ϕΘ1ke ϕΘNke gT : ð2:9Þ
In Eqs. 2.7, and 2.8, mj and Ij are the mass and mass moment of inertia of the jth
floor, respectively. As shown in this figure, the principal directions of the first and
second modal responses of the building model are almost orthogonal; the angle
between the principal directions of the first two modes is 90.6 . This figure also
shows that the first mode is predominantly translational (Rρ1e < 1) and the second
mode is almost purely translational (Rρ2e << 1), while the third mode is predom-
inantly torsional (Rρ3e > 1). Because the first and second modes are predominantly
translational (Rρ1e , Rρ2e < 1) in the building models, this is classified as a torsionally
stiff (TS) system in this study, as is discussed in previous study (Fujii 2014).
20 K. Fujii
Fig. 2.5 Elastic acceleration response spectra for simulated ground motion. (a) “Major” compo-
nent. (b) “Minor” component
2.3.2 Ground Motion
In this study, the seismic excitation was considered to be bi-directional in the X–Y
plane, and three sets of artificial ground motions were generated. Note that these
three sets of artificial ground motions were the same as those used in the previous
study (Fujii 2014).
The elastic response spectra of the artificial ground motions with 5 % critical
damping are shown in Fig. 2.5. In this study, the “minor” component was scaled by
the multiplying parameter γ ¼ 0.0 (unidirectional excitation), 0.5, 0.7 and 1.0, to
investigate the influence of the ratio of the spectra for the horizontal minor to major
components. According to Lopez et al. (2006), this ratio varies between 0.63 and
0.81, and the average of this ratio is 0.70. Therefore, the case γ ¼ 0.7 may be
considered as the most realistic case in real ground motion.
In this study, nonlinear time-history analyses were carried out for various values
of the parameter γ and various directions of incidence of the seismic input. ψ, the
angle of incidence of the “major” component with respect to the X-axis, varied at
15 intervals from (ψ 1 90) to (ψ 1 + 90) , where ψ 1 (¼59.7 ) was the angle of
incidence of the U-axis corresponding to the predicted “largest” peak equivalent
displacement of the first modal response D1U*max shown in the previous study (Fujii
2014).
2.3.3 Comparisons of the Predicted and Time History

Analyses Results
Figure 2.6 compares the predicted “average” peak roof displacement with the
results of time-history analyses, and Fig. 2.7 also compares the predicted “average”
peak storey drift of the flexible-edge frames (frames Y4 and X6) with the time
Fig. 2.6 Comparisons of the peak roof displacement at each frame. (a) γ ¼ 0.0 (unidirectional
excitation). (b) γ ¼ 0.5, (c) γ ¼ 0.7, (d) γ ¼ 1.0
Fig. 2.7 Comparisons of the peak drift at the flexible-edge frame. (a) γ ¼ 0.0. (b) γ ¼ 0.5, (c)
γ ¼ 0.7, (d) γ ¼ 1.0
22 K. Fujii
history analyses results. The predicted results by using the geometric mean
spectrum (predicted (A)), and the square root of the average of squares spectrum
(predicted (B)) are also shown. Note that in case of γ ¼ 0.0 (unidirectional excita-
tion), the predicted results from the geometric mean spectrum (predicted (A)) are
not shown in these figures, because the predicted results show zero responses.
As shown in these figures, the predicted peak responses by using both spectra
(predicted (A) and (B)) agree well with the average of time history analyses results
for γ ¼ 0.5, 0.7 and 1.0; the differences of predicted responses between (A) and
(B) are negligibly small when γ ¼ 0.7, and in case of γ ¼ 0.5, the plots of “predict
(B)” results show approximately 10 % larger (more conservative) responses than
“predict (A)” results. While in case of γ ¼ 0.0 (unidirectional excitation), the plots
of “predict (B)” results are close to the average of time history analyses results.
Therefore, in case of the spectral ratio γ is larger than 0.5, either the geometric mean
spectrum or the square root of the average of squares spectrum may be used as the
“reduced response spectrum”, for the prediction of the “average” peak response of
all possible angle of incidence.
2.4 Conclusions
In this study the reduced response spectrum is introduced for the prediction of the
“average” response of all possible angle of incidence, and then the “average” peak
seismic response of an asymmetric building under horizontal bi-directional ground
motion, acting at an arbitrary angle of incidence, is predicted based on the proce-
dure proposed in Fujii (2014). In this study, two definitions of the “reduced
response spectrum” are examined; one is the geometric mean spectrum of the
horizontal major and minor components, and the other is the square root of the
average of squares spectrum from all possible angle of incidence.
In the numerical example, nonlinear time-history analyses of a four-storey TS
asymmetric building were carried out considering various directions for the seismic
inputs. The results showed that the predicted “average” peak response displacement
for the flexible-side frame satisfactorily agreed with the average of the time-history
analyses results, either definition of the “reduced response spectrum”, when the
spectral ratio of the minor to the major components is larger than 0.5.
For the prediction of the “average” peak response under “real” horizontal
bi-directional ground motion whose spectral shape of both components are differ-
ent, the author think GMRot50, the median of geometric mean spectrum of all
possible angle of incidence, proposed by Boore et al. (2006), may be used as the
“reduced spectrum”. Further investigation is needed for the applicability of this
procedure under “real” horizontal bi-directional ground motion.
Erratum to “Prediction of the largest peak nonlinear seismic

response of asymmetric buildings under bi-directional
excitation using pushover analyses”
Unfortunately, there were some errors in the previous study (Fujii 2014), which is
the key paper for this article. The correct versions of Eqs. 17 and 19 are shown
below.
Γ 2Vie φ2ie T Mn d Γ 2Vie φ2ie T n f R

n D2V
*
¼ *
, n A2V * ¼ ; ð17Þ
M2Vie M2Vie *

PU ¼ MΓ 1Uie φ1ie A1U * max 0:5Γ 2Vie φ2ie A2V * max
: ð19Þ
PV ¼ M 0:5Γ 1Uie φ1ie A1U * max þ Γ 2Vie φ2ie A2V * max
References
Beyer K, Boomer JJ (2006) Relationships between median values and between aleatory variabil-
ities for different definitions of the horizontal component of motion. Bull Seismol Soc Am
96(4A):1512–1522
Beyer K, Boomer JJ (2007) Selection and scaling of real accelerograms for bi-directional loading:
a review of current practice and code provisions. J Earthq Eng 11:13–45
Boore DM, Watson-Lamprey J, Abrahamson NA (2006) Orientation-independent measures of
ground motion. Bull Seismol Soc Am 96(4A):1502–1511
Fujii K. (2010) Seismic assessment of asymmetric buildings considering the critical direction of
seismic input. Paper presented at the 14th European conference on earthquake engineering,
paper no. 0623, Ohrid, Macedonia, 30 Aug–3 Sep 2010
Fujii K (2011) Nonlinear static procedure for multi-story asymmetric frame buildings considering
bidirectional excitation. J Earthq Eng 15:245–273
Fujii K (2014) Prediction of the largest peak nonlinear seismic response of asymmetric buildings
under bi-directional excitation using pushover analyses. Bull Earthq Eng 12:909–938
L
opez A, Hernández JJ, Bonilla R, Fernández A (2006) Response spectra for multicomponent
structural analysis. Earthq Spectra 22:85–113
Chapter 3
Evaluation of Torsional Component
of Ground Motion by Different Methods
Using Dense Array Data
G.R. Nouri, M.R. Ghayamghamian, and M. Hashemifard
Abstract In addition to the structural torsional effects coming from the

horizontal asymmetry of the building structure one can also meet with the direct
torsional ground excitations of the building. The dynamic analyses of structures
have been carried out by neglecting the excitation by the rotational ground
motions in engineering practice largely caused by the lack of recorded torsional
ground motions. Using dense array data is one of the approaches has been
developed to produce torsional component of ground motion. In this paper, by
using the data of Chiba dense array and applying three methods of Time Deriva-
tion, Finite Difference and Geodetic, torsional motion are estimated and com-
pared. In this array, accelerometers are placed densely with separation distances
between 5 and 300 m. The results showed that the peak torsional ground motion
which was computed by Time Derivation method is larger than which was
computed by Geodetic method.
Peak torsional ground motion values that are estimated by Finite Difference
method show smaller values than those computed by Time Derivation for
long separation distances (>20 m). However, the values estimated from the Finite
Difference and Time Derivation methods are relatively close for short separation
distances. Also, the effects of peak ground acceleration and magnitude of
events on the torsional motions have been investigated and the relationship
between PGA and the maximum torsional motion for different methods has
been proposed.
Keywords Torsional motion • Seismic dense array • Geodetic method • Time

derivation method • Finite difference method
G.R. Nouri (*)

Department of Civil Engineering, Faculty of Engineering, Kharazmi University, Tehran, Iran
M.R. Ghayamghamian • M. Hashemifard
International Institute of Earthquake Engineering and Seismology, Tehran, Iran
e-mail: [email protected]; [email protected]

DOI 10.1007/978-3-319-14246-3_3
26 G.R. Nouri et al.
3.1 Introduction
Strong ground motion consists of three components of translation along x, y and z

axes and three components of rotation about these axes and also six components of
strain. In general, three main approaches have been developed to extract rotational
motions: one is a numerical simulation of the field of radiation from the source
mechanism (Aki 1986; Haskell 1969). The second approach is measuring rotational
components indirectly by translational components. In preliminary studies,
researchers tried to propose a relation between translational and rotational compo-
nents using one seismic station (Newmark 1969; Ghafory-Ashtiany and Singh
1986). Also, indirect measurements of rotational motions using a seismo-meter
array have been studied by several investigators (Niazi 1986; Oliveira and Bolt
1989; Spudich et al. 1995; Bodin et al. 1997; Li et al. 2004; Huang 2003;
Ghayamghamian and Nouri 2007).
The third approach is measuring rotational components directly by rotational
sensors. Ring laser gyroscopes have become one of the most important instruments
for measuring rotation (Nigbor 1994; Takeo 1998). Suryanto et al. (2006) present
for the first time a comparison of array-derived rotations with direct measurements
of rotations. Their research showed that the overall fit between direct and array-
derived measurements is surprisingly good.
Ghayamghamian and Nouri (2007) studied the rotational ground motions and
their dependence on seismic parameters, using Chiba dense array data, and a
possibility of estimating torsional ground motion from translational records was
investigated. They computed torsional motion from the difference of two transla-
tional records on the ground. Also, the effect of torsional ground motion on
structural response was studied by Ghayamghamian et al. (2009). The torsional
motion was estimated by a geodetic method using data of Chiba dense array. The
results showed that the increase in the displacements of symmetric or asymmetric
buildings due to torsional excitation of the ground is largest for structures with very
short translational periods (less than about 0.3 s) and small frequency ratios (Ω< 1).
For such structures, the accidental eccentricities ea resulting from torsional ground
motions were found to be larger than those proposed by the design codes.
In this paper, data are collected from the Chiba dense array to estimate the
torsional ground motion. Due to closely spaced instruments in Chiba array and
regular arrangement of instruments at two inner rings, the better estimation of
torsional motion can be achieved using multiple stations up to high frequencies.
To compute torsional motion, three methods, namely, Time Derivation, Finite
Difference and Geodetic are deployed and the results compared to each other. Also,
variations of the maximum torsional motion with peak ground acceleration and
magnitude of earthquake are investigated in these three methods.
3 Evaluation of Torsional Component of Ground Motion by Different Methods. . . 27
3.2 Chiba Dense Array
A 3-D array system has been installed at Chiba, an experimental station of the
Institute of Industrial Science, University of Tokyo, in 1982 (Fig. 3.1). The topo-
graphical conditions of the site are generally simple, with the ground surface being
almost flat. In this array, seismometers and accelerometers are placed, with a
minimum separation distance of 5 m, both on the ground surface and in boreholes.
The array system is composed of 15 boreholes with 44 three-component acceler-
ometers, 9 are densely arranged. Stations C1–C4 and P1–P4 are respectively placed
on circles with radii of 5 and 15 m with respect to station C0, which is placed at the
center of these two rings (Katayama et al. 1990). Nine events recorded with high
signal-to-noise ratios (SNR), and wide ranges of magnitudes and PGAs were
selected (Ghayamghamian and Nouri 2007). Specifications of these events, are
given in Table 3.1.
3.3 Estimation of Torsional Motion
In this paper, three methods have been applied to estimate the torsional motion
using translational components which include Time Derivation, Finite Difference
and Geodetic methods.
Fig. 3.1 Chiba array configuration and reference system

Table 3.1 Specification of selected events

PGA
Event no. Focal depth (km) Distance (km) NS EW MJMA SNR at C0 (%)
33 73.3 104.5 52 60 6.5 98.6
37 57.9 44.7 400 293 6.7 99.8
42 47.6 37.9 117 79 5.2 98.7
46 55.3 47.7 57 71 5.6 98.5
47 55.7 55.2 32 34 6.0 98.1
81 96.0 42.2 71 86 6.0 98.7
82 69.0 62.4 38 51 5.3 97.2
84 50.0 40.2 91 121 5.4 98.8
87 92.0 52.4 91 94 5.9 99.0
3.3.1 Time Derivation Method
Ghafory-Ashtiany and Singh (1986) obtained following expression for computation

of ground rotational acceleration by using the Eq. 3.1.
1 d € € i ðtÞ

ψ k ðtÞ ¼ X j ðt Þ X ð3:1Þ
c jd t
Where, ψ k (t) is the rotation about the k-axis, Xi and Xj are displacements along
the x and y axes and xi ،xj and xk are principle axes. Also cj is shear wave velocity in
direction xj .
3.3.2 Finite Difference Method
The average torsional motions can be approximated from the difference of two
translational records in an array of stations.

1 u€2 ðtÞ €u1 ðtÞ €v2 ðtÞ €v1 ðtÞ
ψz ¼ ð3:2Þ
2 Δy Δx
where, ψ z (t) is torsional acceleration, u€j(t) and €v j ðtÞ are translational accelerations
along the x and y axes respectively related to a station pair (j ¼ 1,2).
This method has been used for computing torsional motion in various researches
(Huang 2003; Ghayamghamian and Nouri 2007; Hao 1996; Ghayamghamian and
Motosaka 2003). This estimation of torsional motion is the first order accuracy and
can be done especially when the stations are distributed regularly in the ideal cross
shaped array.
3.3.3 Geodetic Method
Spudich et al. (1995) introduced a geodetic method that can estimate torsional
motion using multiple stations with higher precision. In the framework of classical
elasticity, and further assuming infinitesimal deformations, displacement of a point
r is related to that of a neighboring point r+δr:
u ðr þ δ rÞ ¼ uðrÞ þ Gδr ¼ uðrÞ þ εδ r þ ω δ r ð3:3Þ
where G, ɛ, ω are the displacement-gradient matrix, strain and rotation respectively.

Using the displacement-gradient matrix G, one can compute strains and rotations.
The G matrix is given by
0 1
∂ x ux ∂ y ux ∂z ux
G ¼ @ ∂x u y ∂yuy ∂z u y A ð3:4Þ
∂ x uz ∂ y uz ∂ z uz
The relation between rotational and translational motions is obtained through the
application of the curl operator (∇) to the displacement by:
0 1 0 1
ωx ∂x uz ∂z u y
1 1
@ ω y A ¼ ∇ uð r Þ ¼ @ ∂ z ux ∂ x uz A ð3:5Þ
2 2
ωz ∂ x u y ∂ y ux
where ωx, ωy, ωz are rotations about x, y and z axes. At least three stations must be
used to determine the horizontal-displacement gradient using this method. This
method was applied to studying the dynamic deformations induced by the Landers
earthquake (M ¼ 7.4, 1992) and recorded by the Parkfield seismic array (UPSAR)
in California. In addition, Suryanto et al. (2006) used the same method to compare
array-derived torsional ground motion with direct-ring laser measurements and
found the two to be in good agreement.
3.4 Comparison Between Results of the Methods
By using the three above mentioned methods, the torsional ground acceleration was
calculated and compared with each other. For this purpose, maximum torsional
motion and standard deviation values have been used. Torsional motion for one
station, pairs of stations with different separation distances and two rings of
accelerometers (C0, C1–C4) and (C0, P1–P4) are calculated by Time Derivation,
Finite Difference and Geodetic methods respectively. Figure 3.2 shows comparison
of standard derivation and the maximum values of torsional motion obtained from
that three methods for selected events. Comparison between the results of various
Fig. 3.2 Comparison of maximum torsional motion and standard deviation in the different
methods for selected events
events demonstrate that torsional component values strongly depend on the earth-
quake characteristics.
Maximum torsional component and standard derivation values belong to the
event 37 that possesses the maximum value of peak ground acceleration
(PGA ¼ 400 gal) and magnitude of earthquake (MJMA ¼ 6.5). As illustrated in
Fig. 3.2, standard derivation and maximum torsional motion values are extremely
sensitive to the separation distance of stations and decrease as the distance increases
in the Finite Difference method. This procedure is independent of the peak ground
acceleration and magnitude of earthquake.
Values obtained from the Time Derivation method are larger than the ones from
Finite Difference method with separation distance larger than 20 m as shown in
Fig. 3.2. The ratio of values computed by these two methods are 6–8 at far
separation distances. However, for short separation distances they demonstrate
close values.
It is clear that the maximum values of torsional motion as estimated by the Time
Derivation method are larger than those computed by the Geodetic method. Also,
larger values of torsional motion for the Geodetic method can be observed com-
pared with those for Finite Difference method that the separation distances are far.
Array-derived torsional motions are subjected to an important limitation. The
instruments should be as close as possible for the calculated linear approximations
to be as close as possible to the true gradients. Bodin et al. (1997) showed that to obtain
array-gradient estimates accurate to within ~90 % of true gradients, the array dimen-
sions must be less than one quarter-wavelength of the dominant energy in the wave
train. Later, Langston (2007a, b) indicated that the accuracy order of finite difference
approximation depends also on the geometry of the array. He found that the station
spacing must be ~10 % of a horizontal wavelength to obtain 90 % accuracy, and these
finite difference estimates are in first- and second-order of accuracy for irregular and
regular arrays, respectively. Regarding estimated large-wave velocity (Yamazaki and
Turker 1992) and the very closely spaced instruments in the Chiba array, the torsional
motions can be accurately evaluated for the two closely spaced rings (C0, C1–C4) and
(C0, P1–P4) up to the high-frequency range (<11 Hz).
3.5 Effects of Peak Ground Acceleration and Magnitude

of Earthquake
It has been tried to establish a relationship between the translational and torsional
components due to existing problems in recording torsional motion. In this section
variations of the maximum torsional motion with peak ground acceleration are
investigated in the three methods. The mean estimated values of both inner and
outer rings in the Geodetic method are used. As shown in Fig. 3.3, with increasing
peak ground acceleration, maximum of torsional motion also increases. However,
these changes are small in events with low peak ground acceleration and maximum
torsional motion will increase significantly with increasing translational accelera-
tion. It is seen that, there is a linear correlation between maximum values of the two
components with an acceptable accuracy by examining the relationship between the
maximum translational and torsional components.
Magnitude of earthquake is also one of its characteristics. Chiba dense array data
covers a wide range of magnitudes and changes of the maximum torsional motion
with events magnitude in the three methods are shown in Fig. 3.3. The results
demonstrate that maximum torsional motion values changes partially until magni-
tude 6.5, however, after this value we can see a sudden increase. As illustrated,
Fig. 3.3 Variations of the maximum torsional motion with peak ground acceleration and magni-
tude of earthquakes
there is no determined relationship between magnitude of earthquake and maxi-

mum torsional motion.
3.6 Conclusions
Translational motions of seismic arrays were applied for estimating torsional

motion. In this study, torsional ground accelerations were estimated by three
methods, namely, Time Derivation, Finite Difference and Geodetic method using
data of Chiba dense array. Then, the maximum torsional motion and standard
deviation values of the three methods were compared. The results showed that
the peak torsional ground motion which was computed by Time Derivation method
is larger than those was computed by Geodetic method. Peak torsional ground
motion values that were estimated by Finite difference method show smaller values
than which were computed by Time derivation for long separation distances
(>20 m). However, the values estimated from the Finite Difference and Time
derivation methods are close to each other for short separation distances. The
results revealed that, there is a linear correlation between maximum values of
translational and torsional components for all three methods.
References
Aki K (1986) Seismic displacement near a fault. Geophys Res 73:5359–5376

Bodin P, Gomberg J, Sing SK, Santoyo M (1997) Dynamic deformations of shallow sediments in
the valley of Mexico, part I: three-dimensional strains and rotations recorded on a seismic
array. Bull Seismol Soc Am 87:528–539
Ghafory-Ashtiany M, Singh MP (1986) Structural response for six correlated earthquake compo-
nents. Earthq Eng Struct Dyn 14:103–119
Ghayamghamian MR, Motosaka M (2003) The effects of torsion and motion cou-pling in site
response estimation. J Earthq Eng Struct Dyn 32(5):691–709
Ghayamghamian MR, Nouri GR (2007) On the characteristics of ground motion rotational
components using Chiba dense array data. Earthq Eng Struct Dyn 36:1407–1427
Ghayamghamian M, Nouri G, Igel H, Tobita T (2009) The effect of torsional ground motion on
structural response: code recommendation for accidental eccentricity. Bull Seismol Soc Am
99:1261–1270
Hao H (1996) Characteristics of torsional ground motion. Earthq Eng Struct Dyn 25:599–610
Haskell NA (1969) Elastic displacements in the near-field of a propagating fault. Bull Seismol Soc
Am 59:865–908
Huang BS (2003) Ground rotational motions of the 1999 Chi-Chi, Taiwan earthquake as inferred
from dense array observations. Geophys Res Lett 30:1307–1310
Katayama T, Yamazaki F, Nagata S (1990) Development of strong motion database for the Chiba
seismometer array, s.l.: University of Tokyo.
Langston A (2007a) Spatial gradient analysis of linear seismic array. Bull Seismol Soc Am
97:268–280
Langston A (2007b) Wave gradiometry in two dimensions. Bull Seismol Soc Am 97:401–416
Li H, Sun L, Wang S (2004) Improved approach for obtaining rotational components of seismic
motion. Nucl Eng Des 232:131–137
Newmark NM (1969) Torsion in symmetrical buildings. In: Proceedings of the 4th World
Conference on Earthquake Engineering, Santiago, Chile
Niazi M (1986) Inferred displacements, velocities and rotations of a long rigid foundation located
at el centro differential array site during the 1979 imperial valley, California, earthquake.
Earthq Eng Struct Dyn 14:531–542
Nigbor RL (1994) Six-degree-of-freedom ground-motion measurement. Bull Seismol Soc Am
84:1665–1669
Oliveira CS, Bolt BA (1989) Rotational components of surface strong ground motion. Earthq Eng
Spudich P et al (1995) Transient stresses at Parkfield, California, produced by the M7.4 Landers
earthquake of June 28, 1992: observations from the UPSAR dense seismograph array.
J Geophys Res 100:675–690
Suryanto W et al (2006) First comparison of array-derived rotational ground motions with direct
ring laser measurements. Bull Seismol Soc Am 96:2059–2071
Takeo M (1998) Ground rotational motions recorded in near-source region of earth-quakes.
Geophys Res Lett 25:789–792
Yamazaki F, Turker T (1992) Spatial variation study on earthquake ground motion observed by the
Chiba array. In: Madrid, Proceeding of the Tenth World Conference on Earthquake
Engineering
Chapter 4
Estimation of Rotational Ground Motion
Effects on the Bell Tower of Parma Cathedral
Zbigniew Zembaty, Andrea Rossi, and Andrea Spagnoli
Abstract The purpose of this paper is to present results of calculations of seismic

response of the old Parma Bell Tower with and without the rotational, rocking
seismic effects as modeled in the European seismic code: Eurocode 8 (part 6:
Towers, Masts and Chimneys). A detailed, finite element model of the tower was
prepared, which led to the solution of its eigen problem and to conventional
dynamic, response modal analysis. Next a simplified rod model, equivalent to the
detailed finite element model was prepared, taking into account soil compliance as
well as the rocking excitation effects. The results of computations show rather
substantial, 30–40 % contribution of the rocking excitations in the overall seismic
response.
Keywords Response spectra • Rotational ground motion • Seismic rocking

excitations • Eurocode 8 • Tower shaped structures
4.1 Introduction
Formally one can define three rotational components of the surface, seismic ground
motion. These are two rotations ψ, θ about the two horizontal axes x, y and rotation
φ about vertical axis z (see Fig. 4.1).
Among them, particularly important can be rotation θ about the horizontal axis y,
perpendicular to the epicentral direction. For this direction there are numerous
Z. Zembaty (*)
Faculty of Civil Engineering, Opole University of Technology, ul. Proszkowska 76,
45-061 Opole, Poland
A. Rossi
Rose School, EUCENTRE, via Ferrata 1, 27100 Pavia, Italy
A. Spagnoli
Department of Civil, Environmental, Land Management, Engineering and Architecture,
Parma University, via Universita, 12, 43121 Parma, Italy

DOI 10.1007/978-3-319-14246-3_4
36 Z. Zembaty et al.
Fig. 4.1 Coordinate system

at the ground surface and
the rotational components
of seismic ground motion
models suggesting the seismic rocking excitations from the body and surface waves
propagation, general spatial seismic effects and including (or excluding) the local
soil conditions (see e.g. Falamarz-Sheikhabadi and Ghafory-Ashtiany 2012; Kalkan
and Grazer 2007; Trifunac 1982; Castellani and Zembaty 1996).
The first attempt of modeling the rotational effects in form of typical response
spectrum format can be found in Eurocode 8, part 6 devoted to chimneys, masts and
towers, 2005 (see also: Zembaty 2009). In this code the rotational response spectra
about Cartesian axes x, y, & z on the ground surface are modeled as functions of
shear wave velocity at the ground surface and the traditional Eurocode 8, transla-
tional design response spectra (2004). The response spectrum of the rocking
component θ about the horizontal axis y (Fig. 4.1) is given by formula A.1 of
Eurocode 8 part 6 (2005):
1:7πSe ðT Þ
R yθ ðT Þ ¼ ð4:1Þ
νS T
where T is the natural period, vS equals average shear wave velocity in the top 30 m
of the local soil, while Se(T ) stands for the elastic translational response spectrum.
In Eurocode 8 (2004), Se(T) it is given by formula 3.2.2.2:
8
>
> S 1þ
T
ð2, 5η 1Þ if 0 T TB
>
>
>
> TB
>
>
>
< S 2, 5 η
> if TB T TC
Se ðTÞ ¼ ag TC ð4:2Þ
>
> S 2, 5 η if TC T TD
>
> T
>
>
>
>
>
> TC TD
: S 2, 5 η if T > TD
T2
in which TB, TC, TD and S represent local site conditions as well as two types of
ground motions (far and near field) while ag is the horizontal design acceleration.
In Fig. 4.2 the plots of response spectra type 1 and type 2 are shown as functions
of natural period T. The type 1 response spectrum is similar to the response
spectrum of the actual Italian seismic code NTC (2008). The type 2 response
spectrum was included in present analysis for comparison.
4 Rotational Excitations of Parma Tower 37
Fig. 4.2 Plots of response spectra type 1 and 2 for 5 % damping ratio
Practical applications of formula 1 with Eq. 4.2 require first to solve the eigen
problem of the structure and next the application of the mode superposition
technique in respective implementation of the response spectrum methodology.
Usually first two to three modes are enough to reach satisfactory accuracy for
symmetric slender towers (e.g. circular chimneys) under horizontal seismic exci-
tations. The inclusion of up to four to five modes makes it possible to ensure that the
required convergence of the mode superposition method was achieved. This calcu-
lation procedure can be carried out using one of many programs implementing
Finite Element Method in structural dynamics. So far however none of the com-
mercial FEM programs allows to include rotational, excitation seismic effects,
in particular in format of an acceleration response spectrum. For this reason, in
this paper, slender tower is first modelled using SAP 2000 program to solve its
eigen problem and compute its response to horizontal seismic excitations using
translational response spectrum. Next an equivalent, simplified model of the tower
is formulated as a Timoshenko beam and calculated for the combined rotational and
horizontal excitations. The Timoshenko beam model allows to include appropriate
combination of shear and bending characteristic for the old masonry towers which
is the case of seismic vibrations of the Parma Cathedral Bell Tower.
4.2 Formulation of the Problem
Consider matrix equation of motion of a structure modeled as a plane, discrete

system under horizontal excitations u(t).
M€q þ Cq_ þ Kq ¼ M€uðtÞ ð4:3Þ
where matrix M is diagonal and contains discrete masses mj of the modelled

structure, K is the stiffness matrix, C is the damping matrix. Its solution can be
obtained using mode superposition method and Duhamel integral as follows:
Xn ðt
qj ¼ w η
i¼1 ji i
€uðτÞhi ðt τÞdτ ð4:4Þ
0
where hi stands for modal impulse response function
1 ωi t
hi ðtÞ ¼ e sin ðωid tÞ ð4:5Þ
ωid
ωi ¼ 2π=T i is the angular natural frequency of mode i-th, Ti natural period, wji j-th

element of the i-th eigen vector, ωid damped natural frequency ωid ¼ ωi √ 1 ξ2j
and ηi modal participation factor:
wTr
i M1
ηi ¼ ð4:6Þ
wTr
i Mwi
in which wi is modal vector i, bold symbol 1 denotes vector containing values of

1 and Tr stands for vector transposition.
Adding rotational excitations θ(t) about horizontal axis y (rocking) leads to
following form of Eq. 4.3 (see e.g. [5]):

M€q þ Cq_ þ Kq ¼ M€uðtÞ þ m j H j Θ€ ðtÞ ð4:7Þ
in which symbol {mjHj} represents a vector containing values of the multiplications

of discrete masses mj by their respective heights Hj above the ground surface.
It should be noted that the structure rests on a compliant soil described by
rotational stiffness kΘ (Fig. 4.2).
Applying the response spectrum method one can assess the displacements qj at
the height Hj for a given mode i, separately for horizontal excitations:

max qhor ffi w ji ηhor 1 Sa ðωi ; ξi Þ ð4:8Þ
ji i
ωi2
and for the rotational excitations

max q rot ffi w ji η rot 1 S€θ ðωi ; ξi Þ ð4:9Þ
ji i
ω 2 a
i
€
in which Sa and SΘa denote horizontal and rotational, acceleration response spectra
respectively, while
Fig. 4.3 A simplified

model of a slender tower
under horizontal-rotational
excitations
wTr
i M1
ηhor ¼ ð4:10aÞ
i
wTr
i Mwi
wTr
i MH
ηirot ¼ Tr
ð4:10bÞ
wi Mwi
represent modal participation factors calculated differently for the horizontal and
the rotational excitations and H is a vector containing the heights Hj of the discrete
masses mj above the ground (Fig. 4.3).
Applying, for the slender tower, with well separated natural frequencies, the
familiar SRSS rule to combine the responses in the succeeding modes, and assum-
ing hypothetically that the same rule also holds when combining the responses from
horizontal and rotational excitations, one can asses max response at Hj height
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Xn 2 X n 2
max q j ffi i¼1
max q hor
ji þ i¼1
max q rot
ji ð4:11Þ
The rocking effects on slender towers will be best observed in the base shear and
overturning moment. They can be calculated from the equivalent pseudo-static,
horizontal forces associated with each discrete mass mj:
P ji ¼ m j w ji ηi Sa ðωi ; ξi Þ ð4:12Þ
These pseudo-static forces are calculated in a different way for the horizontal and
rotational excitations. Thus symbol ηi in Eq. 4.12 is given either by Eq. 4.10a or
4.10b, while symbol Sa stands either for horizontal acceleration response spectrum
€
or for the rotational one SΘ
a.
Summing down all the pseudo-static forces gives the base shear force associated
with the natural mode i. Using basic formulas of the mode superposition method,
after some algebra, respective base shear calculated for mode i equals:
X 2
mk wki
Fbase½i ¼ Xk S a ð ωi ; ξ i Þ ð4:13Þ
m w2
k k ki
Summing down moments of the pseudo-static forces with respect to the ground
surface, gives the overturning moment of the mode i:
X
Mbase½i ¼ H P
j j ji
ð4:14Þ
The total base shear and total overturning moments due to horizontal and rotational
effects are calculated using the SRSS rule the same way as in Eq. 4.11.
4.3 Response of Parma Bell Tower to Horizontal-Rocking

Excitations
4.3.1 Description of Parma Bell Tower and Its Dynamic

Properties
The bell tower of Parma cathedral (Fig. 4.4) was built in the twelfth century.
It is 64.25 m high and has hollow square cross-section of 7.72 m side. The
thickness of the masonry wall decreases along the height, from 1.40 to 1.05 m
(Fig. 4.5). There are two series of openings: one small at mid-height and one large at
the top. The structure is composed of an outer and inner masonry wall of clay bricks
infilled with a sort of conglomerate made of brick and stone rubble. The foundation
is a masonry block about 6.00 m deep and 11.70 m wide. The soil underneath the
tower consists of alluvial deposits with prevailing slimy clays, clayey sandy slimes
and slimy sands. The respective site type has been determined as C according to
Eurocode 8, part 1, (2005), classification.
Fig. 4.4 Parma cathedral and its bell tower. The photograph on the left (Courtesy of photo Vaghi/
CSAC, Universita di Parma/Photography Section)
4.3.2 Dynamic Properties of the Parma Bell Tower
First a 3D, finite element model of the Parma Tower was prepared using SAP 2000.
The masonry walls were modeled using the so called “shell thick element” type to
represent both shear and bending, prevailing in the seismic response of these types
of structures. The foundation block was modeled using the “solid” elements type.
The number of finite elements equalled 488 of the shell type, and 243 of the solid
type. Then the eigen problem was calculated and the fundamental natural period T1
was calculated equal to 1.366 s. This natural period does not differ much from the
experimental results of Roberto Cerioni et al. (1996), using a 3D FEM model, that
is: T1 ¼ 1.374 s.
Fig. 4.5 Detailed cross-sections of the Parma Cathedral Bell Tower

4.3.3 Results of the Analysis of the Response
Since so far, the rotational excitations are not included in any of the commercial
seismic engineering FEM software, respective simplified model of the tower
structure should be build “manually”. To model slender towers one can use a
beam or frame element. Taking into account a combined shear and bending,
constituting seismic response of typical masonry towers it was decided to use a
plane Timoshenko beam model of the tower.
In order to calibrate the beam model of the tower a criterion of equal displace-
ments under horizontal seismic excitations was chosen. The tip displacements of
the tower were calculated using the 3D FEM model and using the Timoshenko
beam and respective tip displacements were compared. After multiple trials and
errors appropriate values of EJ and GA constant along the beam height were
chosen, so that the differences in the tip displacements of the 3D FEM model and
the simplified beam model is minimized. As a result the, values of E¼3·109 Pa and
G¼ 1.36·109 Pa were obtained. The Poisson ratio ν was assumed as equal to 0.1.
The foundation at the base of the structure was built in form of a simple cuboid.
It was introduced to the vibrating system as an additional moment of inertia Iθ,
equal to: 34,782,246 kg m2 and rotational soil compliance which was calculated
taking into account the foundation area and sub-soil properties. Applying classic
formulas for the elastic half space respective rotational stiffness kθ equals
1.94·1012 N·m/rad (Fig. 4.3).
As a result of tuning of the simplified beam model, as described above, the first
natural period was obtained as equal to T1 ¼ 1.366 s – the same as the one from 3D
FEM model of SAP 2000.
To model the horizontal excitations the response spectrum of Eq. 4.2 was used
for type 1 and 2 of the response spectra (Fig. 4.2), while respective rocking
excitations were modeled using Eq. 4.1. Peak ground acceleration ag ¼ 2.146 m/
s2 and the site category C were assumed with following parameters of the response
spectra: TB ¼ 0.2 s, TC ¼ 0.6 s, TD ¼ 2 s, S ¼ 1.15 for type 1 spectrum and TB ¼ 0.1 s,
TC ¼ 0.25 s, TD ¼ 1.2 s, S ¼ 1.5 for type 2 spectrum. Respective damping correction
factor was assumed as equal to 1 (ξ ¼ 5 %).
The numerical analyses were carried out in Matlab for the respective Timo-
shenko beam model. Two specific cases were compared. First, when only horizon-
tal excitations are assumed, and second, when the seismic excitations are assumed
as simultaneously acting horizontal and rocking effects. The base shear and
overturning moments are compared in Table 4.1. It can be seen from Table 4.1
that the effect of rocking excitations on the overall seismic response is substantial in
case of the rather flexible soil representing the Parma tower site. As a result of
including the rocking effects the tip displacements increased of about 24 %, while
the base shear and overturning moments of 30–40 % and about 25 % respectively.
In Figs. 4.6 and 4.7 detailed plots of the shear forces and bending moments along
the height of the tower are shown for type 1 and 2 response spectra.
44
Table 4.1 Seismic horizontal ‘h’ and rocking ‘r’ excitation effects on tip displacements, base shear and overturning moments for the Parma Cathedral Bell
Tower under type 1 and 2 EC-8 response spectra
Value of the response Increase
Type of response Type of excitations Type 1 Type 2 Type 1 Type 2
Tip displacements ‘h’ 0.2711 m 0.132412 m 23.72 % 24.53 %
‘h’ + ‘r’ 0.3354 m 0.164893 m
Base shear ‘h’ 6,674,260 N 5,773,605 N 31.31 % 40.16 %
‘h’ + ‘r’ 8,763,971 N 8,092,285 N
Overturning moments ‘h’ 1.95E+08 N 1.04E+08 Nm 24.37 % 27.84 %
‘h’ + ‘r’ 2.45E+08 N 1.33E+08 Nm
Z. Zembaty et al.
Fig. 4.6 Plot of the shear forces along the height of the tower
It is interesting to note from Figs. 4.6 and 4.7 that the shear and bending moment
responses for type 1 and type 2 response spectra are substantially different. The
main reason is the different contribution of the higher than the first mode for both
spectra. This can clearly be seen in Tables 4.2 and 4.3 showing contributions of the
first five modes in the “total” response at selected heights of the tower. The “total”
response was computed in this case by including in the computations 20 modes. The
higher modes (particularly second and the third) generate greater pseudo-static
Fig. 4.7 Plot of the bending moments along the height of the tower
forces in the upper part of the tower. This difference can be seen in the shapes of the
plots of shear and moments from Figs. 4.6 and 4.7 for type 1 and type 2 spectra.
When slender towers are under a combined horizontal and rocking excitations,
including also the rotational soil compliance, one can argue that the heavy weights
of the tower finite elements, acting on the eccentricities caused by the horizontal
displacements of the tower, may substantially increase respective bending
moments. For this reason this well known second order effect (P-Δ) has been
included here by adding an additional overturning moment at the base of the
Table 4.2 Contribution of the first five modes in the total response of the tower at selected
heights, for type 1 response spectrum (z ¼ height, S ¼ shear force, M ¼ bending moment)
Mode ‘1’ Modes Modes up Modes up Modes up
only ‘1’+’2’ to three to four to five
z [m] F % of total % of total % of total % of total % of total
39.93 S 77.12 99.09 99.09 99.09 99.82
M 58.59 96.54 96.54 99.91 99.93
30.17 S 96.18 97.00 97.00 99.75 99.78
M 79.05 99.79 99.79 99.91 99.99
25.56 S 96.40 97.63 97.63 99.72 99.88
M 87.21 99.93 99.93 99.95 99.99
0 S 72.26 97.45 97.45 99.42 99.82
M 96.97 99.90 99.90 99.99 100.00
Table 4.3 Contribution of the first five modes in the total response of the tower at selected
heights, for type 2 response spectrum (z ¼ height, S ¼ shear force, M ¼ bending moment)
Mode ‘1’ Modes Modes up Modes up Modes up
only ‘1’+’2’ to three to four to five
z [m] F % of total % of total % of total % of total % of total
39.93 S 43.21 96.06 96.06 96.06 99.35
M 26.90 90.77 90.77 99.75 99.83
30.17 S 72.75 76.40 76.40 98.20 98.40
M 46.22 99.05 99.05 99.58 99.97
25.56 S 74.99 80.63 80.63 97.78 99.17
M 58.54 99.59 99.59 99.68 99.95
0 S 37.37 91.19 91.19 97.92 99.41
M 84.58 99.03 99.03 99.89 99.98
structure equal to the sum of the product of the nodal horizontal displacements and
respective weights of the lumped masses. The result was checked for the excitations
of up to the peak ground accelerations ag ¼ 2.146 m/s2 ¼ 0.22 g, which approxi-
mately corresponds to the design for MM intensity of about VIII. It was found that
in the case of the Parma Cathedral Bell Tower the P-Δ effects did not exceed 1.2 %
(Rossi 2012).
4.4 Conclusion
This paper presents results of a numerical analysis of combined horizontal-rocking

seismic effects on an old masonry Parma Bell Tower. Since standard Finite Element
programs are not capable to include rotational excitations, a simplified, Timo-
shenko beam model of the tower was formulated. A numerical analysis in Matlab
was carried out using the Eurocode 8 part 6, (2005), model of the rotational seismic
effects. It was found that for the Parma Bell Tower situated on soil C and at the
design intensity level defined by accelerations ag¼2.146 m/s2 the rocking effects
are substantial, increasing the base shear for 33 % and the overturning moments for
about 25 %. It was also checked that for this level of excitations the second order
effects coming from the P-Δ effects are not substantial (about 1.2 %).
The rotational seismic effects are under early stage investigations (see e.g. Lee
et al. 2009; Igel et al. 2012) so the formulas of EC-8 part 6 should still be calibrated
by more experimental data. Until the rotational data coming from strong intensive
records are gathered using special modern rotational devices (Lee et al. 2009; Igel
et al. 2012), the approximate formulas of EC-8 part 6, (2005), need to be applied in
the engineering practice.
Acknowledgments The second author gratefully acknowledges the Erasmus scholarship which
was used to stay for 1 semester at the Faculty of Civil Engineering, Opole University of
Technology, in Poland when the MSc thesis (Rossi 2012) was prepared.
References
Castellani A, Zembaty Z (1996) Comparison between earthquake rotation spectra obtained by

different experimental sources. Eng Struct 18(4):597–603
CEN, Comité Européen de Normalisation (2004) Eurocode 8: design of structures for earthquake
resistance. Part 1: general rules, seismic actions and rules for buildings. CEN (Comite
Europeen de Normalisation), EN 1998-1. Brussels, Belgium
resistance. Part 6: towers, masts and chimneys. CEN (Comite Europeen de Normalisation), EN
1998-6. Brussels, Belgium
Cerioni R, Brighenti R, Donida G (1996) Seismic analysis of masonry towers: the bell-tower of the
Parma Cathedral, Studi e Ricerche – Vol. 17, Scuola di Specializzazione in Costruzioni in C.
A., Fratelli Pesenti, Politecnico di Milano, Italia
Falamarz-Sheikhabadi G-A (2012) Approximate formulas for rotational effects in earthquake
engineering. J Seismol 16:815–827
Igel H, Brokesova J, Evans J, Zembaty Z (2012) Preface to special issue on “Advances in
rotational seismology: instrumentation, theory, observations and engineering”. J Seismol
16:571–572
Kalkan E, Grazer V (2007) Coupled tilt and translational ground motion response spectra. J Struct
Eng ASCE 133(5):609–619
Lee WHK, Celebi M, Todorovska M, Igel H (2009) Introduction to the special issue on rotational
seismology and engineering applications. Bull Seismol Soc Am 99(2B):945–957
NTC (2008) New Italian design code for constructions, “Nuove Norme Tecniche per le
Costruzioni” of 14 January 2008
Rossi A (2012) Rotational seismic effects on slender towers with the application of Eurocode
8 Part 6 MsC thesis, Parma University. pp. 1–279.
Trifunac M (1982) A note on rotational components of earthquake motions for incident body
waves. Soil Dyn Earthq Eng 1:11–19
Zembaty Z (2009) Rotational seismic load defintion in eurocode 8, part 6, for slender tower-shaped
structures. Bull Seismol Soc Am 99(2B):1483–1485
Chapter 5
FOSREM: Fibre-Optic System for Rotational
Events and Phenomena Monitoring:
Construction, Investigation and Area
of Application
Leszek R. Jaroszewicz, Anna Kurzych, Zbigniew Krajewski,

Jerzy K. Kowalski, and Krzysztof P. Teisseyre
Abstract This paper reviews our expertise with construction, investigation and
simulation of the fibre optic interferometric device named FOSREM (Fibre-Optic
System for Rotational Events and Phenomena Monitoring). The presented device
was designed for a direct monitoring of rotational components emitted during
seismic events as well as existing in irregular and complex structures. The con-
struction of the FOSREM utilizes the Sagnac fibre interferometer in a minimum
optical gyro configuration. This approach causes that FOSREM is complete
insensitivity to linear motions, and it enables to measure directly the rotational
components. In order to make FOSREM mobile and autonomous device we were
focused on decreasing size to the 36 36 16 cm dimension and implementing
special FORS – Telemetric Server which enables to control FOSREM remotely via
Internet. The laboratory investigation of our system indicated that it keeps the
theoretical sensitivity equal to 2·108 rad/s/Hz1/2 and accuracy no less than
6·109–5·105 rad/s in a frequency band from 0 Hz to the upper frequency between
2.56 and 328.12 Hz, respectively. FOSREM protects linear changes of sensitivity in
the above detection frequency bandpass and has the maximum values of rotation
rate possible to record without “overshoot” equal to 10 rad/s due to an innovative
electronic system.
L.R. Jaroszewicz (*) • A. Kurzych • Z. Krajewski

Institute of Applied Physics, Military University of Technology, 2 Gen. Sylwestra Kaliskiego
St., 00-908 Warsaw, Poland
J.K. Kowalski
m-Soft Ltd., 9-4 Sotta “Sokoła” St., 02-790 Warsaw, Poland
K.P. Teisseyre
Institute of Geophysics, Polish Academy of Sciences, 64 Ksie˛cia Janusza St.,
01-452 Warsaw, Poland

DOI 10.1007/978-3-319-14246-3_5
50 L.R. Jaroszewicz et al.
Keywords Rotational phenomena • Sagnac effect application • Seismology

• Interferometry • Rotational seismometer
5.1 Introduction
Investigation of seismic rotational phenomena has been started theoretically for

over 30 years (Lee et al. 2009). The atypical observations which appeared after
earthquakes like strange, rotational deformations of movements or tombs gave
reasons to research the mechanism of their formation (Kozak 2009). They were
interpreted as an interaction of seismic waves with components of media which
they pass through. Further considerations have been focused on the existence of
seismic rotational phenomena in grained rocks in a form of rotational events, as
well as seismic rotational waves (Droste and Teisseyre 1997). These deliberations
have been developed in the case of rocks with microstructure, defects (Eringen
1999; Teisseyre and Boratynski 2002) or without any internal structure (Teisseyre
2005; Teisseyre et al. 2005; Teisseyre and Gorski 2009).
Whereas one of the most important problem of the seismic behaviour of irregular
constructions in-plane is the existence of difficulties with controlling horizontal
rotation of these structures. The existence of such rotations has a direct influence on
torsional effects in structures as well as inter-story drift. Nowadays there are a lot of
advanced materials which enable constructing complicated, innovative and very
high structures. These constructions require unique and precise devices for contin-
uously monitoring structural responses (Cowsik et al. 2009).
The two above paragraphs indicated that new innovative instruments to measure
the rotational components of earthquakes as well as to monitor real rotation effects
in engineering constructions are necessary. Measurements of rotational motions
obtained by a special array or set of conventional seismometers are sensitive to
linear motions (Teisseyre and Nagahama 1999; Jaroszewicz et al. 2003). In our
opinion, the sensors based on the Sagnac effect (Sagnac 1913) seem to be the most
proper in the pointed out application fields. The main advantage of such approach is
possibility to measure rotational motions in a direct way (Jaroszewicz et al. 2014;
Kurzych et al. 2014). Moreover it does not require an external reference frame for
its measurements. The first review about application of the commercial available
fibre optic gyroscope (FOG) for the measurement of rotations in structural engi-
neering can be found in (Schreiber et al. 2009) as well as in (Zembaty et al. 2013)
for stiffness reconstruction.
However, the direct FOG application for such investigation is connected with
data processing problems because FOG output gives angles instead of angular
velocity which parameter is optimized according to minimize the drift phenome-
non. For above reasons we constructed two devices: the Autonomous Fibre-Optic
Rotational Seismograph (AFORS) and Fibre-Optic System for Rotational Events
and Phenomena Monitoring (FOSREM). The first one was used for direct recording
the rotational components of seismic events (Jaroszewicz et al. 2011, 2014;
Kurzych et al. 2014) and initially for continuous monitoring of the ‘horizontal
5 FOSREM‐Fibre‐Optic System for. . . 51
building’s rotation (Jaroszewicz et al. 2013a) while the second will be applied in
area of strong-motion seismology and to monitor engineering construction.
This paper describes and summarizes our broad experience in construction and
application of above mentioned systems.
5.2 Construction of the Rotational Fibre Optic

Interferometric Devices
The constructed rotational fibre optic interferometric devices (AFORS and

FOSREM) contain generally two parts: Optical Head and electronic system. The
Optical Head contains fibre optic interferometer where the Sagnac effect generates
a phase shift. The Sagnac effect is a result of difference between two equal beams
propagating around closed optical path, in an opposite direction. After recombina-
tion of these two beams the Sagnac phase shift induced by a rotational rate Ω
perpendicular to a plane of sensor is equal to (Post 1967):
4πRL 1
Δφ ¼ Ω¼ Ω ð5:1Þ
λc So
where: L – length of the fibre in the sensor loop, R – sensor loop diameter, λ –
wavelength of the source, c – velocity of the light in the vacuum, So – optical
constant of the interferometer.
The above dependence allows to measure rotations in a direct way and without
any external reference frame because the Sagnac phase shift does not depend on a
centre rotation location as well as a sensor plane shape (Post 1967). For above
reasons sensors based on the Sagnac effect seem to be appropriate to measure
rotational effects, perturbations and inter-story drifts in buildings.
We designed AFORS’s/FOSREM’s Optical Head according to the FOG minimal
configuration, as is shown in Fig. 5.1. It consists of the: superluminescent diode
(SLED) (Exalos, Schlieren, Switzerland; with ΔB ¼ 31.2 nm, λ0 ¼ 1305.7 nm and
P ¼ 9.43 mW), two X-type couplers (Phoenix Photonics, Birchington, UK; with
α ¼ 0.20 dB), depolarizer (AFORS: Phoenix Photonics; with DOP <5 % and
α ¼ 0.20 dB), isolator (FCA, Niepołomice, Poland; with α ¼ 0.34 dB and 39 dB
isolation), detector (Optoway Technology, Taiwan; with S ¼ 0.9 A/W), two fibre-
optic polarizers mounted in-line (Phoenix Photonics; with ε ¼ 43 dB and
α ¼ 0.45 dB each) and phase modulator. (Piezomechanik, Germany). The wide-
band, low coherence SLED protects the proper method for eliminating the polar-
ization influence on the sensor work and emits the high optical power, which it
affects the system sensitivity. The second coupler has been added after the polarizer
to ensure that both counter-propagating beams pass through the same path in the
sensor loop (Udd and Spillman 2011). The polarizers assure that the two light
waves return to the first coupler in the same polarization. This configuration
Fig. 5.1 The block diagram of the AFORS/FOSREM construction (upper part presents the optical
head and the lower presents electronic system)
guarantees that the only nonreciprocal effect in the system is the Sagnac effect. In
the next part of this paper we presented all reasonable causes of the selection
exactly above parameters of the optical elements.
The electronic system, presented in lower part of Fig. 5.1, is named Autonomous
Signal Processing Unit – ASPU (m-Soft Ltd., Warsaw, Poland). It allows to
calculate and record information about rotation motions which is obtained directly
from the measured Sagnac phase shift. Generally, the ASPU protects selection of
the first (A1ω) and the second (A2ω) amplitude of the harmonic output signal [u(t)],
which enables to assign rotation rate according to the relation of (Jaroszewicz
et al. 2011):

A1ω
Ω ¼ So arctan½Se uðtÞ ¼ So arctan Se ð5:2Þ
A2ω
where Se – the electrical constant related to parameters of the applied components

and obtained during the sensor calibration together with determining the optical
constant So. The calibration process is wide described in (Krajewski 2005;
Jaroszewicz et al. 2011). The electronic part uses quartz oscillator which deter-
mines the basic detection time equals 4.7104 ms for AFORS. The multiplication of
this time gives opportunity to choose the frequency from 0 Hz to the upper chosen
value between 0.84 and 106.15 Hz, and above band range is named “detection
frequency bandpass” in the next part of this paper. It should be underlined that this
feature of AFORS/FOSREM allows to detect rotational movements in extremely
low frequencies, even equal to 0 Hz, it is significant advantage of optical devices in
comparison to the mechanical devices, which have limited so-called bandwidth,
connected with their mechanical nonlinear frequency characteristic. According to A
Glossary for Rotational Seismology (Lee 2009), the bandwith is a range between
high-pass and low-pass cut-off frequencies and often meant as the portion of
frequency-response amplitude spectrum that is approximately flat.
It should be underlined that it is completely another approach in comparison to
the method used in the FOG where measurement of the angular changes is obtained
by applying the time integration of the angular rates. Moreover, the ASPU sends
records to a GSM/GSP modem connected with WEB special FORS – Telemetric
Server. This server enables to make AFORS and FOSREM remotely control via
internet including needful system corrections, software upgrade, data storage as
well as changes of the working parameters.
The sensitivity of the designed sensors is the most significant parameter due to
apply them to measure rotational effects. Theoretical considerations (Ostrzyżek
1989) reach to conclusion that the optimization of the Optical Head elements,
which a main goal is maximization of the sensor sensitivity, requires selection of
such parameters as: radius of the sensor loop, optical power and wavelength of the
light source, length of the fibre and total losses of the optical path. These parameters
have the most meaningful effect on the system sensitivity.
In our first devices, named AFORS-1,-2,-3, the optical fibre of 15 km – SMF-28e+
was wound on the 0.63 m diameter sensor loop which has been made of a special
composite material with permalloy particles for shielding the sensor from external
magnetic field. Based on all above described configuration and consideration finally
we obtained the accuracy for AFORSs as is shown in Fig. 5.2. As one can see
AFORS-1 and AFORS-2 have similar accuracy, only about half of the order worse
than theoretically calculated their sensitivity for given detection frequency bandpass.
Unfortunately, for AFORS-3 accuracy is more than three times worse and this system
may be used in limited application. Such deterioration is connected with problems of
applied electronic part in AFORS-3. Because all AFORSs use “open configuration”
for rotation rate calculation according to the formula (5.2), the parameters of the
sensor loop determine directly the maximum values of rotation rate possible to record
without “overshoot” which is equal to 0.02 rad/s.
The AFORS-1 (Fig. 5.3) has been installed in the Ksia˛ż seismological observa-
tory in Poland since July 2010 with its continuous monitoring via Internet basing on
the application FORS – Telemetric Server.
In Fig. 5.4 we presented example of the rotational event recorded after the Honshu
earthquake M ¼ 9.0 which took place on March 11th 2011, at 5 h 46 min 23 s.
The system of next two devices AFORS-2 and AFORS-3 have been initially
used to measure rotation response of engineering constructions. We used two
seismographs in order to have one as the reference system. The summarized results
of this investigation are presented in Fig. 5.5 as amplitude of building rotation
measured at different building floors caused by ground moves after tramp pass
through street in distance about 50 m from and parallel to long building wall.
AFORS-3 accuracy
1x10-6
AFORS-2 accuracy
AFORS-1 accuracy
5x10-7
1x10-7
AFORS-2 quantum
5x10-7
Ω[rad/s]
noise limitation
1x10-8
AFORS-3 quantum
5x10-9
noise limitation
AFORS frequency
bandpass
1x10-9
1 2 5 10 20 50 100
Frequency band [Hz]
Fig. 5.2 The accuracy measured in Warsaw, Poland for the chosen detection frequency bandpass
(0.84–106.15 Hz) for AFORSs
Fig. 5.3 GOOGLE map with current devices localization

Fig. 5.4 The plot of the seismic event recorded in Ksia˛ż, Poland on March 11th 2011, starting
from 8 h 46 min, after the Honshu earthquake M ¼ 9.0, times UTC (Jaroszewicz et al. 2013a)
Fig. 5.5 The data recorded on different floors as the response for ground moves (Jaroszewicz
et al. 2013b)
Nevertheless, our investigations which were carried out by AFORS’s pointed out
that we have to construct a new mobile system. The main reasons for such
assumption were the noticeable disadvantages of AFORS’s application in structural
engineering which are: extremely high cost (uses only for limited applications),
problem with verification of their proper operations, single axis rotation measure-
ments and errors under high temperatures (above 60 C) as well as transport
complications due to large dimensions of the devices.
For above-mentioned reasons a new Fibre-Optic System for Rotational Events

and Phenomena Monitoring (FOSREM) has been performed which should give
more practical device. It should be emphasised that FOSREM gives possibilities to
measure rotational effects in a wide range of frequencies and rotational rates.
Moreover, we focused on decreasing size of the sensor to 36 36 16 cm dimen-
sion. Thus we decided to record data about rotation motions during the earthquakes
in seismological observatory by AFORS systems due to high accuracy and suffi-
cient detection frequency bandpass for seismic application while our new system
FOSREM will be applied to investigate engineering construction rotation response
as well as to operate during strong-motions earthquakes.
5.3 Simulated and Experimental Results of FOSREM
The sensitivity of the designed sensors described the lowest value of the measured
angular velocity in a set detection frequency bandpass. It means the velocity in
which the quotient of the output power to the noises is equal to one. To precisely
estimate the main parameters of sensors we have made simulations which enabled
to design FOSREM. A few of our previous investigations (Jaroszewicz et al. 2013b)
carried out by AFORSs were disturbed by high temperature. Thus, in this paper we
included discussion about the influence of the temperature on the system operating
as well.
The dependence between sensitivity and optical power of the light source shows
that sensitivity increases with increasing the source optical power. The relation
presented in Fig. 5.6a shows that the most advantageous value of optical power is
near 10 mW which corresponds to apply in FOSREM SLED diode with a optical
power P ¼ 9.43 mW.
To enlarge the sensitivity one can apply the shorter wavelength of light source
(see Fig. 5.6b). Although higher attenuation is related with shorter wavelength.
Therefore operating in the second telecommunication window seems to be the
optimal solution.
In the FOSREM construction we were also focused on decreasing size of the
sensors. Dimensions of the first two FOSREM devices are equal to the
47 23 36 cm (similar to the size of a flight hand luggage – Fig. 5.7a), however
next devices will have size equals 36 36 16 cm which considerably enlarges
their mobility (Fig. 5.7b). It enables to carry out the planned measurements simply
and without transport complications. Additionally, we designed a mechanical
protection of the Optical Head, as is shown in Fig. 5.7b, printed on 3D Printer
(MakerBot Replicator 2X).
Such approach requires, at first, analysis of the fibre length and the sensor loop
radius selection. To define the optimal fibre length we performed simulation which
is shown in Fig. 5.8. The simulation was prepared for 1305.7 nm wavelength of the
light source and 16.3 dB system losses as well as the sensitivity of the photodiode
equals 0.9 A/W. The simulation shows that the selection fibre length of 5 km is the
Fig. 5.6 Influence of the light source parameters on sensor sensitivity: (a) The sensor sensitivity
vs. optical power of the source; (b) The sensor sensitivity vs. wavelength of the source
Fig. 5.7 Package of the designed sensor: (a) total size of the first two sensors; (b) 3D visualization
of the new sensor with all components, colour: beige – package, dark – depolarizer, grey – sensor
loop, green – electronic parts, yellow – modulator, blue – SLED diode, orange – sensor’s
mechanical protection and acoustic isolation
Fig. 5.8 Influence of fibre length and radius loop on the sensor sensitivity: (a) correlation between
fibre length, detection frequency bandpass and sensor sensitivity; (b) correlation between fibre
length, radius loop and sensor sensitivity
most optimal length for obtaining minimal rate of rotation rates in entire detection
frequency bandpass. The Sagnac effect increases with the length of the optical fibre.
Although the optical fibre introduces an attenuation which decreases the level of
signal. The sensitivity is at the same level in the entire detection frequency
bandpass for fibre length of 5 km. The fibre length of 5 km and 0.12 m radius of
sensor loop enable to achieve sensitivity about 2·108 rad/s/Hz1/2. FOSREM uses
quartz oscillator, which determines the basic detection time equals 1.5234 ms, thus
above sensitivity should protect the accuracy no less than 6·109–5·105 rad/s in a
frequency band from 0 Hz to the upper frequency between 2.56 and 328.12 Hz,
respectively. To perform the sensor loop in FOSREM we wound optical fibre to an
aluminum reel (the grey one in Fig. 5.7b) in a double-quadrupole mode (Dai
et al. 2002) for the thermal stabilization of the sensor’s work. One should noticed
that this method of winding can cause growth of the attenuation. For the reasons
above we decided that 5 km is the most optimal length.
The total losses of the optical path (α) have the main influence on system
sensitivity (see simulation in Fig. 5.9). Due to the above we made assumption that
total losses of the Optical Head should be below value of 17 dB (where 9 dB is
theoretical minimum losses) because it protects the system sensitivity at expected
level. The value of the system losses higher than 17 dB causes rapid increase of the
minimum measured rotation. In first two FOSREM constructions we obtained losses
equals about 16 dB which contain: losses on all fibre splices (about 2 dB), attenuation
in the sensor loop about 1.65 dB, losses on the couplers 9 dB resulting from the
minimum configuration, losses on the other optical elements such as the polarizer and
the depolarizer about 3.65 dB. Moreover, it should be underlined that the selected
radius of the sensor loop does not introduce additional macrobending losses.
Figure 5.10 shows correlation between temperature, detection frequency band
and the minimal measured rotation rate. As one can see Ω is stable in temperature
range between 260 and 333 K. Such results are contradicting our previous
Fig. 5.9 The sensor sensitivity vs. optical system losses
4. × 10-7
W min [rad/s]
3. × 10-7
300
2. × 10-7
1. × 10-7
0 200
260
Detection frequency
280 bandpass [Hz]
100
300
Temperature [K]
320
0
Fig. 5.10 Correlation between temperature, detection frequency bandpass and sensor sensitivity
measurements where recorded data was disturbed by the high temperature

(Jaroszewicz et al. 2013b). However in the AFORS we used optical fibre of
15 km which could introduce higher temperature fluctuation in the system and
polarization instability caused by the temperature perturbations while the presented
simulation in Fig. 5.10 was performed for sensor with optical fibre of 5 km.
The presented simulations indicate that FOSREM sensitivity depends on the
various parameters. Our considerations allow to design very promising device
which will be applied to the investigation of rotation motions, displacements and
inter-story drifts of civil engineering structures.
Finally, in order to carry out some experimental measurements we have placed
the first constructed prototype of FOSREM on a rotational table (Fig. 5.11). The
presented results in Fig. 5.12 show measurement of Earth rotation component in
Warsaw (equal to 4.45·105 rad/s) for different value of detection frequencies
bandpass for tentatively calibrated our device. We carried out measurements
using four selected upper frequencies: 0.01 Hz, 10.25 Hz, 109.38 and maximal –
328.12 Hz, respectively.
Because engineering applications need system which is able to detect rotation
rate as high as 10 rad/s, the reconstruction of ASPU according to special electronic
procedure (prepared as an patent application) gives for FOSREM the expected
maximum value of rotation rate possible to record without “overshoot” up to above
mentioned value, also.
The main purpose of presenting these results was to demonstrate a wide useful-
ness of our instruments. We are convinced that AFORS and FOSREM parameters
give opportunity to record highly interesting rotational events or phenomena.
Fig. 5.11 The laboratory set to measure rotational motions by FOSREM

Fig. 5.12 Example of measured Earth rotation rate component in Warsaw at various detection
frequency bandpass: (a) 0.01 Hz; (b) 10.25 Hz; (c) 109.38 Hz; (d) 328.12 Hz
Fig. 5.12 (continued)
5.4 Conclusions
In this paper we presented the most promising interferometric sensor construction

to monitor continuously the rotational effects of complicated engineering construc-
tions as well as to measure effects in engineering strong-motion seismology area –
FOSREM. It is unique and innovative device which utilizes the Sagnac effect. This
technique allows to measure the rotational effects directly and without any refer-
ence frame. It is very important during the earthquakes when nothing is stable.
The included simulations have enabled to design the FOSREM and have shown
that selected parameters of the optical elements are appropriate to obtain enough
accuracy (no less than 6·109–5·105 rad/s) for the seismic application. Moreover,
the FOSREM operates in the wide range of the detection frequency bandpass from
0 Hz to the chosen upper frequency range from 2.56 to 328.12 Hz with maximum
value of detected rotation rate of the order of 10 rad/s. The designed FOSREM
construction is mobile and the electric part makes it totally remotely controlled via
Internet and it allows to obtain immediate information about recorded events. All
FOSREM advantages indicate that it is proper device to apply it in the investigation
of rotational effects in engineering structures.
Moreover, presented data shows the proper work of AFROSs as a system for
rotational motion detection and they are very promising for the future investigation
of rotational events in Ksia˛ż seismological observatory. We are collecting contin-
uously records about seismic evens and they are under scientific processing and will
be published in papers linked with seismological science.
We believe that our devices can significantly influence seismological science.
Acknowledgments This work has been done under financial support of the National Centre for
Research and Development, project PBS1/B3/7/2012 “Fibre-Optic Monitoring Rotational Events
and Phenomena – FOMREPh” as well as MUT statutory activity PBS-850.
References
Cowsik R, Madziwa-Nussinov T, Wagoner K, Wiens D, Wysossion M (2009) Performance

characteristics of a rotational seismometer for near-field and engineering applications. Bull
Seismol Soc Am 99(2B):1181–1189
Dai X, Zhao X, Cai B, Yang G, Zhou K, Liu C (2002) Quantitative analysis of the Shupe reduction
in a fiberoptic Sagnac interferometer. Opt Eng 41:1155–1156
Droste Z, Teisseyre R (1997) Rotational and displacemental components of ground motion as
deduced from data of the azimuth system of seismographs. Publ Inst Geophys Pol Acad Sci
M-1(97):157–167
Eringen AC (1999) Microcontinuum field theories, vol 1, Foundations and Solids. Springer,
New York
Jaroszewicz LR, Krajewski Z, Solarz L, Marć P, Kostrzyński T (2003) A new area of the fiber-
optic sagnac interferometer application. In: Proceedings of International Microwave and
Optoelectronics Conference IMOC-2003, Iguazu Falls, Brazil, 20–23 Sep 2003
Jaroszewicz LR, Krajewski Z, Kowalski H, Mazur G, Zin owko P, Kowalski JK (2011) AFORS
autonomous fibre-optic rotational seismograph: design and application. Acta Geophys
59:578–596
Jaroszewicz LR, Krajewski Z, Teisseyre KP (2013a) The possibility of a continuous monitoring of
the horizontal buildings’ rotation by the fiber-optic rotational seismograph AFORS type. In:
Lavan O, De Stefano M (eds) Seismic behaviour and design of irregular and complex civil
structures. Springer, Heidelberg, pp 339–351
Jaroszewicz LR, Krajewski Z, Matysik M, Kowalski KJ (2013b) Investigation of building rotation
motion using autonomous fibre rotational seismographs. Paper presented at the 3rd Interna-
tional Workshop on Rotational Seismology – 3IWoRS, Christchurch, 22–25 Sep 2013
Jaroszewicz LR, Krajewski Z, Kowalski KJ, Kurzych A, Raszewski R (2014) AFORS – autono-
mous fiber optic rotational seismograph as a system for continuous monitoring the rotational
seismic events. Adv Mater Res 909:444–449
Kozak JT (2009) Tutorial on earthquake rotational effects: historical examples. Bull Seismol Soc
Am 99:998–1010
Krajewski Z (2005) Fiber optic Sagnac interferometer as device for rotational effect investigation
connected with seismic events. Dissertation, Military University of Technology
Kurzych A, Jaroszewicz LR, Krajewski Z, Teisseyre KP, Kowalski JK (2014) Fibre optic system
for monitoring rotational seismic phenomena. Sensors 14:5459–5469
Lee WHK, Celebi M, Todorovska MI, Igel H (2009) Introduction to the special issue on rotational
seismology and engineering applications. Bull Seismol Soc Am 99:945–957
Leee WHK (2009) A glossary for rotational seismology. Bull Seismol Soc Am 99:1082–1090
Ostrzyżek A (1989) Accuracy analyze of the angle speed measurement in the fibre optic gyro-
scope. Dissertation, Military University of Technology
Post EJ (1967) Sagnac effect. Rev Mod Phys 39:475–493
Sagnac G (1913) The light ether demonstrated by the effect of the relative wind in ether into a
uniform rotation interferometer. Acad Sci 95:708–710
Schreiber KU, Velikoseltsev A, Carr AJ, Franco-Anaya R (2009) The application fibre optic
gyroscope for the measurement of rotations in structural engineering. Bull Seismol Soc Am
99(2B):1207–1214
Teisseyre R (2005) Asymmetric continuum mechanics: deviations from elasticity and symmetry.
Acta Geophys 53(2):115–126
Teisseyre R, Boratynski W (2002) Continuum with self-rotation nuclei: evolution of defect fields
and equations of motion. Acta Geophys 50(2):223–229
Teisseyre R, Gorski M (2009) Transport in fracture processes: fragmentation and slip. Acta
Geophys 57(3):583–599
Teisseyre R, Nagahama H (1999) Micro-inertia continuum: rotations and semi-waves. Acta

Geophys Pol 47:259–272
Teisseyre R, Białecki M, G orski M (2005) Degenerated mechanics in a homogeneous continuum:
potentials for spin and twist. Acta Geophys 53(3):219–230
Udd E, Spillman BW (2011) Fiber optic sensors. Wiley, Hoboken
Zembaty Z, Kokot S, Bobra P (2013) Application of rotation rate sensors in an experiment of
stiffness ‘reconstruction’. Smart Mater Struct 22: doi: 10.1088/0964-1726/22/7/077001
Chapter 6
Application of Rotation Rate Sensors
in Measuring Beam Flexure and Structural
Health Monitoring
Zbigniew Zembaty, Seweryn Kokot, and Piotr Bobra
Abstract This paper presents a new approach to measure and monitor structural
vibrations in civil and seismic engineering which uses new rotational sensors which
can measure flexural vibrations of a beam axis. First a rotation rate sensor
(measuring rotational velocity) is tested with respect to its ability to follow changes
of strains in a beam during its vibrations. Next a system of rotation rate sensors is
applied to effectively reconstruct stiffness variations of a simple, cantilever beam. It
is demonstrated that the rotation rate sensors can be used to effectively reconstruct
three unknown stiffness drops of a cantilever beam under harmonic vibrations. Both
experiments are carried out using small plexi beams in laboratory scale. At this
moment the rotational sensors are still rather expensive and with limited range and
accuracy. However with the time passing by, their quality will improve and prices
decrease making them very effective instruments in seismic engineering and health
monitoring of structural systems.
Keywords Vibrations • Rotation rate sensors • Beam flexure • Inverse problem

• Structural health monitoring
6.1 Introduction
Structural Health Monitoring (SHM) is a rapidly developing, new branch of civil

engineering, which originated in the second half of twentieth century in other fields
of engineering, particularly in the machine and aero industry, due to a particular
need of constant monitoring of important machines and structural systems. One of
the key problems of SHM in civil engineering is to conclude about the state of
structures based on monitoring their dynamic response to various types of excita-
tions (harmonic, wind, seismic, ambient etc.). Using respective response
Z. Zembaty (*) • S. Kokot • P. Bobra

Faculty of Civil Engineering, Opole University of Technology,
ul. Proszkowska 76, 45-758 Opole, Poland
e-mail: [email protected]; [email protected]; [email protected]

DOI 10.1007/978-3-319-14246-3_6
measurements one can retrieve natural frequencies and modes of the structure and
use them in modal analyses leading to damage assessment and localizations.
However when one is trying to locate damages in structures by observing changes
in the natural modes it appears that the observed stiffness losses are much better
reflected in the variations of spatial derivatives of the modes than in the modes
themselves. On the other hand, numerical differentiation of the measured natural
modes introduces such substantial noise, that in spite of using special smoothing
techniques, effective damage localization is still very difficult (see e.g. papers by
Maek and DeRoeck 1999; Ndambi et al. 2002). For some time however, new
techniques of directly measuring angle variations appeared (see e.g. Meydan
1997) and nowadays matured to achieve angle resolution of 103 . Thus, in
addition to transversal accelerations, now it is possible to measure angle variations
along the bar axis during vibrations of the structures. This way, the changes in
curvature of the axes of the bars of the structures can be obtained almost directly.
Numerical simulations aiming at investigations of potential advantages of these
new angular measurements have already appeared in the literature, and the effec-
tiveness of including the rotations in conventional modal analysis was confirmed
(Abdo and Hori 2002; Kokot and Zembaty 2009b). The measurements of beam axis
rotations make it also possible to retrieve respective strains for the same beam in
bending.
This paper investigates the advantages of using the rotation rate sensors in
monitoring beam vibrations in bending. First, selected results of an experiment in
which strains of a beam in harmonic flexural, vibrations are assessed based on the
measurements of rotational velocity of the beam axis are presented. Next, results of
another experiment are presented, in which rotation rate sensors are used to
reconstruct stiffness changes of a beam under harmonic flexural vibrations.
6.2 Beam Under Flexural Vibrations
Consider equation of motion of a uniform cantilever beam

ðEJ ðxÞ ¼ EJ ¼ const, mðxÞ ¼ m ¼ constÞ under kinematic excitations (Fig. 6.1):
" #
4 4
∂ w ∂ w
EJ þ κ 4 w_ þ μmw_ þ μm€
w ¼ m€uðtÞ ð6:1Þ
∂x4 ∂x
where ‘dot’ stands for differentiation with respect to time, u€(t) is the acceleration of
the kinematic motion of the beam, κ, μ are constants reflecting the participation of
the damping proportional to stiffness and mass respectively.
Applying mode superposition method (e.g. Chopra 2011) leads to following
solution
6 Application of Rotation Rate Sensors 67
Fig. 6.1 Cantilever beam under vertical, kinematic excitations
X
1 ð1
wðx; tÞ ¼ ℘ j ψ j ðxÞ h j ðτÞ €uðt τÞdτ ð6:2Þ
j¼1 1
in which
ðL
ψ j ðxÞdx
℘j ¼ ð0L ð6:3Þ
ψ 2j ðxÞdx
0
is the modal participation factor, ψ j(x) stands for the “j-th” mode shape, while hj(t)
is the impulse response function of the “j-th” natural mode of vibration:
1
h j ðtÞ ¼ exp ξ j ω j t sin ω jd t ð6:4Þ
ω jd

where ωjd ¼ √(1-ξj2) is the damped natural frequency and ξ j ¼ μ= 2ω j þ κω j =
2 is the modal damping ratio. Equation 6.2 can be used to obtain its spatial
derivative i.e. the angle of the axis rotation:
X 1 ð1
∂wðx; tÞ 0
ϑðx; tÞ ¼ ¼ ℘ j ψ j ðxÞ h j ðτÞ €uðt τÞdτ ð6:5Þ
∂x j¼1 1
and second derivative, which is proportional to bending moment

2 X1 ð1
0 ∂ wðx; tÞ M 1 00
ϑ ðx; tÞ ¼ ¼ ¼ ¼ ℘ ψ ð x Þ h j ðτÞ €uðt τÞdτ ð6:6Þ
∂x2 EJ ρ j¼1
j j
1
Using the Fourier transform one may re-write Eq. 6.2 in frequency domain as
follows
X1
W ðx; ωÞ ¼ ℘ j ψ j H j ðωÞUaccel ðωÞ ð6:7Þ
j¼1
where Hj(ω) stands for frequency response function of the “j-th” mode of vibration
ð1
1
H j ð ωÞ ¼ €uðτÞ eiωτ dτ ¼ ð6:8Þ
1 ω2j ω2 þ 2iξ j ω j ω
in which i ¼ √(1) and Uaccel(ω) denotes Fourier transform of the excitation

accelerations u€(t)
ð1
Uaccel ðωÞ ¼ €uðτÞ eiωτ dτ ð6:9Þ
1
In case of steady-state vibrations the minus signs in formulas (6.1, 6.5, and 6.7)
can be dropped. Applying the rules of time and spatial differentiation of formula
(6.2) into its frequency domain form (6.7), one can obtain velocity of the angle of
the beam axis ϑ_ denoted in frequency domain as θvel:
X
1
0
θvel ðx; ωÞ ¼ ℘ j ψ j ðxÞiω H j ðωÞ Uaccel ðωÞ ð6:10Þ
j¼1
Assume now that the beam has a rectangular bh cross-section. For such the Euler-
Bernoulli beam the maximum strain (on the beam surface) equals
2
h∂ w
ε¼ ð6:11Þ
2 ∂x2
Thus applying Eq. 6.5 the maximum strain can be obtained in frequency domain as
follows
hX 1
00
εðx; ωÞ ¼ ℘ ψ ðxÞH j ðωÞ U accel ðωÞ ð6:12Þ
2 j¼1 j j
Comparing Eq. 6.12 with 6.10 gives formula for the beam surface strain in terms of
rotational velocity written in frequency domain
X1 00
h ℘ j ψ j ðxÞH j ðωÞ
εðx; ωÞ ¼ θvel ðx; ωÞ X1
j¼1
0 ð6:13Þ
2iω ℘ j ψ j ðxÞH j ðωÞ
j¼1
For harmonic vibrations one may substitute amplitude of rotation rate am (θvel)
obtaining phase shifted maximum strain. Thus the final formula to be used in the
comparisons of the amplitude of rotation rate and the amplitude of maximum strain
takes form
X1
℘ ψ
00
ð ÞH ð ω Þ

h 1 j j x j
amðεÞ ¼ jεðx; ωÞj ¼ amðθvel Þ X1
j¼1
0 ð6:14Þ
2iω ℘ j ψ j ðxÞH j ðωÞ
j¼1
6.3 Experiment 1: Obtaining Strain from Rotational

Velocity of the Beam Axis
The small scale, laboratory models to study structural rotations should be chosen in
such a way, that the resulting rotations are large enough to represent typical
rotations of structures in full scale. After an analysis what was available, it was
decided to perform experiments on small span, cantilever beams made of plexiglass
excited by kinematic harmonic excitations (Fig. 6.2).
It should be noted that such small beams are much easier to excite in laboratory
conditions by using kinematic excitation, than exerting the actuator motion on
them. For this reason a heavy, long span, simply supported steel beam was prepared
to act as a support for the analysed plexiglass beam. The steel beam was excited in
vertical direction by the actuator controlled by HBM Instron system, while the
Fig. 6.2 Schematic view of the plexi beam under vertical, harmonic, kinematic excitations with
four gauges and rotational sensor mounted in its mid span
Fig. 6.3 Illustration of the experimental set-up to measure rotations and strains
Table 6.1 Comparison of measured and calculated amplitudes of rotation rates and strains
Amplitude of strains
Amplitude of Amplitude Calculated (Eq. 6.14
excitation of rotation and column 3 as Calculated
Frequency accelerations rate Measured input) using FEM
[Hz] [m/s2] [deg/s]
1 2 3 4 5 6
4.50 0.5454 2.70 1.45E-05 1.37E-05 1.36E-05
5.25 0.7727 6.88 3.14E-05 3.07E-05 3.05E-05
5.70 0.9655 12.85 5.91E-05 5.36E-05 5.58E-05
analysed, plexiglass beam was clamped at the mid-span of the steel beam (Fig. 6.3).
This solution made it possible to excite kinematic vibrations of the plexiglass beam.
To measure rotations the Systron Donner sensor HZ-100-100 (100 /s range) was
applied. The harmonic motion of the actuator was controlled by Instron system,
while data acquisition was done by multi-channel system “MCG Plus” of Hottinger.
The material data of plexiglass are as follows: density ρ ¼ 1,318.7 kg/m3 and Young
modulus E ¼ 4.51 GPa. In Table 6.1 selected results of this experiment are shown.
The differences among the strains obtained from rotation rate and directly
measured reached about 5–9 % which can be explained by the differences between
simplified, continuous dynamic model of the cantilever beam and its actual exper-
imental realization as well as various forms of noises present in this experiment.
6.4 Reconstruction of Flexural Stiffness of a Beam Under

Harmonic Vibrations from the Rotations of Its Axis
Consider familiar equation of motion of a discrete dynamic system under harmonic

excitations with frequency p [rad/s] and vector of excitation amplitudes P0:
q þ Kd q ¼ P0 ei pt
M€ ð6:15Þ
where superscript ‘d’ indicates that the structure vibrates in a ‘damaged’ state with
its original, stiffness matrix K reduced to Kd due to the accumulated damages,
while the matrix of inertia M stays unchanged. It is also assumed that structural
response is small and out of the resonance frequencies zones, so that damping
effects may be neglected. Such situation occurs for example for the cracked
reinforced concrete structures (e.g. Zembaty et al. 2006). In Fig. 6.4, a fragment
of a beam structure under harmonic excitations is presented. The structure is
divided into finite elements k-1, k, k + 1 etc. The generalized coordinates qi, qi+1,
qi+2, are assumed along the measurement directions and include also rotational
degrees of freedom.
Fig. 6.4 Fragment of a

beam structure under
harmonic excitations with
translational and rotational
degrees of freedom
Equation 6.15 can be solved using familiar algebraic matrix equation with
respect to the unknown response amplitude vector u:

Kd p2 M u ¼ P0 ð6:16Þ
The global stiffness matrix can calculated using the Finite Element Method (FEM)
with contributions from all respective finite elements:
X
n X
n
Kd ¼ i ¼
Kde αi Kue
i ð6:17Þ
i¼1 i¼1
where n is the number of all discretized elements. At this stage one can introduce
the key parameters of this analysis which are non-dimensional stiffness reduction
factors αi which describe relative stiffness loss due to accumulated damages
(0 αi 1). Substituting Eq. 6.17 into Eq. 6.16 one obtains:
!
X
n
αi Kue
i pM
2
u ¼ P0 ð6:18Þ
i
Equation 6.18 can be applied to find the vector of the amplitudes of displacements
u in terms of the vector of driving force amplitudes P0, when the FEM model of the
structure is prepared.
Consider now the norm J measuring difference between vectors of the ampli-
tudes uc calculated using Finite Element Method and the vector of amplitudes um
measured in the actual structure:
!2
X
nd
u cj ðαÞ u mj
J ðαÞ ¼ ð6:19aÞ
j¼1
u mj
where nd is the number of measured displacement amplitudes. Finding minimum of

Eq. 6.19a the vector of stiffness reduction factors α describing the actual state of
damage can be obtained. This reconstruction procedure requires the structure to be
excited and to acquire all the amplitudes along dynamic degrees of freedom of the
structure. Classic FEM analysis of beams and frames defines structural response
usually only in terms of translational coordinates, while the rotational degrees of
freedom are removed from the global stiffness matrix by static condensation. Thus
one can easily include rotational coordinates in the analyses by keeping selected,
required rotational degrees of freedom (not condensing them out). In this case
instead of Eq. 6.19a we will have
!2
X
ndr
u cj ðαÞ u mj
J dr ðαÞ ¼ ð6:19bÞ
j¼1
u mj
where ndr denotes the number of translational and additional, rotational degrees of
freedom. A comparison of the effectiveness of minimization of functionals given by
Eq. 6.19a or 6.19b using Genetic Algorithms and Levenberg-Margquardt local
search (GA L-M) was subject of a detailed numerical analysis in the paper by
Kokot and Zembaty (2009b). When analyzing reconstructions of multiple stiffness
two measures of its effectiveness can be defined:
• Weighted Average Error (WAE):

vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
u ne
uX α a α d 2
WAE ¼ t i i
ð6:20Þ
i¼1
αid
measuring averaged difference between computed and assumed vectors of stiffness

reduction factors (the averaging summation takes place over all the discretized
elements ne).
• Maximum Error (ME) between actual and measured stiffness distribution:

ME ¼ maxαia αid ð6:21Þ
i
The above symbols αi with superscripts a and d denote respectively the ‘assumed’
and ‘detected’ stiffness losses.
6.5 Experiment 2: Reconstruction of Stiffness Variations

of a Cantilever Beam
The tests were carried out using the same experimental set up as previously, but
with two types of plexiglass beams: the ‘intact’ beam and a beam with three drops
of stiffness (Fig. 6.5).
To measure translational vibrations three miniature accelerometers PCB 333B52
were installed underneath the beam. The angular motions were measured using
three rotation rate sensors HZ 100-100 installed on top of the beam. All sensors
were glued at points corresponding to the nodes of the finite element. The mass
losses from the reductions of the beam cross-sections were compensated by gluing
nuts to the beam (see Fig. 6.6).
Fig. 6.5 The geometry of the beam investigated in experiment ‘2’ – in intact and the in the
‘damaged’ state defined by three stiffness reductions
Fig. 6.6 Detail of the beam with angular sensor HZ 100-100 installed at the beam stiffness drop
and the additional, compensating masses (nuts)
Fig. 6.7 Dynamic model of the cantilever beam vibrating under vertical, kinematic, harmonic
excitations
The detailed scheme of the vibrating beam together with all the translational and
rotational degrees of freedom is shown in Fig. 6.7. Two models of the 750 mm long
plexiglass beam were measured and analysed: the first “intact” one, with constant
cross section b ¼ 80 mm and h ¼ 14 mm (see upper part of Fig. 6.5) and the second,
“damaged” one, with decreased cross sections (see lower part of Fig. 6.5), leading
to 15 % drops of stiffness.
The fundamental, natural frequency of the tested cantilever beam (with all the
sensors attached) equalled 6.90 Hz while for the “damaged” one 6.23 Hz. This was
important to know before the experiment started, as the “reconstruction” method
described in Chap. 2, by definition avoids vibrations close to resonances (Kokot and
Zembaty 2009a, b).
The tests were carried out by measuring harmonic vibrations of the plexiglass
beams for various excitation frequencies (from 3.5 to 7.5 Hz), outside the resonance
zones, at low excitation level (amplitude of displacements at the fixed end was
equal to 0.5 mm). The minimization in the stiffness reconstruction process was
carried out using the hybrid optimization procedure described in detail in the
paper by Kokot and Zembaty (2009a). First a specific number of 6,000 generations
in the genetic algorithm allowed to reach the vicinity of the global minimum. Next
the Levenberg-Marquardt local search method was involved to fine-tune the
solution.
Table 6.2 Stiffness reduction factors for a beam with three stiffness reductions under harmonic
kinematic excitations at frequency 4.5 Hz
α1 α2 α3
Actual stiffness reduction 0.85 0.70 0.55
Detected stiffness reduction 0.86 0.67 0.54
Two solutions were studied in detail: with three measured translation accelera-
tions only and with three rotation rates only. Results for excitation frequency
pf ¼ 4.5 Hz and amplitude of the beam tip response equal to about 0.5 mm, are
shown in Table 6.2. The errors of reconstruction equalled: WAE ¼ 2.48 % and
ME ¼ 3.87 %.
6.6 Conclusions
This paper reports the results of early stage, laboratory experiments of the applica-
tion of modern rotation rate sensors to measure and monitor of structural vibrations
in civil and seismic engineering. Following positive recommendations from numer-
ical simulations (Abdo and Hori 2002; Kokot and Zembaty 2009b), both experi-
ments demonstrated the ability of modern rotation rate sensors to be effectively
applied in measuring structural vibrations and in difficult, stiffness “reconstruc-
tions” of practical structural health monitoring.
This is a good prognostic for future application of rotation rate sensors not only
to directly measure rotations of key structural elements, but also to follow strains
and monitor distributed damages of some civil engineering structures. More details
about the reported experiments can be found in the recent paper by Zembaty
et al. (2013).
Acknowledgments This research was supported in part by the Polish NSF Grant N506 289037
“Simulation and experimental investigations of rotation measurements in dynamic identification
of bar structures” and statutory fund of Polish Ministry of Science and Higher Education (NBS
15/14). The Authors wish to thank dr Bronisław Je˛draszak (Opole University of Technology) and
Dariusz Knapek (EC Test Systems, Cracow) for their assistance in carrying out the experiments.
References
Abdo MA-B, Hori M (2002) A numerical study of structural damage detection using changes in
the rotation of mode shapes. J Sound Vib 251:227–239
Chopra AK (2011) Dynamics of structures, theory and application to earthquake engineering.
Prentice Hall, Englewood Cliffs
Kokot S, Zembaty Z (2009a) Damage reconstruction of 3d frames using genetic algorithms with
Levenberg-Marquardt local search. Soil Dyn Earthq Eng 29:311–323
Kokot S, Zembaty Z (2009b) Vibration based stiffness reconstruction of beams and frames by
observing their rotations under harmonic excitations – a numerical analysis. Eng Struct
31:1581–1588
Maeck J, De Roeck G (1999) Dynamic bending and torsional stiffness derivation from modal
curvatures and torsion rates. J Sound Vib 225:153–170
Meydan T (1997) Recent trends in linear and angular accelerometers. Sensors Actuators A Phys
59:43–50
Ndambi J-M, Vantomme J, Harri K (2002) Damage assessment in reinforced concrete beams using
eigen frequencies and mode shape derivatives. Eng Struct 24:501–515
Zembaty Z, Kowalski M, Pospisil S (2006) Dynamic identification of a reinforced concrete frame
in progressive states of damage. Eng Struct 28:668–681
Zembaty Z, Kokot S, Bobra P (2013) Application of rotation rate sensors in an experiment of
stiffness ‘reconstruction’. Smart Mater Struct. doi:10.1088/0964-1726/22/7/077001
Chapter 7
Asymmetric Continuum with Shear
and Rotation Strains Including Quantum
Synchronous Processes
Roman Teisseyre
Abstract We shortly explain why the rotations strains should be included together
with the shear strains in the theory of solid continuum; the related arguments are
based both on the experimental and theoretical levels. In the presented Asymmetric
Continuum Theory the wave propagation mechanism appears in a logical way with
the respective shear and rotation strains. In this logical approach we include, beside
the shear strains, also the rotation strains and their related wave propagation with
shear and rotation strain interactions.
Further, we try to explain how to describe a fracture process with a help of the
quantum theory and with some synchronous motions. The synchronous quantum
processes permit to explain the fracture mechanism and also may include a problem
of the lightning and aurora events.
This attempt might join the quantum processes with a fracture event appearing
inside the elastic continuum with shear and rotation strains and with the synchro-
nous quantum mechanism; this is a new approach joining the improved classic
theory with the quantum events, which lead together to the synchronous processes.
Keywords Quantum process • Shear and rotation strains • Theoretical

investigation • Asymmetric continuum theory • Seismology
7.1 Continuum with Shear and Rotation Strains
In Continuum Asymmetric Theory we may prove only the existence of displace-

ment derivatives. The true displacements would always lead to a slip motions along
some fault, thus, we should underline that the recorded displacements belong only
R. Teisseyre (*)
Institute of Geophysics, Polish Academy of Sciences, Warsaw, Poland

DOI 10.1007/978-3-319-14246-3_7
78 R. Teisseyre
to the displacement derivatives (Fig. 7.1). Further on, we present equations for
deviatoric shear and rotation strains.
For measurements:
∂uk
uk Δxi , i ¼ k 1 ð7:1Þ
∂xi
$
where Δxi is the length of a seismograph rigid platform and displacement, uk , is an
apparent displacement.
The rotation and shear strains – included in our theory, should be measured with
a special recording system.
In order to explain the wave propagation mechanism we can present the time
period sequence of shear and rotation strain maxima; here, we present this sequence
for a radial propagation (Fig. 7.1).
Fig. 7.1 Propagation pattern of the shear and rotation strains; the maxima of the shear and rotation
strains along the sequence of five respective time intervals
7 Asymmetric Continuum and Quantum Processes 79
In the classic definition for the elastic regime we have for strain components: Eik-
total, Eˆ(ik)- shear, and, E˘[ik]- rotation strains:
^ ∂uk
Eik ¼ E^ ðikÞ þ E ½ik ¼ ð7:2aÞ
∂xi
At constant pressure and without external forces we have the relations:
2^ 2^
∂ E^ ðniÞ ∂ E^ ðniÞ
2 2
∂ E ½ni ∂ E ½ni
μ ρ ¼0 , μ ρ ¼0 ð7:2bÞ
∂xs ∂xs ∂t2 ∂xs ∂xs ∂t2
where we remind the summation convention at the repeated indexes:
We may remark that, in general, the rotation strain amplitudes might be quite low,
however, the derivatives of the shear and rotation strains should fulfill the Maxwell-
like relations (Teisseyre 2009, 2011, 2013; Teisseyre and Teisseyre-Jeleńska 2014).
For the energy, related to the shear and rotation strains, we may write at a
constant pressure:
!
X X X
E¼ Sik Eik ¼ 2μ EðikÞ EðikÞ þ E½ik E½ik
i, k i, k
X
2μ EðikÞ EðikÞ , i 6¼ k ð7:3Þ
i, k
as the energy related to rotation strains could be neglected here.

The vector form of the rotation strains, E[ik], could be defined as:

E½i ¼ E½23 , E½31 , E½12 ð7:4aÞ
While the deviatoric shear strains, E^ ðikÞ ¼ EðikÞ 13δik EðssÞ , could be also
presented in a vector form in this specific coordinate system:
1
E^ ðiÞ ¼ E^ ð23Þ , E^ ð31Þ , E^ ð12Þ , for E^ ðikÞ ¼ EðikÞ δik EðssÞ ð7:4bÞ
3
However, in the system {x, ict} we may present this relation using the basic 4D
invariant Dirac tensors:
2 3 2 3
0 0 0 1 0 0 1 0
6 0 7 6 0 1 7
6 0 1 0 7 6 0 0 7
γ1 ¼ 6 7 , γ2 ¼ 6 7,
4 0 1 0 0 5 4 1 0 0 0 5
1 0 0 0 0 1 0 0
2 3 2 3 ð7:5aÞ
0 0 0 1 1 0 0 0
6 0 07 60 1 0 0 7
6 0 1 7 6 7
γ3 ¼ i6 7, γ4 ¼ i 6 7
4 0 1 0 05 4 0 0 1 0 5
1 0 0 0 0 0 0 1
80 R. Teisseyre
2 3
0 1 0 0
6 1 0 0 0 7
γ4 γ2 γ3 ¼ 6
4 0
7 ð7:5bÞ
0 0 1 5
0 0 1 0
Thus, these complex deviatoric shear strains with a help of these Dirac tensors
(7.5a, 7.5b) can be presented as follows::
EðλκÞ ¼ Eð1Þ γ1 þ Eð2Þ γ2 þ Eð3Þ γ4 γ2 γ3 ð7:6aÞ
where for the 4D strain tensor, E(λκ), we shall write:

2 3
0 Eð3Þ Eð2Þ Eð1Þ
6 Eð3Þ 0 Eð1Þ Eð2Þ 7
EðλκÞ ¼6
4 Eð2Þ
7 ð7:6bÞ
Eð1Þ 0 Eð3Þ 5
Eð1Þ Eð2Þ Eð3Þ 0
The related wave equations for these strains, Eλκ ¼ EðλκÞ þ iE½λκ , with the velocity,
pffiffiffiffiffiffiffiffi
κ V μ=ρ, can be presented as follows:
c
2 2
∂ Eλμ κ2 ∂ Eλμ
2 2 ¼ Y λμ $ WEλμ ¼ Y λμ ð7:7aÞ
∂xn ∂xn c ∂t
^
Of course, we may return to the equations in a vector form with Ei ¼ E^ ðiÞ þ i E ½i
(with the similarly defined external vector, Ys):
2 2
∂ κ2 ∂
E s 2 2 Es ¼ Y s , or WEs ¼ Y s ð7:7bÞ
∂xn ∂xn c ∂t
This relation for the complex vectors, can lead us to the Maxwell-like equations
for V ¼ κc, and current JS ¼ 0:
^ ^
∂ E ½q 1 ∂E^ ðsÞ ∂E^ ðqÞ 1 ∂ E ½s
εspq ¼ 0, εspq þ ¼0 , or :
∂x p V ∂t ∂x p V ∂t
^ rffiffiffi ð7:8Þ
^ χ ∂E^ χ ∂E c μ
rot E ¼ 0 , rot E^ þ ¼0 ; ¼V¼
c ∂t c ∂t χ ρ
These relations are quite similar to those for the electro-magnetic fields; the
difference relates only to propagation velocity, which for the electro-magnetic
fields relates to the light velocity.
The presented expression (7.8) leads directly to the Eqns. 7.7a, b.
7.2 Synchronous Quantum Processes
The basic element of quantum theory, at the thermodynamic equilibrium, is the

black body radiation described by the Planck relation:
1 1
8πνh νh 8πν5 h νh
eðλÞ ¼ 4 exp 1 or eðνÞ ¼ exp 1 ð7:9Þ
λ KT c4 KT
where, e, is the energy radiated per unit wave length, λ; ν means frequency, ν ¼ c=λ,
and h is the Planck constant; K – the Boltzmann constant and T – temperature, c –
light velocity.
This energy can be related to the resonance frequency; the classical question of
the black body radiation, that is, the radiation in thermodynamic equilibrium-
fundaments of quantum mechanics.
Of course the synchronous quantum related processes appear in all radiation
processes; the most basic phenomena in our world. However, here we will consider
only these synchronous quantum processes which appear due to the deformation
fields, that is related to the shear and rotation strains, and which lead to the Planck
black body radiation. The resonance frequency means here the minima of fracture
resistances at thermodynamic equilibrium, Fig. 7.2. At this conditions the basic unit
energy, hv, and related black body radiation (7.9) may approach us to some
synchronous processes. First we may imagine that, a distribution of the high
shear strains forming the local micro-fractures, could be approximated by the
parallel and local micro-displacements; thus, searching their continuum descrip-
tion, we might present these micro-displacements as situated along a group of the
parallel lines. Thus, in this way, we may equivalently describe the micro-
displacements, formed due to the shear strains, as a dense group of a continual
distribution of the parallel lines.
Fig. 7.2 Resonance frequency, or frequencies, at thermodynamic equilibrium: these maxima

mean the minima of fracture resistances
82 R. Teisseyre
The presented way leads us from the parallel micro-displacements up to their

line continuum distributions forming the dense group; further in this way, we may
achieve a continuum description of these parallel micro-events.
Thus, we may assume that before any fracture we might have a distribution of
some molecular-displacements along the parallel lines processes, in such way we
may approach to a fracture line and then to a fracture plane. Here, we should remind
that the strain energy (Eq. 7.3) at a constant pressure is related to the sum of the
strain squares, but may be limited only to the shear strains (7.3).
We may present the energy point concentrations along the distributed points at
the almost parallel lines being formed before a fracture. Such energies concentrated
along these lines appear due to the precursory processes and are related mainly to
the shear strains:
X
E^ ðskÞ ! E^ ðSkÞ - a sum related to parallel lines ð7:10Þ
S
which means a concentrations ofXstrains along the almost parallel lines, S, forming a
group leading to a plane, a sum or the parallel lines forming a big group around
S
the central fracture pass plane.
The expression for such energy concentration may be presented as concentration
of the shear strains:
!
X X X^ ^
E¼ Sik Eik ¼ 2μ E^ ðikÞ E^ ðikÞ þ E ½ik E ½ik
i, k i, k
X
2μ E^ ðikÞ E^ ðikÞ ð7:11Þ
i, k
and from the another side, this basic energies along the mentioned above synchro-
nous lines concentrated along an expected fracture plane can be presented as
follows:
2 32
X
xð0s 1
6 7 8πνh νh
E¼4 GSk exp gs xs x0s dx0s 5 ½expfgðt t0 Þg 2 4 exp 1
S, k λ KT
x0s
ð7:12aÞ
where, in order to form the energy concentration lines and further the fracture along
some central plane element, we have introduced the summations along both these
lines, S, and also along the perpendicular direction, k; a possible number of
concentrated lines relate to a space element, that is both to the S and k directions;
thus, the expected fracture plane may be related to the S0 and k directions.
Such approach is introduced to better understand formation of the fracture plane

element, however, a change of the related coefficients, GSk, gS and g, up to the limit,
G0S , g0S , g0 (fracture level), should be observed by the changes of some other
related fields, e.g., electric or magnetic ones.
Thus, when the coefficients increase to a fracture limit, Gs ! G0s and
gs ! g0s , g ! g0 , we arrive to a fracture event, with the concentrated energies

decreasing rapidly, when we drive apart from the fracture domain, X0s , X0s ; t0 :
2 32
ðs 1
0
X
6X 0 0

0
07
0 2 8πνh νh
E0 ¼ 4 GSk exp gs xs xs dxs 5 exp g ðt t0 Þ exp 1
S, k λ4 KT
X0s
ð7:12bÞ
In this way, we describe an increase of the fracture energy in the time and space
up to the final values related to these fracture coefficients; the resonance frequency,
at thermodynamic equilibrium, presents the minimum of fracture resistance.
We assume that these considered phenomena, related to the electron motions,
could remain synchronous in the adequate time and space domains. Moreover, a
total fracture processes might be formed with the time/space synchronous series
one after other.
The introduced synchronization means the synchronous fracture formation; a
real fracture motion starts, when a density of quantum synchronous motions reach
the critical level.
We should remember that the Planck quantum theory introduces the resonance
wave frequency related to the black body radiation effect; basing on this approach
we might try to consider a material strength minimum, as related to such resonance
frequency, under the static vibration parts of load. We may illustrate the possible
resonance frequencies, at thermodynamic equilibrium in Fig. 7.2; in the presented
figures we present the reverse values needed for fracture, that is the reverse values
of the fracture energies, at the synchronous fracture phenomena. Thus, the vertical
axis relates the decrease of the fracture resistance.
Thus, we present a minimum of energy required for the fracture at a thermody-
namic equilibrium. An individual fracture process at the quantum level may lead to
the synchronous and collective fracture events at the critical fracture level: G0s ,
g0S , g0.
Of course, the released energy leads also to the changes of the strains. We should
remind the strain equations (Eq. 7.3), shear and rotational, Eˆ(ni) , E˘[ni], and related
energies.
84 R. Teisseyre
7.3 Precursory Processes
We assume, that at the fracture processes, we should have for their synchronous
motions at the different positions in space and time, thus, there should appear a
similarity of these motions. Therefore, we may deal there with some similarities
related to the central line, along which we expect a fracture process.
Of course, these processes should be governed by the Shr€odinger equation in its
more adequate forms, which mean here the motions related to a center of synchro-
nous field density; this process may lead to the probable fracture. Thus, especially
important for prediction methods, could be the mentioned processes appearing in
time and space domain.
For fracture, before and after the given event, we may have an appearance of
chain of some fields governed by the Shr€odinger like equations; specially important
might be, here, the electric and magnetic signals.
As mentioned above, a real fracture starts when the quantum synchronous
density reaches the critical level. However, already before some fracture event,
there can appear also the collective electron motions under even not great synchro-
nous concentration. In this way there can appear some electric and magnetic pre-
cursors, governed by these quantum processes and joined to the Maxwell relations.
We can believe that some electric or magnetic signals may appear due to these
energy releases appearing at the beginning of these synchronous processes, thus,
even beneath the critical concentration.
We assume that the considered phenomena related to the electron motions could
remain synchronous in the adequate time and space domains. Of course, a total
fracture process may be formed with the time/space synchronous series appearing
one after other with much smaller velocities.
In this way we may obtain also the space/time fracture center at t ¼ t0 , xs ¼ x0s
(which even may move), and where this synchronization means a synchronous
fracture formation; a real fracture motion starts when a density of quantum syn-
chronous motions reach the critical density: GCs .
We may remark that, in general, the rotation strains might be quite low,
however, the derivatives of the shear and rotation strains should fulfill the
Maxwell-like relations (Teisseyre 2013; Teisseyre and Teisseyre-Jeleńska 2014).
In the classic definition for the elastic regime we should remained the strain
energy concentration and the line synchronous energy concentration (7.12a, 7.12b)
For a search related to the energy concentrations we should pay an attention to
the electric and magnetic fields; the related joined expressions may be write as
follows (Teisseyre and Teisseyre-Jeleńska 2014):
rffiffiffiffiffiffiffiffiffiffiffiffiffi
X
SðklÞ ¼ λδkl Ess þ 2μEðklÞ þ ϑδkl G þ ϑGðklÞ þ δkl GE E2k
k ð7:13Þ
S½kl ¼ 2μE½kl þ ϑεkls Gs þ ϑG½kl , ϑ ¼ ð1, iÞ
where for the dielectric materials, we should consider the electrostriction, GE,
related to the random electrical domains; ϑ ¼ ð1, iÞ
The most possible relations between these fields could be combined together in
the following way:

rffiffiffiffiffiffiffiffiffiffiffiffiffi
X
α SðklÞ λδkl Ess 2μEðklÞ ¼ χ ðsklÞ Es þ f ðsklÞ Π s þ d iðklj Þ Π i j þ gδkl GE E2k

k
η S½kl 2μE½kl ¼ χ ½skl Ms þ f ½skl Π s þ d i½klj Π i j
ð7:14Þ
where we have introduced the unknown parameters:
α, χ ðsklÞ , f ðsklÞ , diðklj Þ , g and η, ξ½skl , f ½skl , di½klj
However again, this expression could be written in the more general forms with a
number of independent equations, with α ! α ¼ α1 þ α2 þ α3 þ α4 (see
Eq. 7.14), and η ! η1 þ η2 þ η3 (see Eq. 7.14), as follows:

α1 SðklÞ λδkl Ess 2μEðklÞ ¼ χ ðsklÞ Es , η1 S½kl 2μE½kl ¼ χ ½skl Ms

α2 SðklÞ λδkl Ess 2μEðklÞ ¼ f ðsklÞ Π s , η2 S½kl 2μE½kl ¼ f ½skl Π s

ð7:15Þ
α3 SðklÞ λδkl Ess 2μEðklÞ ¼ diðklj Þ Π i j , η3 S½kl 2μE½kl ¼ d i½klj Π i j

rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
X
α4 SðklÞ λδkl Ess 2μEðklÞ ¼ gδkl GE d E2k
k
where we remind the summation convention at the repeated indexes.

These relations may present the origin of the different waves, which propagate
according to the Maxwell–like equations, but with the different velocities: ones
related to the strain fields and others to the electro/magnetic fields. These separate
waves might inform us on their common source processes, but such task can be
quite difficult. Here, some help may be related to the fact that the related signals,
shifted in time due to the differences between these wave velocities, should present
some similarities and should have similar frequencies. Thus, such task might
inform us on the source processes; nevertheless, we remark that in the most cases
it could be very difficult to achieve the positive results. Here, first of all, we should
mentions the papers and achievements of the prof. Varotsos group (see some related
references, Varotsos et al. 1986, 2006, 2011).
References
Teisseyre R (2009) Tutorial on new development in physics of rotation motions. Bull Seismol Soc
Am 99(2B):1132–1136
Teisseyre R (2011) Why rotation seismology: confrontation between classic and asymmetric
theories. Bull Seismol Soc Am 101(4):1683–1691
86 R. Teisseyre
Teisseyre R (2013) Molecular Transport in in Fracture Processes. Acta Geophys 61(1):18e–25

Teisseyre R, Teisseyre-Jeleńska M (2014) Asymmetric continuum: extreme processes in solids
and fluids. Springer, Berlin/Heidelberg, p 180
Varotsos PA, Alexopoulos KD (1986) Thermodynamics of point defects and their relation with
bulk properties. North-Holland, Amsterdam, p 474
Varotsos PA, Sarlis NV, Skordas ES, Tanaka HK, Lazaridou MS (2006) Entropy of seismic
electric signals: analysis in natural time under time reversal. Phys Rev E 73:031114
Varotsos PA, Sarlis NV, Skordas ES (2011) Natural time analysis: the new view of time. In:
Precursory seismic electric signals, earthquakes and other complex time-series. Springer,
Berlin/Heidelberg, p 205
Part II
Seismic Analysis and Design of Irregular
Structures
Chapter 8
Seismic Assessment of RC Frame Buildings
Klemen Sinkovič, Iztok Peruš, and Peter Fajfar
Abstract In the chapter, the N2 method and the non-linear dynamic analysis
(NDA) are applied for the assessment of seismic performance of two variants of a
4-storey reinforced concrete (RC) frame building. The first variant is represented by
the bare frames, whereas in the second variant infill panels are included in the upper
three storeys, thus creating a soft first storey structure. In one direction, the results
clearly show the detrimental effect of infills which generate a storey mechanism
instead of a global mechanism which occurs in the case of the bare frame structure.
In the other direction, a story mechanism occurs also in the case of the bare frame
structure. The infills shift the mechanism from the third to the first storey. There is
good agreement of the results obtained by the N2 method and NDA. The results
indicate that both variants of the building are able to survive the design ground
motion.
Keywords Seismic performance assessment • Nonlinear analysis • Seismic

demand • Structural capacity • Reinforced concrete frame buildings
8.1 Introduction
For the assessment of seismic performance of buildings subjected to strong earth-

quakes, nonlinear analysis is needed. The most complex analysis method is the
non-linear dynamic analysis (NDA), which is used mainly in research and for
analysis of some important structures. It requires several analyses with different
accelerograms and additional data on the mathematical model (hysteretic and
damping models). For practical application, pushover-based methods, e.g. the N2
method (Fajfar 2000), implemented in Eurocode 8 – Part 1 (EC8-1, CEN 2004a),
are better suited. In this chapter both the NDA and the N2 method are used for the
seismic performance assessment of a four-storey reinforced concrete (RC) frame
K. Sinkovič (*) • I. Peruš • P. Fajfar

Faculty of Civil and Geodetic Engineering, University of Ljubljana, Jamova cesta 2,
1000 Ljubljana, Slovenia
e-mail: [email protected]; [email protected];
[email protected]

DOI 10.1007/978-3-319-14246-3_8
90 K. Sinkovič et al.
structure, representing an existing building built before modern seismic codes have
been adopted. In the first variant, the bare frame structure is analysed, whereas in
the second variant infill panels are included in the upper three storeys, thus creating
a soft first storey structure. In the chapter the main results are presented and
discussed.
8.2 Methodology
The expected seismic performance of a building is estimated by comparing the

demand and the capacity of the structure.
In the case of the N2 method, the determination of demand has been well
established. If the influence of higher modes in elevation and in plan (torsion) is
not important, the basic version of the method, implemented in EC8-1, can be used.
By performing the pushover analysis up to the target displacement, determined
according to a well-known procedure, seismic demand for any relevant quantity can
be determined.
On the other hand, for the time being, a widely accepted approach for the
assessment of the capacity at the level of the whole structure does not exist. Existing
standards and codes do not provide a clear definition of the seismic capacity of the
structure. Usually, it is conservatively assumed that the structure fails when the near
collapse (NC) limit state is attained. This limit state does not represent a physical
collapse of the building (which is extremely difficult to predict) but a complete
economic failure. An option for the definition of the NC limit state at the level of the
structure is a similar definition as in the case of individual elements, i.e. at a 20 %
drop of the lateral resistance of the structure. However, this definition, which seems
to be the most appropriate, cannot be applied in nonlinear dynamic analysis and
also not in a pushover analysis with simplified models, e.g. in the case of models
without strength-degradation. A more practical definition is based on the assump-
tion that the NC limit state of the structure is reached when the first important
vertical element reaches the NC limit state. This definition of capacity was used in
our study. The capacity of the structure in terms of roof displacement and storey
drift was determined based on the ultimate rotation capacity (i.e. the NC limit state)
of the most critical column. The relation between different quantities corresponds
to the results of the pushover analysis at the moment when the ultimate rotation of
the most critical column is attained.
For the ultimate rotation the empirical expressions provided in the current
version of Eurocode 8 – Part 3 (EC8-3, CEN 2005) for the ultimate chord rotation
θum were used. Originally, the formulas were proposed by Panagiotakos and
Fardis (2001).
8 Seismic Assessment of RC Frame Buildings 91
θum ¼ 0:016 ð0:3ν Þ

0:225 0:35
maxð0:01; ω0 Þ Lv αρsx
f yw
f min 9; 25 fc 1:25100ρd ð8:1Þ
maxð0:01; ωÞ c h
For the meaning of symbols and for different reductions of capacity for elements
designed according to out-of-dated codes please consult EC8-3. Since best esti-
mates are used, no safety factor γ el is included in Eq. 8.1. According to the current
version of EC8-3, for the RC members without detailing for earthquake resistance,
the value in Eq. 8.1 is divided by a factor of 1.2. In members with smooth (plain)
longitudinal bars without lapping in the vicinity of the end region where yielding is
expected, the ultimate chord rotation θum is further multiplied by a factor of 0.8. For
columns, the resulting values of θum are between about 3 % and 4 %.
In the case of the NDA, a set of 30 accelerograms was selected with the mean
elastic acceleration spectrum similar to the EC8-1 spectrum, used in the N2 analysis
(Vukobratović and Fajfar 2015, Fig. 8.2). Seismic assessment is based on the same
quantities as in the case of the N2 analysis. Seismic demand is represented by
median values determined by NDA. For the determination of seismic capacities
based on NDA an incremental dynamic analysis (IDA) is needed. Since this is an
extremely time consuming procedure, in our study the same capacities as in the case
of N2 were used also in the case of NDA.
8.3 Assessment of Seismic Performance of 4-storey RC

Building
8.3.1 Description of the Example Building and Mathematical

Modelling
The example structure is a 4-storey RC building that was designed on the basis of
the design practice and codes which were used in Italy between the 1950s and the
1970s (Celarec et al. 2012). The structure is double-symmetric with storey height
3.0 m. In Fig. 8.1 the global geometry and the cross-sections of the building are
presented. Note that the cross-sections of the columns are decreasing along the
height of the building. The masses amount to 182.3 t in the first three storeys, and
193.3 t in the top storey. The total weight of the building is W ¼ 7261 kN. The
variants of the building without infills and with them (in the upper three storeys)
were analysed.
The mean value of concrete compressive strength amounted to fcm ¼ 33 MPa and
the yield strength of the steel reinforcement amounted to fym ¼ 370 MPa (smooth
reinforcement). In RC columns shear reinforcement Asw/s ¼ Φ6/30 cm was used,
which was at the time of building construction a usual requirement for shear
Fig. 8.1 The elevation views, the plan view of the second storey and typical reinforcement of
beams and columns of the example building
reinforcement. The amount of longitudinal reinforcement in the individual columns

and beams is shown in Fig. 8.1.
All the analysis were performed with OpenSees (2013), using the PBEE-toolbox
(2010), which is a simple yet effective tool for the seismic analysis of RC frames by
using simplified non-linear models. A simplified non-linear 3D model of RC frame
structure was made. The effective width of the beams was considered according to
Eurocode 2 (EC2, CEN 2004b). Fixed supports of columns at the bottom level were
assumed. The RC slabs were assumed to be rigid in their planes and completely
flexible out of plane. Masses were concentrated in the centre of mass at each storey.
Accidental eccentricity was not taken into account. Second order (P – Δ) effect was
considered.
The inelastic flexural behaviour of beams and columns was modelled by
one-component lumped plasticity elements. Plastic hinges were defined by
moment-rotation relationship schematically shown in Fig. 8.4b). For beams, the
plastic hinge was used for major axis bending only. For columns, two independent
plastic hinges for bending about the two principal axes were used. The interaction
between axial force and bending moment was not considered. Yield moment My
was determined for each element by analysing the cross-section of the element.
Fig. 8.2 The elastic acceleration spectrum for ag ¼ 0.29 g and soil type B, spectra for individual
accelerograms and mean spectrum
Cracked sections were modelled according to EC8-1 by using the elastic flexural
properties of elements equal to one-half of the corresponding stiffness of the
un-cracked elements. The ultimate rotation θum was determined according to
Eq. 8.1.
The infills were modelled by means of diagonal struts. The details of the
modelling do not influence the results.
For NDA, all analyses were performed on the same MDOF structural model as
the N2 analysis. Stiffness degrading hysteretic behaviour was used (material »Hys-
teretic« in OpenSees). The parameter β, which controls the unloading stiffness, was
equal to 0.5. Rayleigh damping (ξ ¼ 5 %), proportional to mass and instantaneous
stiffness was applied (considering the first two vibration periods).
In the mathematical model used in this study, it was assumed that the elements
do not fail in shear. In order to confirm this assumption, the results should be
checked in order to identify possible shear failures of elements, especially columns.
The discussion on shear capacity is out of the scope of this chapter. However, it
should be mentioned that, in the case of the investigated building, all elements
passed the check.
Seismic demand is defined by EC8-1 for the design ground acceleration
ag ¼ 0.29 g, for soil type B (resulting in peak ground acceleration PGA ¼ 0.35 g)
and for 5 % damping (Fig. 8.2).
8.3.2 Seismic Performance Assessment
The analyses were performed for two variants of the building, each for two
directions of ground motion. The N2 analysis for the Y direction of the bare
Fig. 8.3 The pushover curves with indicated demand and capacity
frame structure is shown more in details, whereas for the other analyses only the
results are shown in Figs. 8.3 and 8.5.
In the N2 method, the lateral loads for the pushover analysis were determined
based on the first mode shape. The pushover curve for the bare frame building in the
Y direction is shown in Fig. 8.3b), where the NC limit state of the building, defined
as the state when the first column attains its NC limit state, is marked. The results of
the pushover analysis indicate that the interior column C4 in the first storey is the
first column that fails in flexure (θum ¼ 0.031) at the roof displacement
DNC ¼ 29.2 cm. The corresponding storey drift ratio of the first storey is equal to
δNC.1 ¼ 9.5 cm/300 cm ¼ 0.032. The spectral acceleration corresponding to the
capacity amounts to Sa.NC ¼ 1.25 g, and PGA ¼ 0.86 g (see Fig. 8.4a). Several
columns in the first three storeys fail soon after the failure of the first column, so
the detailed results are sensitive to the details of the modelling.
Demand is based on the EC8-1 elastic spectrum presented in Fig. 8.2. The
fundamental effective period amounts to Teff ¼ 0.86 s, resulting in the elastic
spectral acceleration Se ¼ 0.50 g. The target displacement of the SDOF system Dt,
SDOF ¼ 9.3 cm is defined by the intersection point of the line representing the period
and the elastic demand spectrum (see Fig. 8.4a). The corresponding target roof
displacement of the MDOF structure is equal to Dt:MDOF ¼ Dt:SDOF Γ ¼ 9:3 cm
1:26 ¼ 11:8 cm (Γ ¼ 1.26), whereas the storey drift ratio of the first storey is equal
to δt:1 ¼ 3:5 cm=300 cm ¼ 0:012. The distribution of storey drifts over the height of
the structure is shown in Fig. 8.5. A global plastic mechanism over the first three
stories is formed. A comparison of capacity and demand in terms of PGA, roof
displacement and storey drift ratio suggest that the building is able to survive the
design earthquake (Table 8.1). The storey drift ratio is shown for the first storey,
where the critical column is located. Very similar values apply for the second and
the third storey.
Fig. 8.4 (a) Comparison of capacity and demand for the equivalent SDOF system in AD format
and (b) the damage in plastic hinges at the NC limit state for bare frame building in Y direction
Storey BARE FRAME Storey FRAME with INFILLS

4 (X direction) 4 (X direction)
3 3
2 2
1 1
δX [%] δX [%]
0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Storey BARE FRAME Storey FRAME with INFILLS
4 (Y direction) 4 (Y direction) acc.
16th
84th
3 3 N2
MEAN
MEDIAN
2 2
1 1
δY [%] δY [%]
0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
Fig. 8.5 Seismic demand in terms of storey drift ratios
The results for other three analyses, presented in Figs. 8.3 and 8.5, demonstrate
that in all three cases storey mechanisms are formed. In the case of the bare frame
structure in X-direction, the critical storey is the third storey (due to the change in
the stiffness of columns). In the case of the infilled frames with the open first storey
the mechanism forms in the first storey, as expected.
Table 8.1 Comparison between capacity (C) and demand (D) in terms of roof displacements and
storey drift ratios (for third storey in the case of bare frame in X direction and for the first storey in
other cases)
Bare frame Frame with infills
C D C/D C D C/D
Analysis X direction
N2 PGA [g] 0.49 0.35 1.40 0.83 0.35 2.39
Roof disp. [cm] 12.4 8.9 1.40 9.5 3.7 2.55
St. drift ratio [%] 3.8 2.6 1.47 3.1 1.1 2.68
NDA Roof disp. [cm] 12.4 7.9 1.56 9.5 3.2 2.96
St. drift ratio [%] 3.8 1.9 1.91 3.1 1.0 3.10
Analysis Y direction
N2 PGA [g] 0.86 0.35 2.48 0.52 0.35 1.48
Roof disp. [cm] 29.2 11.8 2.48 9.8 6.6 1.48
St. drift ratio [%] 3.2 1.2 2.73 3.1 2.1 1.51
NDA Roof disp. [cm] 29.2 11.0 2.64 9.8 6.2 1.57
St. drift ratio [%] 3.2 0.9 3.61 3.1 1.9 1.62
Mean and median values of displacements and storey drift ratios were deter-
mined by NDA. For the demand in terms of storey drift ratios, the results of
individual analyses, the mean and the median values, and the values of 16th and
84th percentiles are shown in Fig. 8.5.
The comparisons between capacity and demand are presented in Table 8.1. In
the case of NDA, the same values for capacity were considered as in the case of the
N2 method. Median values were used for demand.
8.4 Discussion and Conclusions
In the Y direction, the results indicate a detrimental effect of infills, which generate
a storey mechanism instead of a global mechanism which occurs in the case of the
bare frame structure. In the X direction, a story mechanism occurs in both variants
of the structure. The third storey is critical for the bare frame (due to a change of
stiffness and strength), whereas in the case of the partially infilled structure the
mechanism occurs in the first storey. In the X direction, the influence of infills is
beneficial. In all cases the capacity is larger than the demand corresponding to the
design ground motion. However, it should be noted that average values for all
quantities were used in analyses. If uncertainties were taken into account (by safety
factors), the bare frame structure in X-direction and the partially infilled structure in
Y direction would not pass the check.
There is a good agreement of the results obtained by the N2 method and NDA.
Both methods are able to demonstrate the potential weaknesses of the structure.
Acknowledgements The results presented in this chapter are based on work continuously
supported by the Slovenian research Agency. This support is gratefully acknowledged.
References
Celarec D, Ricci P, Dolšek M (2012) The sensitivity of seismic response parameters to the
uncertain modelling variables of masonry-infilled reinforced concrete frames. Eng Struct
35:165–177
CEN (2004a) EN 1998-1:2004 Eurocode 8: design of structures for earthquake resistance – part 1:
general rules, seismic actions and rules for buildings. European Committee for Standardisation,
Brussels, Belgium
CEN (2004b) EN 1992-1-1:2004 Eurocode 2: design of concrete structures – part 1-1: general
rules and rules for buildings. European Committee for Standardisation, Brussels, Belgium
CEN (2005) EN 1998-3:2005/AC:2010 Eurocode 8: design of structures for earthquake resistance
– part 3: assessment and retrofitting of buildings. European Committee for Standardisation,
Brussels, Belgium
Dolšek M (2010) PBEE toolbox user’s manual. The Institute of Structural Engineering, Earth-
quake Engineering and Construction IT (IKPIR). Available via: http://ice4risk.slo2projekt.
info/publications.html
Fajfar P (2000) A nonlinear analysis method for performance-based seismic design. Earthq
Spectra 16(3):573–592
OpenSees (2013) Open system for earthquake engineering simulation. Pacific Earthquake Engi-
neering Research Center (PEER). Available via: http://opensees.berkley.edu/
Panagiotakos TB, Fardis MN (2001) Deformations of reinforced concrete members at yielding and
ultimate. ACI Struct J 98(2):135–148
Vukobratović V, Fajfar P (2015) A method for the direct determination of approximate floor
response spectra for SDOF inelastic structures. Bull Earthq Eng 13(5):1405–1424
Chapter 9
Torsional Index of an Asymmetric Building
Based on Mode Shape
Kenji Fujii
Abstract It is essential to carry out three-dimensional analyses considering all

possible directions of seismic input, when conducting seismic assessment of an
asymmetric building. For this purpose, the author has proposed a simplified proce-
dure to predict the largest peak seismic response of an asymmetric building
subjected to horizontal bi-directional ground motion acting at an arbitrary angle
of incidence. In the proposed simplified procedure, the largest peak response is
predicted from the response of two independent equivalent single-degree-of-free-
dom models representing the first and second modes and combination of pushover
analyses. However, it is unclear which conditions an asymmetric building should
satisfy for the applicability of the proposed procedure. In this study, a torsional
index is defined based on each mode shape, and the applicability of the proposed
procedure is discussed based on the torsional index. The advantages of the proposed
torsional index are as follows: (a) it is related to the effective (equivalent) modal
mass ratio which represents the contribution of each modal response to the whole
response, and (b) it is easily extended for an multi-storey asymmetric building, and
(c) there is the clear relationship between torsional indices of two different modes
and the angle between the principal directions of two modes. The numerical results
show that the proposed simplified procedure is at least applicable to torsionally stiff
systems for both orthogonal directions, while it is questionable for torsionally
flexible systems.
Keywords Asymmetric building • Torsional index • Eigenvalue analysis •

Effective modal mass ratio • Principal direction of modal response
K. Fujii (*)
Department of Architecture and Civil Engineering, Chiba Institute of Technology,
2-17-1 Tsudanuma, Narashino, Chiba 275-0016, Japan

DOI 10.1007/978-3-319-14246-3_9
100 K. Fujii
9.1 Introduction
In designing of new building for earthquake resistance or when conducting seismic

assessment of an asymmetric building, horizontal ground motion is applied to each
of the main orthogonal axes of building. However, this procedure may be inade-
quate because the most critical direction of incidence of the seismic input, which
would produce the largest response, may be different from the direction of the
building’s main orthogonal axis and the major component of the ground motion
may act in any direction. Therefore it is essential to carry out three-dimensional
analyses considering all the possible directions of seismic input. For this purpose,
the author has proposed a simplified procedure to predict the largest peak seismic
response of an asymmetric building subjected to horizontal bi-directional ground
motion acting at an arbitrary angle of incidence (Fujii 2014). In this proposed
procedure, the largest peak response of an asymmetric building is predicted based
on (a) peak responses of two independent equivalent single-degree-of-freedom
(SDOF) models representing the first and second modal response and
(b) combination of pushover analyses considering the bi-directional excitations.
There are two critical assumptions in the procedure (Fujii 2014). The first is that
the building oscillates predominantly in a single mode in each set of orthogonal
directions, and the second is that the principal directions of the first and second
modal responses are almost orthogonal. However, it is unclear which conditions an
asymmetric building should satisfy for the applicability of the procedure.
In this study, a torsional index is defined based on each mode shape, and the
relationship of the parameters of the asymmetric building model (eccentricity ratio,
radius ratios of gyration torsional stiffness) to the proposed torsional index is
investigated. The applicability of the procedure (Fujii 2014) is discussed based on
the torsional index.
9.2 Definition of the Torsional Index Based on Mode Shape
9.2.1 Natural Modes of a Single-Storey Asymmetric Building

Model
The building model investigated in this study is a simple, single-storey asymmetric

building model, shown in Fig. 9.1. The equation of motions for undamped free
vibration of this model is shown as Eq. 9.1, and rotational displacement z (¼ rθ) is
used instead of rotational angle θ.
9 Torsional Index Based on Mode Shape 101
Fig. 9.1 Single-storey asymmetric building model
Fig. 9.2 Natural mode of a single-storey asymmetric building model
8 9 2 38 9 8 9
< €x = ω0X 2 0 ω0X 2 EY < x = < 0 =
€y þ 4 0 ω0Y 2 ω0Y 2 EX 5 y ¼ 0 ; ð9:1Þ
: ; : ; : ;
€z ω0X 2 EY ω0Y 2 EX ω0θ 2 z 0
ω0X 2 ¼ K X =m, ω0Y 2 ¼ K Y =m, ω0θ 2 ¼ K θ =I, EX ¼ eX =r, EY ¼ eY =r; ð9:2Þ
where ω0X, ω0Y are the uncoupled natural circular frequencies of translational
oscillation in the X and Y directions, respectively, ω0θ is the uncoupled natural
circular frequency of rotational oscillation with respect to the centre of mass G (not
to the centre of stiffness C), and EX, EY are the eccentricity ratio, respectively. The
natural mode of a single-storey asymmetric building model φi, obtained from
eigenvalue analysis, is shown in Fig. 9.2.
φ i ¼ f ϕX ϕY ϕZ gT : ð9:3Þ
In Fig. 9.2, Oi (ρXi, ρYi) is the centre of rotation of the ith mode determined from
Eq. 9.4, and therefore the distance from Oi to G is expressed by Eq. 9.5.
ρXi ¼ r ðϕYi =ϕZi Þ, ρYi ¼ r ðϕXi =ϕZi Þ; ð9:4Þ

102 K. Fujii
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ρi ¼ ρXi 2 þ ρYi 2 ¼ r ϕXi 2 þ ϕYi 2 =jϕZi j: ð9:5Þ
Considering a set of orthogonal i-i0 axes in the X-Y plane, where the i-axis is the
principal axis of the ith modal response, the tangent of ψ i, the angle of incidence of
the i-axis with respect to the X-axis, is determined from Eq. 9.6.
tan ψ i ¼ ϕYi =ϕXi : ð9:6Þ
The effective (equivalent) modal mass ratio of ith mode in the X- and Y-directions,
miX* and miY*, respectively, is determined from Eq. 9.7.
ϕXi 2 ϕYi 2
miX * ¼ , miY
*
¼ : ð9:7Þ
ϕXi 2 þ ϕYi 2 þ ϕZi 2 ϕXi 2 þ ϕYi 2 þ ϕZi 2
9.2.2 Torsional Index of Each Mode
Based on the equations developed above, the torsional index based on the mode
shape is formulated. The effective modal mass ratio ith mode in its principal
direction, mi*, is determined from Eq. 9.8.
ϕXi 2 þ ϕYi 2
mi * ¼ miX * þ miY * ¼ : ð9:8Þ
ϕXi 2 þ ϕYi 2 þ ϕZi 2
Equation 9.8 can be rewritten in form of Eq. 9.9, considering Eq. 9.5.
1 1
mi * ¼ ¼ : ð9:9Þ
1 þ ϕZi 2
= ϕXi þ ϕYi
2 2
1 þ ðr=ρi Þ2
It is very important to note that mi* only depends on the ratio (r/ρi). Thus, the
torsional index of the ith mode, Rρi, is defined by Eq. 9.10.
r jϕZi j
Rρi ¼ ¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ffi: ð9:10Þ
ρi
ϕXi 2 þ ϕYi 2
It can be seen that mi* is unity when Rρi is zero (purely translational), while mi* is
close to zero when Rρi is significantly large. Therefore, the terms “predominantly
translational” and “predominantly torsional” can be defined using Rρi; the “pre-
dominantly translational” mode is the mode when Rρi < 1, while the “predomi-
nantly torsional” mode is the mode when Rρi > 1.
Note that this torsional index Rρi is easily extended for a multi-storey building, as
discussed in previous study (Fujii 2014). The definition of Rρi for a multi-storey
building is shown in Appendix.
9.3 Characteristics of the Torsional Index for Each Mode
9.3.1 Relationship of the Torsional Indices for the Three

Modes
The sum of the effective modal mass ratio in each X- and Y-direction for all three
modes is obtained as in Eq. 9.11.
X
3 X
3
miX ∗ ¼ 1, miY ∗ ¼ 1: ð9:11Þ
i¼1 i¼1
Therefore, the sum of mi* of all three modes is obtained from Eq. 9.12 as:
X
3 X
3
1
mi * ¼ ¼2 ð9:12Þ
i¼1 i¼1 1 þ Rρi 2
Assuming the third mode is predominantly the torsional mode (Rρ3 < 1), m3*
must be smaller than 1/2. In which case, the sum m1 * þ m2 * must be larger than
3/2, while m1*, m2* must have a value from 0 to 1. Therefore, the relationship shown
in Eq. 9.13 is obtained.
1 1 1 1
If Rρ3 > 1, 1, and 1: ð9:13Þ
2 1 þ Rρ1 2 2 1 þ Rρ2 2
As a result, we have the interesting findings shown in Eq. 9.14.
If Rρ3 > 1, 0 Rρ1 1, and 0 Rρ2 1: ð9:14Þ
Equation 9.14 implies that, in the case of the single-storey asymmetric building
model, only one mode can be the predominantly torsional mode; if the third mode is
the predominantly torsional mode, the other two modes must be predominantly
translational modes.
9.3.2 Relationship Between the Torsional Indices of the First

and Second Modes and the Angles Between
the Principal Directions of the Two Different Modes
One of the critical assumptions of the procedure, presented in previous study (Fujii
2014), is that the principal directions of the first two modes are almost orthogonal.
In this section, the relationship between Rρi, Rρj (i 6¼ 1) and the angles between the
104 K. Fujii
principal directions of the two different modes, is investigated. A unit vector αi

parallel to the principal direction of ith modal response is defined by Eq. 9.15.
n ϕXi ϕYi
oT
αi ¼ f cos ψ i sin ψ i 0 gT ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ffi pffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ffi 0 : ð9:15Þ
ϕXi þϕYi
2 2
ϕXi þϕYi
2 2
Therefore, the cosine of Δψ ij, the angle between the principal direction of the
first and second modes, can be determined from Eq. 9.16.
αi α j ϕXi ϕX j þ ϕYi ϕY j
cos Δψ i j ¼ ¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ffiqffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi : ð9:16Þ

jαi j α j
ϕ 2þϕ 2 ϕ 2þϕ 2
Xi Yi Xj Yj
The orthogonal conditions of φi and φj can be written in the form of Eq. 9.17.
ϕXi ϕX j þ ϕYi ϕY j þ ϕZi ϕZ j ¼ 0: ð9:17Þ
From Eqs. 9.16 and 9.17, an interesting relationship between Rρi, Rρj and Δψ ij is
obtained as shown in Eq. 9.18.

ϕ ϕZ j
cos Δψ i j ¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiqffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ¼ Rρi Rρ j :
Zi
ð9:18Þ
2
ϕXi 2 þ ϕYi 2 ϕX j þ ϕY j
2
product of the torsional indices in the ith

It is clear from this equation that if the
and jth modes, Rρi, Rρj, is close to zero Rρi Rρ j 1 , the principal directions of the
two modes are almost orthogonal.
9.3.3 Effective Modal Mass Ratios of the First and Second

Modes
Another critical assumption of the presented procedure (Fujii 2014) is that the
building oscillates predominantly in a single mode in each set of orthogonal
directions. In this section, the validity of this assumption is discussed based on
the torsional index. Considering another set of orthogonal U-V axes in the X-Y
plane, where the U-axis is the principal axis of the first modal response, the
effective modal mass ratio of the first mode with respect to the U-axis, m1U*, and
the second modal response with respect to the V-axis, m2V*, can be expressed as
follows:
1
m1U * ¼ m1 * ¼ ; ð9:19Þ
1 þ Rρ1 2
ðϕX2 sin ψ 1 þ ϕY2 cos ψ 1 Þ2 1
m2V * ¼ ¼ sin 2 Δψ 12 : ð9:20Þ
ϕX2 2 þ ϕY2 2 þ ϕZ2 2 1 þ Rρ2 2
From Eq. 9.19, it is evident that if the first mode is predominantly the transla-
tional mode (Rρ1 < 1), the equivalent first modal mass ratio in the U-direction m1U*
is the largest of all three modes in that direction; in such a case, the building
oscillates predominantly in the first mode when the unidirectional excitation acts in
the U-direction. For the second modal response, Eq. 9.20 can be rewritten as
Eq. 9.21 taking into consideration Eq. 9.18.
1 1 n 2 o
m2V * ¼ 1 cos 2 Δψ 12 ¼ 1 Rρ1 Rρ2 : ð9:21Þ
1 þ Rρ2 2
1 þ Rρ2 2
This equation implies that if the first and second modes are predominantly
translational modes (Rρ1 < 1 and Rρ2 < 1), the building oscillates principally in
the second mode when the unidirectional excitation acts in the V-direction.
9.4 Numerical Investigations
9.4.1 Model Parameters
In the numerical investigation, a parametric study is carried out by considering the

various parameters of the asymmetric building model. The radius ratio of gyration
for the torsional stiffness with respect to the X- and Y-axes, JX and JY, respectively,
is defined by Eq. 9.22.
sffiffiffiffiffiffiffiffiffiffiffiffiffiffi sffiffiffiffiffiffiffiffiffiffiffiffiffiffi
j K θ =K X ω0θ jY K θ =K Y ω0θ
JX ¼ X ¼ ¼ , JY ¼ ¼ ¼ : ð9:22Þ
r I=m ω0X r I=m ω0Y
In this parametric study, EX, EY varied from 0.05 to 0.95 at intervals of 0.05,
while JX, JY varied from 0.10 to 1.95 at intervals of 0.05 (with the condition JY
JX). Note that the cases that do not satisfy the conditions shown as Eq. 9.23 are
eliminated because in such cases the torsional stiffness with respect to the centre of
0
stiffness, Kθ shown in Eq. 9.24, will be a negative value.
ðEY =J X Þ2 þ ðEX =J Y Þ2 1: ð9:23Þ

106 K. Fujii
0
K θ ¼ K θ K X eY 2 K Y eX 2 : ð9:24Þ
In this parametric study, all analytical models are categorized into three cases; Case
1 (JY JX 1), Case 2 (JY 1 JX) and Case 3 (1 JY JX).
9.4.2 Results and Discussions
Figures 9.3, 9.4, and 9.5 show the plot of torsional indices of the three modes for
each case. As shown in Fig. 9.3, for all the models categorized in Case 1 the first and
second modes are predominantly translational modes (Rρ1 < 1 and Rρ2 < 1) and the
third mode is predominantly a torsional mode (Rρ3 < 1). It should be pointed out
that all the plots in Fig. 9.3a satisfy the following condition.
Fig. 9.3 Plots of the torsional indices for all modes in Case 1
1 1 3
þ : ð9:25Þ
1 þ Rρ1 2 1 þ Rρ2 2 2
However, in Cases 2 (shown in Fig. 9.4) and 3 (shown in Fig. 9.5), the first mode
is predominantly a torsional mode; in Case 2, either the first or second mode is
predominantly a torsional mode (Rρ1 or Rρ2 may larger than 1), while in Case 3, Rρ1
is always larger than 1, and Rρ2 and Rρ3 are smaller than 1. It should be pointed out
again that all the plots in Fig. 9.5c satisfy the following condition (Eq. 9.26), which
is similar to that shown in Fig. 9.3a.
1 1 3
þ : ð9:26Þ
1 þ Rρ2 2
1 þ Rρ3 2 2
Figure 9.6 shows the relationship between m1U* and m2V* for all three cases. As
expected from the results shown in Figs. 9.3, 9.4, and 9.5 and Eqs. 9.19 and 9.20,
m1U* and m2V* are larger than 0.5 for all models categorized in Case 1 (JY JX 1);
as shown in Fig. 9.6a, the sum of m1U* and m2V* is larger than 3/2 for most models
of Case 1. In contrast, m1U* and m2V* may be smaller than 1/2 in the other cases,
especially in Case 3, where m1U* is smaller than 1/2 for all models and m2V* may be
between 0 and 1. This implies that most models in Cases 2 and 3 may not oscillate
predominantly in the first mode when the unidirectional excitation acts in the
U-direction.
From these results it may be concluded that the two critical assumptions of the
simplified procedure (Fujii 2014) are valid only in Case 1 (JY JX 1).
Note that the classification of systems as either TS or TF systems is based on the
ratio of the uncoupled torsional mode to the lateral frequencies ΩθX, ΩθY of the
corresponding torsionally balanced system, defined by Eq. 9.27 (Hejal and Chopra
1987).
108 K. Fujii
Fig. 9.6 Plots of m1U* and m2V* in the three cases. (a) Case 1. (b) Case 2. (c) Case 3
0 0
ΩθX ¼ ω0θ =ω0X , ΩθY ¼ ω0θ =ω0Y : ð9:27Þ
In Eq. 9.27, ω0θ0 is the uncoupled natural circular frequency of rotational oscillation
with respect to the centre of stiffness. According to Hejal and Chopra (1987), the
system ΩθX, ΩθY > 1 is classified as a TS system in both the X- and Y-directions.
0
Because the torsional stiffness with respect to the centre of stiffness Kθ , defined
by Eq. 9.24, is smaller than Kθ, ΩθX is smaller than JX (and also ΩθY is smaller than
JY). If ΩθX is larger than unity, JX is then larger than unity (and if ΩθY is larger than
unity, JY is then larger than unity). This implies that all the systems classified as TS
systems for both the X- and Y-directions are included in Case 1 in this parametric
study.
Therefore, the simplified procedure (Fujii 2014) may be applicable at least for
TS systems for both orthogonal directions. However the applicability of the proce-
dure is questionable for TF systems, because the two critical assumptions in the
procedure are invalid for most of them.
9.5 Conclusions
In this study, a torsional index is defined based on each mode shape, and the
relationship of the parameters for an asymmetric building model and the proposed
torsional index is investigated. Then the applicability of the simplified procedure
(Fujii 2014) is then discussed based on the torsional index. The numerical results
show that the simplified procedure (Fujii 2014) may be applicable at least for TS
systems for both orthogonal directions, while the applicability of the procedure is
questionable for TF systems.
The advantages of the proposed torsional index are as follows: (a) it is related to
the effective modal mass ratio which represents the contribution of each modal
response to the whole response, and (b) it is easily extended for an multi-storey
asymmetric building, and (c) there is the clear relationship between torsional
indices of two different modes and the angle between the principal directions of
two modes. Therefore, the index proposed herein is suitable to discuss the applica-
bility of the simplified procedure (Fujii 2014).
Appendix: Definition of Torsional Index for a Multi-storey

Building
Consider a N-storey asymmetric building, the torsional index of the ith mode, Rρi,
can be defined by Eq. 9.28.
vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
! !ffi
u N
u X X N X N
Rρi ¼ t I j ϕΘ ji 2 = m j ϕXji 2 þ m j ϕYji 2 ; ð9:28Þ
j¼1 j¼1 j¼1
φi ¼ f ϕX1i ϕXNi ϕY1i ϕYNi ϕΘ1i ϕΘNi gT : ð9:29Þ
In Eq. 9.28, mj and Ij are mass and mass moment of inertia of the jth floor,
respectively, and ϕXji, ϕYji, ϕΘji are the X, Y and rotational component of jth floor of
ith mode vector φi. Note that Eq. 9.10 is equivalent to Eq. 9.28: Eq. 9.10 is easily
obtained from Eq. 9.28 by considering the case N ¼ 1.
References
under bi-directional excitation using pushover analyses. Bull Earthq Eng 12:909–938
Hejal R, Chopra AK (1987) Earthquake response of torsionally-coupled buildings. Earthquake
Engineering Research Center, report no. UCB/EERC-87/20, College of Engineering, Univer-
sity of California at Berkeley
Chapter 10
An Approximate Method for Assessing
the Seismic Response of Irregular
in Elevation Asymmetric Buildings
George K. Georgoussis
Abstract An approximate procedure is presented for assessing the seismic

response of irregular in elevation asymmetric buildings. The procedure is based
on the concept of the equivalent single story system, which has been introduced by
the author in earlier papers for assessing the response of uniform in height build-
ings, and retains the simplicity of the corresponding methodology. Frequencies and
peak values of base resultant forces of multi story buildings with large setbacks,
classified by Eurocode 8 as irregular structures, may be predicted by the proposed
approximate method when the excitation is characterized by response spectra. The
center of stiffness of the equivalent single story system defines the modal center of
rigidity, and its coincidence with the mass axis of the real building specifies a
structure which sustains an almost translational response into the elastic phase.
Besides, such a structural configuration retains this response into the inelastic phase
when the strength assignment of all lateral load resisting bents is stiffness propor-
tional. The accuracy of the proposed procedure is first illustrated in mixed-bent-
type eight-story elastic structures, which are characterized by EC8-2004 as irreg-
ular in elevation structures, and comparisons are made with the accurate results
obtained by response spectrum analyses using the SAP2000 computer program.
The inelastic response of these structures, when the strength of various bents is
determined by a planar static analysis under a code lateral loading, is investigated
under the Loma Prieta (1989) and Imperial Valley (1940) ground motions.
Keywords Irregular buildings • Inelastic response • Modal centre of rigidity
G.K. Georgoussis (*)

Department of Civil Engineering Educators, School of Pedagogical and Technological
Education (ASPETE), N. Heraklion, 14121 Attica, Greece

DOI 10.1007/978-3-319-14246-3_10
112 G.K. Georgoussis
10.1 Introduction
It has been shown in author’s earlier papers (Georgoussis 2009, 2010, 2012) that
basic dynamic data (frequencies, resultant base shears and torques) of uniform multi-
story mixed-bent-type eccentric structures may be found by analyzing equivalent
single story modal systems. This type of analysis represents an extension of a similar
procedure presented in past for the analysis of proportionate multi-story buildings
(Kan and Chopra 1977; Hejal and Chopra 1989; Athanatopoulou et al. 2006). The
stiffness of the elements of the equivalent single story modal system is based on the
element frequencies of the bent-subsystems that provide the lateral resistance of a
given structure. These frequencies are determined from the corresponding individual
bents when they are assumed to carry, as planar frames, the mass of the complete
structure. In the case of uniform structures composed by very dissimilar bents, a
higher accuracy of the aforementioned analysis can be attained with the use of the
effective element frequencies, which are based on the element frequencies, but, also,
take into account the ratio of the effective modal mass of the individual bent to the
corresponding mass of the uncoupled multistory system (Georgoussis et al. 2013a;
Georgoussis 2014). The efficiency of the aforementioned effective frequencies is now
examined in eccentric buildings with large setbacks.
The main property of the centre of the element stiffnesses of the equivalent
single story modal system, which defines the modal center of rigidity (m-CR), is
that when it lies on the mass axis, the response of elastic uniform building
structures, asymmetric in plan, is basically translational. Besides, this response is
preserved into the post-elastic phase, where the structure is stressed beyond the
elastic limits, provided that the strength assignment of its resisting bents is stiffness
proportional. In other words, this response is obtained when the building is detailed
as a planar structure under a code load (Georgoussis 2014; Georgoussis
et al. 2013b). This is attributed to the almost concurrent yielding of all resisting
elements, which preserves the translational response, attained at the end of the
elastic phase, to the post elastic one. This response is evident in eccentric single
story systems. Reviewing the literature, it can be seen that systems, with coincident
the centres of mass and rigidity and elasto-plastic elements having a strength
distribution proportional to the stiffness distribution (usually called torsionally
balanced (TB) models) present a purely translational inelastic response under
strong ground excitations. For this reason they are used as ‘reference’ models in
relevant studies (Correnza et al. 1994; Chandler et al. 1996; Wong and Tso 1994).
This behaviour is attained because yielding is initiated at the same instant for all
elements and the element force balance about CM is preserved into the inelastic
phase, leading to a translational response throughout the ground shaking
(De Stefano and Pintucchi 2008; Anagnostopoulos et al. 2013).
The first objective of this study is to demonstrate that the same type of analysis
can be accurately applied for the assessment of frequencies and base shears of
asymmetric multistory buildings with abrupt mass discontinuities, which are clas-
sified by EC8-2004 as irregular structures. As shown in the following sections, the
10 An Approximate Method for Seismic Response 113
prediction of base torques is not as accurate as in the case of uniform buildings, but,
the predicted optimum location of m-CR is quite reasonable. Note here, that
EC8-2004 and other modern country codes specify a full three-dimensional
dynamic analysis for irregular in elevation structures. There are no recommenda-
tions of how the practicing engineer can assess the fundamental frequency by a
simple formula or methodology and there are not provisions which allow the
structural detailing by a pseudo-static structural design against an equivalent lateral
load. Only in the case of buildings with a fairly even distribution of mass (regular
buildings) the codes provide simple expressions for calculating the fundamental
frequency and allow for a pseudo-static structural design.
The second objective is to demonstrate that when (i) the mass axis of setback
buildings passes through (or in a close distance from) m-CR, and (ii) the strength
assignment of the various bents is stiffness proportional (that is, it is based on a
planar static analysis under a set of lateral forces simulating an equivalent ‘seismic
loading’), the response of the mentioned structures into the inelastic phase is
basically translational. Such a response has already been shown in uniform struc-
tures (Georgoussis 2014; Georgoussis et al. 2013b), but at present it is demonstrated
in structures with a mass irregularity. Common 8-story setback buildings are
examined under two characteristic ground motions (Loma Prieta (1989) and Impe-
rial Valley (1940)), selected from the strong ground motion database of the Pacific
Earthquake Engineering Research (PEER) Center (http://peer.berkely.edu) and
scaled to a PGA¼0.5 g.
10.2 Asymmetric Setback Buildings and Modelling
A typical mono-symmetric multistory building with a setback is shown in Fig. 10.1.

The building is uniform over the height Hb, which defines the base structure and has
a uniformly distributed mass, equal to mb per floor, and a radius of gyration equal to
rb. Above this level, it has a setback forming a uniform tower structure of a reduced
floor plan with a height equal to Ht, a mass per floor equal to mt and a radius of
gyration equal to rt. Each floor consists of a rigid slab (deck) and at present the
centers of mass (CM) at each floor are assumed to lie on the same vertical line (mass
axis) which is passing through the centroids of all decks. All bents (rigid frames,
shear walls, coupled wall systems, etc.) extend up to the top of the building.
The methodology to analyze elastic setback buildings, like that of Fig. 10.1, is
outlined in an author’s earlier paper (Georgoussis 2011). The backbone of this
method is similar to that applied to uniform over the height systems (Georgoussis
2009, 2010, 2012). In brief, in mono-symmetric systems, the peak elastic response
of medium height buildings can be derived by analyzing two equivalent single-
story modal systems, each of which has a mass equal to the k-mode effective mass,
M∗ k ðk ¼ 1, 2Þ, of the uncoupled multi-story structure, and it is supported by ele-
ments with a stiffness equal to the product of M∗k with the first mode (when k ¼ 1) or
TOWER
STRUCTURE
Ht
y
b
H
i-bent
yi
x
Hb
j-bent
CM
BASE xj
STRUCTURE
TOWER STRUCTURE
BASE STRUCTURE
Fig. 10.1 (a) Multistory setback building with (b) an asymmetric structural configuration
second mode (when k ¼ 2) squared element frequencies of the corresponding real

bents of the assumed multi-story structure. For an excitation along the y-direction,
the effective mass M∗
k can be simply taken equal to that corresponding to the same

∗
direction M yk . At present this procedure is also applied and in a slightly modified
version, by using the effective element frequencies (instead of the element frequen-
cies), as they are defined in earlier papers (Georgoussis et al. 2013a; Georgoussis
2014). Under these assumptions, the undamped equation of motion of the k-mode
equivalent single story system, in a coordinate system with the origin at the center
of mass (Fig. 10.1b), is as follows:
M∗ € ∗ ∗
k Uk þ Kk Uk ¼ Mk ι€
ug ð10:1Þ
where
1 0
M∗ ∗
k ¼ M yk is the effective k-mode mass matrix,
0 r 2ek
Uk ¼ h uk θk iT is the corresponding modal displacement vector at CM
∗
k y k∗
K∗ ¼ yw
is the effective k-mode stiffness matrix
k k∗
wy k∗
w k
ιT ¼ h 1 0 iT is the influence vector, and
k∗ ∗ ∗
yk ¼ Σk jk ¼ M yk Σω jk
2

k∗
wk ¼ Σx 2 ∗
k
j jk þ Σ y 2 ∗
k
i ik ¼ M ∗
yk Σ x ω
2 2
j jk þ y ω
2 2
i ik ð10:2Þ
k∗
ywk ¼ k∗
wyk ¼ Σx j k∗jk ¼ M∗
yk Σx j ω jk
2
The quantities ωjk and ωik are the element frequencies of the j and i-bents,
aligned along the y- and x-directions at distances xj and yi respectively. The
accuracy of this procedure is also investigated by using the effective element

frequencies ω jk and ωik (instead of ωjk and ωik), which are defined as follows:

ω2jk ¼ ω2jk M∗jk =M∗ ∗ ∗
yk , ωik ¼ ωik Mik =Mxk
2 2
ð10:3Þ
It is evident that when the lateral stiffness of a given building is composed by the
same type of bents (e.g. flexural shear walls), the aforementioned effective element
frequencies are respectively equal to the element frequencies ωjk and ωik. The
radius of gyration of the equivalent single story system rek is taken equal to r ek r b ,
where r ek represents a Rayleigh’s quotient, given as (Georgoussis 2011):

r ek ¼ ω yk =ωr yk ðor r ek ¼ ðωxk =ωrxk ÞÞ ð10:4Þ
where ωyk (or ωxk) is the k-mode frequency of the uncoupled multistory structure in
the y-direction (or x-direction) and ωryk (or ωrxk) the corresponding frequency of the
same structure when the mass in the floors of the tower section is reduced to
mrt ¼ ðr t =r b Þ2 mt . It has been shown (Georgoussis 2011) that in common setback
buildings, the ratio r ek is very little dependent on the type of the lateral load resisting
system (frame, wall, dual system). Therefore, any of the expressions of Eq. 10.4
may be used for practical applications, but it is advisable to use the mean value of
these expressions, since this averaging procedure utilizes the response of the
structure in both directions.
Note here that the coupled Eq. 10.1, for the first mode (k ¼ 1) single-story
system, provides the response quantities of the first two modes of vibration.
Therefore, when the plan configuration produces an uncoupled stiffness matrix in
Eq. 10.1, the first two modes of vibration (translational and rotational) are
decoupled and the response for a low height building will be practically transla-
tional. In fact, this condition specifies that the first mode center of rigidity (m-CR)
of the corresponding single-story system coincides with CM. As, in general,
the
x-coordinate of m-CR can be determined from the condition: k∗ ∗
ywk ¼ kwyk ¼ 0,
i.e.:

xmCR ¼ Σ x j ω2j1 =Σ ω2j1 ð10:5Þ
minimum seismic torsion is expected in low or medium height structures (where the
first two modes of vibration virtually determine their response) when the location of
m-CR coincides with CM. Such structural configurations may be seen as torsionally
balanced systems. The second objective of the paper is to demonstrate that such
setback buildings retain this translational response into the post elastic phase when
they are detailed as planar structures under a code horizontal load.
a Setback model structure

y
b EC8-2004 spectrum
A/g
Tower structure 15x10m
Flat
CW 1.00
FR
Wa Wb
0.75
x Linear Hyperbolic
CM
0.50
x 0.4
CW 0.25
4m 6.5m
0.15 T(s)
Base structure 22x15m
0.3 0.6 0.9 1.2 1.5
Fig. 10.2 (a) Example setback building, (b) EC8-2004 design spectrum
10.3 Buildings Studied
To illustrate the application and accuracy of the proposed method, the example
setback building shown in Fig. 10.2a was analyzed. This is an 8-story mono-
symmetric building, which is divided in two substructures: the base structure
represents a uniform building system composed by floors of 22 15 m, while the
tower substructure is composed by floors of reduced dimensions 15 10 m. The
lateral load resisting system, extending up to the top of the building, is composed by
dissimilar bents: two structural walls (Wa and Wb) and a moment resisting frame
(FR) are aligned along the y-direction and a pair of coupled-wall bents (CW) is
oriented along the x-axis of symmetry. The structural walls Wa and Wb are of cross
sections 30x500cm, the moment resisting frame FR consists of two 75 75 cm
columns, 5 m apart, connected by beams of a cross section 40 70 cm, while the
CW bents are composed two 30 300 cm walls, 5 m apart, connected by lintel
beams of a cross section 25 90 cm at the floor levels. The latter bents are located
symmetrically to CM at the edges of the floors of the tower structure, that is at
distances equal to 5 m. The mass of the base floors is mb ¼ 264kNs2/m, the radius
of gyration about CM is rb ¼ 7.687 m and the corresponding quantities of the tower
structure are equal to mt ¼ 120 kNs2/m and rt ¼ 5.204 m respectively. The story
height is 3.5 m and the modulus of elasticity (E) is assumed equal to 20 106 kN/
m2, typical for concrete structures. The centers of mass of the floor slabs lie on a
same vertical line, which passes through the centroids of all the orthogonal floor
plans of the example structure.
Three models of the aforementioned example building are examined. In the first
model (T2–B6) the tower structures consists of two floors, in the second model
(T4–B4) of four floors and, finally, in the third model (T6–B2) the tower structure
consists of six floors. For each model different structural configurations are exam-
ined as follows: wall Wa and frame FR are assumed to be located at fixed positions,
the first on the left of CM in a distance equal to 4 m and the second on the right of
CM at a distance of 6.5 m, while the second wall Wb is taking all the possible
locations along the x-axis within the limits of the tower section.
At first the periods/frequencies of the assumed models, for all possible locations
of Wb, are examined. The accuracy of the proposed approximate procedure to
predict periods of vibrations is investigated by comparison with the results derived
from the computer program SAP2000-V11. In the computer analyses, the out of
plane stiffness of the bents was neglected and in the wide column analogy used to
simulate the CW bents the clear span of the coupling beams was increased by the
depth of the beams (Coull and Puri 1968). To apply the proposed method, the first
pair of element frequencies of the various bent-subsystems is required, and also the
effective modal masses of the uncoupled system. For example, denoting with M the
total mass of the second model structure (M ¼ 4mb + 4mt ¼ 1536 kNs2/m for T4–
B4), these quantities for the bents of this model were found by means of the
SAP2000 program as follows:
Walls Wa, Wb: ωw1 ¼ 5.092/s, ωw2 ¼ 23.994/s, frame FR: ωf1 ¼ 3.033/s,
ωf2 ¼ 8.483/s, coupled walls CW: ωcw1 ¼ 5.372/s, ωcw2 ¼ 19.592/s, and the first
∗
two effective modal masses of the uncoupled system are: M y1 ¼ M∗ y1 =M ¼ 0:588
∗
and M y2 ¼ 0:265. The radius of gyration rek (k ¼ 1,2), computed as described in the
previous section, was found equal to re1 ¼ 0.742*rb ¼ 5.704 m and
re2 ¼ 0.871*rb ¼ 6.695 m for the first and second mode equivalent single story
systems respectively.
When the proposed method is based on the effective element frequencies, these
quantities are found: ωw1 ¼ 5:044=s, ωw2 ¼ 24:442=s for walls Wa, Wb and: ω f 1
¼ 3:412=s, ω f 2 ¼ 6:122=s for frame FR (for the coupled walls CW the effective
frequencies are equal to the element frequencies as above).
The inelastic response of the assumed model structures was also investigated
under two characteristic ground motions (Loma Prieta (1989) and Imperial Valley
(1940)), selected from the strong ground motion database of the Pacific Earthquake
Engineering Research (PEER) Center (http://peer.berkely.edu) and scaled to a
PGA ¼ 0.5 g (unidirectional excitations along the y-axis). For all the possible
locations of Wb, inelastic analyses, by means of the computer program SAP2000-
V11, were performed to evaluate top rotations and base shears and torques. The
strength assignment of all bents of the assumed models is based on a planar static
analysis, along the x and y directions, under an external lateral loading with the
floor forces determined from Eq. 4.11 of EC8-2004 and summing to a base (design)
shear equal to 20 % of the total weight of the structure.
10.4 Model Frequencies and Observed Nonlinear Seismic

Response
The first four periods of vibration of the model setback structures, computed by the
proposed method on the grounds of the element frequencies (red lines) and, also, on
the grounds of the effective element frequencies (green lines) for different locations
of the Wb (indicated by the normalized coordinate x ¼ x=r b ), are shown in
Fig. 10.3, together with the accurate SAP2000 computer values (black lines). It is
evident that the first pair of frequencies is closer to the accurate computer results
when the proposed method is based on the effective element frequencies (the green
lines are closer to the black ones for the first and second periods of vibration), while
the predicted second pair of frequencies is more accurate when the proposed
method is based on the element frequencies (the red lines are closer to black ones
for the second pair -third and fourth- periods of vibration).
Interpreting these observations, it seems that best data are provided by the
proposed method when the first mode equivalent single-story system is formulated
on the basis of the effective element frequencies, while best results are obtained
from the second mode system when it is based on element frequencies. The results
of this analysis (base shears and torques), in relation to the EC-8 acceleration
spectrum (Fig. 10.2b), are shown in Fig. 10.4 by green lines, together with the
results obtained by analyzing the 3D structures with the SAP2000 software (black
lines). For comparison reasons the data obtained from the proposed method on the
basis of only the element frequencies (for both equivalent single story modal
systems) are also shown in Fig. 10.4 with red lines.
The response of the inelastic setback structures under the Loma Prieta (compo-
nent Corralitos 000, 1989) and Imperial Valley (component ElC180, 1940) excita-
tions, are shown in Fig. 10.5. Note here that the strength of all models is determined
by a planar static analysis under a set of floor forces determined from Eq. 4.11 of
EC8-2004. In order to compare elastic and inelastic behaviors, the elastic responses
of the assumed models under the same excitations are also presented in this figure.
Three response parameters, obtained by time history analyses assuming a 5 %
damping ratio, are shown: top rotations, normalized base shears and normalized
base torques. The red lines represent the peak elastic response (top rotations: θe, are
shown by dashed lines, normalized base shears: Ve ¼ Ve =Vd (Vd: the design base
shear equal to 20 % of the total weight) by solid lines and normalized base torques:
Te ¼ Te =rb Vd by dotted lines). The corresponding black lines represent the peak
inelastic behavior θin , Vin ¼ Vin =Vd , Tin ¼ Tin =rb Vd . Envisaging Fig. 10.5 it can
be seen that minimum values of base torques and top rotations are observed when
the location of wall Wb receives values close to those predicted from Eq. 10.5, by
equating xm-CR to zero. For the assumed models T2-B6, T4-B4 and T6-B2, the
predicted optimum locations of FR are found equal to x ¼ 0:133, 0.168 and 0.12
respectively.
10.5 Conclusions
In this work, the torsional seismic behaviour of mass irregular asymmetric build-
ings was assessed using an approximate simplified procedure which has been used
in the past for the analysis of uniform multi-story building. Frequencies of mono-
symmetric, medium height setback buildings, composed by dissimilar bents, can be
1st..mode..periods
sec
Model T2-B6
1.2
1
2nd..mode..periods
0.8
0.6
3rd..mode..periods 0.4
4th..mode..periods 0.2
x
0
-1 -0.75 -0.5 -0.25 0 0.25 0.5 0.75 1
sec
Model T4-B4
1st..mode..periods
1
0.8
2nd..mode..periods
0.6
3rd..mode..periods 0.4
x
0
-1 -0.75 -0.5 -0.25 0 0.25 0.5 0.75 1
sec
Model T6-B2
1st..mode..periods
1
0.8
2nd..mode..periods
0.6
0.4
3rd..mode..periods
x
0
-1 -0.75 -0.5 -0.25 0 0.25 0.5 0.75 1
Fig. 10.3 Periods of vibration of setback models

1
Model T2-B6, 0.8
Model T2-B6, Vea
base shears base torques
0.8 Tcom
0.6
0.6 0.4
Vm
Vcom
0.4 0.2
Tea
Tm
x 0
x
0.2
-1 -0.75 -0.5 -0.25 0 0.25 0.5 0.75 1 -1 -0.75 -0.5 -0.25 0 0.25 0.5 0.75 1
0.8
1 Model T4-B4,
Model T4-B4,
Vea
Vm 0.8 0.6
Vcom Tcom
Tm
0.6 0.4
0.4 0.2
Tea
x x
0.2 0
-1 -0.75 -0.5 -0.25 0 0.25 0.5 0.75 1 -1 -0.75 -0.5 -0.25 0 0.25 0.5 0.75 1
1 0.8
Model T6-B2, Model T6-B2,
Vcom 0.6
0.8
Tcom
0.6 0.4
Tm
Vm
Vea 0.4 0.2
Tea
x x
0.2 0
-1 -0.75 -0.5 -0.25 0 0.25 0.5 0.75 1 -1 -0.75 -0.5 -0.25 0 0.25 0.5 0.75 1
Fig. 10.4 Peak base shears and torques of setback models
estimated with reasonable accuracy from the analysis of two equivalent, single-
story modal systems, the masses of which are determined from the first two
vibration modes of the uncoupled multi-story structure and the stiffnesses of the
resisting elements are determined from the corresponding individual bents when
they are assumed to carry, as planar frames, the mass of the complete structure. This
3.5 (a)Model T2-B6, Imperial 2 (b) Model T2-B6, Loma

Ve
3 Ve θe
1.6
θe 2.5 Te Vin
Te 1.2
2
Vin
θin
1.5
θin 0.8
1
Tin Tin
0.4
0.5
x x
0 0
-1 -0.75 -0.5 -0.25 0 0.25 0.5 0.75 1 -1 -0.75 -0.5 -0.25 0 0.25 0.5 0.75 1
(a)Model T4-B4, Imperial

3.5 (b)Model T4-B4 Loma
2.5
3 Ve
2
2.5 Ve
θe Te
2 Vin θin 1.5 Vin
Te
θin
1.5 θe
1
1 Tin
Tin 0.5
0.5
0
x x
0
-1 -0.75 -0.5 -0.25 0 0.25 0.5 0.75 1 -1 -0.75 -0.5 -0.25 0 0.25 0.5 0.75 1
4
(a)Model T6-B2, Imperial
2.5 (b) Model T6-B2, Loma
3.5
3 Ve 2 Ve
2.5
Vin
Te 1.5
θe 2
Vin θin θe
1.5 1
θin 1 Te
Tin 0.5 Tin
0.5
x x
0 0
-1 -0.75 -0.5 -0.25 0 0.25 0.5 0.75 1 -1 -0.75 -0.5 -0.25 0 0.25 0.5 0.75 1
2
Fig. 10.5 Top rotations (*10 , rads), base shears and torques of setback models
simple analysis provides also quite reasonable estimates of the resultant base shears
of the setback buildings and, to some extent, of the base torques. Quite reasonable is
also the prediction of the location of the first mode center of rigidity. The main
property of this point is that when it is within a close distance from the mass axis,
the response of elastic buildings is virtually translational. This behavior is preserved
in the inelastic phase, when the strength assignment of the lateral load resisting
bents is derived from a planar static analysis, as a consequence of the almost
concurrent yielding of these bents. This is demonstrated in common 8-story setback
buildings under two characteristic ground motions.
Acknowledgements This research has been co-financed by the European Union (European
Social Fund – ESF) and Greek national funds through the Operational Program “Education and
Lifelong Learning” of the National Strategic Reference Framework (NSRF) – Research Funding
Program: ARCHIMEDES III. Investing in knowledge society through the European Social Fund.
The author is grateful for this support.
References
Anagnostopoulos SA, Kyrkos MT, Stathopoulos KG (2013) Earthquake induced torsion in

buildings: critical review and state of the art. The 2013 World Congress on Advances in
Structural Engineering and Mechanics (ASEM13), Jeju, Sept 8–12
Athanatopoulou AM, Makarios T, Anastassiadis K (2006) Earthquake analysis of isotropic
asymmetric multistory buildings. Struct Des Tall Spec Build 15:417–443
Chandler AM, Duan XN, Rutenberg A (1996) Seismic torsional response: assumptions, contro-
versies and research progress. Eur Earthq Eng 1:37–51
Correnza JC, Hutchinnson GL, Chandler AM (1994) Effect of transverse load-resisting elements
on inelastic earthquake response of eccentric-plan buildings. Earthq Eng Struct Dyn 23:75–89
Coull A, Puri RD (1968) Analysis of pierced shear walls. J Struct Eng ASCE 94(1):71–82
De Stefano M, Pintucchi B (2008) A review of research on seismic behaviour of irregular building
structures since 2002. Bull Earthq Eng 6:285–308
Georgoussis GK (2009) An alternative approach for assessing eccentricities in asymmetric
multistory structures 1: elastic systems. Struct Des Tall Spec Build 18(2):181–202
Georgoussis GK (2010) Modal rigidity center: its use for assessing elastic torsion in asymmetric
buildings. Earthq Struct Int J 1(2):163–175
Georgoussis GK (2011) Simplified dynamic analysis of eccentric buildings with a setback. 1: the
effect of mass irregularity. Struct Des Tall Spec Build 20:911–927
Georgoussis G (2012) Seismic analysis of non-proportionate eccentric buildings. Adv Mater Res
450–451:1482–1488
Georgoussis G (2014) Modified seismic analysis of multistory asymmetric elastic buildings and
suggestions for minimizing the rotational response. Earthq Struct Int J 7(1):039–052
Georgoussis G, Tsompanos A, Makarios T, Papalou A (2013a) Optimum structural configuration
of irregular buildings. 1: elastic systems. The 2013 World Congress on Advances in Structural
Engineering and Mechanics (ASEM13), Jeju, Sept 8–12
Georgoussis G, Tsompanos A, Makarios T, Papalou A (2013b) Optimum structural configuration
of irregular buildings. 2: inelastic systems. The 2013 World Congress on Advances in Struc-
tural Engineering and Mechanics (ASEM13), Jeju, Sept 8–12
Hejal R, Chopra AK (1989) Earthquake analysis of a class of torsionally-coupled buildings. Earthq
Eng Struct Dyn 18:305–323
Kan CL, Chopra AK (1977) Elastic earthquake analysis of torsionally coupled multistorey
buildings. J Struct Div ASCE 103(4):821–838
Wong CM, Tso WK (1994) Inelastic seismic response of torsionally unbalanced systems designed
using elastic dynamic analysis. Earthq Eng Struct Dyn 23:777–798
Chapter 11
Application of Nonlinear Static Procedures
for the Seismic Assessment of a 9-Storey
Asymmetric Plan Building
André Belejo and Rita Bento
Abstract The Seismic Assessment of a 9-storey asymmetric plan building is

performed through recent Nonlinear Static Procedures (NSPs). Among these
methods two multimode methods are included: the Modal Pushover Analysis
(MPA) and the Improved Modal Pushover Analysis (IMPA). The IMPA is a
multimode procedure that has the advantage of redefining the lateral load applied,
when comparing with multimode current methods; hence, instead of consider the
elastic deformed shape when applying the pushover load pattern, it is possible to
consider the deformed shape of the structure when behaving inelastically. The
IMPA was proposed in the past and was successfully applied in the seismic
assessment of bridges. Thus, the main objective of this work is to test IMPA in
buildings. For this purpose the seismic demands of an asymmetric plan building,
considering both components of ground motion acting simultaneously, are herein
estimated by means of IMPA and compared with Nonlinear Dynamic Analyses
(NDA), MPA and two other NSPs that are proposed in American and European
seismic codes (ASCE/SEI 41-06 NSP and N2 method respectively). The results are
obtained in terms of lateral displacements profiles, interstorey drifts, normalized top
displacements and shear forces.
Keywords Improved modal pushover analysis • Torsion • Asymmetric-plan

building • Nonlinear static procedures • Nonlinear dynamic analysis
A. Belejo (*)
Oregon State University, Corvallis, OR, USA
R. Bento
ICIST, Instituto Superior Técnico, Universidade de Lisboa, Av. Rovisco Pais,
1049-001 Lisbon, Portugal

DOI 10.1007/978-3-319-14246-3_11
124 A. Belejo and R. Bento
11.1 Introduction
The IMPA, introduced by Paraskeva and Kappos (2010) for bridges, is an improved
version of the MPA procedure, which has, as main characteristic, the ability to
overcome the invariability of the lateral force distribution.
The IMPA is based on the MPA proposed by Chopra and Goel (2002), known as
a complete version of a multi-mode pushover analysis. The MPA is a multi-run
method, wherein several pushover curves are obtained from different load patterns
proportional to each mode of vibration (as many as modes considered) and the final
response is obtained combining the results that correspond to each pushover curve
using an appropriate combination rule. This method has continuously been
improved and updated: MPA was first extended to asymmetric buildings again by
Chopra and Goel (2004), and recently adapted to consider both components of
ground motion acting simultaneously in buildings by Reyes and Chopra (2011).
The IMPA is herein applied to assess the seismic behavior of an asymmetric plan
building, along with three more NSPs: the MPA, the NSP proposed in ASCE/SEI
41-06 (ASCE 2007) and the most recent extension of N2 method developed by
Kreslin and Fajfar (2012).
These procedures are applied within different seismic intensities, considering
both components of ground motion acting simultaneously.
The objectives of this paper are: (i) to evaluate the accuracy of IMPA on the
prediction of seismic demands of an asymmetric plan building, especially when the
structure exhibits inelastic behavior; (ii) to evaluate the number of modes required
by the multimode procedures in order to obtain reliable results; (iii) to compare the
two multimode methods, IMPA and MPA and; (iv) to comparatively assess the
accuracy of all methods herein applied: IMPA, MPA, the NSP proposed in ASCE/
SEI 41-06 and the extended N2 method, which is evaluated by comparison with
nonlinear dynamic analyses (NDAs).
11.2 Nonlinear Static Procedures (NSPs)
The MPA considers a non adaptive force based pushover analysis based on modal
proportional load patterns. The method takes into account the higher mode effects
since in each run a different load pattern proportional to each mode of vibration of
the structure is applied, and the results computed from each pushover curve are
combined to obtain the final results. The complete methodology as a whole is
described step by step in Reyes and Chopra (2011).
The key idea of the IMPA procedure is to use the deformed shape of the structure
responding inelastically to the considered earthquake level in lieu of the elastic
mode shape. So the IMPA, following the guidelines of the work performed for
bridges by Paraskeva and Kappos (2010), is divided in two phases: (i) in the first
phase, the seismic response is computed for each mode, as MPA does; (ii) in the
11 NSPs Applied to an Asymmetric Plan Building 125
second phase, the procedure is restarted using a lateral load pattern proportional to
the displacement shape vector correspondent to the peak deformation obtained in
the first phase, and the process is repeated. In Belejo and Bento (2015), the method
is described step by step.
The most recent version of N2 method corresponds to an extended version of the
original N2 method proposed by Fajfar and Fichinger (1988) which is also
described in Eurocode 8 (CEN 2004) that takes into account both the torsional
and the higher modes effects, by adjusting the pushover results, computed with the
original N2 method, by means of correction factors based on linear dynamic
response spectrum analysis.
The ASCE/SEI 41-06 NSP is based on the Displacement Coefficient method
(FEMA 1997) and provides a direct numerical process for calculating the displace-
ment demand. The method has been updated over time in American codes like
FEMA 356 (ASCE 2000) and FEMA440 (ATC 2005) ending up with its final
version in ASCE/SEI 41-06.
11.3 Case Study: 9-Storey Building
The building selected for this work is an asymmetric plan 9-storey frame steel
building that represents the older buildings designed according to the 1985 Uniform
Building Code (UBC85) and was defined and studied by Reyes (2009).
The building structure is characterized by moment resisting frames, detached in
Fig. 11.1 by the identified columns, gathered with the beams that connect them;
gravity frames whose function is to support the gravity loads and braced frames in
the alignments C1-C8, C3-C9, C9-C12 and C14-C17 that provide a “panel effect”
in those localizations. More information about the building is provided in Reyes
(2009) and Belejo and Bento (2015).
11.4 Modelling Issues
The analysis software adopted in this work was SeismoStruct v6.0 (Seismosoft
2006), a downloadable fibre element based finite element software.
The 3D model representing the building under analysis was built using space
frames assuming the centerline dimensions. The nonlinear behavior of the members
was modeled through the use of fibre element models with each fibre characterized
by the respective material relationship.
In terms od damping, the hysteretic damping was already implicitly included in
the nonlinear fibre model formulation of the inelastic frame elements. On the hand,
the viscous damping was modeled on the Rayleigh type with its two constants
selected to give 2 % damping ratio at the fundamental period of vibration T1 and a
9.14
C14 C15 C16 C17
9.14
C13
9.14
CM
C8 C10 C11 C12
C9
9.14
9.14 9.14 9.14 9.14

C6 C7
9.14
C4 C5
9.14
C1 C2 C3
9.14 9.14
b
3.96
3.96
3.96
3.96
3.96
C3 C5 C7 C12 C13 C17

3.96
3.96
3.96
3.96
Fig. 11.1 9-storey building configuration: (a) in plan; (b) in elevation

period of 0.2T1, following the work performed by Reyes (2009) on a similar

building.
A rigid diaphragm effect was modeled with nodal constraints under a penalty
functions option. The penalty function exponent used was 107.
The weight of each floor was assigned as lumped mass in the nodes,
corresponding to the respective tributary area.
11.5 Seismic Features
Fourteen bi-directional ground motion records were selected in order to perform

NDA. Their selection was performed among the set of records used by Reyes
(2009). The earthquake motion selection of these 14 records was performed by
first plotting the original, non-scaled response spectrum for each earthquake
motion, and comparing the spectral accelerations to the design spectral accelera-
tions. Each earthquake motion was then scaled linearly by a scaling factor (SF),
within the range of 0.2<SF<5.0, in a range of spectral periods, 0.2–1.5T1, with the
objective of obtaining a match between the scaled response spectrum and the design
spectrum in a Period range of interest. The 14 motions with the smallest root-mean-
square-error (RMSE) values, considering the geometric mean of both orthogonal
directions, and their corresponding scale factor SF were recorded. The RMSE,
proposed by Kottke and Rathje (2008) measures the goodness-of-fit for each
response spectrum to the “target” spectrum. In this work all the records (geometric
mean) were matched to the seismic hazard spectrum with the corresponding site
located in Los Angeles, and considering 2 % probability of occurrence in 50 years
(P50 ¼ 2 %). The 14 matched spectra are plotted in Fig. 11.2, and Table 11.1
identifies the Earthquakes and the station where recorded.
Due to the uncertainty of knowing the position of the building relatively to the
components of the records, all ground motions were assigned to the building in two
Fig. 11.2 Seismic hazard

spectrum and the median
response spectrum of
14 scaled ground motions
Table 11.1 Ground motion Earthquake Station

records considered
Northridge 1994 Canyon Country
Northridge 1994 Santa Monica
Northridge 1994 Beverly Hills
Duzce 1999 Bolu
Hector Mine Oct 1999 Hec
Imperial Valley 1979 Calexico Fire Station
Kobe 1995 Shin-Osaka
Kocaeli 1999 Duzce
Landers 1992 Joshua Tree
Loma Prieta 1989 Gilroy Array #3
Loma Prieta 1989 Oakland
Loma Prieta 1989 Hollister City Hall
Loma Prieta 1989 Hollister Diff Array
Chi-Chi 1999 Chy006
Table 11.2 Periods (in seconds) and effective modal mass percentages of both buildings studied
1st set of modes 2nd set of modes
Mode Period (sec) [Ux] (%) [Uy] (%) Mode Period (sec) [Ux] (%) [Uy] (%)
1 1.87 8.2 26.4 4 0.69 1.4 2.6
2 1.69 59.3 21.1 5 0.60 6.7 4.0
3 1.57 13.1 32.8 6 0.55 2.3 4.4
different ways: “fault parallel” component of the record according to the X com-
ponent of the building and “fault normal” component of the record assigned to the Y
component of the building; and the opposite. Therefore the final seismic response is
determined by the median of the 28 results obtained.
The median response spectrum obtained when both components of ground
motion are considered is also shown in Fig. 11.2.
11.6 Numerical Results and Discussion
In this section, seismic responses obtained through the procedures abovementioned,

are presented in terms of pushover curves, lateral displacement profiles, interstorey
drifts, normalized top displacements and shear forces for different levels of seismic
intensities, and considering both components of ground motion acting
simultaneously.
The modal properties of the studied building are presented in Table 11.2, which
shows the periods and the effective modal mass percentages in both X and Y
directions, for the first and second set (three modes) of modes.
The results obtained attest to the torsion-dominated characteristics of the struc-
ture. The first mode is a torsional mode, the second mode shows translation along
both axes, but predominantly in the X direction, and the third mode shows trans-
lational behavior in Y direction coupled with torsion; which means that the building
is torsional flexible in both directions. The second set of modes follows the modal
behavior of the first set of modes.
11.6.1 Capacity Curves
In Fig. 11.3 the pushover curves obtained for the MPA (consequent first phase of
IMPA for the first set of modes) are displayed for each mode considered together
with the peak displacements obtained. The building is tested considering 10 % and
2 % probability of occurrence in 50 years (P50 ¼ 10 % and P50 ¼ 2 % respectively).
Fig. 11.3 Pushover curves used in MPA procedure: (a1) 1st triplet of nodes in X direction; (a2) 1st
triplet of nodes in Y direction; (b1) 2nd triplet of nodes in X direction; (b2) 2nd triplet of nodes in Y
direction
Fig. 11.4 Pushover curves of the second phase of IMPA: (a) X direction (b) Y direction
In IMPA procedure, due to the fact of being a double-run method, a second

pushover curve, corresponding to each intensity and direction of ground motion, is
obtained and shown in Fig. 11.4. This procedure is not performed for the second set
of modes (modes 4–6) for the reason that the higher mode equivalent Single-
degree-of-freedom (SDOF) systems do not have any contribution to the inelastic
response when reaching the peak deformation in the first phase.
11.6.2 Number of Modes Required for MPA and IMPA
The number of modes selected has an important role in the accuracy of these
methods. For that reason, MPA and IMPA were applied changing the number of
modes considered: (i) only one set of modes was considered (Table 11.2) and;
(ii) two sets of modes were adopted. Figure 11.5 shows how the accuracy of MPA
results, in terms of displacements and internal forces, can be influenced by the
number of modes considered.
Fig. 11.5 (a) Lateral displacement profiles; (b) Interstorey drifts; (c) Shear forces; (d) Normalized
top displacements
According to the results showed in Fig. 11.5, one can observe that, when
considering only one set of modes, the interstorey drifts and mainly shear forces
are underestimated. On the other hand, considering two sets of modes, MPA leads
to conservative results.
11.6.3 Comparison Between NSPs and NDA
In this section, the seismic demands obtained through all NSPs applied are herein
presented, wherein two sets of modes were considered for the multimode methods.
To study the behavior along the height of the building in both directions, lateral
displacement profiles and interstorey drifts were obtained. Such results are respec-
tively displayed in Figs. 11.6 and 11.7.
In Fig. 11.7 can be observed that all methods lead to adequate results in terms of
lateral displacement profiles, being all of them slightly conservative. The accuracy
shown in the results, regarding the multimode methods, is explained due to the fact
that the higher modes were considered. Moreover, it is worth to mention that
previous results obtained when adopting only the first set of modes for the MPA
and IMPA led to less accurate results.
Extended N2 and ASCE/SEI 41-06 NSPs are generally the most conservative
methods in terms of interstorey drifts.
The torsional behavior of the building was addressed through the analysis of a
trend of normalized top displacements which are displayed in Fig. 11.8, and from its
observation, can be stated that IMPA and MPA generally show accurate torsional
behavior when compared with NDA.
With respect to Shear Forces, an extension of MPA, proposed by Reyes and
Chopra (2011), based on imposing a set of displacements that are compatible with
Fig. 11.6 Lateral displacement profiles: (a) X direction; (b) Y direction
Fig. 11.7 Inter-storey drifts: (a) X direction; (b) Y direction
Fig. 11.8 Normalized top displacements: (a) X direction; (b) Y direction

Fig. 11.9 Shear forces: (a) X direction; (b) Y direction
the estimate of story drifts obtained, is applied in both multimode methods. In

Fig. 11.9 the shear forces are displayed and a good match between all NSPs applied
and NDA is observed.
11.7 Conclusions
In this paper, the seismic performance of an asymmetric plan building with nine
stories is accessed by applying IMPA, which results obtained were evaluated and
compared with the one evaluated by means of NDA. Comparisons with other
current and well known NSPs as MPA, extended N2 and ASCE/SEI 41-06 NSP,
were also performed. The individual performance of each one was additionally
evaluated.
When compared with NDA while observing both methods, in a general view,
one can conclude that there is a slight improvement with IMPA in relation to MPA
when capturing torsional response of the building.
This set of results show a good match with NDA for all procedures.
Regarding to the multimode procedures, the fact of considering a greater number
of modes, improved the results obtained with only one set. And it is notable that the
higher modes consideration in the multi-mode procedures has an important role
capturing the response of the 9-Storey building, mainly in terms of shear forces.
Regarding to the methods proposed (or extensions) in the Seismic Codes: the
extended N2 captures with accuracy the torsional amplification in the buildings,
contrasting with ASCE/SEI 41-06 NSP which does not capture the torsional
amplification in the buildings, since it estimates linearly the response from one
side of the building to the other.
Further investigation about the effectiveness of IMPA should be performed by

applying the method to more buildings where the deformed shape changes sub-
stantially during pushover analysis.
Acknowledgments The authors would like to acknowledge the financial support of the Portu-
guese Foundation for Science and Technology through the research project PTDC/ECM/100299/
2008. In addition, the authors would like to thank Prof. Juan Carlos Reyes for all the support and
material provided whenever it was requested.
References
ASCE (2000) American Society of Civil Engineers. Prestandard and commentary for the seismic
rehabilitation of buildings, FEMA-356. Washington, DC
ASCE (2007) Seismic rehabilitation of existing buildings. ASCE/SEI 41-06, Reston
ATC (2005) Applied Technology Council. Improvement of nonlinear static seismic analysis
procedures, FEMA440 report. Redwood City
Belejo A, Bento R (2015) Improvements in modal pushover analysis for the seismic assessment of
asymmetric plan buildings. Version submitted to Soil Dynamics and Earthquake Engineering
CEN (2004) Comité Européen de Normalisation. Eurocode 8: design of structures for earthquake
resistance. Part 1: general rules, seismic actions and rules for buildings. EN 1998-1:2004.
Brussels
Chopra A, Goel R (2002) A modal pushover analysis procedure for estimating seismic demands
for buildings. Earthq Eng Struct 31:561–582
Chopra A, Goel R (2004) A modal pushover analysis procedure to estimate seismic demands for
asymmetric-plan buildings. Earthq Eng Struct Dyn 33:903–927
Fajfar P, Fischinger M (1988) N2—a method for non-linear seismic analysis of regular buildings.
In: Proceedings of the 9th World conference in earthquake engineering Tokyo-Kyoto Japan 5:
111–116
FEMA (1997) Federal Emergency Management Agency. NEHRP guidelines for the seismic
rehabilitation of buildings (FEMA 273). Washington, DC
Kottke A, Rathje E (2008) A semi-automated procedure for selecting and scaling recorded
earthquake motions for dynamic analysis. Earthq Spectra 24(4):911–932
Kreslin M, Fajfar P (2012) The extended N2 method considering higher mode effects in both plan
and elevation. Bull Earthq Eng 10:695–715
Paraskeva T, Kappos A (2010) Further development of a multimodal pushover analysis procedure
for seismic assessment of bridges. Earthq Eng Struct Dyn 39(11):211–222
Reyes J (2009) Estimating seismic demands for performance-based engineering buildings. PhD
dissertation, University of California, Berkeley
Reyes J, Chopra A (2011) Three-dimensional modal pushover analysis of buildings subjected to
two components of ground motion, including its evaluation for tall buildings. Earthq Eng
SeismoSoft (2006) SeismoStruct—a computer program for static and dynamic nonlinear analysis
of framed structures. SeismoSoft Ltd, Pavia
Chapter 12
Seismic Assessment of an Existing Irregular
RC Building According to Eurocode
8 Methods
Alessandra La Brusco, Valentine Mariani, Marco Tanganelli, Stefania Viti,

and Mario De Stefano
Abstract The seismic assessment of existing buildings is an essential issue of seismic

engineering. This work is focused on the evaluation of the seismic performance of
existing RC buildings according to the current European Technical Codes. Alternative
types of analyses, all consistent to the Code provisions, have been performed with
reference to a case-study, that is a real, RC hospital building. An accurate knowledge
of the building has been achieved, as a result of a collaboration between the University
of Florence and the Regional Government. Both elastic and inelastic modeling, as well
as static and dynamic one, have been adopted in the analysis. The global response –
with special attention to torsional effects – and the seismic performance of each single
member have been found with all the performed analyses. The comparison among the
analyses has been performed in terms of both global and local response parameters,
and the reliability of each analysis has been pointed out.
Keywords Seismic performance of existing buildings • Seismic analyses • Seismic

assessment of RC structures
12.1 Introduction
The evaluation of the seismic performance of existing buildings is a crucial issue of

seismic engineering. Due to its intrinsic complexity, the evaluation of seismic perfor-
mance is affected by many uncertain factors even more in case of irregular buildings.
Among these factors, special attention has to be paid to the intrinsic variability of the
mechanical properties of materials, which can also become source of irregularity
(De Stefano et al. 2013a, b, 2014a), and to the analytical procedure adopted in analysis
A. La Brusco (*) • V. Mariani • M. Tanganelli • S. Viti • M. De Stefano

Department of Architecture (DiDA), University of Florence, Piazza Brunelleschi, 6,
50121 Florence, Italy
e-mail: [email protected]; [email protected]; [email protected];
[email protected]; [email protected]

DOI 10.1007/978-3-319-14246-3_12
136 A. La Brusco et al.
(De Luca et al. 2011; De Luca and Verderame 2011). The current European Technical
Code, Eurocode 8 (EC8-3 2005), in fact, allows the designer to choose the type of
analysis, according to some guidelines. It is well known that the choice about material
behavior (linear or nonlinear) and type of performed analysis (static or dynamic)
strongly affects the analytical reproduction of the seismic behavior of structures and,
consequently, their performance prediction. In this framework, this work deals with
the effects of the type of performed analysis on the seismic performance evaluation of
existing buildings. The research is carried out with reference to a case-study, i.e. a real
RC building, lightly irregular in plan, currently used as a hospital. The evaluation of its
seismic performance is based on a wide knowledge process, that is the result of a joint
agreement with the Regional Government of Tuscany.
In the first part of the paper the response of the case-study under a seismic
excitation defined according to EC8 provisions has been found. Special attention
has been paid to its torsional response. For slightly irregular structures, indeed,
seismic codes allow to neglect torsional effects in the analysis, despite they can
possibly induce not negligible amplifications in the seismic response.
In the second part of the paper the seismic performance has been evaluated
according to EC8 provisions. Two different limit states, i.e. a serviceability (Dam-
age Limitation, DL) and a ultimate one (Life Safety, LS), have been considered in
the analysis. The results obtained for the seismic assessment of the case-study by
performing three different types of analytical procedure are shown and compared.
12.2 Case Study
The building, having a RC skeleton, has been designed in 1976, i.e. just after the
introduction of the first seismic Italian Technical Code. Therefore, the building
presents some efficient design criterions, like column section reduction from
foundation level to the top storey, or solid connection of the beam-column joints,
although it is far away from complying the current seismic design criteria. The
3-storey building, shown in Fig. 12.1, has a regular (rectangular) plan, with one
symmetric axis only.
Fig. 12.1 Plan and sections (measures in meters)

12 Seismic Assessment of an Existing Irregular RC Building According to. . . 137
Table 12.1 Cross section dimensions (in cm) of columns and beams
Columns Beams
Name b h Name bh
1 2
1 st. 2 st. 3 st. st. st. 3 st, a 3 st, b
c1–c9 30 50 30 40 30 30 x1,2 x2,3 x7,8 x8,9 30 60 30 20
y3,6 y6,9 Z-shape
x4,5 x5,6 30 80 30 20
30 60
y1,4, y2,5, y4,7; y5,8 30 80 –
30 60
In Table 12.1 cross section dimensions of columns and beams are reported, while
the area of the reinforcements is around 1 % of the cross section area. Details about
number and type of rebars have been illustrated in De Stefano et al. (2014b). Some
beams, having a Z-shape, have been modeled by means of a rectangular section
with equivalent inertia. It should be noted that the third floor of the building consists
of two different structural layers, partially coinciding. In fact two different floors,
40 cm far away each other, constitute the last storey of the building. In some
alignments (X3, Y1 and Y3) two layers of beams separately support the two different
floors, whilst in the other alignments (X1, X2 and Y2) a single beam supports both
floors. Therefore the third floor and the related beams will be in the following
distinguished with a subscript a or b, depending on weather they refer to the lower
or upper layer respectively.
Mechanical properties of structural materials, i.e. concrete and steel, have been
determined through destructive and not-destructive tests, according to the Italian
Seismic Code provisions (NTC 2008). Three destructive tests on concrete have
been performed on the columns, which have been integrated by other SonReb
(sclerometric + ultrasonic methods) tests, extended even to the beams. The final
compressive strength has been found by combining both destructive and SonReb
results, by adopting an ad hoc expression (Cristofaro 2009; Cristofaro et al. 2012),
which has provided a final cylindrical strength, fc,mean equal to 10.2 MPa. The
global knowledge level achieved for the concrete structural elements has been
conservatively evaluated as Knowledge Level 2 (KL2, CF ¼ 1.20), according to
EC8, associated to a Confidence Factor (CF) equal to 1.20, leading to a design
(reduced) value of strength, fcd, equal to 8.5 MPa.
Two different types of rebars, respectively ribbed and not, have been adopted for
reinforcement. The reinforcement steel has been classified as FeB32K class, with a
yield stress over 320 MPa and a ultimate stress over 500 MPa. Three destructive
tests, one for each storey, have been done on rebars samples, according to the
standard procedure (UNI EN ISO 6892 2009), returning a mean value, fs,mean, equal
to 385.7 MPa. Since a Knowledge Level KL1 (CF equal to 1.35) has been assumed
for steel, a design strength, fsd, equal to 285.7 MPa has been used for analysis.
Ground mechanical properties have been determined through a geophysical site

investigation, consisting in seismic refraction and down hole techniques. The
results, processed using both Generalized Reciprocal Method and tomographic
one, have an average shear wave velocity Vs,30 (Sirles and Viksne 1990) equal to
390 m/s. Therefore the soil has been assumed as B-type, according to NTC (2008).
12.3 Seismic Analysis
All analyses have been performed by using the computer code SAP2000 (2009).
The two floors of the third level have been modeled according to the real geometry
as regards the stiffness and strength distribution, while the mass (both translational
and rotational) of the storey has been considered applied at the center of the storey
package. The effect of the joint stiffness has been considered by introducing a rigid
offset at each element end. A 50 % reduced value of the Young modulus
(Ec, red ¼ 11,072 MPa) of the concrete has been assumed, as suggested by NTC
(2008). The floor stiffness has been introduced by assigning the diaphragm con-
straint to all nodes belonging to the same floor. Each member is modeled by an
elastic finite element with terminal plastic hinges, whose properties have been
defined by assigning a bi-linear moment-rotation relationship, defined by the
yield and the ultimate points. Limit values of bending moment and rotation have
been made according to EC8 prescriptions.
The expected maximum seismic intensity of the area, measured in terms of Peak
Ground Acceleration (PGA), is provided by NTC, as well as the shape of the elastic
spectrum of the case-study. To perform the dynamic analysis, two different sets of
ground motions have been considered, whose average spectra closely approach the
NTC (2008) one for the two considered limit states. They have been provided by
Working Group Itaca (Itaca 2008), on the basis of a PGA equal to 0.25 g, a nominal
life of the structure of 50 years and a magnitude between 5.5 and 6.5.
The ultimate capacity associated to LS limit state of each member is evaluated in
terms of chord rotation or bending moment, depending on the type of performed
analysis, for ductile mechanisms, and in terms of ultimate shear for brittle ones.
Regarding the serviceability (DL) limit state, instead, a limit storey drift equal to
5 ‰ has been assumed.
The structural response of the case-study has been found by performing three
different analyses, i.e. the linear pseudo-dynamic, the nonlinear static and the
nonlinear dynamic ones. As regards the nonlinear static analysis, in recent years
different improvements have been introduced (Fajfar et al. 2005; D’Ambrisi
et al. 2009; Bhatt and Bento 2014) to account for structural irregularities. In the
current work, anyway, the standard N2 method, as provided by EC8, has been
applied.
The seismic behavior of the case-study has been described by checking both
global and local response parameters. The considered global parameters are the Top
Displacement (TD) and the torsional effects, measured as Normalized Top
Displacement (NTD), i.e. the top displacement at each column-line normalized to

the one at the Mass Center (MC). The local response parameters are the ones
provided by EC8 to define the seismic performance, depending on the performed
analysis.
In the pseudo-dynamic linear analysis the seismic response of the case-study has
been evaluated by considering the values of the spectral acceleration for the first
five periods, found through a preliminary modal analysis, and then combining the
consequent effects according to their participation factors. Twelve different ana-
lyses have been run for each direction, since the lateral force must be applied at MC
with no eccentricity (e ¼ 0) and with an eccentricity equal to 5 %. In all cases the
response must be combined to a quote, equal to 30 %, of the response in the
orthogonal direction. The nonlinear static (pushover) analysis has been performed
by considering two alternative heightwise horizontal patterns, respectively propor-
tional to masses and to the first vibrational mode. For each pattern the two ways and
a 5 % eccentricity must be considered, with a total number of analyses equal to
12 for each direction. The nonlinear dynamic procedure consists of seven different
analyses, one for each ground motion, for each direction. According to EC8 the
seismic response to be considered is the maximum one for pseudo-dynamic and
pushover analyses, and the mean one for the nonlinear dynamic analysis.
12.4 Seismic Response of the Case-Study
In order to quantify the in-plan irregularity, the eccentricity along the two main
directions has been found in terms of mass, stiffness and strength. At each storey,
the MC has been found by considering the mass of the floors and of the infill panels;
the strength center has been found as a function of the concrete strength in the
columns, while the center of stiffness has been determined by applying the simpli-
fied relationship proposed by Anagnastopoulos (Anagnastopoulos et al. 2013) and
applied in De Stefano et al. (2015). In Table 12.2 the eccentricities between MC and
the strength (estr) and stiffness (estiff) centers are listed for each storey. It can be
noted that, despite the building is structurally symmetric about the Y-direction, it
has an irregular infill panels distribution, and therefore it presents an eccentricity in
both directions.
Figure 12.2 shows the structural response obtained by performing the three types
of analysis in terms of Top Displacement and total Base Shear. The capacity curves
Table 12.2 Strength and stiffness eccentricity at each storey

X-direction Y-direction
Level estr (%) estiff (%) estr (%) estiff (%)
1st storey 9.00 10.71 3.33 2.23
2nd storey 7.48 9.41 5.61 5.00
3rd storey 6.08 7.95 0.76 3.17
800 800
LINEAR_q=1.5
700 700
Base Shear (kN)

Base Shear (kN)
600 600 LINEAR_q=3.0

500 500
400 400 PUSHOVER_mass
300 300 proportional
200 200 PUSHOVER_1mode
proportional
100 100
X-direcon Y-direcon Time History_LS
0 0
0 20 40 60 80 100 120 0 20 40 60 80 100 120
Time History_DL
Top Displacement (mm) Top Displacement (mm)
Fig. 12.2 Seismic response of the case-study
found by performing the nonlinear static (pushover) analysis have been compared
to the response found by performing the nonlinear dynamic analysis. For each
ground motion, the maximum Base Shear and Top Displacement values have been
extracted and reported as a point in Fig. 12.2. The nonlinear static and dynamic
analyses gave consistent results, since the points fit very closely the capacity curves.
The response found by performing the linear analyses refers to two different q-
values, respectively assumed equal to 3.0 (ductile mechanisms) and equal to 1.5
(brittle mechanisms). The two families of points lye on the same line (the blue
dashed line in Fig. 12.2), so evidencing the linearity of the response.
The elastic stiffness related to the elastic analysis coincides to the pushover one
(first mode proportional). The families of capacity curves found by assuming the
two different heighwise patterns slightly differ each other for elastic stiffness, shear
and displacement capacity.
The torsional effects due to the in-plan irregularity has been checked by plotting
the maximum Top Displacement at the MC and at each side of the case-study.
Figure 12.3 shows the NTD at each side of the building for each considered limit
state. As it was expectable, the torsional effects are larger in the serviceability limit
state than in the ultimate one. The comparison of the torsional effects found by the
three analysis types shows that the elastic analysis provides the larger effects in all
cases. The two inelastic analyses provide similar results along the X-direction
(larger eccentricity), while in the Y-direction the pushover analysis provides values
of NTD much larger than the ones obtained by the dynamic analysis.
12.5 Seismic Performance of the Case-Study
In this section the seismic performance of the case-study has been checked
according to EC8 prescriptions. The DL limit state has been checked in terms of
maximum drift, by comparing the drift distribution provided by each analysis to the
5 ‰ limit imposed by EC8. Since the larger drifts occur at the flexible edge of the
structure (De Stefano et al. 2014b, c), for sake of brevity in Fig. 12.4 only the drift
values found at the flexible side have been shown. The dashed lines represent the
maximum drift obtained by each analysis. The results provided by the linear
analysis refer to a behavior factor equal to 3.0, since the maximum drift has to be
1.30 1.30
DL, X-direcon DL, Y-direcon
1.20 1.20 Linear
1.10 1.10
Pushover, mass-
NTD
NTD
1.00 1.00 proportional

0.90 0.90 Pushover, 1-mode
0.80 0.80 distribution
0.70 0.70 Dynamic

Y1 MC Y3 X1 MC X3
1.30 1.30
LS, X-direcon LS, Y-direcon
1.20 1.20
1.10 1.10
NTD
NTD
1.00 1.00
0.90 0.90
0.80 0.80
0.70 0.70
Y1 MC Y3 X1 MC X3
Fig. 12.3 Normalized top displacement
3 3
2 2
Floor
Floor
1 1
column-line Y1 column line X3

0 0
0.0% 0.5% 1.0% 0.0% 0.5% 1.0%
Interstory Dri (%) Interstory Dri (%)
Linear Pushover, 1-mode distribution

Pushover, mass-proportional Dynamic
Fig. 12.4 Maximum drift at the flexible side of the structure at the DL limit state according to the
three performed analyses
related to the flexural (ductile) behavior of the case-study. Concerning the nonlinear
dynamic analysis, the maximum response is the mean value of the seven ground
motions, according to EC8 prescriptions. It should be noted that the drift values
provided by the dynamic analysis are very scattered, much more than the ones
provided by the alternative methods.
The 5 ‰ limit provided by EC8 for the DL limit state is respected in all cases.
The LS limit state has been studied by checking the ratio between the capacity (C)
of each member (beams and columns) and the corresponding demand (D). The
results found by performing each analysis are shown for the first storey only, where
the maximum response has been found. In Figs. 12.5, 12.6, and 12.7 the C/D values
exceeding the value of 5 have been shown with an arrow at the border of the figure.
1st level, ductile 1st level, brittle

5 5
columns_dir X
4 4
C/D 3 3
C/D
2 2
1 1
0 0
c1 c2 c3 c4 c5 c6 c7 c8 c9 c1 c2 c3 c4 c5 c6 c7 c8 c9
COLUMNS COLUMNS
5 5
4 4
beams_dir X
3 3
C/D
C/D
2 2
1 1
0 0
x 2-3
x 4-5
x 5-6
x 7-8
x 2-3
x 4-5
x 5-6
x 7-8
x 8-9
x 8-9
x 1-2
x 1-2
BEAMS BEAMS
5 5
columns_dir Y
4 4
3 3
C/D
C/D
2 2
1 1
0 0
COLUMNS COLUMNS
5 5
4 4
beams_dir Y
3 3
C/D
C/D
2 2
1 1
0 0
y 2-5
y 3-6
y 4-7
y 5-8
y 2-5
y 3-6
y 4-7
y 5-8
y 6-9
y 1-4
y 6-9
y 1-4
BEAMS BEAMS
L+E0 L-E0 L+E+ L-E+ L+E- L-E-
Fig. 12.5 Seismic performance of the first storey of the structure along the X and Y directions
obtained by pseudo-dynamic linear analysis
Figure 12.5 shows the C/D values obtained by performing the pseudo-dynamic
linear analysis for ductile and brittle mechanisms at the first storey beams and
columns. It can be observed that some structural members exhibit C/D values lower
than 1 both in the analyses along X and Y directions, not in compliance with the EC8
limit conditions; in particular the columns are more vulnerable to ductile failure,
while beams to the brittle one.
Figure 12.6 shows the results obtained by non linear static analysis in terms of
C/D values for each structural member. Once again it can be noted that the columns
result more sensitive to flexural failure, whilst beams suffer more for shear failure,
even if, in this case, only few beams (in the Y-direction) present C/D values lower
than one.

5 5
columns_dir X
4 4
3 3
C/D
C/D
2 2
1 1
0 0
COLUMNS COLUMNS
5 5
4 4
beams_dir X
3 3
C/D
C/D
2 2
1 1
0 0
x 2-3
x 4-5
x 5-6
x 7-8
x 8-9
x 1-2
x 2-3
x 4-5
x 5-6
x 7-8
x 8-9
x 1-2
BEAMS BEAMS
5 5
columns_dir Y
4 4
3 3
C/D
C/D
2 2
1 1
0 0
COLUMNS COLUMNS
5 5
4 4
beams_dir Y
3 3
C/D
C/D
2 2
1 1
0 0
y 2-5
y 3-6
y 4-7
y 5-8
y 6-9
y 1-4
y 2-5
y 3-6
y 4-7
y 5-8
y 6-9
y 1-4
BEAMS BEAMS
PM+E0 PM-E0 PM+E+ PM-E+ PM+E-
P1+E0 P1-E0 P1+E+ P1-E+ P1+E-
Fig. 12.6 Seismic performance of the first storey of the case study along the X and Y-directions
obtained by nonlinear static analysis
Figure 12.7 shows the C/D values obtained by non linear dynamics analysis. In
this case the results are very scattered, as a consequence of the influence of the
seismic input on the response. Finally, Fig. 12.8 shows a comparison among the
minimum values of seismic performance (ratio C/D) found by the performed
analyses for columns (c) and beams (b) for brittle (brit) and ductile (duc) mecha-
nisms respectively.
The linear analysis proves to be the most conservative, providing C/D values
below the unity in many cases. The nonlinear analyses provide very similar results
in some cases, as the brittle mechanism in columns, while they differ each other in
the performance evaluation of beams. As it can be noted by diagrams in Fig. 12.8,
the seismic performance found by performing the dynamic analysis satisfies the

5 5
columns _dir X
4 4
3 3
C/D
C/D
2 2
1 1
0 0
COLUMNS COLUMNS
5 5
4 4
beams_dir X
3 3
C/D
C/D
2 2
1 1
0 0
x 2-3
x 4-5
x 5-6
x 7-8
x 8-9
x 1-2
x 2-3
x 4-5
x 5-6
x 7-8
x 8-9
x 1-2
BEAMS BEAMS
5 5
columns_dir Y
4 4
3 3
C/D
C/D
2 2
1 1
0 0
c1 c2 c3 c4 c5 c6 c7 c8 c9 c1 c2 c3 c4 c5 c6 c7
COLUMNS COLUMNS
5 5
4 4
beams_dir Y
3 3
C/D
C/D
2 2
1 1
0 0
y 2-5
y 3-6
y 4-7
y 5-8
y 6-9
y 1-4
y 2-5
y 3-6
y 4-7
y 5-8
y 6-9
y 1-4
BEAMS BEAMS
Fig. 12.7 Seismic performance of the first storey of the case study along the X and Y-directions
obtained by nonlinear dynamic analysis
6 6 6
1st level, X-dir. 5 2nd level, X-dir. 3rd level, X-dir.
5 5
4 4 4
C/D
C/D
C/D
3 3 3
2 2 2
1 1 1
0 0 0
c_brit c_duc b_brit b_duc c_brit c_duc b_brit b_duc c_brit c_duc b_brit b_duc
6 6 6
1st level, Y-dir 2nd level, Y-dir 3rd level, Y-dir.
5 5 5
4 4 4
C/D
C/D
C/D
3 3 3
2 2 2
1 1 1
0 0 0
c_brit c_duc b_brit b_duc c_brit c_duc b_brit b_duc c_brit c_duc b_brit b_duc
linear pushover time-history
Fig. 12.8 Comparison among the performance level found by the performed analyses
EC8 requirements in all cases. When the nonlinear static analysis is performed, the
brittle mechanism in the beams does not satisfy the EC8 limit conditions, while all
the other requirements are respected.
12.6 Conclusive Remarks
In this work the seismic assessment of a real RC existing building has been carried
out, by performing three different types of analysis, according to the current
European Technical Code (EC8): the pseudo-dynamic elastic, the non-linear static
and the nonlinear dynamic ones. The seismic response of the case-study has been
investigated by checking both global and local response parameters. Concerning
the global response of the building, TD and NTD have been checked. The nonlinear
analyses provided similar results in terms of TD-Base Shear trends. Even the elastic
analysis has approached very well the elastic results found by the nonlinear
analysis, providing seismic responses (i.e. Base Shear and TD) which very much
differ each other as a function of the assumed behavior factor, respectively equal to
1.5 (brittle mechanisms) and 3.0 (ductile mechanisms); all results, anyway, lie on
the ideal axis representing the elastic response of the structure, coincident to the
elastic branch of the pushover curves for horizontal forces proportional to the first
mode of vibration.
The torsional response proved to be more sensitive to the type of performed
analysis. As it was expectable, the torsional effects resulted larger in the service-
ability limit state than in the ultimate one; the linear analysis provided larger top
storey rotation than the other two analyses, in all cases. Along the X-direction, the
one with a larger eccentricity, the two nonlinear analyses have exhibited similar
results, while in the Y-direction the pushover analysis is more conservative than the
dynamic one, since the 5 % eccentricity provided by EC8 largely covers the
effective eccentricity. Concerning the serviceability limit state, the three analyses
lead to similar results, providing a satisfactory response of the case-study. The
seismic performance referred to the ultimate limit state, instead, has turned out to be
more sensitive to the type of adopted analysis. The linear analysis resulted more
conservative than the other ones, providing C/D values below the unity in many
cases.
The two nonlinear analyses (pushover and dynamic) have provided very similar
results in some cases, e.g. for the brittle mechanism in columns, while they differ
each other in the performance evaluation of beams. Results obtained by the
dynamic analysis satisfy all the EC8 requirements. When the nonlinear static
analysis is adopted, instead, the brittle mechanism in the beams does not satisfy
the EC8 limit values, while all the other requirements are fulfilled.
As a conclusion, the three analyses have evidenced important differences both in
terms of global representation of the seismic response, including torsional effects,
and in terms of local seismic performance according to the EC8 procedure. Despite
each analysis can be assumed to evaluate the seismic assessment of existing

buildings, the choice of the analysis type can largely affect the final evaluation of
the safety level of the building.
Acknowledgements The financial support provided by ReLUIS within the project “ReLUIS-
DPC 2014” (Progettazione e valutazione della sicurezza e della vulnerabilita di edifici ed opere) is
gratefully acknowledged.
References
Anagnastopoulos SA, Kyrkos MT, Stathopoulos KG (2013) Earthquake induced torsion in build-
ings: critical review and state of art. Proc. ASEM13, Jeju Korea. Thecno-Press Journals
Bhatt C, Bento R (2014) The extended adaptive capacity spectrum method for the seismic
assessment of plan asymmetric buildings. Earthq Spectra (in-press)
Cristofaro MT (2009) Metodi di valutazione della resistenza a compressione del calcestruzzo di
strutture in c.a. esistenti, tesi di dottorato, Universita degli Studi di Firenze (in Italian)
Cristofaro MT, D’Ambrisi A, De Stefano M, Pucinotti R, Tanganelli M (2012) Studio sulla
dispersione dei valori di resistenza a compressione del calcestruzzo di edifici esistenti.
Giornale Prove Non Distruttive Monitoraggio Diagnostica 2/2012:32–39, ISSN: 1721-7075
(in Italian)
D’Ambrisi A, De Stefano M, Tanganelli M (2009) Use of pushover analysis for predicting seismic
response of irregular buildings: a case study. J Earthq Eng 13:1089–1100, Taylor & Francis
De Luca F, Verderame GM (2011) A practice-oriented approach for the assessment of brittle
failures in existing reinforced concrete elements. Eng Struct 48:373–388
De Luca F, Verderame GM, Manfredi G (2011) La verifica di edifici esistenti in cemento armato:
criticita dell’attuale approccio normativo italiano. 26 convegno nazionale AICAP, Padova
19–21 May 2011 (in Italian)
De Stefano M, Tanganelli M, Viti S (2013a) On the variability of concrete strength as a source of
irregularity in elevation for existing RC buildings: a case study. Bull Earthq Eng 11(5):
1711–1726. doi:10.1007/s10518-013-9463-2, ISSN: 1573-1456
De Stefano M, Tanganelli M, Viti S (2013b) Effect of the variability in plan of concrete
mechanical properties on the seismic response of existing RC framed structures. Bull Earthq
Eng 11(4):1049–1060. doi:10.1007/s10518-012-9412-5, ISSN: 1573-1456
De Stefano M, Tanganelli M, Viti S (2014a) Variability in concrete mechanical properties as a
source of in-plan irregularity for existing RC framed structures. Eng Struct Vol 59:161–172
De Stefano M, La Brusco A, Mariani V, Tanganelli M, Viti S (2014b) The role of structural
modeling on the seismic assessment of existing RC buildings according to Eurocode 8. In: 2nd
European conference on earthquake engineering and seismology. Istanbul, 25–29 Aug
De Stefano M, Tanganelli M, Viti S (2015) Torsional effects related to concrete strength variabil-
ity in existing buildings: a numerical analysis. Earthq Struct 8(2):379–399
EC 8-3 (2005) Design of structures for earthquake resistance, part 3: strengthening and repair of
buildings, European standard EN 1998-3. European Committee for Standardization (CEN),
Brussels
Fajfar P, Marusic D, Perus I (2005) Torsional effects in the pushover-based seismic analysis of
buildings. J Earthq Eng 9(6):831–854
Itaca (2008) Database of the Italian strong motions data. http://itaca.mi.ingv.it
NTC (2008) Norme tecniche per le costruzioni. D.M. Ministero Infrastrutture e Trasporti
14 gennaio 2008, G.U.R.I. 4 Febbraio 2008, Roma (in Italian)
SAP2000 Advanced 14.0.0, Structural analysis program (2009) Analysis reference manual. Com-
puter and Structures, Berkley
Sirles C, Viksne A (1990) Site-specific shear wave velocity determinations for geotechnical
engineering applications. In: Geotechnical and enviromental geophysics, vol. 3, Soc. Expl.
Geophys. (Tulsa, Oklahoma), pp. 121-131
UNI EN ISO 6892-1 (2009) Materiali metallici – Prova di trazione – Parte 1: Metodo di prova a
temperatura ambiente, International Organization for Standardization, Geneva
Chapter 13
The Concrete Strength Variability as Source
of Irregularity for RC Existing Buildings
Stefania Viti, Marco Tanganelli, and Marco De Stefano
Abstract Existing buildings can easily present material mechanical properties

which can largely vary even within a single structure. As a consequence of the
high strength variability, at the occurrence of seismic events the structure may
evidence unexpected phenomena, like torsional effects, with larger experienced
deformations and, in turn, with reduced seismic performance. This work is focused
on the reduction in seismic performance due to the concrete strength variability.
The analysis has been performed on a case-study, i.e. a 3D RC framed 4 storey
building. A normal distribution, compatible to a large database, has been taken to
represent the concrete strength domain. Due to the introduced strength variability, a
stiffness and strength eccentricity arises at the first storey of the structure, with a
consequent increase in its seismic response. The capacity (C) of each column of the
case study, found according to EC8 prescriptions, has been compared to the demand
(D), for three different limit states. The seismic performance found by accounting
for the strength variability has been compared to the one provided by the EC8
standard procedure, which has resulted to be conservative in the seismic response
estimation and in the evaluation of the torsional effects, while it is not conservative
in quantifying the seismic performance of the case-study.
Keywords Existing RC buildings • Concrete strength variability • Concrete

mechanical properties • Seismic response of irregular structures • Torsional effects
13.1 Introduction
One of the most crucial technical issues of seismic engineering is the evaluation of
the seismic safety of existing buildings, which involves a suitable characterization
of actual material properties. Especially in RC buildings, the homogeneity of the
material within each structure cannot be assumed, since the strength variability
inside single buildings can result in CoV values over 30 % (Cristofaro 2009).
S. Viti (*) • M. Tanganelli • M. De Stefano


DOI 10.1007/978-3-319-14246-3_13
150 S. Viti et al.
Furthermore, an exhaustive characterization of the concrete mechanical properties

is hard to achieve, since it would require a large number of destructive and
not-destructive tests (Cristofaro et al. 2012).
International technical legislation provides different criteria to identify the
required mechanical properties of concrete. The European technical Code,
Eurocode 8 (EC8-3 2005) indicates the mean value of the strength domain for
analysis, while it prescribes a reduced strength value (Franchin et al. 2007; Rajeev
et al. 2010; Jalayer et al. 2008; Monti et al. 2007) for verification. The reduction is
made by introducing a Confidence Factor (CF), ranging between 1.00 and 1.35
(Italian Annex) depending on the knowledge level of the structure. The safest
approach provided by EC8, therefore, consists of evaluating the structural response
by using the mean concrete strength, and to compare member forces and deforma-
tions to their limit values found by assuming a CF ¼ 1.35. The amount of strength
variability, as represented by CoV, does not affect anyway the analysis.
In this paper the concrete strength variability has been investigated on a case-
study, i.e. a 4-storey RC building, as possible source of in-plan irregularity, and the
torsional effects related to such irregularity have been evaluated. The concrete
strength has been characterized on the base of a large database provided by the
Regional Government of Tuscany (Cristofaro 2009). The strength variability has
been introduced at the columns belonging to first storey only, while the other
columns, as well as all the beams, are characterized by the mean value of the
strength domain. This investigation follows some previous works made by the
authors (De Stefano et al. 2013a, b, 2014) on a similar case-study, which have
proved the strength variability to largely affect both the seismic demand and
performance of existing buildings. In De Stefano et al. (2014) only two “extreme”,
very skewed, in-plan strength distributions have been considered. Therefore, only
the largest possible effects related to the assumed strength variability have been
found, while no information was found on the probability of occurrence of such
effects. In this work, instead, a significant number (180) of different in-plan
distributions have been considered, qualitatively representing all the possible com-
binations among the values of the strength domain. Depending on the considered
in-plan strength distributions, different amounts of strength and stiffness eccentric-
ity (Bosco et al. 2012, 2013) have been found. The maximum seismic demand
erasing by the introduced eccentricity has been compared to the demand provided
by EC8, assuming an accidental eccentricity of 5 %.
The capacity of each column, expressed in terms of chord rotation and shear
force, has be determined with reference to three different limit states, both
according to EC8 and as a function of the strength variability.
Finally, the performance of the case-study has been investigated by comparing
capacity and demand in terms of chord rotation and shear force. While the chord
rotation is the reference quantity for all limit states, the shear force has been
considered only for the Near Collapse limit state. The comparison evidenced that
EC8 approach for seismic performance evaluation, including the 5 % accidental
eccentricity, does not cover the effects due to the considered strength variability.
13 Concrete Strength Variability 151
13.2 The Analysis
The sample structure is a 4 story 3D reinforced concrete frame with two 4.5 m long
bays in the y-direction and five bays 3.5 m long in the x-direction, as shown in
Fig. 13.1. The building, designed for vertical loads only, is symmetric along both
x and y directions. All the columns have a cross dimension of 30 30 cm while
longitudinal beams have a dimension of 30 50 cm in both directions. Details of
joints and reinforcements can be found in De Stefano et al. (2015a). A mean
concrete strength ( fc), equal to 19.36 MPa, has been assumed, while the reinforce-
ment is assumed to have the same mechanical properties as the Italian FeB38k steel
(yield stress over 375 MPa, ultimate stress over 430 MPa). The strength variability
has been introduced in the columns of the first storey only, by assuming a Normal
distribution and three different levels of Coefficient of Variation (CoV), equal to
15 %, 30 % and 45 % respectively. Each strength domain is made by seven values,
corresponding to different percentiles (5 %, 10 %, 20, 50 %, 80 %, 90 % and 95),
which have been given to the 18 columns of the first storey in order to obtain the
expected distribution; in each model the strength values are assorted according to
the specifications listed in Fig. 13.1.
The first vibrational period of the case-study (see Table 13.1), corresponding to a
translational vibration along the y-direction, is equal to 0.777 s. The analysis has
been performed by considering one direction only for the seismic excitation, i.e. the
y-direction. Therefore, the variability has been given along the x-direction.
Six groups of 30 schemes each have been considered, having the weakest
position in each of the column lines of frame 2 (see Fig. 13.1). For each position
of the weakest column, all the most significant strength combinations have been
considered. In the assumption of the 180 layouts special attention has been paid to
the “extreme” considered strength values, i.e. the values corresponding to the
percentiles k05, k10, k90 and k95, that have been exhaustively combined. A
more complete description of the models can be found in De Stefano et al. (2015a).
The analysis has been performed by using the computer code Seismostruct
(Seismosoft 2006) and describing the cross sections through a fiber model, which
has allowed to adopt different models for the concrete of the confined core (Mander
et al.), the unconfined concrete of the cover (three-linear) and the reinforcement
(bilinear). Contribution of floor slabs has been considered by introducing a rigid
diaphragm.
The inelastic response of the case-study has been found by applying the standard
N2 method, as provided by EC8. Three different limit states, i.e. Damage Limita-
tion (DL), Severe Damage (SD) and Near Collapse (NC) have been considered.
Each limit state has been associated to a single seismic intensity, with PGAs equal
to 0.15 g, 0.20 g and 0.25 respectively. Such values of PGA have been assumed on
the basis of the highest seismicity occurring in Tuscany, i.e. the region where the
investigation on the concrete strength (Cristofaro 2009) has been made.
The seismic input has been assumed to be represented by the elastic spectrum
provided by EC8 for a soil-type B. Since in De Stefano et al. (2013a) has been found
152 S. Viti et al.
a b c d e f percentile No of columns
K05 1
y K10 2
1 K20 3
K50 6
2
x K80 3
K90 2
K95 1
3
Fig. 13.1 Case-study: plan configuration and fc assumption at the first storey columns
Table 13.1 Periods and participating masses of the structure

Period Mode Participating mass
1st period 0.777 s Translational, y 87.42 %
2nd period 0.735 s Translational, x 88.14 %
3rd period 0.694 s Torsional, z 87.87 %
that the assumed horizontal pattern distribution does not significantly affect the
results, in this work only one force pattern, proportional to the first vibration mode,
has been considered.
13.3 The Introduced Eccentricity
Due to the introduced strength variability, each considered plan layout presents
both a strength and a stiffness irregularity, since the Young modulus Ec is defined as
a function of fc (EC2 2002). Both strength and stiffness eccentricities have been
found; the strength eccentricity, eV, has been expressed in terms of the ultimate
shear of the columns while the stiffness eccentricity, eK, has been expressed through
a simplified stiffness (K) expression (Anagnastopoulos et al. 2009), i.e. K ¼ (My H )/
(6 θy). In Table 13.2 the ranges obtained for the two eccentricities are listed as a
function of the assumed CoV.
Another quantity adopted to characterize the strength plan layout is the radius of
gyration (Anagnastopoulos et al. 2013), which gives a measure of the strength
centrifugation. Two radii of gyration have been found, i.e. based on the ultimate
shear (ρV), and the simplified shear type behavior (ρK). Figure 13.2 reports their
values, nondimensionalized with respect to the mass radius. Values of nondimen-
sionalized radii of gyration, being larger than the unity, show that the building
models can be classified as moderately torsionally stiff and strong.
Table 13.2 Ranges of eccentricities (eK, eV) due to the introduced strength variability
CoV ¼ 15 % CoV ¼ 30 % CoV ¼ 45 %
Min Max Min Max Min Max
eK 1.7 % 1.7 % 3.6 % 3.6 % 6.1 % +6.1 %
eV 0.6 +0.6 % 1.4 % +1.4 % 2.8 % +2.8 %
Fig. 13.2 Radius of 1.30

K_max
gyration for the considered
radius of gyration
CoV strength 1.25 K_min
1.20 V_max
1.15 V_min
1.10
15% 30% 45%
CoV_fc
13.4 Results
The seismic response at the first storey of the case study has been studied in terms of
maximum drift and shear force at each column. In Fig. 13.3 the results obtained
from analysis and from EC8 standard approach, including the 5 % eccentricity, are
shown for three different PGAs. Since the columns belonging to the frame 2 evi-
dence an higher seismic response, only the results referred to such columns have
been shown in this paper. A more exhaustive description of the results can be found
in De Stefano et al. (2015b).
The variability in the drift response increases both with the seismic intensity and
the strength variability as represented by CoV. Moreover, at the PGA increasing,
the torsional effects arising from the introduced eccentricity increase as well,
achieving their maximum amount at the side columns (column lines a and f ).
The structural demand provided by EC8, when the 5 % eccentricity is introduced,
covers the increase in the side drift due to torsional effects. It should be reminded,
anyway, that the 5 % eccentricity provided by EC8 is aimed to cover all the
accidental irregularities and not the one related to the strength variability only.
The shear response is not affected by torsional effects, despite it presents a large
scatter due to the introduced strength variability.
Figures 13.4 and 13.5 show, in terms of chord rotation and shear force respec-
tively, the capacity of the columns for the considered limit states. The limit values
have been found by considering, for the concrete strength, both the assumed
samples and the EC8 instructions. According to EC8, only one ultimate limit
state, namely the NC one, has been considered for shear verification. The seismic
performance is measured as the ratio between demand (D) and capacity (C). The
seismic intensity to assume for each limit state has been selected according to the
seismicity of Tuscany, since the concrete strength characterization refers to
154 S. Viti et al.
Fig. 13.3 Seismic demand (first storey, frame 2)
buildings of such area. When the ratio D/C is lower than unity the structure respects
the limit provisions, while for values over unity the structure results to be not
compatible with the safety requirements.
In Fig. 13.6 the obtained values of D/C are shown for the considered limit states.
Both the D/C values obtained by considering the strength variability and the EC8
procedure, by accounting for the 5 % eccentricity and not, are shown for each
column. It can be noted that the D/C ratio is sensitive both to PGA and CoV.
As regards the DL limit state, the seismic performance of the case-study has been
evaluated for a PGA equal to 0.15 g. The ratio D/C found by considering the
strength variability is below the unity only for CoV ¼ 15 %. For higher CoVs the
limit value is exceeded, approaching the value of 2 for CoV ¼ 45 %. It should be
noted that the EC8 approach provides D/C always below the unity if the 5 %
eccentricity is not considered, while it achieves 1.25 when the eccentricity is taken
into account.
The SD limit state has been associated to a PGA equal to 0.20 g. The perfor-
mance limits are exceeded both when the EC8 approach is applied and when the
Fig. 13.4 Limit chord rotations (frame 2, first storey columns)
Fig. 13.5 Limit shear force (frame 2, first storey columns)
strength variability is introduced. It should be noted, anyway, that the maximum

D/C value coming from the higher strength variability is 3.5 times higher than the
corresponding value provided by EC8, even when the 5 % eccentricity is consid-
ered. Similar observations can be done for the NC limit state, associated to a PGA
equal to 0.25 g.
The shear force, checked with reference to the NC limit state, is less sensitive to
the strength variability than the chord rotation. At the increasing of the strength
variability the obtained D/C ranges increase, but the final assessment of the building
safety found from the analysis is similar to one provided by the EC8 approach. In all
the considered limit states, the most conservative results provided by EC8,
156 S. Viti et al.
Fig. 13.6 Seismic performance of the first storey columns
i.e. obtained by assuming KL ¼ 3 (CF ¼ 1.35) and the 5 % eccentricity, are conser-
vative, compared to the ones found by analysis, only for CoV below 30 %.
13.5 Conclusions
This work deals with the effects of the concrete strength variability on the seismic
response of existing RC buildings. The concrete strength has been described
through a 7-sample domain, having three different amounts of variability
(CoV ¼ 15 %, 30 %, 45 %), consistent with the experimental results found for
existing building in Tuscan (Italy). 180 in-plan layouts, comprehending all the most
significant strength combinations, have been considered to represent the strength
distribution at the columns of the first storey. The effects of such variability have
been evaluated in terms of induced eccentricity, seismic response and capacity and,
therefore, in terms of seismic performance.
Due to the introduced strength variability, in fact, the building experiences both
strength and stiffness eccentricity, and, consequently, a torsional response, with an
increase in the demand. The increase in chord rotation is maximum at the building
flexible side, with an increase of 50 % when a high PGA (PGA ¼ 0.25 g) and CoV
(CoV ¼ 45 %), are considered. The shear force, instead, lightly increases at the
stiffer side of the structure; such increase is scarcely sensitive to the considered
PGA, while it is largely affected by the amount of the considered CoV; in fact, for
CoV ¼ 45 %, it achieves a maximum of 40 % for all considered PGAs.
The seismic demand found by considering the strength variability has been
compared to the one provided by the standard EC8 approach, which has resulted
to be conservative as regards the chord rotation, while it does not cover the increase
in shear force. The seismic capacity of the first storey columns has been found, both
considering the assumed strength variability and the conventional EC8 approach,
for three considered limit states in terms of chord rotation and for the only CP limit
state in terms of shear force. Finally, the ratio between demand (D) and capacity
(C) of the case-study has been investigated. As regards the chord rotation perfor-
mance, the comparison to the EC8 previsions shows different results depending on
the considered limit states. For the serviceability limit state (DL), in fact, the EC8
evaluation provides similar results to those from the analysis. For a strength
variability below 30 %, the maximum D/C values are close with the two
approaches; when the highest value of CoV (CoV ¼ 45 %) is considered, the
analysis provides D/C values higher than the EC8 approach, but the final evaluation
about the building safety does not change. When the ultimate limit states (SD, NC)
are considered, instead, the analysis provides D/C values even three times larger
than the EC8 previsions, with a consequent different evaluation of the performance.
As regards EC8 previsions, the 5 % role introduced by EC8 to take into account of
accidental eccentricity proved to be essential for response and performance evalu-
ation, while the adoption of different CF values does not significantly affect the
response and the performance of the case-study. When the seismic performance is
evaluated in terms of shear force, the strength variability reduces the seismic
performance of the structure, even if it does not affect the final evaluation of the
performance acceptance of the case-study.
As a conclusion, the performed analysis proved the strength variability to be a
significant source of in-plan irregularity for the case-study, affecting its seismic
performance, and leading to an evaluation of its seismic assessment different from
the one provided by the standard EC8 approach, even when all the most conserva-
tive assumptions have been considered (i.e. 5 % eccentricity and CF equal to 1.35).
Acknowledgements The financial support provided by ReLUIS within the project “ReLUIS-
DPC 2014” (Progettazione e valutazione della sicurezza e della vulnerabilita di edifici ed opere) is
gratefully acknowledged.
158 S. Viti et al.
References
Anagnastopoulos SA, Alexopoulou C, Kyrkos MT (2009) An answer to an important controversy

and the need for caution when using simple models to predict inelastic earthquake response of
buildings with torsion. Earthq Eng Struct Dyn 39:521–540
Anagnastopoulos SA, Kyrkos MT, Stathopoulos KG (2013) Earthquake induced torsion in build-
ings: critical review and state of art. Proc. ASEM13, Jeju
Bosco M, Ghersi A, Marino EM (2012) Corrective eccentricities for assessment by the nonlinear
static method of 3D structures subjected to bidirectional ground motions. Earthq Eng Struct
Dyn 41:1751–1773
Bosco M, Marino EM, Rossi PP (2013) An analytical method for the evaluation of the in-plan
irregularity of non-regularly asymmetric buildings. Bull Earthq Eng 11:1423–1445
strutture in c.a. esistenti. PhD Dissertation, Universita di Firenze
dispersione dei valori di resistenza a compressione del calcestruzzo di edifici esistenti. Il
giornale delle prove non distruttive, monitoraggio, diagnostica., vol. 2/2012, pp. 32–39,
ISSN: 1721-7075
De Stefano M, Tanganelli M, Viti S (2013a) Effect of the variability in plan of concrete
Eng. doi:10.1007/s10518-012-9412-5
De Stefano M, Tanganelli M, Viti S (2013b) On the variability of concrete strength as a source of
irregularity in elevation for existing RC buildings: a case study. Bull Earthq Eng. doi:10.1007/
s10518-013-9463-2
De Stefano M, Tanganelli M, Viti S (2014) Variability in concrete mechanical properties as a
source of in-plan irregularity for existing RC framed structures. Eng Struct. doi:10.1016/j.
engstruct.2013.10.027
De Stefano M, Tanganelli M, Viti S (2015a) Torsional effects due to concrete strength variability
in existing buildings. Earthq Struct 8(2):379–399
De Stefano M, Tanganelli M, Viti S (2015b) Seismic performance sensitivity to concrete strength
variability: a case-study. Earthq Struct 9(2):321–337
EC 2 (2002) Eurocode 2: design of concrete structures. European Committee for Standardization
(CEN), Brussels
EC 8–3 (2005) Design of structures for earthquake resistance, part 3: strengthening and repair of
buildings, European standard EN 1998-3. European Committee for Standardization (CEN),
Brussels
Franchin P, Pinto PE, Rajeev P (2007) Confidence factor? J Earthq Eng 14(7):989–1007
Jalayer F, Iervolino I, Manfredi G (2008) Structural modeling: uncertainties and their influence on
seismic assessment of existing RC structures. Struct Saf 32:220–228
Monti G, Alessandri S, Goretti A (2007) Livelli di conoscenza e fattori di confidenza. XII
Convegno ANIDIS, L’ingegneria sismica in Italia. Pisa 10–14 Giugno 2007
Rajeev P, Franchin P, Pinto PE (2010) Review of confidence factor in EC8-Part 3: a European
Code for Seismic Assessment of Existing Buildings, International Conference on Sustainable
Built Environment ICSBE 2010, Kandy
Seismosoft (2006) Seismostruct version 5.2.2 – a computer program for static and dynamic
nonlinear analysis of framed structures. Available online from URL: www.seismosoft.com
Chapter 14
The Influence of Axial Load Variation
on the Seismic Performance of RC Buildings
Valentine Mariani, Marco Tanganelli, Stefania Viti, and Mario De Stefano
Abstract It is well known that the axial load can largely vary during a seismic
event, playing an important role in the seismic performance of RC columns. In
existing buildings this problem can be even more significant than in new ones, since
the material can easily present poor mechanical properties. The paper is aimed at
investigating the role of the axial load variation on the seismic capacity of RC
columns, evaluated in terms of limit chord rotation and shear force, according to
Eurocode 8. The research is performed with reference to a case-study, which is a
doubly symmetric 4-storey RC framed building. The axial load variation affects
both the seismic response and the capacity of the columns of the case-study, and,
therefore, their seismic performance. Special attention has been paid to the role of
the effective concrete strength of columns on the sensitivity of the seismic perfor-
mance to the axial load variation.
Keywords Axial load variability • RC framed structures • Seismic assessment •

Concrete mechanical properties • Seismic performance of existing RC buildings
14.1 Introduction
It’s common knowledge that the axial load, N, plays an important role in the
evaluation of the structural performance of RC columns (Abbasnia et al. 2011;
Saadeghvaziri 1997). The axial load, indeed, largely affects both the seismic
demand and capacity of RC buildings. Concerning the seismic demand, when the
structure is subjected to a horizontal loading, e.g. a seismic action, it necessary
experiences an axial load variation in its vertical elements. At the occurring of
severe ground motion, in particular, some columns can be subjected to significant
axial load reduction, experiencing in extreme cases traction, or conversely, they can
experience a large increase in compression. Even the capacity is affected by the
V. Mariani (*) • M. Tanganelli • S. Viti • M. De Stefano

[email protected]

DOI 10.1007/978-3-319-14246-3_14
160 V. Mariani et al.
axial load. For axial load higher than the “balanced” value, in fact, the ultimate
bending moment progressively decreases. If the section is subjected to traction, its
flexural and shear capacity is largely reduced, as well as when the section is
subjected to high levels of compression. The seismic performance, being the ratio
between demand (D) and capacity (C), should therefore be carefully checked taking
into account the axial load variation, while the current European Technical Code
(Eurocode 8, EC8), prescribes to define the capacity of the structural sections
referring to the static axial load only. Furthermore, in existing buildings the
sensitivity of the seismic performance to axial load variation can be even more
significant than in new ones, since the material can easily present poor and
uncertain mechanical properties (Cristofaro 2009; Cristofaro et al. 2012). This
paper is aimed at investigating the role of the axial load variation on the seismic
performance of an RC case-study building. Two different limit states, respectively a
serviceability (Damage Limitation, LD) one and an ultimate (Severe Damage, SD)
one, have been considered, and the chord rotation has been assumed, in both cases,
as control parameter. The case-study is a doubly symmetric 4-storey RC building,
designed to vertical load only, representing a typical example of pre-seismic code
structure. The role of the axial load variation has been investigated in terms of
seismic performance with special attention to the role of the effective strength of the
columns concrete.
14.2 The Case-Study
The sample structure (De Stefano et al. 2015) is a 4-storey 3D reinforced concrete
frame, symmetric along both x and y directions, with two 4.5 m long bays in the y-
direction and 5 bays 3.5 m long in the x-direction, as shown in Fig. 14.1. All the
columns have cross section dimensions of 30 30 cm, with 8 ϕ14 rebars as
longitudinal reinforcement and ϕ6 stirrups with a spacing of 20 cm. Longitudinal
beams have constant cross section dimensions of 30 50 cm in both directions.
The concrete has been assumed to have a mean strength equal to 19.36 MPa,
while for the reinforcement the Italian FeB38k steel (yield stress over 375 MPa,
a b c d e f
2
y
1
x
Fig. 14.1 Case-study: 3D view and plan configuration

14 Influence of Axial Load Variation 161
ultimate stress over 430 MPa) has been assumed. The building is designed for
vertical loads only (dead load equal to 5.9 KN/m2, live load equal to 2.0 KN/m2),
ignoring seismic loads.
14.3 The Analysis
A nonlinear dynamic (time-history) analysis has been performed by using the com-
puter code Seismostruct (Seismosoft 2006). A two-dimensional fiber model has been
adopted for the cross sections. Each member (both columns and beams) has been
described through an inelastic frame element, and a displacement-based approach has
been selected for the analytical solution. In order to overcome the nonlinearity related
to high amount of deformation, each member has been subdivided into four segments.
The Mander et al. model (Mander et al. 1988) has been assumed for the core concrete,
a three-linear model has been assumed for the unconfined concrete, and a bilinear
model has been assumed for the reinforcement steel. The stiffness of floor slabs has
been considered by introducing a rigid diaphragm.
The seismic input has been defined by assuming a set of seven ground motions
whose mean spectrum closely fits the elastic one provided by EC8 for a soil-type
B. The records were provided by Working Group Itaca (Itaca 2008), on the basis of
a PGA equal to 0.25 g, a nominal life of the structure of 50 years and a magnitude
between 5.5 and 6.5. Two different limit states have been considered: the Damage
Limitation (DL) and the Severe Damage (SD) limit states. According to EC8
provision, a limit value equal to the yield chord rotation has been assumed for the
DL limit state, while a limit value based on the ultimate rotation has been consid-
ered for the LS limit state. Both (DL and SD) limit values depend on the amount of
axial load in the member. The concrete characterization, exhaustively explained in
De Stefano et al. (2013a, b, 2014a) and applied in De Stefano et al. (2014b, 2015)
on the same case-study, is based on a large database of experimental values of
compressive strength (Cristofaro 2009); a single mean strength, equal to
19.36 MPa, and three different CoV, respectively equal to 15 %, 30 % and 45 %
have been considered in the analysis. For each CoV, the concrete distribution has
been represented by a sample of seven concrete strength values, corresponding to
the percentiles of 5 %, 10 %, 20 %, 50 %, 80 %, 90 % and 95 % respectively.
14.4 Results
14.4.1 Response Sensitivity to the Axial Load Variation
The maximum and minimum axial load experienced by the case-study after the
assumed set of ground motions have been checked for each frame and column line.
Figure 14.2 shows the axial load variation, normalized to the value of the
Axial Load Variation (%)

Axial Load Variation (%) 80 80 80
Axial Load Variation (%)

frame 1 frame 2 frame 3
60 60 60
40 40 40
20 20 20
0 0 0
-20 .00 .05 .10 .15 .20 .25 -20 .00 .05 .10 .15 .20 .25 -20 .00 .05 .10 .15 .20 .25
-40 -40 -40
-60 -60 -60
-80 -80 -80
PGA (g) PGA (g) PGA (g)
internal column lines side column lines
Fig. 14.2 Axial load variation in each frame of the case-study
corresponding static axial load, experienced by each column for the different
considered PGAs. Since in each frame, the internal columns (column lines b, c,
d and e) and the side ones (column lines a and f ) experience almost the same axial
load variation (De Stefano et al. 2014b), in Fig. 14.2 only two different trends have
been shown, referred to the side and internal columns. The axial load variation
affects more the side frames (frame 1 and 3) than the central one, since the frame
2 supports a larger amount of gravity loads. It should be noted that, despite the
building is symmetric along both main axes, the axial load variation is not the same
in the two side frames (frames 1 and 3); besides, in each frame, the increase and the
reduction in axial load are not exactly the same.
14.4.2 Capacity Sensitivity to the Axial Load Variation
The structural capacity of existing structure in terms of ductile mechanisms is

measured with reference to chord rotation. Figure 14.3 shows the capacity, in
terms of chord rotation, for DL and SD limit states as a function of N. The limit
chord rotations been found according to EC8, assumming the yield chord rotation
for the DL limit state an ¾ of the ultimate chord rotation for the SD limit state. In
each graph, the gray area represents the range of capacities provided by EC8, which
defines the design compressive strength as the mean one reduced by the Confidence
Factor CF (equal to 1.00, 1.20 and 1.35 with relation to the acheived knowledge
level). As can be noted, the range of capacities provided by EC8 is smaller than the
one obtained when the strength variability is considered, even for the lowest
assumed CoV (CoV ¼ 15 %). For higher values of CoV, more likely to be found
in existing buildings (De Stefano et al. 2013a,b, 2014a), the difference between the
two ranges is even larger.
14.4.3 Seismic Performance Sensitivity to the Axial Load

Variation
Figures 14.4 and 14.5 show, for all the considered PGAs, the ratio between demand
(D) and capacity (C), defined by considering the different strength values provided
DL limit state
1.4E-02 1.4E-02 1.4E-02
CoV=15% CoV=30% CoV=45%
chord rotaon (rad)
chord rotaon (rad)
chord rotaon (rad)

1.2E-02 1.2E-02 1.2E-02
1.0E-02 1.0E-02 1.0E-02
8.0E-03 8.0E-03 8.0E-03
6.0E-03 6.0E-03 6.0E-03
4.0E-03 4.0E-03 4.0E-03
2.0E-03 2.0E-03 2.0E-03
0.0E+00 0.0E+00 0.0E+00
0 200 400 600 800 1000 0 200 400 600 800 1000 0 200 400 600 800 1000
Axial Load (kN) Axial Load (kN) Axial Load (kN)
SD limit state
3.0E-02 3.0E-02 3.0E-02
CoV=15% chord rotaon (rad) CoV=30% CoV=45%
chord rotaon (rad)
chord rotaon (rad)

2.5E-02 2.5E-02 2.5E-02
2.0E-02 2.0E-02 2.0E-02
1.5E-02 1.5E-02 1.5E-02
1.0E-02 1.0E-02 1.0E-02
5.0E-03 5.0E-03 5.0E-03
0.0E+00 0.0E+00 0.0E+00

0 200 400 600 800 1000 0 200 400 600 800 1000 0 200 400 600 800 1000
Axial Load (kN) Axial Load (kN) Axial Load (kN)
K_05 K_10 K_20 K_50 K_80 K_90 K_95
CF=1.00 CF=1.20 CF=1.35
Fig. 14.3 Chord rotation for the DL and SD limit states on varying of axial load
by the domains of Fig. 14.3. For sake of brevity, results are shown for a side column
and a internal one, for each frame. İn the performance evaluation the capacity has
been found with reference to both the limit values of axial load (i.e. minimum and
maximum) experienced by each column due to the seismic excitation; therefore, for
each PGA value, the range in D/C is shown as a function of the axial load. İn
Figs. 14.4 and 14.5 the results obtained for the 50 % percentile can be assumed as
representative of the EC8 approach for the Knowledge Level 3 (CF ¼ 1.00).
As can be noted, the seismic performance of the structure is very much affected
by the assumed strength values, with a large increase in D/C when low percentiles
(K05, K10) have been considered together with high CoVs. As regards the DL limit
state, the case-study complies the EC8 requirements in most of the cases, for PGAs
below 0.15 g, i.e. for all the seismic intensities consistent to the limit state. As
regards the SD limit state, the assumed performance index shows to be very
sensitive to the concrete strength characterization. When a low strength variability
is considered (CoV ¼ 15 %), in fact, the scatter in D/C is almost the same with the
seven considered percentiles. On the contrary, for higher values of CoV, reliable
with concrete strength distribution typical of real existing buildings, the scatter in
D/C due to the axial load variation is very large, suggesting a significant correlation
between the sensitivity to axial load and the effective strength of the column.
Figures 14.6 and 14.7 show, for the DL and SD limit states, the normalized
variation ranges found by comparing each range of seismic performance to the one
found by assuming the corresponding static axial load. The trend of the curves
families confirms the observation made for the previous figures, i.e. the effects of
the axial load variation on the seismic performance are very sensitive to the
3.0 3.0 3.0

CoV = 15% CoV = 30% CoV = 45%
side column
2.0 2.0 2.0
D/C
D/C
D/C
1.0 1.0 1.0
0.0 0.0 0.0

frame 1
0.05 0.10 0.15 0.20 0.25 0.05 0.10 0.15 0.20 0.25 0.05 0.10 0.15 0.20 0.25
3.0 3.0 3.0
CoV = 15% CoV = 30% CoV = 45%
int. column
2.0 2.0 2.0
D/C
D/C
D/C
1.0 1.0 1.0
0.0 0.0 0.0

0.05 0.10 0.15 0.20 0.25 0.05 0.10 0.15 0.20 0.25 0.05 0.10 0.15 0.20 0.25
3.0 3.0 3.0
CoV = 15% CoV = 30% CoV = 45%
side column
2.0 2.0 2.0

D/C
D/C
D/C
1.0 1.0 1.0
0.0 0.0 0.0

frame 2
0.05 0.10 0.15 0.20 0.25 0.05 0.10 0.15 0.20 0.25 0.05 0.10 0.15 0.20 0.25
3.0 3.0 3.0
CoV = 15% CoV = 30% CoV = 45%
int. column
2.0 2.0 2.0
D/C
D/C
D/C
1.0 1.0 1.0
0.0 0.0 0.0

0.05 0.10 0.15 0.20 0.25 0.05 0.10 0.15 0.20 0.25 0.05 0.10 0.15 0.20 0.25
3.0 3.0 3.0
CoV = 15% CoV = 30% CoV = 45%
side column
2.0 2.0 2.0

D/C
D/C
D/C
1.0 1.0 1.0
0.0 0.0 0.0

frame 3
0.05 0.10 0.15 0.20 0.25 0.05 0.10 0.15 0.20 0.25 0.05 0.10 0.15 0.20 0.25
3.0 3.0 3.0
CoV = 15% CoV = 30% CoV = 45%
int. column
2.0 2.0 2.0

D/C
D/C
D/C
1.0 1.0 1.0
0.0 0.0 0.0

0.05 0.10 0.15 0.20 0.25 0.05 0.10 0.15 0.20 0.25 0.05 0.10 0.15 0.20 0.25
K_05 K_10 K_20 K_50 K_80 K_90 K_95
Fig. 14.4 Seismic performance: DL limit state
assumed concrete strength, especially for strong ground motions (high values of
PGA).
14.5 Conclusions
In this paper the effects of the axial load variation on the seismic performance of RC
structures have been investigated with reference to a case-study, i.e. a 4-storey RC
framed building.
4.0 4.0 4.0

CoV = 15% CoV = 30% CoV = 45%
side column 3.0 3.0 3.0
D/C
D/C
D/C
2.0 2.0 2.0
1.0 1.0 1.0
0.0 0.0 0.0

frame 1
0.05 0.10 0.15 0.20 0.25 0.05 0.10 0.15 0.20 0.25 0.05 0.10 0.15 0.20 0.25
4.0 4.0 4.0
CoV = 15% CoV = 30% CoV = 45%
int. column
3.0 3.0 3.0
D/C
D/C
D/C
2.0 2.0 2.0
1.0 1.0 1.0
0.0 0.0 0.0

0.05 0.10 0.15 0.20 0.25 0.05 0.10 0.15 0.20 0.25 0.05 0.10 0.15 0.20 0.25
4.0 4.0 4.0
CoV = 15% CoV = 30% CoV = 45%
side column
3.0 3.0 3.0

D/C
D/C
D/C
2.0 2.0 2.0
1.0 1.0 1.0
0.0 0.0 0.0

frame 2
0.05 0.10 0.15 0.20 0.25 0.05 0.10 0.15 0.20 0.25 0.05 0.10 0.15 0.20 0.25
4.0 4.0 4.0
CoV = 15% CoV = 30% CoV = 45%
int. column
3.0 3.0 3.0

D/C
D/C
D/C
2.0 2.0 2.0
1.0 1.0 1.0
0.0 0.0 0.0

0.05 0.10 0.15 0.20 0.25 0.05 0.10 0.15 0.20 0.25 0.05 0.10 0.15 0.20 0.25
4.0 4.0 4.0
CoV = 15% CoV = 30% CoV = 45%
side column
3.0 3.0 3.0

D/C
D/C
D/C
2.0 2.0 2.0
1.0 1.0 1.0
0.0 0.0 0.0

frame 3
0.05 0.10 0.15 0.20 0.25 0.05 0.10 0.15 0.20 0.25 0.05 0.10 0.15 0.20 0.25
4.0 4.0 4.0
CoV = 15% CoV = 30% CoV = 45%
int. column
3.0 3.0 3.0

D/C
D/C
D/C
2.0 2.0 2.0
1.0 1.0 1.0
0.0 0.0 0.0

0.05 0.10 0.15 0.20 0.25 0.05 0.10 0.15 0.20 0.25 0.05 0.10 0.15 0.20 0.25
K_05 K_10 K_20 K_50 K_80 K_90 K_95
Fig. 14.5 Seismic performance: SD limit state (chord rotation)
The seismic performance, i.e. the ratio between demand (D) and capacity (C),
has been evaluated for the columns of the first storey by considering (i) the current
axial load in the columns which varies during the analysis and (ii) the static value
associated to the vertical loads, according to the current practice and to the technical
codes. Special attention has been paid to the role of the effective concrete strength
of columns, which has been described through a Gaussian distribution, with a mean
equal to 19.36 MPa and three different values of Coefficient of Variation. Each
strength domain has been represented by seven values, corresponding to the
percentiles of 5, 10, 20, 50, 80, 90 and 95 %. The assumed strength domains have
80% 80% 80%

CoV = 15% CoV = 30% CoV = 45%
variaon range
variaon range
variaon range
60% 60% 60%
side column
40% 40% 40%
20% 20% 20%
0% 0% 0%
K05 K10 K20 K50 K80 K90 K95 K05 K10 K20 K50 K80 K90 K95 K05 K10 K20 K50 K80 K90 K95
frame 1
percenle percenle percenle

80% 80% 80%
CoV = 15% CoV = 30% CoV = 45%
variaon range
variaon range
variaon range
60% 60% 60%
int. column
40% 40% 40%
20% 20% 20%
0% 0% 0%
80% 80% 80%
CoV = 15% CoV = 30% CoV = 45%
variaon range
variaon range
variaon range
60% 60% 60%
side column
40% 40% 40%
20% 20% 20%
0% 0% 0%
frame 2

80% 80% 80%
CoV = 15% CoV = 30% CoV = 45%
variaon range
variaon range
variaon range
60% 60% 60%

int. column
40% 40% 40%
20% 20% 20%
0% 0% 0%
80% 80% 80%
CoV = 15% CoV = 30% CoV = 45%
variaon range
variaon range
variaon range
60% 60% 60%

side column
40% 40% 40%
20% 20% 20%
0% 0% 0%
frame 3

80% 80% 80%
CoV = 15% CoV = 30% CoV = 45%
variaon range
variaon range
variaon range
60% 60% 60%

int. column
40% 40% 40%
20% 20% 20%
0% 0% 0%
PGA = 0.05g PGA = 0.10g PGA = 0.15g PGA = 0.20g PGA = 0.25g
Fig. 14.6 Normalized variation ranges: DL limit state

100% 100% 100%

CoV = 15% CoV = 30% CoV = 45%
variaon range
variaon range
80%
variaon range
80% 80%
side column
60% 60% 60%
40% 40% 40%
20% 20% 20%
0% 0% 0%
frame 1

100% 100% 100%
CoV = 15% CoV = 30% CoV = 45%
variaon range
variaon range
80% 80%
variaon range
80%
int. column
60% 60% 60%
40% 40% 40%
20% 20% 20%
0% 0% 0%
100% 100% 100%
CoV = 15% CoV = 30% CoV = 45%
variaon range
variaon range
80% 80%
variaon range
80%
side column
60% 60% 60%
40% 40% 40%
20% 20% 20%
0% 0% 0%
frame 2

100% 100% 100%
CoV = 15% CoV = 30% CoV = 45%
variaon range
variaon range
80%
variaon range
80% 80%
int. column
60% 60% 60%
40% 40% 40%
20% 20% 20%
0% 0% 0%
100% 100% 100%
CoV = 15% CoV = 30% CoV = 45%
variaon range
variaon range
80% 80%
variaon range
80%
side column
60% 60% 60%
40% 40% 40%
20% 20% 20%
0% 0% 0%
frame 3

100% 100% 100%
CoV = 15% CoV = 30% CoV = 45%
variaon range
variaon range
80% 80%
variaon range
80%
int. column
60% 60% 60%
40% 40% 40%
20% 20% 20%
0% 0% 0%
PGA = 0.05g PGA = 0.10g PGA = 0.15g PGA = 0.20g PGA = 0.25g
Fig. 14.7 Normalized variation ranges: SD limit state

been compared to the conventional (reduced) values provided by EC8 as a function

of the achieved Knowledge Level.
The capacity of each column for different amounts of axial load has been found
in terms of chord rotation with reference to two different limit states, i.e. Damage
Limitation (DL) and Life Safety (SD). Each strength value, i.e. the three values
consistent to EC8 (CF ¼ 1.00, 1.20 and 1.35 respectively) and the seven values of
the assumed distribution have been considered and the scatter due to the axial load
variation has been assessed in all cases.
The comparison has shown that the EC8 evaluation of the capacity of columns is
not conservative when compared to the one found by assuming a concrete strength
variability with a CoV over 15 %.
The effects of axial load variation have been finally studied in terms of seismic
performance, i.e. the D/C ratio. The SD limit state resulted more sensitive to axial
load variation, since the D/C ranges found for each strength value are larger than the
ones found in the DL limit state. The strength variability largely affects the obtained
D/C ratios in both cases.
In the DL limit state the effects of strength variability depend on the associated
axial load. In fact, for low amount of axial load, like in the side frames, an increase
in the D/C ratio can be observed at the increasing of the concrete strength. For
higher level of axial load (internal frame), instead, the relationship between con-
crete strength and performance is difficult to predict, since the chord rotation
domains have not a monotonic trend, and both the reduction in stiffness and in
strength play an important role in the seismic response.
In the SD limit state, instead, low values of strength always induce a perfor-
mance reduction. When the lowest percentiles (5 %, 10 %) are considered, together
to the highest CoV (CoV ¼ 45 %), the D/C values associated to the axial load
variation largely overcome the unity. In some cases, the D/C values coming from
the lowest strength percentile is four times the one found adopting the mean
strength value, i.e. by assuming EC8 procedure.
The obtained results confirmed the importance of axial load variation in the
evaluation of the seismic performance, and underlined the relationship between the
effects of axial load variation and real inhomogeneous in-plan distributions of
concrete strength. Both factors, i.e. the effective amount of axial load during the
seismic response of the structure, and a realistic description of the concrete strength
should be carefully evaluated in order to assess the seismic performance of existing
buildings.
References
Abbasnia R, Mirzadeh N, Kildashti K (2011) Assessment of axial force effect on improved

damage index of confined RC beam-column members. Int J Civ Eng 9(3):237–246
strutture in c.a. esistenti. PhD dissertation, Universita di Firenze

dispersione dei valori di resistenza a compressione del calcestruzzo di edifici esistenti. Il
giornale delle prove non distruttive, monitoraggio, diagnostica., vol. 2/2012, p 32–39, ISSN:
1721–7075
De Stefano M, Tanganelli M, Viti S (2013a) Effect of the variability in plan of concrete
Eng 11:1049–1060. doi:10.1007/s10518-012-9412-5
De Stefano M, Tanganelli M, Viti S (2013b) On the variability of concrete strength as a source of
irregularity in elevation for existing RC buildings: a case study. Bull Earth Eng. doi:10.1007/
s10518-013-9463-2
De Stefano M, Tanganelli M, Viti S (2014a) Variability in concrete mechanical properties as a
source of in-plan irregularity for existing RC framed structures. Eng Struct. doi:10.1016/j.
engstruct.2013.10.027
De Stefano M, Mariani V, Tanganelli M, Viti S (2014b) The influence of the axial load variation
on the seismic capacity of existing RC buildings. In: Proc. 2 European Conference on
Earthquake Engineering, Istanbul 2014
De Stefano M, Tanganelli M, Viti S (2015) Torsional effects due to concrete strength variability in
existing buildings. Earthquake and Structures, 8(2):379–399
Itaca (2008) Database of the Italian strong motions data. http://itaca.mi.ingv.it
Mander JB, Priestley MJN, Park R (1988) Theoretical stress-strain model for confined concrete.
J Struct Div ASCE 114:1804–1826
Saadeghvaziri MA (1997) Nonlinear response and modeling of RC columns subjected to varying
axial load. Eng Struct 19(6):417–424
Seismosoft (2006) Seismostruct version 5.2.2 – a computer program for static and dynamic
nonlinear analysis of framed structures. Available online from URL: www.seismosoft.com
Chapter 15
Parametric Study of Inelastic Seismic
Response of Reinforced Concrete Frame
Buildings
Asimina M. Athanatopoulou, Grigorios E. Manoukas,

and Amfilohios Throumoulopoulos
Abstract The objective of this paper is to investigate the influence of strength and
stiffness distribution in-plan to the nonlinear seismic behaviour of reinforced concrete
frame buildings. For this purpose, two 2-storey buildings, a two-way symmetric and a
one-way symmetric with eccentricity relative to the y axis, are designed applying the
Modal Response Spectrum Method of analysis according to Eurocode 8. Then, the
calculated reinforcement of selected elements of the two structures is modified
properly in order to generate seven building models with different strength distribu-
tions. Each model is subjected to seven strong ground motions acting in two
orthogonal directions. The inelastic dynamic analysis is performed for three levels
of seismic intensity corresponding to minor damages, moderate damages and severe
damages. The structural response is evaluated by means of local Damage Indices for
each element as well as of an Overall Structural Damage Index for the whole
building. Useful conclusions concerning the nonlinear response are derived.
Keywords Nonlinear Dynamic Analysis • Inelastic seismic response of R/C

buildings • Center of resistance • Eccentricity • Damage index
15.1 Introduction
It is well-known that the seismic response of buildings for a given excitation

depends on various factors such as the modal characteristics, the damping ratio,
the eccentricity between the centers of stiffness (CS) and mass (CM), etc.
Concerning the linear systems, the influence of the aforementioned factors is clearly
identified and quantified. On the contrary, concerning the nonlinear range of
behaviour, despite the extensive investigations conducted by several researchers
A.M. Athanatopoulou (*) • G.E. Manoukas • A. Throumoulopoulos

Department of Civil Engineering, Aristotle University, University Campus,
54124 Thessaloniki, Macedonia, Greece

DOI 10.1007/978-3-319-14246-3_15
172 A.M. Athanatopoulou et al.
worldwide, no concrete conclusions are derived so far (Paulay 1996, 1997;

Myslimaj and Tso 2002; Stathopoulos and Anagnostopoulos 2005; Peruš and Fajfar
2005; De Stefano et al. 2006a, b; Anagnostopoulos et al. 2010). The inelastic
deformations of the structural elements during earthquakes cause continuous mod-
ifications of the properties of the structures and as a consequence, parameters such
as the structural eccentricity are no more representative of the seismic response.
Thus, many researchers tried to identify additional factors that could influence the
inelastic behaviour. Paulay (1996, 1997) introduced the concept of the center of
resistance (CR), i.e. the point from which passes the resultant of the strengths of all
vertical structural elements and examined the role of strength eccentricity (distance
between CR and CM). Similar investigations were conducted by other researchers
(Myslimaj and Tso 2002; Anagnostopoulos et al. 2010) too. However, no secure
generalized conclusions were obtained. Furthermore, the results of some of the
studies presented were contradictory.
The objective of this paper is to investigate the influence of strength and stiffness
distribution in-plan to the nonlinear seismic behaviour of reinforced concrete frame
buildings. For this purpose, two 2-storey buildings, a two-way symmetric and a
one-way symmetric with eccentricity relative to the y axis, are designed applying
the Modal Response Spectrum Method of analysis according to Eurocode 8 (CEN
2004). Then, the calculated reinforcement of selected elements of the two structures
is modified properly in order to generate seven building models with different
strength distributions. Each model is subjected to seven strong ground motions
acting in two orthogonal directions. The inelastic dynamic analysis is performed for
three levels of seismic intensity corresponding to minor damages, moderate dam-
ages and severe damages. The structural response is evaluated by means of local
Damage Indices for each element as well as of an Overall Structural Damage Index
for the whole building.
15.2 Elastic Analysis and Design
In the framework of the present study, two 2-storey prototype reinforced concrete
buildings are designed according to Eurocodes. In particular, a two-way symmetric
(SB) and a one-way symmetric building with structural eccentricity es ¼ 2.40 m
relative to the y axis (AsBCS-CR) are examined. Both buildings are frame systems
according to the Eurocode 8 classification (CEN 2004, Section 5.2.2.1). Their floor
plans are shown in Figs. 15.1 and 15.2, along with the cross sections dimensions.
The buildings are regular in plan and in elevation, while all storey heights are 3 m.
The concrete is of class C20/25 (fck ¼ 20 MPa) and the reinforcement steel bars
B500C (fyk ¼ 500 MPa) according to the Greek standards.
In addition to the self weight, distributed dead and live slab loads equal to 1.0
and 2.0 kN/m2 respectively are considered. Besides, distributed dead load equal
1.0 kN/m2 is applied to the slabs of the ground floor, in order to take into account
the weight of masonry infill.
15 Inelastic Response of R/C Frame Buildings 173
B5.325/50 K1 45/45 K2 60/35 K3 45/45 K4 60/35 K5 45/45
B9.325/50
5,00m
K6 45/45 K7 45/45 K8 45/45 K9 45/45 K10 45/45

Y
CR(SBStrEc) CM,CS,CR(SB/SBCoIR/SBbIn)
5,00m
X
1,50m
K11 45/45 K12 45/45 K13 45/45 K14 45/45 K15 45/45
5,00m
K16 45/45 K17 60/35 K18 45/45 K19 60/35 K20 45/45
6,00m 6,00m 6,00m 6,00m
Fig. 15.1 Floor plan of the two-way symmetric building
K1 50/50 K2 80/35 K3 45/45 K4 60/35 K5 45/45

B5.325/50
B9.325/50
5,00m
K6 35/90 K7 35/80 K8 45/45 K9 45/45 K10 45/45

CS,CR(AsBCS-CR) Y
CR(AsBCS#CR)
CM CR(AsB CSMCR)
5,00m
2,40m X
0,50m
K11 35/90 K12 35/80 1,20m K13 45/45 K14 45/45 K15 45/45
5,00m
K16 50/50 K17 80/35 K18 45/45 K19 60/35 K20 45/45
6,00m 6,00m 6,00m 6,00m
Fig. 15.2 Floor plan of the one-way symmetric building
The structural analysis and the detailing of the cross sections are conducted with
the aid of appropriate software widely used by engineering practitioners in Greece.
The buildings are analyzed applying the Modal Response Spectrum Analysis
method (CEN 2004, Section 4.3.3.3). The behaviour factor q is taken equal to
3. Despite this low value of q, the buildings are designed to meet the Ductility Class
High requirements and the capacity design provisions. The seismic hazard level
Table 15.1 Modal periods of the analyzed buildings

Modal periods (s)
Mode number Two-way symmetric building One-way symmetric building
1 0.352 0.324
2 0.335 0.323
3 0.276 0.227
4 0.118 0.104
5 0.108 0.102
6 0.090 0.065
adopted corresponds to seismic hazard level zone II of Greek territory possessing

Peak Ground Acceleration equal to 0.24 g with a probability of exceedance 10 % in
50 years. The ground type is of class B (CEN 2004, Section 3.1.1). Modal analysis
is also conducted using the structural analysis program SAP 2000 v.11.0.4. In
Table 15.1 the modal periods of the analyzed buildings are tabulated.
15.3 Inelastic Analysis
Seven structural models – four variants of the two-way symmetric and three
variants of the one-way symmetric building – are generated. In particular, the
analyzed models are as follows (Figs. 15.1 and 15.2):
1. SB: the two-way symmetric prototype building as designed, without any mod-
ifications. The centers of mass (CM), stiffness (CS) and resistance
(CR) coincide.
2. SBColR: the two-way symmetric prototype building with a 35 % reduction of
the reinforcement of columns K7, K8, K9, K12, K13 and K14. CM, CS and CR
coincide.
3. SBbmIn: the two-way symmetric prototype building with a 75 % increase of the
reinforcement of all beams. CM, CS and CR coincide.
4. SBStrEc: the two-way symmetric prototype building with a 20 % reduction of
the reinforcement of columns K5, K9, K10, K14, K15 and K20, as well as a 40 %
increase of the reinforcement of K1, K6, K7, K11, K12 and K16. CM and CS
coincide. An eccentricity er ¼ 1.50 m between CM and CR results.
5. AsBCS-CR: the one-way symmetric prototype building as designed, without any
modifications. CS and CR coincide. The eccentricity between CM and CR is
equal to the structural eccentricity, i.e. er ¼ es ¼ 2.40 m.
6. AsBCS#CR: the one-way symmetric prototype building with a 25 % reduction
of the reinforcement of columns K1, K2, K6, K7, K11, K12, K16 and K17. The
eccentricity between CM and CR is er ¼ 1.20 m, while es remains constant,
i.e. es ¼ 2.40 m.
7. AsBCSMCR: the one-way symmetric prototype building with a 35 % reduction
of the reinforcement of columns K1, K2, K6, K7, K11, K12, K16 and K17, as
Table 15.2 List of seismic excitations

Station Magnitude Peak ground
No Earthquake name Date name (Ms) acceleration (g)
1 Imperial Valley, 15/10/ Calexico 6.53 0.25
USA 1979 Fire
2 Imperial Valley, 15/10/ El Centro 6.53 0.12
USA 1979
3 Friuli, Italy 15/09/ Buia 5.91 0.12
1976
4 Kocaeli, Turkey 17/08/ Arcelik 7.51 0.21
1999
5 Loma Prieta, USA 18/10/ Bran 6.93 0.65
1989
6 Kobe, Japan 16/01/ Takatori 6.90 0.69
1995
7 Lefkada, Greece 14/08/ LEF1 6.20 0.48
2003 (ITSAK)
well as a 50 % increase of the reinforcement of K4, K5, K9, K10, K14, K15, K19
and K20. er is equal to 0.50 m, while es remains constant, i.e. es ¼ 2.40 m.
The whole investigation comprises a suite of seven pairs of horizontal ground
motion records obtained from the PEER and the European strong motion databases
(Table 15.2). The ground motions are recorded on Soil Type B according to
Eurocode 8 (CEN 2004) and their magnitudes (Ms) range between 5.91 and 7.51.
The accelerograms are scaled in order to achieve three predefined levels of
structural damage, as expressed through the Overall Structural Damage Index
(OSDI). In particular, three different scaling factors for each ground motion are
determined after successive tests, in order to achieve OSDI values equal to 0.3, 0.6
and 0.9 for the two-way symmetric prototype building (SB). It is considered that
these values correspond to minor, moderate and severe damages of the structural
elements respectively. The applied scaling factors, which are the same for both
components of each ground motion, are tabulated in Table 15.3.
The seven generated building models are analyzed for the three intensity levels
of the selected ground motions by means of Nonlinear Dynamic Analysis (NDA).
All analyses are performed using the program Ruaumoko (Carr 2005).
For each seismic excitation a representative Damage Index for each critical
section is calculated. In particular, the widely used Park and Ang damage index
(Park and Ang 1985) modified by Kunnath et al. (1992) is used. At a given cross
section the local Damage Index (DI) is given by the following equation:
ϕm ϕ y β
DI ¼ þ ET ð15:1Þ
ϕu ϕ y M y ϕ u
where φm is the maximum curvature observed during the load history, φu is the
ultimate curvature capacity, φy is the yield curvature, ET is the dissipated hysteretic
Table 15.3 Scaling factors OSDI value for model 1

applied to the accelerograms
No Earthquake name 0.3 0.6 0.9
1 Imperial Valley, USA 1.35 2.30 3.35
2 Imperial Valley, USA 2.90 5.25 7.50
3 Friuli, Italy 3.50 5.50 7.10
4 Kocaeli, Turkey 2.80 4.60 7.15
5 Loma Prieta, USA 0.53 0.88 1.18
6 Kobe, Japan 0.53 0.90 1.03
7 Lefkada, Greece 0.70 1.05 1.69
energy, My is the yield moment of the cross section and β is a dimensionless

constant determining the contribution of cyclic loading to damage, which is taken
equal to 0.05 for the analyses conducted in the present study.
Moreover, the Overall Structural Damage Index (OSDI) of the building is
computed as a weighted average of the local Damage Indices at the ends of each
element. The dissipated energy is used as a weight factor:
0 1
X
n B Ε C
B Τi C
OSDI ¼ DIi B n C ð15:2Þ
@X A
i¼1
ΕΤi
i¼1
where DIi is the local Damage Index at cross section i determined by Eq. 15.1, ETi is
the energy dissipated at cross section i and n is the number of cross sections.
In Fig. 15.3 the local Damage Indices (DI) of representative structural elements
of the analyzed models are shown. In particular, the mean values for the seven
ground motions of DIs of four columns (K1, K5, K7, K9) and two beams (B5.3,
B9.3) of the ground floor are presented. The DIs are calculated for both ends of each
structural element: bottom (bot) – top (top) for columns and left (l) – right (r) for
beams. From Fig. 15.3 the following observations can be made:
1. SB: DI values, which translate to ductility demands, are quite uniformly distrib-
uted along the plan. This becomes clear comparing DI values for structural
elements lying at symmetrical positions (K1-K5, K7-K9, B5.3-B9.3). DI values
of beams are significantly higher than those of columns, especially those of the
top. Obviously, this is due to the application of the capacity design provisions. In
general, the seismic response of the structure meets the codes objectives.
2. SBColR: in comparison with SB, a significant increase of the DIs of columns K7
and K9 due to the reduction of their reinforcement is observed. This is reflected
to the OSDI too (see also Fig. 15.4). The uniformity of ductility demands
distribution along the plan remains.
3. SBbmIn: in comparison with SB, a significant reduction of the DIs of beams due
to the increase of their reinforcement is observed. This is reflected to the OSDI
1.4 1.4
1.2 1.2
1 1
0.8 0.8
DI
DI
0.6 0.6
0.4 0.4
0.2 0.2
0 0
Seismic Intensity Seismic Intensity Seismic Intensity Seismic Intensity Seismic Intensity Seismic Intensity
Level I Level II Level III Level I Level II Level III
a) SB b) SBColR
1 1.4
0.9 1.2
0.8
0.7 1
0.6 0.8
DI
DI
0.5
0.4 0.6
0.3 0.4
0.2
0.1 0.2
0 0
c) SBbmIn d) SBStrEc
1.6 1.4
1.4 1.2
1.2 1
1
0.8
DI
DI
0.8
0.6
0.6
0.4 0.4
0.2 0.2
0 0
e) AsBCS-CR f) AsBCS#CR
1.6
1.4
1.2 K1bot K1top K5bot K5top
1
DI
0.8
0.6
K7bot K7top K9bot K9top
0.4
0.2
0
Seismic Intensity Seismic Intensity Seismic Intensity B5.31 B5.3r B9.31 B9.3r
Level I Level II Level III
g) AsBCSMCR
Fig. 15.3 Damage indices of structural elements

1
0.9 SB
0.8 SBColR
0.7
0.6 SBbmIn
OSDI
0.5 SBStrEc
0.4 AsBCS-CR
0.3
0.2 AsBCS#CR
0.1 AsBCSMCR
0
Seismic Intensity Seismic Intensity Seismic Intensity
Level I Level II Level III
Fig. 15.4 Overall structural damage indices
too (see also Fig. 15.4). On the contrary, the DIs of columns are considerably
increased, especially at the upper ends.
4. SBStrEc: in comparison with SB, a significant increase of the DIs of the
weakened columns (K5, K9) and an analogous reduction of the DIs of the
strengthened columns (K1, K7) occurred. With increasing intensity of earth-
quake excitation, the damages of the columns at the strong side increase in a
larger extent with regard to those lying at the opposite side. Concerning the
beams, the ductility demands of those lying at the strong side (B5.3) are
increased, while of those lying at the weak side (B9.3) are reduced, especially
for higher seismic intensity level.
5. AsBCS-CR: due to the capacity design, DI values of beams are significantly
higher than those of columns, especially those of the top. For low seismic
intensity, DI values of columns of the flexible-weak side are higher than those
of the stiff-strong one. However, with increasing intensity an inversion of this
trend is observed. This finding is consistent to previous studies which demon-
strated that in structures experiencing extensive inelastic deformations the
ductility demands are higher at the columns of the stiff side. Concerning the
beams, regardless the seismic intensity, the ductility demands are higher at the
flexible-weak side.
6. AsBCS#CR: in comparison with AsBCS-CR, a significant increase of the DIs of
the weakened columns (K1, K7) due to the reduction of their reinforcement is
observed. The transposition of the center of resistance (CR) towards the flexible-
weak side implies an increase of the ductility demands of columns lying at this
side too. Just like AsBCS-CR, the ductility demands of beams are significantly
higher at the flexible-weak side. However, a uniform reduction of DI values is
observed.
7. AsBCSMCR: in comparison with AsBCS-CR, a significant increase of the DIs
of the weakened columns (K1, K7) and an analogous reduction of the DIs of the
strengthened columns (K5, K9) occurred. With increasing intensity of earth-
quake excitation, the damages of the columns at the stiff (but no more strong)
side increase in a larger extent with regard to those lying at the opposite side.
Concerning the beams, just like AsBCS-CR and AsBCS#CR, the ductility
demands are significantly higher at the flexible side.
In Fig. 15.4 the mean OSDIs for the seven ground motions are shown.
Concerning the models based on the two-way symmetric prototype building (SB,
SBColR, SBbmIn, SBStrEc), it is apparent that the increase of the beams reinforce-
ment (SBbmIn) leads to considerable modification of the seismic behaviour,
although CR remains constant. On the contrary, the transposition of CR towards
the stiff side (SBStrEc) does not affect significantly the overall structural response.
This means that CR, as determined in relevant studies, i.e. without taking into
account the strength of beams, is not representative of the inelastic seismic behav-
iour. Concerning the models AsBCS-CR, AsBCS#CR and AsBCSMCR, the trans-
position of CR towards the flexible side leads to an increase of the overall damage
index. This finding is not consistent to previous studies (Myslimaj and Tso 2002).
15.4 Conclusions
The main conclusions derived from this study are as follows:

1. Concerning the two-way symmetric building, the ductility demands are – as
expected – quite uniformly distributed in plan.
2. Concerning the one-way symmetric building for high level of seismic intensity
the ductility demands are higher at the columns of the stiff side, in relevance
with those lying at the flexible side. Concerning the beams, regardless the
seismic intensity the ductility demands are higher at the flexible side.
3. The strength of the beams can affect significantly the inelastic seismic response
of buildings. As a consequence, CR as determined in relevant studies,
i.e. without taking into account the strength of the beams, is not representative
of the inelastic seismic behaviour of structures.
4. In general, the ductility demands of the columns lying at the stiff side are higher
than of those lying at the flexible side regardless the position of CR. When CM
and CS coincide, the higher demands occur at the weak side.
5. When CR lies at the stiff side, with reducing eccentricity between CR and CM,
the ductility demands of beams are uniformly reduced. When CR is transposed at
the flexible side, an increase of ductility demands is observed. Anyway, the
ductility demands of beams are higher at the flexible side.
References
Anagnostopoulos SA, Alexopoulou C, Stathopoulos KG (2010) An answer to an important

controversy and the need for caution when using simple models to predict inelastic earthquake
response of buildings with torsion. Earthq Eng Struct Dyn 39(5):521–540
Carr AJ (2005) Theory and user guide to associated programs. Ruaumoko manual, vol 1, 3. Uni-
versity of Canterbury, Canterbury
resistance. Part 1: general rules, seismic actions and rules for buildings. EN 1998-1:2004.
Brussels
De Stefano M, Faella G, Ramasco R (2006a) Inelastic seismic response of one-way plan-asym-
metric systems under bi-directional ground motions. Earthq Eng Struct Dyn 27(4):363–376
De Stefano M, Marino EM, Rossi PP (2006b) Effect of overstrength on the seismic behaviour of
multi-storey regularly asymmetric buildings. B Earthq Eng 4(1):23–42
Kunnath SK, Reinhorn AM, Lobo RF (1992) DARC version 3.0: a program for inelastic damage
analysis of reinforced concrete structures. Technical Representative NCEER-92-0022. State
University of New York at Buffalo, New York
Myslimaj B, Tso WK (2002) A strength distribution criterion for minimizing torsional response of
asymmetric wall-type systems. Earthq Eng Struct Dyn 31(1):99–120
Park YJ, Ang AH-S (1985) Mechanistic seismic damage model for reinforced concrete. J Struct
Eng 111(4):722–739
Paulay T (1996) Seismic design for torsional response of ductile buildings. B N Z Natl Soc Earthq
Eng 29(3):178–198
Paulay T (1997) Seismic torsional effects on ductile structural wall systems. J Earthq Eng
1(4):721–745
Peruš I, Fajfar P (2005) On the inelastic torsional response of single-storey structures under
bi-axial excitation. Earthq Eng Struct Dyn 34(8):931–941
Stathopoulos KG, Anagnostopoulos SA (2005) Inelastic torsion of multistorey buildings under
earthquake excitations. Earthq Eng Struct Dyn 34(12):1449–1465
Chapter 16
Seismic Upgrading of Vertically Irregular
Existing r.c. Frames by BRBs
Francesca Barbagallo, Melina Bosco, Edoardo M. Marino, Pier Paolo Rossi,

and Paola R. Stramondo
Abstract In several earthquake-prone countries a large part of existing

r.c. buildings had been designed before seismic codes entered into force and,
therefore, considering gravity loads only. The lateral strength and stiffness of
these structures are inadequate to sustain seismic action and their seismic response
is aggravated by deficiencies due to vertical irregularity. In particular, the ratio of
the demanded storey shear force to the lateral strength presents an irregular
distribution along the height of the building. This promotes the development of a
storey collapse mechanism. In this chapter, the insertion of Buckling Restrained
Braces (BRBs) is proposed both for the seismic upgrading of r.c. frames designed
for gravity loads and for the reduction of their vertical irregularity. The effective-
ness of a design method previously proposed by the authors is investigated. The
method is ruled by two parameters: the behaviour factor and the design storey drift.
The design method is applied for the seismic retrofitting of a 6-storey r.c. frame
considering the values of the ruling parameters in several combinations. Then,
nonlinear dynamic analyses are performed to assess their influence on the seismic
response of the upgraded frames.
Keywords Seismic upgrading • Design method • Buckling restrained braces •

Vertical irregularity • r.c structures
16.1 Introduction
The existing r.c. buildings designed to sustain gravity loads usually have main
structural elements disposed along a single direction and this makes them very
flexible and weak in the orthogonal direction. Furthermore, the heightwise distri-
butions of the lateral stiffness and shear strength of these frames are not suitable to
F. Barbagallo • M. Bosco • E.M. Marino (*) • P.P. Rossi • P.R. Stramondo

Department of Civil Engineering and Architecture, University of Catania, V.le A. Doria, 6,
95125 Catania, Italy
e-mail: [email protected]; [email protected]; [email protected]; prossi@dica.
unict.it; [email protected]

DOI 10.1007/978-3-319-14246-3_16
182 F. Barbagallo et al.
make the drift demand widespread along the height and consistent with the capacity
of the frame. In this chapter, the introduction of Buckling Restrained Braces
(BRBs) is proposed for the seismic rehabilitation of these framed structures in
order to increase to proper values both the lateral stiffness and the shear strength.
BRBs typically consist of a ductile steel core confined by a steel tube (Uang and
Nakashima 2004; Xie 2005). The brace is joined to the frame by means of the
connection segments, while the yielding core and the connection segments are
linked by the transition segments. The transition and connection segments have to
remain elastic during cyclic loading. This is obtained by adopting for these seg-
ments cross-sectional areas larger than that of the yielding core. Therefore, the
stiffness and the strength of BRBs can be varied almost independently, by choosing
appropriate cross-sectional areas and lengths of the segments of the BRB, and yield
stress of the steel.
BRBs, as well as other types of dampers (Daniel et al. 2013), are very promising
devices for seismic protection of buildings. Indeed, BRBs can modify the distribu-
tion of the shear strength along the height so as to promote a widespread yielding of
the structure and, therefore, a more favourable collapse mechanism during strong
ground motions. Moreover, they can modify the distribution of the lateral stiffness
along the height so that the displacements demanded by the ground motion can
better fit the displacement capacity of the structure. This research develops and
validates a design method for BRBs, which is applied for the seismic upgrading of
an existing r.c. frame. The proposed design method aims at obtaining a more regular
structural behaviour along the height and avoiding the soft storey mechanism.
Afterwards the seismic response of the upgraded frames is determined by nonlinear
dynamic analysis, to evaluate the increase of vertical regularity of the structure.
16.2 Proposed Design Method
The method is ruled by two parameters. The first one is the behaviour factor q,
which determines the lateral strength to be provided by BRBs. The second param-
eter is the design storey drift Δd,i, which is given as a fraction of the nominal storey
drift capacity Δl,i. This assumption is to take into account that, in a few storeys,
drifts may be larger than those obtained by the design elastic analysis, due to some
damage concentrations. The method deals with two requirements. The first one is a
strength requirement aiming at providing the r.c. frame upgraded by BRBs with
sufficient lateral strength, which is distributed along the height of the frame
proportionally to the required storey shear. The second one is a displacement
requirement aiming at reducing the displacement demand below the design value.
A flowchart of the design procedure is shown in Fig. 16.1.
16 Seismic Upgrading of r.c. Frames by BRBs 183
Fig. 16.1 Flowchart of the design procedure
16.2.1 Determination of the Displacement and Strength

Demands
Because of the insertion of BRBs, which promotes the simultaneous yielding of all
the storeys and a widespread demand along the height, the demanded drifts Δi can
be determined for the i-th storey by the elastic analysis of the structure. This
analysis is based on the elastic (unreduced) spectrum of the reference ground
motion, here assumed as that having a 10 % probability of exceedance in
50 years. Afterwards, Δi is modified according to European seismic code EC8
(CEN 2004), to take into account that the equal displacement rule does not apply for
structures whose fundamental period T1 is smaller than TC. After BRBs are inserted,
and their stiffness is determined in compliance with the displacement requirement,
the required lateral strength of the frame VEd,i is calculated at each storey through
the elastic analysis of the structure based on the elastic spectrum reduced by the
behaviour factor q.
16.2.2 Determination of the Displacement and Strength

Capacity
In this chapter, the displacement capacity is defined in terms of storey drifts

corresponding to the achievement of the Significant Damage limit state in columns.
The provisions of EC8 quantify the seismic performance in terms of chord rotation.
EC8 defines the chord rotation capacity corresponding to Near Collapse limit state
θum as the chord rotation at yielding plus plastic rotation at column failure θumpl
. In
this chapter θum and θum are evaluated according to EC8-Part 3 (CEN 2010), while
pl
the chord rotation at yielding is obtained as the difference between these two terms.
Moreover, EC8 stipulates that the Significant Damage limit state is achieved when,
somewhere in the structure, the plastic part of the chord rotation is equal to 75 % of
the θumpl
. Thus, the displacement capacity Δl corresponding to the Significant
Damage limit state is evaluated as:

Δl ¼ 0:75 θum
pl
; þ θum θum
pl
H ð16:1Þ
where H is the length of the column equal to the inter-storey height of the frame.
The drift Δl is evaluated for the ends of all the columns of the storey and the
minimum value obtained is assumed as displacement capacity of the storey Δl,i.
The available strength VRd,i at each storey of the frame upgraded by BRBs is
evaluated as the summation of two contributions. The first one is the storey shear
strength provided by BRBs VRd,b,i, calculated as the sum of the two horizontal
components of the axial forces of the braces inserted in the frame at the considered
storey. The second one is the shear strength of the bare r.c. frame VRd,f,i given by the
sum of the shear forces transmitted by the columns of that storey when they are
yielded in flexure at the two end cross-sections.
16.2.3 Design of BRBs
BRBs are designed firstly to fulfil the displacement requirement. To this end, the
drift demand and the drift capacity are compared each other at every storey. Where
the demand exceeds the capacity, the introduction of BRBs provides the structure
with the lacking stiffness.
With regards to the total stiffness of the structure Kreq, it is given by the
summation of the stiffness of the bare r.c. frame model KFR,b and the stiffness of
the truss model KTruss. The stiffness Kreq is evaluated as the ratio of the total shear
force to the design storey drift Δd, while KFR,b is calculated as the ratio of the sum of
the shear forces carried by the columns to the storey drift. Both the storey shear
force and drift are determined by the elastic analysis. Given the values of these
terms the stiffness of the truss model KTruss can be determined as the difference
between Kreq and KFR,b.
Fig. 16.2 Drifts of the truss model caused by (a) BRB axial deformation and (b) column axial
deformation
Furthermore, it is considered that the total storey drift of the truss model Δ is
equal to the sum of the drift caused by the axial deformations of BRBs ΔBRBs
(Fig. 16.2a) and that caused by the axial deformation of columns ΔCOLax
(Fig. 16.2b). The total storey drift of the truss model Δ is calculated as the ratio
between the shear force sustained by BRBs VBRBs, determined by the elastic
analysis, and the stiffness KTruss. Focusing on the drift caused by the axial defor-
mation of columns (Fig. 16.2b), it is evident that shortening/elongation of columns
would cause also a variation in length of BRBs. But, to avoid this, columns rotate
rigidly and provide the drift ΔCOLax that is determined by simple geometrical
considerations. Finally, the drift ΔBRBs is calculated as Δ minus ΔCOLax and the
additional stiffness to be provided by BRBs to fulfil the displacement requirement is
determined as:
V BRBs
K BRBs ¼ ð16:2Þ
ΔBRBs
Given the value of KBRBs, the procedure determines the value of the equivalent
cross-section area of BRBs Aeq by the following relation:
1 K BRBs LBRBs
Aeq ¼ ð16:3Þ
2 E cos 2 α
where LBRBs is the length of BRBs, E is Young’s modulus of steel and α is the angle
of inclination of the BRBs with respect to the beam longitudinal axis.
The design procedure has to be ran iteratively until convergence, because the
insertion of the BRBs increases the frame stiffness and modifies the vibration
periods and the seismic response of the frame.
After the stiffness of BRBs is calculated, their yield strength Ny is determined to
fulfil the strength requirement. To this end, the required strength of the BRBs at the
i-th storey, VreqBRBs,i, is obtained as the difference between the total required
strength VEd,i (obtained by the elastic analysis of the frame with the spectrum
reduced by q) and the shear strength of the bare r.c. frame VRd,f,i. The yield strength
of the BRBs is calculated by the following relation:
V req, BRBs, i
N y, i ¼ ð16:4Þ
2 cos α
For each BRB configuration satisfying the displacement requirement, the yield
strength Ny can be obtained adopting a proper yield stress fy, which can be obtained
as the ratio between Ny and the cross-sectional area Ac of the yielding core of the
BRB. However, fy is not lower than a minimum value that is determined at each
storey in order to have the BRB ductility demand not larger than 19 for the design
seismic action (Merritt et al. 2003; Bosco and Marino 2013).
16.3 Case Study Structure
The described design method has been applied for the seismic upgrading of a
r.c. frame extracted from a six-storey building designed to resist gravity loads
only and described in Barbagallo et al. (2014). The frame has three spans 4.0 m
long and the interstory height is equal to 3.2 m. The floor mass is equal to 102.37 t
and the gravity loads are considered in the seismic design situation. Table 16.1
shows the dimensions of cross-sections of beams and columns. Diagonal BRBs are
inserted in the lateral spans.
Beams and columns are simulated by De Saint Venant members and BRBs are
modelled as trusses. All the nodes belonging to the same floor are constrained to
have the same horizontal displacement. For modal response spectrum analysis, the
seismic input is given by the elastic spectrum proposed by EC8 for soil type C,
characterised by a peak ground acceleration ag equal to 0.35 g. The confidence
factor FC is assumed equal to one for both concrete and steel. The yield stress of
BRBs fy is determined by the proposed design procedure within a range of values
deemed acceptable. Mechanical parameters describing the materials are
summarised in Table 16.2. The design of the analysed frames is carried out
considering three values of Δd,i, equal to 0.6 Δl,i, 0.7 Δl,i, 0.8 Δl,i, and three values
of q, equal to 5, 6 and 7.
Table 16.1 Cross-sections dimensions (b h) of the r.c. frame members

Storey 1st 2nd 3rd 4th 5th 6th
Beams 30 60 30 60 30 60 30 60 30 60 30 60
Perimetral columns 50 30 40 30 35 30 30 30 30 30 30 30
Central columns 30 60 30 50 30 40 30 30 30 30 30 30
Table 16.2 Mechanical parameters adopted for the design of the analysed frame
Concrete Reinforcing steel bars BRBs
Compressive strength 29 MPa
Yielding strength 400 MPa 100–275 MPa
Young’s modulus 30,280 MPa 210,000 MPa 210,000 MPa
Poisson’s ratio 0.5 0.3 0.3
16.4 Numerical Analyses
16.4.1 Numerical Model
Nonlinear dynamic analysis of the bare and the upgraded r.c. frames have been
carried out by means of the OpenSees program (Mazzoni et al. 2003), in order to
evaluate the influence of the parameters q and Δd on the seismic performance of the
frames, with particular regard to their vertical regularity.
A 2-D frame model with masses concentrated at the floor levels is set, with rigid
diaphragms at each level. Because the strength contribution given by the deck to
beams cannot be exactly quantified, an Elastic Beam model (EB model) and a
Plastic Beam model (PB model) have been used to define a bounded behaviour. In
PB model beams and columns are modelled as members constituted by elastic
elements with plastic hinges at their ends. A fibre cross-section, with both concrete
and steel components, is assigned to each plastic hinge whose length is equal to the
depth of the cross-section. The Mander constitutive law is assigned to the concrete
fibres, while an elasto-plastic with strain kinematic hardening constitutive law is
assigned to the steel fibres. The parameters used for materials are summarised in
Table 16.3. Instead, in EB model beams are modelled as infinitely resistant.
BRBs are modelled as truss elements with the cross-sectional area equal to the
equivalent area Aeq obtained in design. The material model proposed by Zona and
Dall’Asta (2012), which takes into account both kinematic and isotropic hardening
(Rossi 2014), is used to simulate the cyclic behaviour of BRBs. Further details
about the model are available in Barbagallo et al. (2014). In compliance with EC8,
the nominal dead loads plus quasi-permanent live loads are assigned as initial
gravity loads in the analysis. A Rayleigh viscous damping is used and set at 5 %
for the first and the third mode of vibration. The P-Δ effect is considered in the
analysis.
For nonlinear time-history analysis, a set of ten artificial accelerograms com-
patible with the EC8 elastic spectrum for soil type C and characterised by 5 %
Table 16.3 Mechanical parameters adopted for the dynamic analysis of the frames
Concrete Reinforcing steel bars
Compressive strength 29 MPa Yielding strength 400 MPa
Young’s modulus 30,280 MPa Young’s modulus 210,000 MPa
Strain at maximum strength 2 103 Strain-hardening ratio 0.0066
Tensile strength ftm ¼ 2.28 MPa
Ultimate strain in tension 7.5 105
damping ratio and peak ground acceleration ag equal to 0.35 g is defined. Details
about the set of accelerograms may be found in Amara et al. (2014).
16.4.2 Seismic Response of the Designed Frames
The seismic performance of the analysed frames is evaluated in terms of maximum

drift demand Δi and ratio of storey drift demand Δi to capacity Δl,i. The values of Δi
and Δi/Δl,i are determined for each of the N storeys of the frame and for each
accelerogram. Their median values over the ten accelerograms are determined for
each storey and their heightwise distributions are shown in Figs. 16.3 and 16.4,
respectively. The results are presented for the design storey drifts Δd,i equal to 0.6
Δl,i and 0.8 Δl,i considering both the Plastic Beam model (PB model) and the Elastic
Beam model (EB model). The four curves refer to the performance of the bare
frame and to that of the upgraded frames designed by q equal to 5, 6 and 7.
Figure 16.3 shows that the insertion of BRBs leads to a reduction of the
maximum drift demand and, therefore, to the improvement of the seismic perfor-
mance. If Δd,i ¼ 0.6 Δl,i, the design of BRBs is basically ruled by the stiffness
requirement. Therefore, the three considered values of q lead to similar BRBs and
the upgraded frames exhibit similar seismic performances. Instead, if Δd,i ¼ 0.8 Δl,i,
the strength requirement controls the design of BRBs and the reduction of the storey
drifts is higher for lower values of q. Furthermore, if the PB model is adopted,
regardless of the adopted Δd,i, some damage concentration is obtained at the lowest
storeys. Instead, the EB model leads to a more regular distribution of storey drifts.
Figure 16.4 allows the determination of the values of the behaviour factor q that
can be used to avoid the exceedance of the Significant Damage limit state. These
values are those corresponding to maximum demand to capacity ratios Δi/Δl,i not
larger than 1. When Δd,i ¼ 0.6 Δl,i, both for PB and for EB model, all the considered
values of q lead to an acceptable seismic response. Instead, if Δd,i ¼ 0.8 Δl,i, the
maximum acceptable value of q is 6 for the EB model and 7 for the PB model. The
most restrictive of the two values (q ¼ 6) is suggested for the design of BRBs when
Δd,i ¼ 0.8 Δl,i is adopted, because it is assumed that the results obtained by the PB
model and EB model bound the real behaviour of the frame.
For each value of q and Δd,i, the response parameters Δi and Δi/Δl,i are
normalised with respect to their maximum value along the height and then the
Bare r.c. frame Upgraded frames: q=5 q=6 q=7
Plastic Beam Model Elastic Beam Model

N Δ d,i = 0.70
? 0.60 Δ
? l,i N Δ d,i = 0.70
? Δ l,i
0.60 ?
6 6
4 4
2 2
0 0
0 20 40 60 80 Δ (mm)
100 0 20 40 60 80 Δ
100
(mm)
N Δ 0.70 Δ
? d,i = 0.80 ? l,i N Δ
? d,i = 0.80 Δ l,i
0.70 ?
6 6
4 4
2 2
0 0
0 20 40 60 80 Δ
100
(mm) 0 20 40 60 80 Δ
100
(mm)
Fig. 16.3 Maximum storey drift demand
N Δd,i = 0.60 Δl,i N Δd,i = 0.60 Δl,i

6 6
4 4
2 2
0 0
0.0 1.0 Δ/Δl 0.0 1.0 Δ/Δl
N Δd,i = 0.80 Δl,i N Δd,i = 0.80 Δl,i
6 6
4 4
2 2
0 0
0.0 1.0 Δ/Δl 0.0 1.0 Δ/Δl
Fig. 16.4 Maximum storey drift demand to capacity ratio
means of the normalised values over the storeys Δm and (Δ/Δl)m are determined. If
the considered Δi is the same at each storey, the response in terms of storey drift is
perfectly regular in elevation and the mean value Δm of the normalised Δi is equal
to 1. Instead, values of Δm smaller than 1 denote a lower level of vertical regularity
of the seismic response. The same considerations apply to (Δ/Δl)m and to regularity
in elevation of the response in terms of storey drift demand to capacity ratio.
Figures 16.5 and 16.6 show the obtained values of Δm and (Δ/Δl)m, which quantify
the regularity of the seismic performance along the height, limited to the values of
q leading to an acceptable seismic response. Each curve refers to frames designed
with a fixed value of Δd,i. The comparison between each curve and the value
Bare r.c. frame Δd,i = 0.6Δl,i Δd,i = 0.8Δl,i

Δm ? d,i =Beam
Plastic 0.70 model
? l,i Δm ? d,i =Beam
Elastic 0.70 model
? l,i
0.9 0.9
0.8 0.8
0.7 0.7
0.6 0.6
0.5 0.5
5 6 7 q8 5 6 7 q8
Fig. 16.5 Mean value of maximum storey drift demand
(Δ/Δ l)m ? d,i =Beam

Plastic 0.70 model
? l,i (Δ/Δl)m ? d,i =
Elastic 0.70 model
Beam ? l,i
0.9 0.9
0.8 0.8
0.7 0.7
0.6 0.6
0.5 0.5
5 6 7 q 5 6 7 q
Fig. 16.6 Mean value of maximum storey drift demand to capacity ratio
obtained for the bare r.c. frame allows the evaluation of the increase of vertical
regularity obtained by the upgrading by BRBs.
In all the considered cases, the values of Δm and (Δ/Δl)m obtained for the
upgraded frames are higher than those of the bare frame. If Δd,i ¼ 0.6 Δl,i, the
higher values of Δm and (Δ/Δl)m are obtained for q ¼ 7. If Δd,i ¼ 0.8 Δl,i, with the
exception of the EB model, which leads to slightly different values of Δm
(Fig. 16.5), the obtained values of Δm and (Δ/Δl)m are the same for the two
considered values of q. In conclusion, the proposed design procedure for the
seismic upgrading of r.c. frames always increases the vertical regularity of the
frames. The improvement of regularity due to the insertion of BRBs is higher when
the seismic response is expressed in terms of storey drifts. Among the considered
pairs of Δd,i and q, which avoid exceeding the Significant Damage limit state, the
pairs Δd,i ¼ 0.6 Δl,i, q ¼ 7 and Δd,i ¼ 0.8 Δl,i, q ¼ 5 are those corresponding to the
most regular behaviour in elevation.
16.5 Conclusions
This chapter investigates the effectiveness of a design method for seismic

upgrading of existing r.c. frames by means of BRBs. The proposed design method
controls the shear strength and lateral stiffness to be provided by BRBs by means of
the behaviour factor q and the design storey drift Δd. Furthermore, it determines the
heightwise distributions of the additional strength and stiffness aiming at promoting

(1) a widespread yielding of the frame and (2) a uniform exploitation of the
displacement capacity. In particular, at each storey, the additional shear strength
to be provided by BRBs is determined so that the ratio of the shear force demand to
the total shear strength is constant along the height of the frame. Furthermore, the
additional lateral stiffness is determined in such a way that the drift demand to
capacity ratio is the same at all the storeys.
The effectiveness of the proposed design method is assessed on a r.c. frame
originally designed for gravity loads only. Nine retrofit solutions with BRBs
designed by the proposed method are analysed. In particular, Δd is assumed equal
to 0.6, 0.7 and 0.8 times of the storey drift Δl corresponding to the Significant
Damage limit state (drift capacity). Instead, q is assumed equal to 5, 6 and 7. The
results of the nonlinear dynamic analysis demonstrate that the seismic response of
the bare r.c. frame is irregular in elevation because of a significant concentration of
the drift demand at the 4th storey. As a consequence, the drift demand to capacity
ratio is larger than one at this storey. The insertion of the BRBs within the r.c. frame
makes its seismic response more regular in elevation, mitigates the concentration of
the storey drift and improves the seismic performance in terms of drift demand to
capacity ratio. The obtained benefits depend on the values assumed for Δd and q. In
particular, if Δd is assumed equal to 0.6 Δl, the storey drift demand never exceeds
the capacity independently of the value of q. Instead, when Δd is equal to 0.8 Δl, a
behaviour factor q not larger than 6 has to be adopted to avoid the exceedance of the
Significant Damage limit state.
References
Amara F, Bosco M, Marino EM, Rossi PP (2014) An accurate strength amplification factor for the
design of SDOF systems with P-Δ effects. Earthq Eng Struct Dyn 43(4):589–611
Barbagallo F, Bosco M, Ghersi A, Marino EM, Rossi PP, Stramondo PR (2014) Calibration of a
design method for seismic upgrading of existing r.c. frames by BRBs. In: Proceedings of the
2nd European conference on earthquake engineering and seismology. Istanbul. 25–29 August
2014
Bosco M, Marino EM (2013) Design method and behavior factor for steel frames with buckling
restrained braces. Earthq Eng Struct Dyn 42(8):1243–1263
CEN (2004) EN 1998-1, EuroCode 8: design of structures for earthquake resistance – Part 1:
General rules, seismic actions and rules for buildings. European Committee for Standardiza-
tion, Bruxelles
CEN (2010) EN 1998-1, EuroCode 8: design of structures for earthquake resistance – Part 3:
Assessment and retrofitting of buildings. European Committee for Standardization, Bruxelles
Daniel Y, Lavan O, Levy R (2013) A simple methodology for the seismic passive control of
irregular 3D frames using friction dampers. In: Lavan O, De Stefano M (eds) Seismic
behaviour and design of irregular and complex civil structures, vol 24. Springer, Dor-
drecht, Netherlands: 285–295
Mazzoni S, McKenna F, Scott MH, Fenves GL, Jeremic B (2003) OpenSEES manual. Pacific
Earthquake Engineering Research Center, University of California, Berkeley
Merritt S, Uang CM, Benzoni G (2003) Subassemblage testing of CoreBrace buckling restrained
braces. Structural Systems Research Project, Report no. TR-2003/01. University of California,
San Diego
Rossi PP (2014) Importance of isotropic hardening in the modeling of buckling restrained braces.
J Struct Eng-ASCE 141(4), doi:10.1061/(ASCE)ST.1943-541X.0001031
Uang CM, Nakashima M (2004) Steel buckling-restrained braced frames. In: Bozorgnia Y,
Bertero VV (eds) Earthquake engineering from engineering seismology to performance
based engineering. CRC Press LLC, Boca Raton, Florida, USA
Xie Q (2005) State of the art of buckling restrained braces in Asia. J Constr Steel Res
61(6):727–748
Zona A, Dall’Asta A (2012) Elastoplastic model for steel buckling restrained braces. J Constr Steel
Res 68(1):118–125
Chapter 17
Application of Nonlinear Static Method
with Corrective Eccentricities to Steel
Multi-storey Braced Buildings
Melina Bosco, Giovanna A.F. Ferrara, Aurelio Ghersi, Edoardo M. Marino,

Abstract Nonlinear static methods may be not very effective in the assessment of
3D building structures because sometimes they do not provide an accurate estimate
of the deck rotations. In order to overcome this shortcoming, the authors of this
chapter proposed a nonlinear static approach for asymmetric structures that is
performed applying the lateral force with two different eccentricities (named
corrective eccentricities) with respect to the centre of mass of the deck. In this
chapter the effectiveness of the corrective eccentricity method is verified with
reference to four five-storey mass eccentric steel buildings in which the seismic
force is sustained by frames equipped with buckling restrained braces.
Keywords Asymmetric buildings • Multi-storey systems • Steel structures •

Seismic assessment • Corrective eccentricity method
17.1 Introduction
The nonlinear static method of analysis has gained considerable popularity in recent
years because it is a fair compromise between the simplicity of the linear method of
analysis and the effectiveness of the nonlinear time-history analysis.
The nonlinear static method suggested in EuroCode 8 (CEN 2004; Fajfar 1999)
generally provides reasonable results for planar frames (Bosco et al. 2009; Giorgi
and Scotta 2013) while, in the case of in-plan irregular structures, it is effective only
for torsionally-rigid systems (Bento et al. 2010; De Stefano et al. 2013, 2014; Fajfar
et al. 2005; Fujii 2014; Kreslin and Fajfar 2012). Based on this consideration,
methods specifically improved for 3D structures have been proposed. For instance,
M. Bosco • G.A.F. Ferrara • A. Ghersi • E.M. Marino • P.P. Rossi (*)

Department of Civil Engineering and Architecture, University of Catania, Viale A. Doria 6,
[email protected]; [email protected]

DOI 10.1007/978-3-319-14246-3_17
194 M. Bosco et al.
Kreslin and Fajfar (2012) developed an improved version of the N2 method called
extended N2 method. The authors of this chapter have recently proposed a new
approach based on a double application of the nonlinear static analysis (Bosco
et al. 2012). In particular, for each considered direction of the seismic action, the
nonlinear static analysis is applied with reference to two different points of the
deck. The points of application of the force have been properly defined in Bosco
et al. (2012) through mathematical relations based on the investigation of the
seismic response of a large set of asymmetric single-storey structural systems.
The resisting elements of these systems were endowed with lateral stiffness and
strength in their plane only and subjected to bidirectional ground motions. The
distances e1 and e2 between the points of application of the forces and the centre of
mass CM are named corrective eccentricities and the proposed nonlinear static
method is called corrective eccentricity method. The corrective eccentricities e1 and
e2 are calculated by simple relations as a function of the parameters that mostly
influence the lateral-torsional coupling of the seismic response of asymmetric
buildings. These parameters are: the rigidity eccentricity er (distance between the
centre of rigidity CR and CM), the strength eccentricity es (distance between the
centre of strength CS and CM), the ratio Ωθ of the torsional to lateral frequencies of
the corresponding torsionally balanced system (obtained by shifting CM into CR)
and the ratio Rμ of the elastic shear force to the actual strength of the corresponding
planar system. This latter system is obtained from the asymmetric system by
restraining the deck rotation.
The effectiveness of the corrective eccentricity method in the prediction of the
peak response of in-plan irregular buildings has been mainly tested with reference
to single-storey systems (Bosco et al. 2013a). In this chapter, its effectiveness is
investigated by multi-storey steel buildings. The seismic response in terms of floor
displacements and storey drifts of each building is evaluated by nonlinear dynamic
analysis, by the corrective eccentricity method and by the standard nonlinear static
method (i.e. without corrective eccentricity). The errors committed by the two
considered nonlinear static methods in the prediction of the maximum dynamic
response are determined and compared.
17.2 Case Studies
The considered five storey buildings have rectangular decks with maximum dimen-
sion L and minimum dimension B equal to 30.0 and 18.0 m, respectively (Fig. 17.1).
The interstorey height is equal to 3.30 m. The seismic force is sustained by frames
with buckling restrained braces (BRBs) arranged in the chevron configuration. The
location of the bays with buckling restrained braces is highlighted in Fig. 17.1 by
double line. In order to analyse both torsionally-flexible (TF) and torsionally-rigid
(TR) systems, two different plan layouts of the resisting elements are considered. In
the first layout (Fig. 17.1a), all the frames with BRBs arranged along the y-axis are
close to the centre of rigidity (frames Y3, Y4, Y5 and Y6). Instead, in the second
17 Corrective Eccentricity and Braced Buildings 195
y y
a b Y1
Y1 Y2 Y3 Y4 Y5 Y6 Y7 Y8 Y2 Y3 Y4 Y5 Y6 Y7 Y8
X5 X5
X4 X4
B = 18.0 m
X3 X3
x O=G x
O=G
X2 X2
X1 X1
L = 30.0 m L = 30.0 m
Fig. 17.1 Plan layout of the buildings (a) torsionally-flexible and (b) torsionally-rigid structures
layout (Fig. 17.1b), two of the frames with BRBs are located on the two sides of the
building (frames Y1 and Y8). For all the considered systems, the braced frames are
symmetrically arranged. Consequently, the centres of rigidity and those of strength
are always coincident with the geometrical centre of the deck.
The storey mass is calculated according to EC8 by taking into account the dead
and live loads present in the building in the seismic design situation (5.0 kN/m2).
The mass radius of gyration is equal to 10.1 m (0.337 L ). Masses are lumped at deck
level and have the same distribution within every deck. The centres of mass of the
storeys are lined up on a vertical axis and belong to the x-axis because masses are
assumed symmetric with respect to this axis. Structural systems with different
values of the rigidity eccentricity er are obtained considering different positions
of CM. Two values of the rigidity eccentricity are considered. These values are
equal to 0.05 L and 0.15 L and are later named low (-L) and high (-H) rigidity
eccentricities. Centres of mass are assumed to be on the right of the origin O.
Therefore, the left side of the deck is the stiff side while the right one is the flexible
side. The strength eccentricity es is always equal to the rigidity eccentricity.
The frames with BRBs are designed applying the seismic force to two separate
planar models along the x- and y-directions. The design method adopted is that
proposed in Bosco and Marino (2013). Steel grade S235 with characteristic value of
the yield stress fy ¼ 235 MPa is used for all the members. The design seismic force
is evaluated by the elastic spectrum proposed in EC8 for soil type C, characterised
by a peak ground acceleration ag equal to 0.35 g, and reduced by the behaviour
factor q equal to 4. Beam and column cross-sections are selected among the
European wide flange shapes (HEA for beams and HEB for columns). More details
about the design method adopted for the frames with BRBs may be found in Bosco
and Marino (2013). The design axial force of the gravity columns is evaluated
according to the tributary area concept considering the load per square meter for
non-seismic design situation equal to 9.2 kN/m2.
The periods of vibration of the corresponding torsionally-balanced systems
(uncoupled periods) are grouped in triplets. Each triplet contains the periods of
the modes of vibration (translational and rotational) characterized by the same
196 M. Bosco et al.
Table 17.1 Uncoupled and coupled periods of the systems

Uncoupled periods [s] Coupled periods [s]
Building Ty1 Tθ1 Ty2 Tθ2 Tyθ1 Tθy1 Tyθ2 Tθy2
TF – L 0.73 0.88 0.25 0.30 0.71 0.92 0.25 0.32
TF – H 0.73 0.88 0.25 0.30 0.62 1.06 0.22 0.37
TR – L 0.73 0.58 0.25 0.20 0.74 0.57 0.26 0.20
TR – H 0.73 0.58 0.25 0.20 0.81 0.52 0.28 0.18
number of changes in the sign of the modal components. The uncoupled periods of
the translational mode of vibration in the y-direction and that of the rotational mode
are reported in Table 17.1 along with the (coupled) periods of vibration of the
asymmetric buildings.
17.3 Modelling of the Structures
The numerical analyses are carried out by means of the OpenSees program
(Mazzoni et al. 2003). The buildings are represented by means of a 3D centreline
model with rigid diaphragms. Beams of the braced frames and columns are
modelled by means of elements (Beam With Hinges Elements) which are elastic
in the middle and inelastic at the ends within parts of finite length. All beam-to-
column connections are pinned. BRBs are modelled by means of a single truss
element. To take into account the variability of the cross-section of the BRB along
the longitudinal axis, these latter members are characterised by an equivalent cross-
section area Aeq equal to
Ac
Aeq ¼ L j Ac ð17:1Þ
Lw A j þ LLwt AAct þ LLwc
where Lw is the length of the brace, Lc ¼ 0.5 Lw is the length of the yielding core of
BRB, Lj/2 ¼ 0.65 m is the length of each part of the connection segment, Lt is the
length of the transition segment and Ac/At ¼ 0.5 and Ac/Aj ¼ 0.3 are the ratios of the
cross-sectional area of the yielding core to the cross-sectional areas of the transition
and connection segments, respectively.
The Giuffré-Menegotto-Pinto uniaxial material is assigned to the truss element.
The plastic resistance of the BRBs is equal to the axial force corresponding to the
yielding of the core. Therefore, the yielding stress is equal to fy,eq.
Ac
f y, eq ¼ f ym ð17:2Þ
Aeq
fym being the mean value of the yield stress of the BRB (235 MPa). The post-yield
stiffness ratio kh is fixed equal to 0.053.
17.4 Seismic Analyses
The seismic analyses are performed in order to predict the seismic response of the
buildings in terms of the floor displacements and storey drifts along the y-direction,
which is the one affected by lateral-torsional coupling. Both nonlinear dynamic
analysis (NDA) and nonlinear static (NS) analysis are performed. The seismic
response obtained by the nonlinear dynamic analysis is assumed as benchmark to
be predicted by the nonlinear static method of analysis. This analysis is also used to
determine the seismic response of the corresponding planar systems. All the
seismic analyses are carried out starting from a structural configuration in which
the structural members have already been subjected to the gravity loads of the
seismic design situation. P-Δ effects are not considered in the analysis because the
interstorey drift sensitivity coefficient is lower than 0.1 for all the considered
buildings (CEN 2004).
17.4.1 Dynamic Analysis
Seven ground motions are considered for the nonlinear dynamic analyses. The
single ground motion consists of two accelerometric components acting along the
x- and y-directions. The components of each ground motion are different from each
other and compatible with the elastic response spectrum proposed in EC8 for soft
soil (type C) and equivalent viscous damping ratio equal to 0.03. The accelerograms
have been generated by means of the SIMQKE program (1976). The total length of
the accelerogram is equal to 20 s while that of the strong motion phase is equal to
7 s, i.e. slightly lower than the minimum value suggested in EC8. Details about the
choice of these accelerograms are described in Amara et al. (2014). Three levels of
seismic excitation are considered. Specifically, the peak ground acceleration is set
equal to 0.25 g, 0.35 g or 0.43 g.
The Rayleigh formulation is used to introduce damping. Mass and stiffness
coefficients are defined so that two modes of vibration of the structures are
characterised by an equivalent viscous damping ratio equal to 0.03. For planar
systems, the considered modes of vibration are the first and second modes of
vibration in the y-direction (i.e. Ty1 and Ty2 in Table 17.1). For the asymmetric
systems, the first mode of vibration is that with the maximum period between Tyθ1
198 M. Bosco et al.
and Tθy1 and the second mode is that with the maximum period between Tyθ2 and
Tθy2.
17.4.2 Pushover Analysis
For each building, the pushover analysis is performed by a set of horizontal forces
applied in the y-direction and distributed along the height according to an inverted
triangular load pattern. For each considered seismic excitation level, the pushover
analysis is stopped when the displacement of the centre of mass of the top floor of
the asymmetric system is equal to the average of the seven maximum top dynamic
displacements of the corresponding planar system (target displacement). The
present research study is intended to investigate solely the improvement caused
by the corrective eccentricities to the prediction of the torsional response. Owing to
this, the errors committed by the pushover analysis and regarding the translational
response of the structures have been eliminated (Bosco et al. 2009). In particular,
when predicting the floor displacements, the displacements provided by the push-
over analysis at each storey of the building are scaled so that the displacement at the
centre of mass equals the average of the maximum dynamic displacements of the
corresponding planar system. When predicting the storey drifts, a similar scaling
procedure is applied to the storey drifts.
17.5 Evaluation of the Corrective Eccentricities
The calculation of the corrective eccentricities e1 and e2 requires that the parame-
ters er, es, Ωθ and Rμ be determined. The rigidity eccentricity and the strength
eccentricity are equal to 4.5 m (i.e. 0.15 L ) for systems TF-H and TR-H, while
they are equal to 1.5 m (i.e. 0.05 L ) for systems TF-L and TR-L.
Among the methods available in literature (Bosco et al. 2013b; Calderoni
et al. 2002; Doudomis and Athanatopoulou 2008; Makarios and Anastassiadis
1998a, b; Makarios 2008; Georgoussis 2009), the one proposed by Makarios and
Anastassiadis (1998a, b) is used for the determination of the uncoupled torsional to
lateral frequency ratios. According to this method, the uncoupled ratio Ωθy is
calculated as the ratios of the rigidity radius of gyration rkx to the mass radius of
gyration. This ratio is equal to 0.822 and 1.258 for the torsionally-flexible and the
torsionally-rigid structures, respectively.
The parameter Rμ is determined as the ratio of the required elastic base shear Vb,
el to the actual lateral strength Vby,u of the corresponding planar system. The elastic
base shear is obtained by modal response spectrum analysis. The elastic response
spectrum is that given in EC8 for soil C with peak ground acceleration ag depending
on the considered seismic excitation level (0.25 g, 0.35 g or 0.43 g). The actual
lateral strength is calculated by pushover analysis, as the base shear corresponding
to the target displacement of the top floor required by the assigned seismic level.
For all the considered structures, the ratios Rμ corresponding to the three considered
values of ag are equal to 2.69, 3.44 or 4.01.
The corrective eccentricities calculated on the basis of the relations proposed in
Bosco et al. (2012) are reported in Table 17.2.
17.6 Results of the Numerical Analysis
In this section, a comparison is carried out between the seismic response obtained
by means of the considered nonlinear static methods, i.e. the corrective eccentricity
method and the standard nonlinear static method, and that obtained by nonlinear
dynamic analysis. Note that the nonlinear static method without corrective eccen-
tricity will be named as “EC8 method” for brevity even if it does not exactly
coincide with the method reported in the code. Indeed, as remarked in Sect. 17.4.2,
the results presented in this section correspond to the average top displacement
obtained by means of the nonlinear dynamic analyses.
The in-plan distributions of the roof displacements obtained by the considered
methods are compared in Fig. 17.2 with reference to systems with high eccentricity
and peak ground acceleration equal to 0.35 g.
To provide a measure of the capability of the nonlinear static methods to predict
the results of nonlinear dynamic analysis, the percentage error committed in the
estimate of the dynamic displacement is calculated as
usti uidyn
Err i ð%Þ ¼ , i ¼ 1 or 2 ð17:3Þ
uidyn
where, uidyn is the mean value of the maximum displacements obtained by

nonlinear dynamic analysis, uist is the displacement obtained by means of the
nonlinear static method of analysis, and the subscript i equal to 1 or 2 denotes the
stiff or the flexible side of the building. The equation above is also used to evaluate
the percentage errors committed in the prediction of the storey drifts of the stiff and
flexible sides. Positive values of the error indicate overestimates of the dynamic
response (conservative errors), while negative values indicate underestimates of the
dynamic response (unconservative errors). The variability of the errors committed
at each storey is finally represented in Fig. 17.3. In particular, for each building, a
bar limited by the minimum and the maximum errors is reported. Black and gray
bars represent errors committed by the standard nonlinear static method or by the
method of the corrective eccentricity, respectively.
Figures 17.2 and 17.3 show that when the nonlinear static method is applied
according to the EC8 provisions (i.e. without corrective eccentricities), the dis-
placements and drifts of the flexible side are well predicted for all the considered
200
Table 17.2 Corrective eccentricities

TF – L TF – H TR – L TR – H
ag/g 0.25 0.35 0.43 0.25 0.35 0.43 0.25 0.35 0.43 0.25 0.35 0.43
e1 (m) –1.97 –1.87 –1.79 –5.91 –5.63 –5.38 –0.86 –0.77 –0.65 –2.59 –2.30 –1.93
e2 (m) –1.10 –0.98 –0.86 –3.30 –2.93 –2.57 –0.22 0.120 0.53 –0.65 0.36 1.60
M. Bosco et al.
NDA Corrective Eccentricity Method Standard NS method

u u
(cm) TF-H (cm) TR-H
15 15
10 10
5 5
0 0
- L /2 CR CM L /2 - L /2 CR CM L /2
Fig. 17.2 Comparison of the top displacements obtained by the considered methods
buildings. Instead, the displacements and drifts of the stiff side are well predicted
only for the structure TR-L.
The application of the corrective eccentricity method maintains the effectiveness
of the EC8 method in the prediction of the displacement of the flexible side and
provides a better estimate on the stiff side. In fact, in the worst case the maximum
unconservative errors committed by the EC8 and corrective eccentricity methods
are about 90 and 12 % in terms of floor displacements (Fig. 17.3a), and about
100 and 35 % in terms of storey drifts (Fig. 17.3b). Even if it is not shown in the
figure, when the corrective eccentricity method is applied, the above mentioned
maximum unconservative error in the estimate of storey drifts is committed at a
single storey while the errors at the other storeys are significantly lower.
17.7 Conclusions
The chapter presents the validation of the corrective eccentricity method on a set of
four multi-storey braced structures. The set of buildings includes torsionally-
flexible and torsionally-rigid structures characterised by low or high rigidity and
strength eccentricities. Further, three different seismic excitation levels are consid-
ered to analyse the effectiveness of the corrective eccentricity method in case of
structures with moderate or high inelastic response. The effectiveness of the
corrective eccentricity method is demonstrated by comparing its accuracy in
predicting the nonlinear dynamic response of the analysed asymmetric buildings
to that obtained by the standard application of the nonlinear static method,
i.e. without corrective eccentricities. This comparison shows that the standard
nonlinear static method provides a suitable prediction of the floor displacement
and storey drifts only of the flexible side of the building. The corrective eccentricity
method significantly enhances the prediction of floor displacement and storey drift
on the stiff side of the building and the error committed in the prediction is always
reasonable.
202 M. Bosco et al.
Corrective Eccentricity Method Standard NS Method

100% 100%
75% Stiff Side Flexible Side 75% Stiff Side Flexible Side
50% 50%
25% 25%
0% 0%
-25% -25%
-50% -50%
-75% 0.25 g -75%
-100% -100%
TF-L
TF-H
TR-L
TR-H
TF-L
TF-H
TR-H
TR-L
TF-L
TF-H
TR-L
TR-H
TF-L
TF-H
TR-L
TR-H
100% 100%
50% 50%
25% 25%
0% 0%
-25% -25%
-50% -50%
-75% 0.35 g -75%
-100% -100%
TF-L
TF-H
TR-L
TR-H
TF-L
TF-H
TF-L
TF-H
TR-L
TR-L
TR-H
TF-L
TF-H
TR-L
TR-H
TR-H
100% 100%
50% 50%
25% 25%
0% 0%
-25% -25%
-50% -50%
-75% 0.43 g -75%
-100% -100%
TF-L
TF-H
TR-L
TR-H
TF-L
TF-H
TF-L
TF-H
TR-L
TR-H
TF-L
TF-H
TR-L
TR-H
TR-L
TR-H
a b
Fig. 17.3 Percentage errors in the evaluation of (a) floor displacements and (b) storey drifts
References
Amara F, Bosco M, Marino EM, Rossi PP (2014) An accurate strength amplification factor for the
design of SDOF systems with P-Δ effects. Earthq Eng Struct Dyn 43(4):589–611
Bento R, Bhatt C, Pinho R (2010) Verification of nonlinear static procedures for a 3D irregular
SPEAR building. Earthq Struct 1(2):177–195
Bosco M, Marino EM (2013) Design method and behavior factor for steel frames with buckling
restrained braces. Earthq Eng Struct Dyn 42(8):1243–1263
Bosco M, Ghersi A, Marino EM (2009) On the evaluation of seismic response of structures by
nonlinear static methods. Earthq Eng Struct Dyn 38(13):1465–1482
Dyn 41(13):1751–1773
Bosco M, Ghersi A, Marino EM, Rossi PP (2013a) Comparison of nonlinear static methods for the
assessment of asymmetric buildings. Bull Earthq Eng 11(6):2287–2308
Bosco M, Marino EM, Rossi PP (2013b) An analytical method for the evaluation of the in-plan
irregularity of non-regularly asymmetric buildings. Bull Earthq Eng 11(5):1423–1445
Calderoni B, D’Aveni A, Ghersi A, Rinaldi Z (2002) Static vs. modal analysis of asymmetric
buildings: effectiveness of dynamic eccentricity formulations. Earthq Spectra 18(2):219–231
CEN, European Committee for Standardization (2004) Eurocode 8: design of structures for
earthquake resistance. Part 1: general rules, seismic actions and rules for buildings. EN
1998-1:2004. Brussels, Belgium
De Stefano M, Tanganelli M, Viti S (2013) On the variability of concrete strength as a source of
irregularity in elevation for existing RC buildings: a case study. Bull Earthq Eng 11(5):
1711–1726
source of in-plan irregularity for existing RC framed structures. Eng Struct 59(2):161–171
Doudomis IN, Athanatopoulou AM (2008) Invariant torsion properties of multistorey asymmetric
buildings. Struct Des Tall Spec Build 17(1):79–97
Fajfar P (1999) Capacity spectrum method based on inelastic demand spectra. Earthq Eng Struct
Dyn 28(9):979–993
Fajfar P, Marusic D, Peruš I (2005) Torsional effects in the pushover-based seismic analysis of
buildings. J Earthq Eng 9:831–854
under bi-directional excitation using pushover analyses. Bull Earthq Eng 12(2):909–938
Giorgi P, Scotta R (2013) Validation and improvement of N1 method for pushover analysis. Soil
Dyn Earthq Eng 55:140–147
Makarios T (2008) Practical calculation of the torsional stiffness radius of multi-storey tall
buildings. Struct Des Tall Spec Build 17(1):39–65
Makarios T, Anastassiadis A (1998a) Real and fictitious elastic axes of multi-storey buildings:
theory. Struct Des Tall Spec Build 7(1):33–55
Makarios T, Anastassiadis A (1998b) Real and fictitious elastic axes of multi-storey buildings:
applications. Struct Des Tall Spec Build 7(1):57–71
Mazzoni S, McKenna F, Scott MH, Fenves GL, Jeremic B (2003) OpenSEES command language
manual. Pacific Earthquake Engineering Research Center, University of California, Berkeley
SIMQKE User Manual (1976) NISEE software library. University of California, Berkeley
Georgoussis GK (2009) An alternative approach for assessing eccentricities in asymmetric
multistory buildings: 1. Elastic Systems. Struct Des Tall Spec Build 18(2):181–202
Chapter 18
Influence of the Interaction Yield
Domain on Lateral-Torsional Coupling
of Asymmetric Single-Storey Systems
Melina Bosco, Aurelio Ghersi, Edoardo M. Marino, and Pier Paolo Rossi
Abstract Single-storey models are widely adopted because they are able to
describe the main aspects of the torsional coupling of asymmetric buildings and,
at the same time, they are simple enough to enable extensive parametric analysis.
Generally, single-story systems are constituted by a rigid deck, where the mass is
lumped, supported by vertical resisting elements with lateral stiffness and strength
in their plane only (uni-axial resisting elements). Thus, these models neglect the
interaction phenomena that can be observed in actual framed structures, where the
vertical resisting elements provide lateral stiffness and strength in all the horizontal
directions. In this chapter, the influence of the interaction phenomena on the
torsional coupling of the seismic response of asymmetric buildings predicted by
nonlinear methods of analysis is investigated. Single storey systems with bi- and
uni-axial resisting elements are used to simulate structures whose seismic response
is affected or not by interaction phenomena. Finally, both nonlinear dynamic and
nonlinear static methods of analysis are investigated.
Keywords Asymmetric buildings • Single-storey systems • Seismic assessment •

Extended N2 method • Corrective eccentricity method
18.1 Introduction
The seismic response of asymmetric buildings depends on many parameters. In

order to evaluate the influence of each parameter, the studies on this topic are often
carried out by means of extensive parametrical analyses. An extensive parametrical
study requires the adoption of simple mathematical models. For this reason, single-
storey models are widely adopted to investigate the torsional behaviour of buildings
M. Bosco (*) • A. Ghersi • E.M. Marino • P.P. Rossi

[email protected]

DOI 10.1007/978-3-319-14246-3_18
206 M. Bosco et al.
(Bosco et al. 2012). In fact, single-storey systems represent the extreme schemati-
zation of a real building and, although they do not cover some peculiarities related
to the heightwise distribution of structural properties (Anagnostopoulos et al. 2010;
De Stefano et al. 2013, 2014; Ghersi et al. 2007), they are able to describe the main
aspects of the torsional seismic behaviour of actual asymmetric buildings which are
regular in elevation (Bosco et al. 2013; De Stefano and Pintucchi 2010; Goel and
Chopra 1990; Hejal and Chopra 1987; Palermo et al. 2013; Peruš and Fajfar 2005).
Generally, single-storey systems are constituted by a rigid deck, where the mass
is lumped, supported by vertical resisting elements with lateral stiffness and
strength in their plane only (uni-axial resisting elements). Based on the response
of these models, four structural parameters seem to influence mostly the seismic
response of asymmetric systems (Goel and Chopra 1990; Hejal and Chopra 1987;
Palermo et al. 2013). These parameters are the rigidity eccentricity er (distance
between the centre of rigidity CR and the centre of mass CM), the ratio Ωθ of the
torsional to lateral frequencies of the corresponding torsionally balanced system
(obtained by shifting CM into CR), the strength eccentricity es (distance between the
centre of strength CS and CM) and the ratio Rμ of the elastic strength demand to the
actual strength of the system.
The single storey models with uni-axial resisting elements may be considered
representative of multi-storey buildings in which the seismic forces are sustained by
braced frames or shear walls. Instead, these models are less effective in simulating
the response of r.c. and steel framed structures (Bosco et al. 2015). The vertical
resisting elements of these structures, i.e. the columns, provide lateral stiffness and
strength in all the horizontal directions and are characterised by a bi-axial yield
domain such that the presence of bi-axial bending reduces their strength capacity
(interaction phenomena). These aspects are neglected by uni-axial resisting
elements.
In this chapter, the influence of interaction phenomena on torsional coupling of
the seismic response of asymmetric buildings is investigated. Specifically, two
different sets of single-storey models are considered. In the first set, the resisting
elements provide lateral stiffness and strength along only one horizontal direction
(either x- or y-direction of an orthogonal coordinate reference system). In the
second set, the resisting elements provide lateral stiffness and strength in all the
horizontal directions and possess a bi-axial yield domain (bi-axial resisting ele-
ments) to take into account the interaction phenomena. The seismic response of the
considered models is expressed in terms of in plan distribution of displacements and
is evaluated by both nonlinear dynamic and nonlinear static analyses.
18.2 The Single-Storey Models
All the analysed systems have a deck that is rectangular in plan and has dimensions,
denoted as B and L, equal to 12.5 m and 29.5 m, respectively. The mass m of the
deck is equal to 1416 t, while the mass radius of gyration rm about the centre of
18 Influence of the Interaction Yield Domain 207
mass CM is equal to 0.312 L. The resisting elements are located symmetrically with
respect to the geometrical centre of the deck G.
The considered systems have values of Ωθ which range from 0.6 to 1.4 with step
0.05 so as to analyse both torsionally flexible (Ωθ < 1) and torsionally stiff systems
(Ωθ > 1). For each value of Ωθ, several values of er from 0.1 L to 0 with step
0.025 L are considered to include both systems with small and large rigidity
eccentricity. The total lateral strength S of the systems is determined assuming
values of Rμ from 1.0 to 6.0 with step 1.0. As a consequence, the considered systems
may experience either moderate or large plastic deformations during the earth-
quake. Finally, for each value of Rμ, the total lateral strength is distributed among
the resisting elements considering values of es from 0.1 L to 0.1 L with step
0.025 L.
The influence of interaction phenomena is evaluated by comparing the seismic
response of systems characterised by the same values of er, es, Ωθ and Rμ and
belonging to the sets of systems described in the two following Sections.
18.2.1 Systems with Uni-axial Resisting Elements
The resisting elements of the first set of systems (Fig. 18.1a) are arranged along the
axes of the assumed reference system (4 along the x-axis and 8 along the y-axis).
Each vertical resisting element has lateral stiffness and strength in its plane only
(Fig. 18.1b) and is characterised by a bi-linear force-displacement relationship. The
contribution γx of the resisting elements arranged along the x-axis to the total
torsional stiffness of the system about the centre of rigidity CR is equal to 20 %.
This value is proper for buildings with a rectangular plan.
Structural systems with prefixed values of er are generated by modifying the
ð1Þ
position of CM. The lateral stiffness kxi of the i-th resisting element along the x-
ð1Þ
direction and that kyj of the j-th element along the y-direction are modified
a y (1) b y
kxi
syj
Side 2 (flexible)
Side 1 (stiff)
(1)
kyj
sxi
B CR=G CS CM x x
es
er
L
Fig. 18.1 (a) Plan layout of the systems of the first set; (b) uni-axial resistance of the resisting
elements
208 M. Bosco et al.
a (2) (2)
y b y
kxi,j,kyi,j Fy
s(2)
yi
Side 2 (flexible)
Side 1 (stiff)
(2)
sxi
B CM Fx
CR=G CS x x
es
er
L
Fig. 18.2 (a) Plan layout of the systems of the second set; (b) bi-axial yield domain of the
resisting elements
according to the procedure described in Ghersi and Rossi (2000) to obtain prefixed
values of Ωθ. Symmetry is maintained with respect to the x-axis.
The global lateral strength Sx ¼ Sy of the systems is assigned so as to have
prefixed values of the ratio Rμ of the elastic strength demand to the actual strength
of the corresponding planar system (i.e. of the system obtained by restraining the
deck rotation of the basic scheme).
Then, in order to have assigned values of the strength eccentricity es, the lateral
strength is distributed among the resisting elements according to the procedure
described in Bosco et al. (2012). Furthermore, in order to consider the influence of
the strength distribution on the inelastic seismic response of the systems, ten
random distributions of strength sxi, syi are generated for each value of es.
18.2.2 Systems with Bi-axial Resisting Elements
The resisting elements of the systems of the second set (Fig. 18.2a) provide lateral
stiffness and strength in all the horizontal directions. The resisting elements are
located at the points of the deck, which are the intersection of the axes of the
resisting elements of the first set of systems.
In order to have systems which are characterised by the same values of the global
parameters (er, es, Ωθ and Rμ) of the corresponding systems of the first set, the
ð2Þ ð2Þ
components along the x- and y-direction of the lateral stiffness (kxi;j , kyi;j ) and
ð2Þ ð2Þ
strength (sxi;j , syi;j ) of the resisting element which is located at the intersection of
the i-th and j-th element are obtained by the relations:
ð1Þ ð1Þ
ð2Þ kxi ð2Þ ky j
kxi, j ¼ ð1Þ
k yi, j ¼ ð1Þ
ð18:1Þ
ny nx
ð1Þ ð1Þ
ð2Þ sxi, j ð2Þ sy j
sxi, j ¼ ð1Þ
s yi, j ¼ ð1Þ
ð18:2Þ
ny nx
ð1Þ ð1Þ
where n y ¼ 8 and nx ¼ 4 are the number of resisting elements arranged along the
y- or x-direction in the systems of the first set.
An elastic-perfectly plastic constitutive relation following the normality rule is
adopted, and interaction phenomena arising in the nonlinear range of behaviour are
accounted for by means of an ellipse yield domain (Fig. 18.2b).
18.3 Numerical Analyses
In order to investigate the influence of interaction phenomena on torsional coupling

of the seismic response of asymmetric buildings, both pushover and nonlinear
dynamic analyses are carried out.
A Rayleigh viscous damping is used for nonlinear dynamic analysis and set at
5 % for the first and the third modes of vibration. The Newmark method with
parameters α and δ set at 0.25 and 0.5, respectively, is used to evaluate the dynamic
response by step-by-step integration of the equations of motion. Bi-directional
ground motions are used for nonlinear dynamic analysis. The accelerograms,
generated by the SIMQKE computer program, are compatible with the elastic
spectrum reported in Eurocode 8 for soil type C, 5 % damping ratio and peak
ground acceleration ag equal to 0.35 g. Each accelerogram is modeled by a
trapezoidal envelope function with initial, central (stationary part) and final parts
of 3, 22.5 and 5 s, respectively. The mean of the zero period spectral response
acceleration values of the generated accelerograms is not lower than the value
stipulated in Eurocode 8 and no value of the mean response spectrum is lower than
90 % of the corresponding value proposed in Eurocode 8.
The pushover analysis is carried out by applying a force along the direction of
the displacements to be estimated, e.g. the y-direction. The force is applied at the
centre of mass of the deck. The pushover analysis is stopped when the displacement
of the centre of mass of the asymmetric system is equal to the average of the
10 maximum displacements of the corresponding planar system evaluated by
nonlinear dynamic analysis.
18.4 Comparison of the Seismic Response of the Systems
Figure 18.3a shows the distribution of the maximum dynamic displacements

(udyn,y) normalised to the maximum displacements of the corresponding planar
system (upl,y). The results refer to a system characterised by er ¼ es ¼ 0.10 L,
210 M. Bosco et al.
Uni-axial resisting elements Bi-axial resisting elements

a b
u dyn,y /u pl,y u st,y /u pl,y
1.5 1.5
1.0 1.0
0.5 0.5
=1.0, R µ=3, e r = e s= −0.10L
0.0 0.0
-L /2 CR=CS CM L /2 -L /2 CR=CS CM L /2
Fig. 18.3 Comparison between systems with uni- or bi-axial resisting elements (a) dynamic
displacements; (b) static displacements
Rμ ¼ 3, and Ωθ ¼ 1.0. The solid line represents the displacements of the system with
uni-axial resisting elements while the dashed line represents those of the
corresponding system with bi-axial resisting elements. The figure shows that the
increase in the dynamic displacements due to torsional effects on the flexible side is
equal for the two systems. Instead, at the stiff side, the torsional effects cause an
increase in the displacements obtained for systems with bi-axial resisting elements
(udyn,y/upl,y ¼ 1.09) and a reduction in the displacements for the system with
uni-axial resisting elements (udyn,y/upl,y ¼ 0.88).
A similar comparison is reported in Fig. 18.3b in terms of displacements
obtained by the pushover analysis (ust,y). Once again, the displacements are
normalised to the maximum displacements of the corresponding planar system.
The figure shows that, for the considered system, the torsional effects are more
significant in the systems with bi-axial resisting elements. In fact, at the stiff side
the ratio ust,y/upl,y is equal to 0.29 for the system with uni-axial resisting element
and equal to 0.14 for the system with bi-axial resisting elements. Similarly, the ratio
above is equal to either 1.48 or 1.58 at the flexible side.
Note that the results referring to Ωθ ¼ 1.0 have been plotted because this value of
Ωθ is proper of buildings with a framed structure, i.e. of buildings for which the
bi-axial interaction phenomena may be significant.
To summarise the results obtained for the large number of structures analysed,
the ratios udyn,y/upl,y and ust,y/upl,y are first calculated at the stiff (udyn1/upl, ust1/upl)
and flexible (udyn2/upl, ust2/upl) side for an assigned system (identified by fixed
values of Ωθ, Rμ es, er and by an assigned distribution of strength among the
resisting elements). As the values of these ratios corresponding to the 10 strength
distributions are always very close to one another, the average of the 10 values is
considered as representative of the results corresponding to the different strength
distributions.
Then, the ratios udyn/upl and ust/upl are plotted in Figs. 18.4 and 18.5 as a function
of er and es for assigned values of Ωθ and Rμ. The grey surface refers to the ratios
obtained for systems with uni-axial resisting elements; the white surface refers to
Rμ = 3.0 - Ωθ = 0.7 Rμ = 3.0 - Ωθ = 1.0 Rμ = 3.0 - Ωθ = 1.3

2 2 2
udin1/upl
udin1/upl
udin1/upl
1 1 1
0 0 0
-0.10 -0.10 -0.10
0.10 0.10 0.10
-0.05 0.00 -0.05 0.00 -0.05 0.00
er 0.00 -0.10
es er 0.00 -0.10
es er 0.00 -0.10
es
Rμ = 3.0 - Ωθ = 0.7 Rμ = 3.0 - Ωθ = 1.0 Rμ = 3.0 - Ωθ = 1.3

2 2 2
udin2/upl
udin2/upl
udin2/upl
1 1 1
0 0 0
-0.10 -0.10 -0.10
0.10 0.10 0.10
-0.05 0.00 -0.05 0.00 -0.05 0.00
er 0.00 -0.10
es er 0.00 -0.10
es er 0.00 -0.10
es
Rμ = 6.0 - Ωθ = 0.7 Rμ = 6.0 - Ωθ = 1.0 Rμ = 6.0 - Ωθ = 1.3

2 2 2
udin1/upl
udin1/upl
udin1/upl
1 1 1
0 0 0
-0.10 -0.10 -0.10
0.10 0.10 0.10
-0.05 0.00 -0.05 0.00 -0.05 0.00
er 0.00 -0.10
es er 0.00 -0.10
es er 0.00 -0.10
es
Rμ = 6.0 - Ωθ = 0.7 Rμ = 6.0 - Ωθ = 1.0 Rμ = 6.0 - Ωθ = 1.3

2 2 2
udin2/upl
udin2/upl
udin2/upl
1 1 1
0 0 0
-0.10 -0.10 -0.10
0.10 0.10 0.10
-0.05 0.00 -0.05 0.00 -0.05 0.00
er 0.00 -0.10
es er 0.00 -0.10
es er 0.00 -0.10
es
Fig. 18.4 Effects of the torsional coupling on the dynamic displacements at the stiff or flexible
side of the systems
systems with bi-axial resisting element. Figure 18.4 shows that, if nonlinear
dynamic analyses are carried out, the variations in the torsional response between
systems belonging to the two considered sets are moderate, independently of the
considered values of Ωθ and Rμ. Instead, the interaction phenomena affect the
prediction of the torsional response obtained by nonlinear static analysis signifi-
cantly (Fig. 18.5). Further, the differences between the surfaces representing the
two considered sets of systems increase as the ratio Rμ increases.
212 M. Bosco et al.
Rμ = 3.0 - Ωθ = 0.7 Rμ = 3.0 - Ωθ = 1.0 Rμ = 3.0 - Ωθ = 1.3

2 2 2
ust1/upl
ust1/upl
ust1/upl
1 1 1
0 0 0
-0.10 -0.10 -0.10
0.10 0.10 0.10
-0.05 0.00 -0.05 0.00 -0.05 0.00
er 0.00 -0.10
es er 0.00 -0.10
es er 0.00 -0.10
es
Rμ = 3.0 - Ωθ = 0.7 Rμ = 3.0 - Ωθ = 1.0 Rμ = 3.0 - Ωθ = 1.3

2 2 2
ust2/upl
ust2/upl
ust2/upl
1 1 1
0 0 0
-0.10 -0.10 -0.10
0.10 0.10 0.10
-0.05 0.00 -0.05 0.00 -0.05 0.00
er 0.00 -0.10
es er 0.00 -0.10
es er 0.00 -0.10
es
Rμ = 6.0 - Ωθ = 0.7 Rμ = 6.0 - Ωθ = 1.0 Rμ = 6.0 - Ωθ = 1.3

2 2 2
ust1/upl
ust1/upl
ust1/upl
1 1 1
0 0 0
-0.10 -0.10 -0.10
0.10 0.10 0.10
-0.05 0.00 -0.05 0.00 -0.05 0.00
er 0.00 -0.10
es er 0.00 -0.10
es er 0.00 -0.10
es
Rμ = 6.0 - Ωθ = 0.7 Rμ = 6.0 - Ωθ = 1.0 Rμ = 6.0 - Ωθ = 1.3

2 2 2
ust2/upl
ust2/upl
ust2/upl
1 1 1
0 0 0
-0.10 -0.10 -0.10
0.10 0.10 0.10
-0.05 0.00 -0.05 0.00 -0.05 0.00
er 0.00 -0.10
es er 0.00 -0.10
es er 0.00 -0.10
es
Fig. 18.5 Effects of the torsional coupling on the static displacements at the stiff or flexible side of
the systems
18.5 Conclusions
The chapter analyses the influence of the interaction phenomena on torsional

coupling of the seismic response of single-storey asymmetric
systems. Specifically, the seismic response is determined by nonlinear dynamic
and nonlinear static analysis for two sets of systems: systems with uni-axial
resisting elements and bi-axial resisting elements, which neglect and take into
account the interaction phenomena, respectively. Then, the displacements of the

two sides of the deck of the asymmetric systems normalised to the maximum
displacement of the corresponding planar system, which are used to quantify the
importance of the torsional coupling of the seismic response, are evaluated for the
two sets of systems and compared.
The investigation shows that the influence of the interaction phenomena on
torsional coupling effect is different depending on the method of analysis used
for the prediction of seismic response. The predictions of the seismic response
obtained by nonlinear dynamic analysis for systems with uni- and bi-axial resisting
elements are always similar. Instead, the torsional responses predicted by nonlinear
static analysis are similar only in case of systems with moderate inelastic response,
while they are significantly different in case of systems that sustain a large inelastic
displacement demand (large value of Rμ). Based on this consideration, nonlinear
static methods for seismic response prediction calibrated by means of systems with
uni-axial resisting elements should be used with caution on structures that are
affected by strong interaction phenomena (for instance r.c. framed structures) and
experience large inelastic deformation.
References
Anagnostopoulos SA, Alexopoulou C, Stathopoulos KG (2010) An answer to an important

controversy and the need for caution when using simple models to predict inelastic earthquake
response of buildings with torsion. Earthq Eng Struct Dyn 39(5):521–540
Dyn 41(13):1751–1773
Bosco M, Ghersi A, Marino EM, Rossi PP (2013) Comparison of nonlinear static methods for the
Bosco M, Ferrara GAF, Ghersi A, Marino EM, Rossi PP (2015) Seismic assessment of existing
r.c. framed structures with in-plan irregularity by nonlinear static methods. Earthq
Struct 8(2):401–422
De Stefano M, Pintucchi B (2010) Predicting torsion-induced lateral displacements for
pushover analysis: influence of torsional system characteristics. Earthq Eng Struct Dyn
39(12):1369–1394
De Stefano M, Tanganelli M, Viti S (2013) On the variability of concrete strength as a
source of irregularity in elevation for existing RC buildings: a case study. Bull Earthq Eng
11(5):1711–1726
source of in-plan irregularity for existing RC framed structures. Eng Struct 59(2):161–171
Ghersi A, Rossi PP (2000) Formulation of design eccentricity to reduce ductility demand in
asymmetric buildings. Eng Struct 22(7):857–871
Ghersi A, Marino EM, Rossi PP (2007) Static versus modal analysis: influence on inelastic
response of multi-storey asymmetric buildings. Bull Earthq Eng 5(4):511–532
Goel RK, Chopra AK (1990) Inelastic seismic response of one-story asymmetric-plan systems,
Report No. UCB/EERC-90/14. Earthquake Engineering Research Center, Berkeley
214 M. Bosco et al.
Hejal H, Chopra AK (1987) Earthquake response of torsionally-coupled buildings, Report

No. UBC/EERC-87/20. Earthquake Engineering Research Center, Berkeley
Palermo M, Silvestri S, Gasparini G, Trombetti T (2013) Physically-based prediction of the
maximum corner displacement magnification of one-storey eccentric systems. Bull Earthq
Eng 11(5):1467–1491
Chapter 19
Improved Nonlinear Static Methods
for Prediction of the Seismic Response
of Asymmetric Single-Storey Systems
Melina Bosco, Aurelio Ghersi, Edoardo M. Marino, and Pier Paolo Rossi
Abstract Traditional nonlinear static methods, e.g. the original version of the N2
method implemented in Eurocode 8, are not always effective in the assessment of
asymmetric structures. To overcome this shortcoming, two methods have been
recently suggested by Kreslin and Fajfar and by Bosco et al. In this chapter, the two
improved nonlinear static methods and the original N2 method are applied to
predict the maximum seismic response of three groups of single-storey systems.
Further, the systems of each group are schematised by means of two different
single-storey models. In the first model, a rigid deck is sustained by resisting
elements, which provide lateral stiffness and strength along only one horizontal
direction. In the second model, the resisting elements provide lateral stiffness and
strength in all the horizontal directions and possess a bi-axial yield domain.
Keywords Nonlinear dynamic response • Nonlinear static response • Asymmetric

buildings • Single-storey systems • Modeling of resisting elements
19.1 Introduction
A proper seismic assessment of existing structures requires the comparison between

the displacement capacity, i.e. the displacement corresponding to the achievement
of a given limit state, and the displacement demand of the structure, i.e. the
displacement caused by the ground motion. To this end, nonlinear static methods
represent a powerful tool because they explicitly evaluate the inelastic response of
structures and do not require the computational effort of nonlinear time-history
analyses.
M. Bosco • A. Ghersi (*) • E.M. Marino • P.P. Rossi

[email protected]

DOI 10.1007/978-3-319-14246-3_19
216 M. Bosco et al.
The nonlinear static method included in Eurocode 8, i.e. the N2 method pro-
posed by Fajfar and his research team (Fajfar and Gašperšič 1996; Fajfar 1999),
considers that the structure vibrates predominantly in a single mode. For this
reason, it is suitable for predicting the response of regular planar frames, particu-
larly if these structures are low-rise. The response of asymmetric structures, instead,
always sees important contributions from more than one mode of vibration and is
characterised by the simultaneous translation and rotation of the decks.
To make the nonlinear static methods reliable for the estimation of the displace-
ment demand of plan-asymmetric structures, several researchers have proposed
improvements to the standard procedure (Bento et al. 2010; Bosco et al. 2012;
Chopra and Goel 2004; Fujii 2014; Kreslin and Fajfar 2012; Peruš and Fajfar 2005).
These new procedures are called improved nonlinear static methods hereinafter.
In this chapter, the effectiveness of two improved nonlinear static methods and
that of the original N2 method is investigated. The first improved method is the
“extended N2 method” (Kreslin and Fajfar 2012). This method is based on the
assumption that the results of the elastic analysis of an asymmetric structure,
properly normalised, represent the upper bound of the torsional amplification of
the inelastic response. The second improved method is the “corrective eccentricity
method” (Bosco et al. 2012). According to this method, two nonlinear static
analyses have to be performed for each direction of the seismic action. The lateral
forces are applied to two points of the deck that are different from the centre of mass
(CM). The distances between the points where the lateral force is applied and CM are
named corrective eccentricities.
The two improved nonlinear static methods and the original N2 method are
applied here to predict the maximum dynamic response of three groups of single-
storey systems. Further, the systems of each group are schematised by means of two
different single-storey models. In the first model, a rigid deck is sustained by
resisting elements, which provide lateral stiffness and strength along only one
horizontal direction (either x- or y-direction of an orthogonal coordinate reference
system). In the second model, the resisting elements provide lateral stiffness and
strength in all the horizontal directions. In particular, an elastic-perfectly plastic
constitutive relation following the normality rule has been adopted, and interaction
phenomena arising in the nonlinear range of behaviour are accounted for by means
of an ellipse yield domain.
19.2 The Improved Non-linear Static Methods
19.2.1 The Extended N2 Method
According to the extended N2 method (Kreslin and Fajfar 2012) the inelastic
seismic response of asymmetric structures can be conservatively predicted by
adjusting the results of the original N2 method by means of correction factors
19 Improved Nonlinear Static Methods 217
derived by a standard modal response spectrum analysis. The original N2 method is

first applied to evaluate the displacement demand of the building. The pushover
analyses can be performed on either two planar models (one for each direction of
the seismic forces) or one 3-D model. The single pushover analysis is carried out
until the displacement of the centre of mass of the roof reaches the displacement
demanded by the assigned peak ground acceleration (target roof displacement).
Second, two modal response spectrum analyses of the 3-D model, one for each
of the two horizontal directions of the seismic action, are performed. While the
effects of the modes of vibration of the single modal response spectrum analysis are
combined according to the CQC rule, those of the two modal response spectrum
analyses are combined by the SRSS rule. The results of the modal response
spectrum analysis are used to define two sets of correction factors to be applied to
the results of the pushover analyses. One of the sets of correction factors refers to
the heightwise distribution of the storey drifts and is not considered here because
only single-storey systems are analysed. The second set of corrections factor refers
to in-plan distribution of the displacements. Each correction factor is calculated as
the ratio of the displacement at an arbitrary location to the displacement at the
centre of mass CM. In this chapter, the latter displacement is evaluated by the modal
response spectrum analysis performed on two planar models (one for each direction
of the seismic forces). The minimum value admitted for the correction factors is 1.0
because any favourable torsional effect arising from the elastic analysis is reduced
in the inelastic range of behaviour.
19.2.2 The Corrective Eccentricity Method
The corrective eccentricity method predicts the displacement demand of asymmetric

buildings through the envelope of the results of two nonlinear static analyses for each
direction of the seismic action. The lateral forces are applied to two points of the deck
that are generally different from the centre of mass CM. The difference between the
abscissa of the point where the force F is applied and the abscissa of CM is named
corrective eccentricity ei. The corrective eccentricities are expressed as a function of
the four parameters that mainly affect the torsional response of asymmetric structures
(Goel and Chopra 1990; Hejal and Chopra 1987; Palermo et al. 2013). These
parameters are the rigidity eccentricity er (distance between the centre of rigidity
CR and CM), the ratio Ωθ of the torsional to lateral frequencies of the corresponding
torsionally balanced system (i.e. of the system obtained by shifting CM into CR), the
strength eccentricity es (distance between the centre of strength CS and CM) and the
ratio Rμ of the elastic strength demand to the actual strength of the corresponding
planar system. This system is obtained by restraining the rotations of the deck.
The corrective eccentricities are calculated by means of analytical expressions
(Bosco et al. 2012), which have been calibrated based on the results of a
parametrical analysis. This analysis was performed on single-storey systems con-
stituted by a rigid deck sustained by resisting elements, which provide lateral
stiffness and strength along only one horizontal direction.
218 M. Bosco et al.
19.3 Analysed Systems
The analysed single-storey systems are mono-symmetric and are constituted by a

rigid deck, which is rectangular in plan and is sustained by vertical resisting
elements arranged along two orthogonal directions. The deck has dimensions,
denoted as B and L in Fig. 19.1, equal to 12.5 m and 29.5 m, respectively. The
systems are mass eccentric (MES). The mass m of the deck is equal to 1416 t and the
mass radius of gyration rm about CM is equal to 0.312 L. The systems are subdivided
into three groups. The systems belonging to a single group have the same value of
the uncoupled periods (i.e. the periods of the corresponding torsionally balanced
system) along the x- and y-directions, as reported in Table 19.1.
Further, two models are adopted to represent the systems of each group. In the
first model, which is suitable to simulate the behaviour of structures with braced
frames or shear walls, the deck is supported by four resisting elements arranged
along the x-axis of the assumed reference system (Fig. 19.1a) and by eight resisting
elements arranged along the y-axis. All the resisting elements have lateral stiffness
and strength in their plane only (uni-axial resisting elements). The force-
displacement relationship of each vertical resisting element is bilinear with no
strain hardening.
In the second model, which is suitable to simulate the behaviour of buildings
with framed structure, the resisting elements are located at the points of the deck
that are the intersection of the axes of the resisting elements adopted in the first
model (Fig. 19.1b). Each resisting element provides lateral stiffness and strength in
all the horizontal directions (bi-axial resisting elements). In particular, an elastic-
perfectly plastic constitutive relation following the normality rule has been adopted,
and interaction phenomena arising in the nonlinear range of behaviour are
accounted for by means of an elliptical yield domain.
a y b y
Side 2 (flexible)
Side 1 (stiff)
B CM
CR=G CS CM x CR=G CS x
es es
er er
L L
Fig. 19.1 Layout of systems with (a) uni-axial resisting elements; (b) bi-axial resisting elements
Table 19.1 Classification of Group

the considered systems
Resisting elements 01 02 03
Uni-axial (U-) T ¼ 0.6 s T ¼ 1.0 s T ¼ 1.4 s
Bi-axial (B-) T ¼ 0.6 s T ¼ 1.0 s T ¼ 1.4 s
For all the systems, the contribution γx of the resisting elements arranged along
the x-axis to the total torsional stiffness about CR is equal to 20 %. The systems
belonging to each group are characterised by different values of the parameter er, es,
Ωθ and Rμ. Specifically, er is in the range from 0.1 L to 0 (step 0.025 L ), es is in the
range from 0.1 L to 0.10 L (step 0.025 L ), Ωθ is in the range from 0.6 to 1.4 (step
0.05) and Rμ is in the range 1.0–6.0 (step 1.0). A further value Rμ ¼ 0.5 is considered
to obtain also systems which exhibit an elastic behaviour.
19.4 Numerical Investigation
In this section, the effectiveness of the nonlinear static methods in estimating the
maximum dynamic displacements of single-storey asymmetric systems is investi-
gated. To this end, the single-storey systems described in the previous section are
subjected to ten pairs of artificially generated ground motions and their seismic
response is determined by nonlinear dynamic analysis. The Newmark method is
used to evaluate the dynamic response of the analysed systems. The Rayleigh
formulation is considered for damping. In particular, the viscous damping ratio is
set equal to 5 % for the first and the third modes of vibration of the system. The
average of the displacements along the y-direction obtained for the ten
bi-directional ground motions is assumed as the benchmark for the nonlinear static
methods. A computer program developed by the authors is used for the numerical
analyses. This program can perform both the pushover and nonlinear dynamic
analysis of 3D single-storey systems supported by uni- or bi-axial vertical resisting
elements.
The errors committed by the nonlinear static methods (original N2 method,
extended N2 method and corrective eccentricity method) in the estimate of the
distribution of the maximum dynamic displacements are determined and analysed.
As the object of this chapter is to test and compare the effectiveness of the analysed
nonlinear static methods in predicting the in-plan distribution of the displacement
demand, the errors in the estimate of the target displacement of the centre of mass
(Bosco et al. 2009) are eliminated. To achieve this goal, the target displacement of
the centre of mass of the asymmetric system is set equal to the average of the ten
maximum (nonlinear) dynamic displacements of the corresponding planar system.
The effectiveness of the nonlinear static methods is quantified by means of the
percentage difference between the displacement evaluated by the nonlinear static
method and the corresponding value obtained by the dynamic analysis. Positive
percentage differences indicate a conservative estimate provided by the nonlinear
static method while negative values correspond to an unconservative estimate.
Three types of errors are calculated with regard to the whole system: (i) the
maximum unconservative error (indicated as max UCo error in figures), (ii) the
maximum conservative error (max Co error in figures), and (iii) the average
absolute error, i.e. the average of the absolute values of the errors committed for
all the resisting elements (AAv error in figures).
220 M. Bosco et al.
19.5 Discussion of the Results
To summarise the results obtained for the large number of structures analysed, each
of the errors defined in the previous section is processed according to a procedure
described in (Bosco et al. 2013) and reported in a compact form in Figs. 19.2 and
19.3. Grey, dark and white surfaces represent the errors committed by the original
Original N2 method Corrective eccentricity method
U -01 +100
+100 0
max UCo Err.(%)

max Co Err.(%)
AAv Err.(%)
+50 +50 -50
0 0 -100
01 01 01
1.4 23 1.4 23 1.4
23 1.2 1.2 1.2
45 0.8
1.0 45 0.8 1.0 45 0.8
1.0
6 0.6
Rμ 6 0.6
Ωθ Rμ Ωθ Rμ 6 0.6
Ωθ
U-03
+100 +100 0
max UCo Err.(%)
max Co Err.(%)
AAv Err.(%)
+50 +50 -50
0 0 -100
01 01 01
23 1.4 23 1.4 23 1.4
1.2 1.2 1.2
45 1.0 45 1.0 45 1.0
0.8 6 0.6 0.8 6 0.6 0.8
6 0.6
Rμ Ωθ Rμ Ωθ Rμ Ωθ
B-01
+100 +100 0
max UCo Err.(%)
max Co Err.(%)
AAv Err.(%)
+50 +50 -50
0 0 -100
01 01 01
23 1.4 23 1.4 23 1.4
1.2 1.2 1.2
45 1.0 45 1.0 45 1.0
0.8 0.8 6 0.6 0.8
6 0.6 6 0.6
B-03
+100 +100 0
max UCo Err.(%)
max Co Err.(%)
AAv Err.(%)
+50 +50 -50
0 0 -100
01 01 01
23 1.4 23 1.4 23 1.4
1.2 1.2 1.2
45 1.0 45 1.0 45 1.0
0.8 0.8 0.8
6 0.6 6 0.6 6 0.6
Fig. 19.2 Comparison between the errors committed by the original N2 method and the corrective
eccentricity method (systems U-01, U-03, B-01, B-03)
Corrective eccentricity method Extended N2 method
U-01
max UCo Err.(%)

+100 +100 0
max Co Err.(%)
AAv Err.(%)
+50 +50 -50
0 0 -100
01 01 01
23 1.4 23 1.4 23 1.4
1.2 1.2 1.2
45 1.0 4 1.0 4 1.0
0.8 56 0.8 56 0.8
6 0.6 0.6 0.6
U-03
max UCo Err.(%)

+100 +100 0
max Co Err.(%)
AAv Err.(%)
+50 +50 -50
0 0 -100
01 01 01
23 1.4 23 1.4 23 1.4
1.2 1.2 1.2
45 1.0 4 1.0 4 1.0
0.8 56 0.8 56 0.8
6 0.6 0.6 0.6
B-01
0
max UCo Err.(%)
+100 +100
max Co Err.(%)
AAv Err.(%)
+50 +50 -50
0 0 -100
01 01 01
1.4 1.4 23 1.4
23 1.2 23 1.2 1.2
45 1.0 4 1.0 4 1.0
0.8 56 0.8 56 0.8
0.6
Rμ 6 0.6
Ωθ Rμ 0.6
Ωθ
Rμ Ωθ
B-03
+100 +100 0
max UCo Err.(%)
max Co Err.(%)
AAv Err.(%)
+50 +50 -50
0 0 -100
01 01 01
23 1.4 23 1.4 23 1.4
1.2 1.2 1.2
45 1.0 4 1.0 4 1.0
0.8 56 0.8 56 0.8
6 0.6 0.6 0.6
Fig. 19.3 Comparison between the errors committed by the improved nonlinear static methods
(systems U-01, U-03, B-01, B-03)
N2 method, the extended N2 method and the corrective eccentricity method,

respectively. Figure 19.2 shows that the standard N2 method is effective in the
prediction of the seismic response of torsionally stiff structures (systems with Ωθ
larger than 1.2). In fact, for these systems all the considered errors are very low.
Instead, the errors committed by this method of analysis (max Co, AAv and max
UCo) are significant for all the other systems. For example, the maximum
222 M. Bosco et al.
unconservative errors committed for torsionally flexible systems are about 80 %

(Ωθ ¼ 0.6, high value of the parameter Rμ). The use of the corrective eccentricity
method enhances significantly the prediction of the displacement demand. In fact,
this method leads to small unconservative and average errors for systems with
uni-axial resisting elements. The errors are lower than those committed by the
original N2 method but still significant for systems characterised by Ωθ > 1.0, high
value of the parameter Rμ and with bi-axial resisting elements. These errors are
committed because the equations that provide the corrective eccentricities were
calibrated neglecting the interactions phenomena that (1) modify significantly the
prediction of the torsional response provided by nonlinear static analysis and that
(2) modify only slightly the dynamic response (Bosco et al. 2015). Even if not
shown in any figure, similar considerations apply to systems U-02 and B-02.
Figure 19.3 shows that the two improved nonlinear static methods are equally
effective in the prediction of the seismic response of systems with uni-axial
resisting elements even if the conservative errors provided by the extended N2
method are generally greater for torsionally stiff structures. The extended N2
method maintains its effectiveness for systems with bi-axial resisting elements. In
fact, the prediction of the displacements provided by this method is equal for
systems with uni- or bi-axial resisting elements.
19.6 Conclusions
The chapter investigates the effectiveness of two improved nonlinear static

methods, namely the extended N2 method and the corrective eccentricity method,
in predicting the seismic response of asymmetric single-storey systems. Their
effectiveness is measured by the improvement that they can provide over the
standard N2 method adopted in EC8. The investigation is conducted on a wide
set of single-storey systems with uni- and bi-axial resisting elements, which are
representative of buildings with braced frames (or shear walls) and framed struc-
ture, respectively.
The numerical investigation demonstrates that the standard N2 method is effec-
tive only in the prediction of the seismic response of very torsionally stiff structures
(systems with Ωθ larger than 1.2), but it leads to significant errors in the estimation
of the displacement demand of the stiff side for other systems. For all the analysed
systems, both the extended N2 method and the corrective eccentricity method
provide an estimation of the displacement demand much closer to that determined
by nonlinear dynamic analysis than that of the standard N2 method. When the two
improved nonlinear static methods are compared each other, the extended N2
method, which is also easier to apply, generally appears more reliable than the
corrective eccentricity method for systems with bi-axial resisting elements. Instead,
the displacement demand predicted by the corrective eccentricity method is gener-
ally more accurate to that obtained by the extended N2 method for systems with
uni-axial resisting elements.
References
Bento R, Bhatt C, Pinho R (2010) Verification of nonlinear static procedures for a 3D irregular
SPEAR building. Earthq Struct 1(2):177–195
Bosco M, Ghersi A, Marino EM (2009) On the evaluation of seismic response of structures by
nonlinear static methods. Earthq Eng Struct Dyn 38(13):1465–1482
Dyn 41(13):1751–1773
Bosco M, Ghersi A, Marino EM, Rossi PP (2013) Comparison of nonlinear static methods for the
Bosco M, Ghersi A, Marino EM, Rossi PP (2015) Influence of the interaction yield domain on
lateral-torsional coupling of asymmetric single-storey systems. In: Zbigniew Zembaty and
Mario De Stefano (Eds) Seismic Behaviour and Design of Irregular and Complex Civil
Structures II, vol 40, Chapter 18. Springer
Chopra AK, Goel RK (2004) A modal pushover analysis procedure to estimate seismic demands
for unsymmetric-plan buildings. Earthq Eng Struct Dyn 33(8):903–927
Fajfar P (1999) Capacity spectrum method based on inelastic demand spectra. Earthq Eng Struct
Dyn 28(9):979–993
Fajfar P, Gašperšič P (1996) The N2 method for the seismic damage analysis of rc buildings.
Earthq Eng Struct Dyn 25(1):31–46
under bi-directional excitation using pushover analyses. Bull Earthq Eng 12(2):909–938
Goel RK, Chopra AK (1990) Inelastic seismic response of one-story, asymmetric plan systems.
Report No. UCB/EERC-90/14. University of California at Berkeley, Berkeley
Hejal R, Chopra AK (1987) Earthquake response of torsionally-coupled buildings. Earthquake
Engineering Research Center, Report n UCB/EERC-87/20. College of Engineering, Univer-
sity of California at Berkeley
Eng 11(5): 1467–1491
Chapter 20
Influence of the Rotational Mass Inertia
on the Torsional Seismic Response
Dietlinde K€
Abstract Structures with uneven distributions of mass and stiffness (in the elastic
range of behavior) and of mass and strength (in the inelastic range of behavior) are
subjected to general torsion. These structures are called plan irregular ones. From a
static elastic point of view general torsion may be described as a rotational
movement around the center of stiffness, due to a given eccentricity (a distance
between the center of stiffness, CR and the center of mass, CM). In order to describe
more accurate the seismic behavior of plan irregular structures, at least a dynamic
elastic point of view is needed. Under dynamic seismic input rotational mass inertia
tries to overcome the rotation generated by the given eccentricity. Therefor a
rotational inertial moment appears, as counterpart for the rotation around the
CR. When moving to the nonlinear range of behavior, the rotational movement
around CM takes place as counterpart of the rotational movement around the center
of strength, CF. The authors aim to investigate how the rotational moment, due to
mass inertia, changes for different structural layouts and seismic inputs. Because
the rotational inertial moment is a measure for the elastic as well as for the inelastic
structural behavior, it turns out to be a consistent way of describing the seismic
response of plan irregular structures.
Keywords Rotational inertial moment • Plan irregularity • Inelastic behaviour •

Torsional restrained structure • Dynamic nonlinear analysis
20.1 Introduction
Several past studies concerning plan irregular structures (K€ober and Zamfirescu
2009, 2010; De La Llera and Chopra 1995; Garcia et al. 2004; Goel and Chopra
1990) showed a great sensitivity of their seismic response to the nature and intensity
of the seismic input. According to former work in this field, the displacement
demand turned out to be very different when changing the accelerogram applied
D. K€ober (*) • D. Zamfirescu

Reinforced Concrete Department, Technical University of Bucharest, B-dul Lacul Tei
no. 122-124, 020396 Bucharest, Romania

DOI 10.1007/978-3-319-14246-3_20
226 D. K€
ober and D. Zamfirescu
Fig. 20.1 Definition of

Mdynamic
for dynamic calculation. Therefore consistent trends of behaviour for plan irregular
structures are difficult to define in terms of structural displacement and rotation. A
response measure having limited variation with respect to the seismic input and to
the structural layout is needed. Former research in the field of plan irregular
structures mentioned the rotary inertia of mass as an important parameter of the
dynamic response (Paulay 2001). This study focuses on the effect of the inertial
rotational moment, a measure suitable for the elastic range of behaviour as well as
for the inelastic one, that may be described as follows (see Fig. 20.1):
Mstatic Mdynamic ¼ Mtotal
where:
Mstatic ¼ F·e – static torsional moment equal to the seismic equivalent load
(F) multiplied by the given eccentricity (e); gives a rotational movement around
the centre of stiffness, CR;
Mtotal ¼ Σ(Fyi·di + Fyj·dj) – torsional capacity defined as product of the yield forces
of the structural elements and the distance between the structural elements and
the centre of mass CM; “i” and “j” are the main directions of the layout;
Mdynamic – counterpart of the static torsional moment, due to rotational mass inertia.
A given eccentricity, e, produces the torsional moment Mstatic, which initiates a
torsional movement around CM, Mdynamic. The two torsional components are
withstand by the structural torsional capacity. Mdynamic may be reflected only by
dynamic calculation. In static calculation mass inertia is not mobilized.
The authors analyzed two different types of single-story plan irregular structures
(twist restricted as well as twist unrestricted) subjected to a natural accelerogram
and three spectrum compatible records, considering the design spectra of the
Romanian Seismic Design Code and of the Eurocode 8. Several corner periods of
the seismic input and structural eigen periods were considered. Dynamic nonlinear
20 Mass Inertia and Seismic Torsion 227
calculation was applied for earthquake intensity values with a PGA of 0,2 g and of
0,4 g). The aim of the study is to investigate if the inertial torsional moment may be
a predictable measure for the seismic response of plan irregular structures.
The study focuses especially on torsional restrained structures (having also
structural elements perpendicular to the direction of the seismic input because
they are practice suitable) and on structures with concentrated stiffness and strength
(walls as lateral resisting elements).
Results from 720 cases (considering the variation of structural stiffness, corner
period of the ground motion and seismic intensity level) were compared.
20.2 Analysed Structures
In this chapter two single story structures, having plan layouts of 11 44 m

respective 22 22 m were investigated. Especially torsional stiff (twist restrained,
TI) structures were analyzed due to the fact that most real structures have lateral
load resisting elements on both main directions. For comparison also few torsional
flexible structure (twist unrestrained, TL), were included in this study (see
Fig. 20.2). Both structures in Fig. 20.2 are idealized structures with a rigid dia-
phragm floor and columns and walls as vertical structural elements. The vertical
structural elements are disposed symmetric about the x and y axis. The structural
mass is lumped at the center of mass (CM).
The structures of Fig. 20.2 are symmetric structures, being characterized through
a coincidence between the center of stiffness (CR), the center of mass (CM) and the
center of strength (CF). The corresponding eccentric systems are obtained by
translating gradually CR and CF along the y axis, from its initial position up to
20 % of the plan dimension of the structure normal to the direction of the seismic
input.
a X b X
6
B/2
B/2
CR=CM=CF CM=CF
4 5
B
1 3
B
1 3
B/2
B/2
2 2
Y 4
Y
Fy Fy
A/2 A/2 A/2 A/2
A A
Fig. 20.2 Layout of symmetric structures: (a) torsional flexible (TL); (b) torsional stiff (TI)
228 D. K€
a
Se/ag b 3
b0= 2.5
3 2.5 x=0,05
b (T)
1.5 1.75/T
1.2 1 5.25/T 2
0.5 TC=0.7s
TB = 0.15s TD = 2.0s TB=0.14 TD=3
0
TC = 0.5s T
Perioada T, s 0 0.5 1 1.5 2 2.5 3 3.5 4
Fig. 20.3 Seismic design spectra: (a) EC8 – Tc ¼ 0.5 s; (b) Romanian Seismic Design Code –
Tc ¼ 0.7 s
The total weight (G) of the floor is 4840 kN (considering a uniform load
p ¼ 10 kN/m2). The structural walls were modeled as elastic-perfectly plastic
springs acting on x and y direction (Zamfirescu 2000).
The stiffness of the structural elements was chosen for both main directions so
that the initial translational period of the structure equals 0.7 s or 1.6 s. The stiffness
and the strength of walls P2 and P4 (for TL) and of walls P2 and P5 (for TI) remain
constant.
Structural static as well as dynamic rotations were computed.
20.3 Seismic Input
The seismic input for this study is unidirectional (along x direction) and is given by
original records as well as by spectrum compatible accelerograms, acting along the
x axis. Therefor design spectra from Eurocode 8 and the Romanian Seismic Design
Code were used. Design spectra for corner periods equal to 0.5 s and 0.7 s were
considered.
The results were obtained for two intensity levels of seismic input, scaling each
accelerogram to a PGA of 0.2 g (Ultimate Limit State, SLU) and 0.4 g (Surviv-
ability Limit State, SVLS). For each seismic intensity level the authors considered
an original record and three spectrum compatible accelerograms (see Fig. 20.3).
20.4 Case Study
The target of this case study was to identify a response measure suitable for plan
irregular structures, that has a controlled variation with respect to the seismic input
and to the structural layout. The authors interest focused on the torsional moment
Fig. 20.4 Total torsional moment versus static torsional moment: (a) TI structure; (b) TL
structure
around CM, Mdynamic (due to rotational mass inertia), as counterpart of the static
torsional moment, Mstatic (given by the static eccentricity). This choice may be
explained by the fact that the torsional moment Mdynamic is a structural response
measure suitable for the elastic as well as for the inelastic range of behaviour and
due to the fact that it may be calculated from static analysis.
Due to the fact that the torsional rotation around CM is reflected only by
dynamic calculation the authors have chosen the name of Mdynamic.
Figure 20.4 shows the variation of Mtotal with respect to Mstatic for the 11 44 m
structural layout and for the range of eccentricities considered (from 20 % up to
20 % of the plan dimension perpendicular to the direction of the seismic input in
steps of 1 %). As expected, Mdynamic is opposite to Mstatic (greater values for Mstatic
than for Mtotal) and has greater values (twice in this case) for the TL structure
compared to the TI structure.
In order to connect the computed Mdynamic to the real behaviour of the analysed
plan irregular structures, dynamic nonlinear calculation was applied and the vari-
ation of Mdynamic with respect to the maximum rotation was investigated (see
Fig. 20.5).
In Fig. 20.5 the variation of Mdynamic with respect to the maximum rotation is
shown for an original seismic input and three spectrum compatible accelerograms.
The straight red line indicates the variation of Mdynamic with the maximum rotation,
having as slope the residual torsional stiffness (computed for the TI structure taking
into account the lateral stiffness of the structural walls perpendicular to the direc-
tion of the seismic input). This estimation is sustained by the fact that the walls
parallel to the direction of the seismic input yield because of translation and do not
contribute to the structural torsional stiffness.
Graphics in Fig. 20.5 are representative for all analyzed cases and show the
following main trends:
(a) Mdynamic has a linear variation with respect to the maximum rotation;
(b) the residual torsional stiffness gives the slope of this variation (until yielding
of all structural elements is achieved);
230 D. K€
Fig. 20.5 Dynamic torsional moment versus maximum rotation; TI structure; T ¼ 0.7 s;
Tc ¼ 0.5 s: (a) 11 44 ULS; (b) 11 44 SLSV; (c) 22 22 ULS; (d) 22 22 SLSV
Fig. 20.6 Dynamic torsional moment versus maximum rotation for the 11 44 TL structure;
T ¼ 0.7 s; Tc ¼ 0.5 s
(c) remarks (a) and (b) are valid for both analyzed structural layouts, both limit
states and all seismic inputs considered.
Figure 20.6 shows an example of variation of Mdynamic with the maximum
rotation for a TL structure. Remarks made for the TI structures remain valid.
As expected, “yielding” of structural elements is reached earlier compared to the
TI structure because walls parallel to the direction of the seismic input have to
Table 20.1 Eccentricity Eccentricity [%]

values for the 11 44 layout
T [s] Tc [s] ULS SLSV
11 44 TI 0.7 0.5 8 8
0.7 8 9
1.6 0.5 15 9
0.7 14 8
11 44 TL 0.7 0.5 5 –
Table 20.2 Eccentricity Eccentricity [%]

values for the 22 22 layout
T [s] Tc [s] ULS SLSV
22 22 TI 0.7 0.5 11 8
0.7 11 9
1.6 0.5 17 12
0.7 15 11
withstand translation and torsion. Once these walls yield, the structure has very low
residual torsional stiffness (curves in Fig. 20.6 become asymptotic).
For the red line in Fig. 20.6 a residual stiffness of 5 % has been considered.
The linear variation of Mdynamic with the maximum rotation is a first
encourageable result of our study. Also the slope of this variation may be estimated
by the residual torsional stiffness. Nevertheless, in order to use those results for
current practice, the eccentricity range for which this estimation may be suitable
should be determined. After “yielding” of walls in both main directions (for TI
structures) this estimation is no more valid and explicit dynamic nonlinear analysis
should be applied.
Tables 20.1 and 20.2 show eccentricity values up to which the residual torsional
stiffness offers a good estimation of the variation of Mdynamic with the maximum
rotation. The eccentricity values are given as percentage from the plan dimension
perpendicular to the direction of the seismic input.
Values in Tables 20.1 and 20.2 show that for T Tc a suitable eccentricity value
may be 10 % irrespective of the limit state, the seismic input or the structural layout.
For T Tc the eccentricity value may be 10 % for SVLS and may rise to 15 %
for ULS.
This eccentricity limit may represent a suitable alternative to nowadays code
provisions regarding the threshold between regular and irregular structures.
232 D. K€
20.5 Concluding Remarks
This study showed that the inertial rotational moment may be a measure of plan
irregularity suitable for practical design. It may be computed by static analysis and
turned out to have a controlled variation with respect to the seismic input, the
structural layout and the seismic intensity.
The residual torsional stiffness gives a good estimation of the variation of the
inertial rotational moment with respect to the maximum structural rotation.
Further investigations need to be conducted considering also a bidirectional
seismic input.
References
De La Llera JK, Chopra AK (1995) Understanding the inelastic seismic behavior of asymmetric
plan – buildings. Earthq Eng Struct Dyn 24(4):549–572
EN 1998-1: 2004. Design of structures for earthquake resistance. General rules, seismic actions
and rules for buildings, pp 45–69
Garcia O, Islas A, Ayala AG (2004) Effect of the in-plan distribution of strength on the non-linear
seismic response of torsionally coupled buildings. In: Proceedings of the 13th WCEE, Lisbon,
24–28 September 2012, p 1891
Goel RK, Chopra AK (1990) Inelastic seismic response of one – story, asymmetric – plan systems.
College of Engineering, University of California at Berkeley. UBC/EERC – 90/14
K€ober D, Zamfirescu D (2009) Simplified methods used for evaluation of the displacement gain
due to general torsion. Sci J Math Model Civ Eng 5(2):32–51
K€ober D, Zamfirescu D (2010) Effects of general torsion on structural displacements. In:
Proceedings 14 ECEE. Technical University Bucharest. ISBN 978-608-65185-1-6
Paulay T (2001) Some design principles relevant to torsional phenomena in ductile buildings.
J Earthq Eng 5(3):273–308
P100 -1/ 2006. Cod de proiectare seismică. Prevederi de proiectare pentru clădiri, 32–61
Zamfirescu D (2000) TORSDIN – dynamic nonlinear analysis computer program. Reinforced
Concrete Department. Technical University Bucharest
Chapter 21
Seismic Response Trends of SDOF Plan
Irregular Structures. Simplified Approach
Dietlinde K€
Abstract Plan irregular structures experience a torsional seismic response even if the
seismic input is unidirectional, due to an uneven distribution of mass and stiffness
(in the elastic range of behavior) and of mass and strength (in the inelastic range of
behavior). The seismic behavior of plan irregular structures is difficult to preview due
to coupling between the translational and the rotational movement and to the change of
the importance of stiffness and strength when moving from the elastic to the inelastic
range of behavior. Past studies have also shown a high sensitivity of the seismic
response of plan irregular structures to the characteristics of the ground motion.
Therefor establishing trends of behavior for plan irregular structures is a goal very
difficult to achieve but still it is a basic requirement for defining simplified approaches
useful in current praxis. Dynamic nonlinear analysis and the simplified SESA method
(Kober D, Zamfirescu D, Effects of general torsion on structural displacements. In:
Proceedings 14 ECEE. ISBN 978-608-65185-1-6, 2010) were used for investigation.
Structural displacements, rotations and rotational moments were checked and com-
pared. The aim of the study is to establish behavior trends for plan irregular structures,
looking especially at the variation of the rotation around CM, due to mass inertia.
Keywords General torsion • Inertial torsional moment • Response scattering

dynamic nonlinear analysis • Statistic evaluation
21.1 Introduction
General torsion is a complex phenomenon that may be explained accurate by

dynamic nonlinear analysis.
Because dynamic nonlinear analysis is too complicated for current practical
design, research on general torsion focused during the last decades on defining
simplified methods (as the N2 method or the MPA method) for computing the
displacement amplification due to general torsion.
D. K€ober (*) • D. Zamfirescu

Reinforced Concrete Department, Technical University of Bucharest,
B-dul Lacul Tei no. 122-124, 020396 Bucharest, Romania

DOI 10.1007/978-3-319-14246-3_21
234 D. K€
The authors proposed a simplified method for the estimation of the effects of
general torsion on single story plan irregular structures under seismic action. The
method (called SESA) is based on superposition of modal effects but it is extended
to nonlinear behavior of the structure by using overdamped displacement response
spectra. The method can be analytically applied for the estimation of the displace-
ment amplification due to torsion (compared to translational only behavior) using
the same steps of a regular spectral analysis.
Past studies (K€ober and Zamfirescu 2009, 2010; De La Llera and Chopra 1995;
Garcia et al. 2004) showed a high sensitivity of results from simplified design
methods, with respect to the seismic intensity. Particularly, the observed variation
of the accuracy of results from the SESA method compared to dynamic nonlinear
analysis results is not monotone with the seismic intensity and differs for different
structural assemblies. In order to apply SESA in practical design, the uncertainty of
the accuracy of results should be investigated. Therefor the authors considered a
study concerning the influence of the amount of nonlinearity on the response of plan
irregular structures to be of interest.
Therefore the authors analyzed a large range of single-story plan irregular
structures (twist restricted as well as twist unrestricted) subjected to a natural
input and three spectrum compatible records, considering the design spectra of
the Romanian Seismic Design Code and of the Eurocode 8 (P100 -1/2006; EN
1998-1, 2004). Three corner periods of the ground motion (0,5 s; 0,7 s; 1,6 s) were
taken into account together with the variation of the structural stiffness (eigen
periods of 0,3 s; 0,7 s; 1,6 s) and with the variation of the earthquake intensity
(from elastic behavior until a PGA of 0,4 g). Eccentricity values (distances between
the center of mass and the stiffness center, which coincides with the resistance
center) of up to 20 % from the plan dimension perpendicular to the direction of
the seismic input were considered.
Results from 324 cases (considering the variation of structural stiffness, corner
period of the ground motion and seismic intensity level) were compared in order to
answer the following questions:
1. How does the accuracy of results (from SESA compared to DNA) change with
the seismic intensity, the structural type (TL or TI) and the corner period of the
seismic input?
2. Is the accuracy of results equal for structural displacements, rotations and the
inertial torsional moment?
21.2 Simplified Method for the Estimation of the Effects

of General Torsion (SESA)
SESA is based on the estimation of the structural response under seismic action of
an irregular single story system, by modal response spectrum analysis. In order to
take into account the inelastic behavior, the capacity spectrum method is used, by
equating the nonlinear system to an elastic one, equivalent in translation. The
21 Plan Irregularity ‐ Trends of Behaviour 235
resulting linear equivalent system is defined by the secant to maximum displace-

ment stiffness, and the viscous damping properties are set through equivalence with
the hysteretic damping properties of the initial system. The simplified method can
be used to assess the displacement amplification due to general torsion maintaining
the simplicity of the spectrum analysis (Goel and Chopra 1990).
The values of the equivalent damping ratio are set in order to obtain the same
displacement of the equivalent linear system with the nonlinear displacement of the
inelastic system.
It is important to mention that the simplified method is entirely consistent with
the assumptions used for the capacity spectrum method. The substitution of the
inelastic behavior by a translation equivalent elastic structure has shortcomings
particularly for periods lower than the corner period of the ground motion (Tc). In
order to minimize this influence the equivalent damping coefficient was determined
by trial and error, iteratively, from the computed inelastic and elastic displacement
spectra.
By applying the SESA method, results show a relatively good match to the
structural response determined by dynamic nonlinear analysis and a better estima-
tion of the structural response of irregular structures (influenced by general torsion)
than the ones that can be obtained by using code provisions, for most of the cases
(K€ober and Zamfirescu 2009).
21.3 Analysed Structures
In this chapter two single story structures (layout 11 44 m), a twist unrestrained
structure (TL) and a twist restrained structure (TI), were analysed (see Fig. 21.1).
Both are idealized structures with a rigid diaphragm floor and columns and walls as
vertical structural elements. The vertical structural elements are disposed symmet-
ric about the x and y axis. The structural mass is lumped at the center of mass (CM).
The structures of Fig. 21.1 are symmetric structures, being characterized through
a coincidence between the center of stiffness (CR), the center of mass (CM) and the
a X b X
6
B/2
B/2
CR=CM=CF CM=CF
4 5
B
1 3
B
1 3
B/2
B/2
2 2
Y 4
Y
Fy Fy
A/2 A/2 A/2 A/2
A A
Fig. 21.1 Layout of symmetric structures: (a) twist unrestrained (TL); (b) twist restrained (TI)
236 D. K€
center of resistance (CF). The corresponding eccentric systems are obtained by

translating gradually CR and CF along the y axis, from its initial position up to
20 % of the plan dimension of the structure normal to the direction of the seismic
input.
The total weight (G) of the floor is 4840 kN (considering a uniform load
p ¼ 10 kN/m2). The structural walls were modeled as elastic-perfectly plastic
springs acting on x and y direction.
The stiffness of the structural elements was chosen for both main directions so
that the initial translational period of the structure equals 0.3 s, 0.7 s or 1.6 s. The
stiffness and the strength of walls P2 and P4 (for TL) and of walls P2 and P5 (for TI)
remain constant.
21.4 Seismic Input
The seismic input for this study is unidirectional (along x direction) and is given by
original records as well as by spectrum compatible accelerograms, acting along the
x axis. Therefor design spectra from Eurocode 8 and the Romanian Seismic Design
Code were used. Design spectra for corner periods equal to 0.5 s, 0.7 s and 1.6 s
were considered.
The results were obtained for elastic behavior (Serviceability Limit State, SLS)
and four intensity levels of seismic input for the inelastic behavior. Therefor each
accelerogram was scaled for four levels of strength: 0.1 g, 0.2 g (Ultimate Limit
State, SLU), 0.3 g and 0.4 g (Survivability Limit State, SLSV). For each seismic
intensity level the authors considered an original record and three spectrum com-
patible accelerograms.
21.5 Comparative Study
The target of this comparative study was to identify how well the SESA method
based on modal analysis and overdamped response spectra can estimate the seismic
response obtained by dynamic nonlinear calculation (using the Torsdin program),
for different seismic intensity levels. The results of the SESA method were com-
pared to the ones obtained by three-dimensional dynamic analysis (Zamfirescu
2000) in terms of displacement values at characteristic points of the structure
(total displacement of the center of mass (uCMx ), structural rotation (θ), displace-
ments of walls P1 and P3 (ux1, ux3)) and of inertial torsional moments (Fig. 21.2).
The inertial torsional moment was computed as difference between the static
moment and the torsional capacity (torsional moment for yielding of structural
elements).
Fig. 21.2 Definition of

Mdynamic
Results in terms of displacement values at characteristic points of the structure

were gathered with respect to the seismic intensity level. Results in terms of inertial
torsional moments were gathered with respect to the structural layout.
21.5.1 Structural Displacements and Rotations Gathered

with Respect to the Intensity Level of the Seismic
Input
For SLS, the SESA method mostly overestimates dynamic nonlinear analysis
results. As expected, results are better for uCMx and θ as for ux1 and ux3.
Figures 21.3 and 21.4 show results for the nonlinear range of behavior, sepa-
rately for torsional unrestrained (TL) and for torsional restrained (TI) structures.
The horizontal axes shows whether the SESA method overestimates (graphics right
side, positive percentage) or underestimates (graphics left side, negative percent-
age) dynamic nonlinear analysis results. The vertical axes represents the percentage
of results (from the total number of results) that fit into a range of accuracy.
For TL it was considered that the torsional stiffness of all walls is affected in the
same way as their translational stiffness. For TI, the preliminary results showed that
the walls situated perpendicular to the direction of seismic input yield also, and the
consideration of their full lateral stiffness to the rotational stiffness of the structure
leads to unconservative results. Consequently, for the comparative study the per-
pendicular walls participate with half of their lateral stiffness to the rotational
stiffness of the structure to take into account the yielding effect.
For TL structures accuracy seems to drop with the rising of seismic intensity,
although it is not a monotone variation. Structural displacements are estimated far
better by the SESA method than structural rotations. Due to the fact that for
238 D. K€
Fig. 21.3 Results for TL nonlinear behavior: (a) displacements of the mass center; (b) structural
rotations; (c) displacements of walls P1 and P3
Fig. 21.4 Results for TI nonlinear behavior: (a) displacements of the mass center; (b) structural
rotations; (c) displacements of walls P1 and P3
240 D. K€
practical design usually displacements are needed, the rotations loss of accuracy is
not inconvenient.
According to the Eurocode 8, the TL structures analyzed in this chapter expe-
rience torsional sensitivity for eccentricity values greater than 12 % of the plan
dimension perpendicular to the direction of seismic input. By restraining the
statistic evaluation of results to eccentricities up to 12 %, structural displacements
are overestimated by less than 10 % by the SESA method in an amount of 85 % for
0.1 g seismic intensity and 65 % for 0.2 g, 0.3 g and 0.4 g seismic intensity.
Accuracy of results is better for TI structures due to the positive influence of the
structural walls perpendicular to the direction of seismic input. The amount of
overestimated results is greater than for TL structures. For 0.1 g seismic intensity
and eccentricity values up to 20 %, up to 70 % of the structural displacements are
overestimated by less than 10 % by the SESA method. This percentage changes into
50 % for 0.2 g, 0.3 g and 0.4 g seismic intensity.
For this comparative study constant accuracy is obtained for seismic intensities
greater than 0.1 g. This may be explained by the constancy of the structural
response for the entire eccentricity range considered when getting more and more
into the nonlinear range of behavior.
21.5.2 Inertial Torsional Moment for TL and TI
Following former research (Paulay 2001) the authors investigated the effect of the
inertial rotational moment, a measure suitable for the elastic range of behaviour as
well as for the inelastic one (Fig. 21.5).
Fig. 21.5 Accuracy of Mdynamic

The relation between Mdynamic and the maximum rotation showed complete
yielding of all walls. For TL structures a residual stiffness of walls (after yielding)
between 5 and 25 % turned out to be realistic.
By using Mdynamic as behaviour measure, results show a much better accuracy
than for displacements at different points of the structure. All results are estimated
by less than 5 % for TI structures and 75 % of results are estimated by less than
10 % for the TL structure.
The amount of nonlinearity (defined in this chapter as seismic intensity level)

influences the accuracy of results from the SESA method for plan irregular struc-
tures, compared to dynamic nonlinear analysis results.
This study gives following answers to the questions in Chap. 1:
1. The accuracy of structural displacements and rotations (from SESA compared to
DNA) doesn’t have a monotonic variation with respect to the seismic intensity,
is better for TI structures compared to TL structures and for T > Tc.
For strong nonlinear behavior structural rotations remain nearly constant and
the structural response tends to an upper limit.
Structural displacements are estimated far better by the SESA method than
structural rotations. For practical design usually displacements are needed, so
the rotations loss of accuracy is not inconvenient.
2. The inertial torsional moment turned out to be the only behavior measure with
controlled variability with respect to the seismic intensity, the structural layout
or the corner period of the seismic input.
The results obtained in this and former studies make the authors confident in
proposing the SESA method for practical design, as a simplified, analytical appli-
cable design method for single story plan irregular structures.
References
De La Llera JK, Chopra AK (1995) Understanding the inelastic seismic behavior of asymmetric
plan – buildings. Earthq Eng Struct Dyn 24(4):549–572
EN 1998-1 (2004) Design of structures for earthquake resistance. General rules, seismic actions
and rules for buildings, 45–69
P100 -1/2006. Cod de proiectare seismică. Prevederi de proiectare pentru clădiri, 32–61
Garcia O, Islas A, Ayala AG (2004) Effect of the in-plan distribution of strength on the non-linear
seismic response of torsionally coupled buildings. In: Proceedings of the 13th WCEE, Lisbon,
24–28 September 2012, p 1891
Goel RK, Chopra AK (1990) Inelastic seismic response of one – story, asymmetric – plan systems.
College of Engineering, University of California at Berkeley. UBC/EERC – 90/14
242 D. K€
K€ober D, Zamfirescu D (2009) Simplified methods used for evaluation of the displacement gain
due to general torsion. Sci J Math Model Civ Eng 5(2):32–51
K€ober D, Zamfirescu D (2010) Effects of general torsion on structural displacements. In: Proceedings
14 ECEE. Technical University Bucharest. ISBN 978-608-65185-1-6
Paulay T (2001) Some design principles relevant to torsional phenomena in ductile buildings.
J Earthq Eng 5(3):273–308
Zamfirescu D (2000) TORSDIN – dynamic nonlinear analysis computer program. Reinforced
Concrete Department. Technical University Bucharest
Chapter 22
Maximum Corner Displacement
Amplifications for Inelastic One-Storey
In-Plan Asymmetric Systems Under
Seismic Excitation
Michele Palermo, Stefano Silvestri, Giada Gasparini,

and Tomaso Trombetti
Abstract The seismic behaviour of one-storey asymmetric structures has been

extensively studied since 1970s by a number of research studies which identified
the coupled nature of the translational-to-torsional response of this class of systems
leading to severe displacement magnifications at the perimeter frames and to
significant increase of local peak seismic demand to the structural elements. In
previous research works, some of the authors introduced a simplified method, the so
called “Alpha” method, for the estimation of the maximum torsional response of
linear-elastic one-storey in-plan asymmetric systems. The present paper provides
an extension of the “Alpha” method to inelastic systems. The main objective is to
evaluate how the excursion of the structural elements in the inelastic range affects
the seismic displacement demand. In detail, 3-D surfaces of corner displacement
magnification factors for different force reduction factors are showed and compar-
isons with the results obtained from linear analysis are provided. The 3-D surfaces
could be useful to assess the seismic displacement demand of asymmetric systems
in a preliminary design phase.
Keywords One-storey asymmetric systems • Corner displacement amplifications •

Elastic seismic response • Inelastic seismic response • Period shifting
M. Palermo (*) • S. Silvestri • G. Gasparini • T. Trombetti

Department DICAM, University of Bologna, Viale Risorgimento 2, 40136 Bologna, Italy
[email protected]

DOI 10.1007/978-3-319-14246-3_22
244 M. Palermo et al.
22.1 Introduction
Since the late 1970s it is well known that in-plan asymmetric systems, when
subjected to dynamic excitation, develop a coupled lateral-torsional response that
may considerably increase their local peak response, such as the corner displace-
ments (Kan and Chopra 1987; Hejal and Chopra 1987; Rutenberg 1992).
In order to effectively apply the performance-based design approach to seismic
design there is a growing need for code-oriented methodologies aimed at predicting
deformation parameters. Thus, the estimation of the displacement demand at
different locations, especially for eccentric structures, appears a fundamental
issue. Furthermore, the ability to predict the torsional response of eccentric systems
can be also useful to improve the capability of one of the most actually used seismic
design approaches (i.e. push-over analysis, Perus and Fajfar (2005)).
Since the early 1990s, Nagarajaiah et al. (1993) observed that, for the specific
class of torsionally-stiff asymmetric structures, the maximum center mass displace-
ment can be well approximated by the maximum displacement of the equivalent
not-eccentric system.
In previous research works, some of the authors identified a structural parameter,
called “Alpha”, related to the attitude of one-storey asymmetric linear systems to
develop a rotational response in free vibration, and proposed a simplified proce-
dure, called “Alpha” method, for the estimation of the maximum torsional response
(Trombetti and Conte 2005; Trombetti et al. 2008). In its original formulation,
based on the assumption of equal maximum center mass displacement response
between the eccentric system and its not-eccentric counterpart, the “Alpha” method
has been also used for the prediction of maximum corner side displacements. The
assumption has been proven valid for torsionally-stiff systems only. More recently,
the “Alpha” method has been extended to all classes of one-storey asymmetric
systems, thus including both torsionally-stiff and torsionally-flexible systems,
thanks to the introduction of some correction coefficients (Palermo et al. 2013).
In detail, a coefficient accounting for the so-called “period shifting effect”, leading
to an increase in the center mass displacement between the eccentric system and its
not-eccentric counterpart, has been introduced (Palermo et al. 2013).
In this paper a further generalization of the “Alpha” method is proposed by
considering non-linear asymmetric systems. First, linear-elastic analyses are
performed varying the damping ratio ξ from 0.02 to 0.30. Then, non-linear seismic
analyses are performed by assuming an elastic-perfectly plastic response of the
structural elements for various force reduction factors R. The main objective is to
evaluate how the excursion of the structural elements in the inelastic field affects
the displacement demand of one-storey in-plan asymmetric structures.
22 Max Displacements for Asymmetric Systems 245
22.2 Problem Formulation and Summary

of Previous Results
Let us consider the one-storey eccentric structure (i.e. a system characterized by

non-coincident center of mass, CM, and center of stiffness, CK, leading to a
one-way eccentricity Ex ¼ E, Ey ¼ 0) of Fig. 22.1 (the origin of the reference
system is located at CM). It is assumed that the diaphragm is infinitely rigid in its
own plane, and that the lateral-resisting elements are massless and axially
inextensible. The self torsional stiffness (kθ) of each lateral-resisting element is
neglected. Under these assumptions, the system can be modelled as a 3-dof system:
longitudinal center mass displacement, uy,CM,; transversal center mass displace-
ment, ux,CM; center mass rotation, uθ,CM (coincident with the floor rotation, uθ). The
case of one-way dynamic excitation (e.g. free vibrations or seismic input) along the
longitudinal direction (namely, the y-direction) is treated in the next sections.
The longitudinal corner side displacements, i.e. the displacement of the so called
stiff side uy,s (e.g. point B or B0 , the closer to CK) and of the flexible side uy,f,
(e.g. point A or A0 , the farther from CK), at any instant of time t, are given by:
L
u y, f ðtÞ ¼ u y, CM ðtÞ uθ ðtÞ
2 ð22:1Þ
L
u y, s ðtÞ ¼ u y, CM ðtÞ þ uθ ðtÞ
2
Estimating corner displacements under seismic excitation according to Eq. 22.1

needs time-history analyses to be performed. Nonetheless, the practical engineer is
Fig. 22.1 Plan view of the studied system

mainly interested in the peak response, which, generally speaking, can be expressed
as an appropriate combination of the longitudinal and rotational maximum
responses. In a recent research work, the following expression for the estimation
of the maximum corner side displacements, uy, corner, max, has been introduced,
leading to the so called “Alpha” method:
u y, corner, max ¼ δ u y, CM, max, NE ð1 þ B A αu ϕÞ ð22:2Þ
The symbols have the following meaning:

u y, CM, max
• δ¼ is the translational contribution providing the center mass
u y, CM, max, NE
displacement amplification with respect to that of the equivalent not-eccentric
system (N-E), also referred to as “period-shifting effect”.
uθ, max
• A αu ¼ ρ gives the rotational contribution (αu is a parameter given in
u y, CM, max
closed-form by Trombetti and Conte (2005) for the case of undamped free
vibration response; ρ is the mass radius of gyration of the system, A is a
parameter which accounts for the seismic nature of the input and has to be
calibrated through numerical simulations).
• B is the correlation coefficient between the maximum displacement response and
the maximum rotational response.
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2 h 2 i
• ϕ ¼ 3 Lx =L y = 1 þ Lx =L y is a geometric shape factor.
A closed-form estimation of δ has been provided by Palermo et al. (2013) by

making use of the analytical expression of the natural periods and of a simplified
response spectrum. It has been noted that torsionally-flexible systems are poten-
tially prone to high center mass displacement amplifications which are caused by
the shift of the fundamental period (the so-called “period shifting effect”).
Since the late 1990s the ratio between the maximum rotational response and the
maximum translational response in undamped free vibrations, the so-called
“Alpha” parameter, has been derived in closed-form as a function of the normalized
eccentricity (e ¼ E/De, where De is the equivalent diagonal of the floor plan equal to
pffiffiffiffiffi
12 ρ) and torsional-to-lateral frequency ratio Ωθ (Goel and Chopra 1991) giving
the name to the “Alpha” method (Trombetti and Conte 2005):
pffiffiffi
ρuθ, max 4e 3
αu ¼ ¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2 ffi ð22:3Þ
u y, max
Ω2θ þ 12e2 1 þ 48e2
The correlation coefficient Bu (subscript u standing for “undamped”) is defined such

as:
L
u y, flex ¼ u y, max þ Bu uθ, max ð22:4Þ
2
leading to the following closed form expressions derived in the case of undamped
free vibration:
4A3
Bu ¼ 1 ρ for A1 A3
αu Lx m ð22:5Þ
4A1
Bu ¼ 1 ρ for A1 < A3
αu Lx m
Note that 1 B 1 α2ρ

u Lx
. A1 and A3 are coefficients depending on e and Ωθ. See
Trombetti and Conte (2005) for further details.
22.3 Linear Seismic Response
An ensemble of 50 ground motions selected from the PEER database has been used
to perform the seismic analyses. The ground motions are selected with shear wave
velocity Vs,30 in the range of 360–800 m/s (i.e. soil type B according to EC8 2002).
The following response parameters are evaluated and analyzed (subscript eq stands
for earthquake):
uθmax
αd, eq ¼ ρ ð22:6Þ
u y, CM max
u y, u y, CM, max
Bd, eq ¼ max ð22:7Þ
uθ, max ðLx =2Þ
u y, s , max
MCM, s ¼ ð22:8Þ
u y, CM , max
u y, f , max
MCM, f ¼ ð22:9Þ
u y, CM, max
MCM,s and MCM,f are the magnification factors (with respect to the center mass
displacement) at the stiff side and flexible side, respectively. Figures 22.2 and 22.3
show 3-D surfaces of selected response quantities. The following observations
arise:
• Peak values of αd,eq are larger than 1.0 for small values of TL and Ωθ and for
e around 0.2. The results indicate that the closed-form given in free vibration
(αu) is not always an upper bound. However, once the longitudinal period TL
increases, the peaks decrease and even vanish for TL larger than 1.5 s.
• For fixed values of TL, values of αd,eq decrease as the damping ratio ξ increases.
• Torsionally-stiff systems are more sensitive to damping ratios than torsionally-
flexible systems.
Fig. 22.2 “Alpha” rotational parameter as function of e and Ωθ. (a) TL ¼ 0.5 s; (b) TL ¼ 3.0 s
• Values of MCM,s are between 0.4 and 1.8. The MCM,s surfaces are quite smooth
for torsionally-stiff systems (values are less than 1.0), while they rapidly vary for
torsionally-flexible systems (values are also larger than 1.0).
• For fixed values of TL and ξ, the magnification factor at the stiff side MCM,s
decreases once the eccentricity increases, except within an area characterized by
small values of e and Ωθ (high torsionally-flexible systems), where it exhibits
peaks larger than 1.0 (maximum values are around 1.8). As the damping ratio
increases, this area reduces.
Fig. 22.3 Magnification factors of the stiff and flexible side as function of e and Ωθ. (a) TL ¼ 0.5 s;
(b) TL ¼ 3.0 s
• Values of MCM,f are between 0.9 and 2.2. MCM,f surfaces appear quite smooth
within the entire domain.
• For fixed values of ξ and TL, MCM,f increases with the eccentricity e, while the
peak decreases as TL increases.
For the sake of conciseness, the plots of the correlation coefficients Bd,eq are not
here displayed and discussed.
22.4 Nonlinear Seismic Response
The simple 3-dof system idealization adopted in the linear analyses cannot be used
to study the non-linear response of planar asymmetric systems since, within the
inelastic field, also the location of each resisting element and the number of
resisting elements may affect the response. For this reason, the specific, but still
simple, non-linear asymmetric model as proposed by Goel and Chopra (1991) is
used to perform the non-linear analyses. It is clear that, in order to extend the results
to realistic structures, a wide parametric analysis carried out on several idealised
models is needed. The system consists of a roof diaphragm, assumed to be rigid in
its own plane, supported by three frames, namely A, B and C (see Fig. 22.4).
Frame A is oriented along the y-direction, at a distance E from the center of
mass (CM). Frames B and C are oriented along the x-direction, located at the same
distance D/2 from the CM. Frames B and C are assumed to have the same lateral
stiffness (k/2) so that the system is not eccentric along the y-direction. Frame A is
assumed to have a lateral stiffness equal to k. Along the x-direction the
eccentricity is equal to E. The rigid motion of the roof can be described by the
three degrees of freedom defined at the CM of the slab: displacements ux (in the
x-direction) and uy (in the y-direction), and torsional rotation uθ (about the vertical
axis). The equations of motion of this specific system in the initial elastic range
are given by:
Fig. 22.4 Idealized non linear eccentric system used to perform non-linear analyses
2 32 3 2 3
m 0 0 ú€x ðtÞ ux ðtÞ
40 m 0 54 ú€y ðtÞ 5 þ ½C4 u y ðtÞ 5
0 0 m ρú€θ ðtÞ ρuθ ðtÞ
2 3
2 32 3 px ðtÞ
1 0 0
pffiffiffiffiffi u ðtÞ 6 7
54 u y ðtÞ 5 ¼ 6 p y ðtÞ 7
x
þ mω2L 4 0 pffiffiffiffiffi
1 12e 4 5 ð22:10Þ
pθ ðtÞ
0 12e 12e2 þ d =4
2
ρuθ ðtÞ
ρ
where: m is the mass of the system; [C] is the damping matrix (classical damping is
assumed); d ¼ D/ρ. Values of parameter Ωθ distinguish in-plan asymmetric systems
into: (i) torsionally-stiff systems: Ωθ 1.0; (ii) torsionally-flexible systems: Ωθ <
1.0.
Each frame is characterized by an elastic-perfectly plastic response. The yield
strength Fy is obtained by imposing a force reduction factor R between 2 and
5. Non-linear time-history analyses have been developed using the same ground
motion ensemble used in the linear analyses. The response parameters computed
from the non-linear analysis are indicated with subscript NL and represented in
Figs. 22.5 and 22.6 through 3-D surfaces. The following observations arise:
• The αd,NL,eq surfaces are qualitatively similar to those observed in the linear case.
The largest values of αd,NL,eq are observed for small values of TL and Ωθ and for
e around 0.2. Again, peaks larger than 1.0 are observed (i.e. αu is not always an
upper bound estimation).
• As the longitudinal period TL increases, the peaks decrease and they even vanish
for TL ¼ 1.5 s.
• For a fixed longitudinal period TL, the maximum rotational response tends to
decrease as R increases. The dependence of αd,NL,eq on R is larger for the class of
torsionally-stiff structures.
• The surfaces of both MCM,NL,s and MCM,NL,f are qualitatively similar to those
observed in the linear case. Peak values are slightly reduced.
• MCM,NL,s appears quite smooth for torsionally-stiff systems (values are less than
1.0), while it rapidly varies for torsionally-flexible systems (values are also
larger than 1.0).
• For fixed values of TL, the peaks of MCM,NL,s gradually decrease as R increases.
On the contrary, for torsionally-stiff systems, MCM,NL,s tends to slightly increase
(the surface becomes almost flat for R ¼ 5 with values between 0.8 and 1.0).
• The magnification factor MCM,NL,f increases as the eccentricity increases. The
peaks are observed for high torsionally-flexible systems and eccentricities e
around 0.2.
• For fixed values of R, the peaks of MCM,NL,f decrease as TL increases and the
surface tends to flatten out.
For the sake of conciseness, the graphs of Bd,NL,eq are not here displayed and
discussed.
Fig. 22.5 “Alpha” rotational parameter as function of e and Ωθ. (a) TL ¼ 0.5 s; (b) TL ¼ 3.0 s
22.5 Conclusions
This paper provides a further insight into the non-linear coupled lateral-torsional
dynamic response of one-storey asymmetric systems through the development of a
systematic parametric analysis aimed at exploiting the influence of the fundamental
system parameters (e, Ωθ, TL, ξ, R) on the seismic displacement demand. In detail,
3-D surfaces of response parameters such as the normalized maximum rotational
Fig. 22.6 Magnification factors of the stiff and flexible side as function of e and Ωθ. (a) TL ¼ 0.5 s;
(b) TL ¼ 3.0 s
response and the magnification factors of the maximum stiff and flexible side
displacement have been obtained for both the linear and the non-linear case.
These 3-D response surfaces could be used in the preliminary structural design
phase in order to estimate the seismic displacement demand as a function of the
normalized eccentricity and the torsional-to-lateral frequency ratio.
Acknowledgments The authors would like to thank former graduate students Elisa Sammarco
and Susanna Casacci for their valuable contributions in the development of the numerical analyses.
References
Eurocode 8 (2002) Design of structures for earthquake resistance. prEN 1998-1

Goel RK, Chopra AK (1991) Inelastic seismic response of one‐storey, asymmetric‐plan systems:
effects of system parameters and yielding. Earthq Eng Struct Dyn 20(3):201–222
Hejal R, Chopra AK (1987) Earthquake response of torsionally coupled buildings. Report
UCB/EERC 87/20, Earthquake Engineering Research Center, University of California,
Berkeley, CA
Kan CL, Chopra AK (1987) Linear and non linear earthquake responses of simply torsionally
coupled systems. Earthquake Engineering Research centre, University of California,
Berkeley, CA
Nagarajaiah S, Reinhorn AM, Constantinou MC (1993) Torsion in base isolated structures with
elastomeric isolation systems. J Struct Eng ASCE 119:2932–2951
Eng 11(5):1467–1491
Perus I, Fajfar P (2005) On the inelastic torsional response of single-storey structures under
bi-axial excitation. Earthq Eng Struct Dyn 34:931–941
Rutenberg A (1992) Nonlinear response of asymmetric building structures and seismic codes: a
state of the Art. Eur Earthq Eng VI(2):3–19
Trombetti T, Conte JP (2005) New insight into and simplified approach to seismic analysis of
torsionally coupled one-storey, elastic systems. J Sound Vib 286:265–312
Trombetti T, Silvestri S, Gasparini G, Pintucchi B, De Stefano M (2008) Numerical verification of
the effectiveness of the “Alpha” method for the estimation of the maximum rotational elastic
response of eccentric systems. Eur Earthq Eng 12(2):249–280
Chapter 23
Earthquake-Induced Pounding Between
Asymmetric Steel Buildings
Barbara Sołtysik and Robert Jankowski
Abstract Earthquake-induced pounding between buildings has been the subject of

numerous numerical and experimental studies in the resent years. The phenomenon
may cause severe damage to structural elements as well as may also lead to the total
collapse of colliding structures. A major reason leading to pounding between
adjacent, insufficiently separated buildings results from the differences in their
dynamic properties. The aim of this paper is to show the results of the numerical
analysis focuses on pounding between two L-shaped asymmetric steel buildings
under earthquake excitation. In order to identify the dynamic characteristics of
analyzed structures, the modal analysis has been first conducted. Then, the detailed
dynamic analysis of interacting structures under earthquake excitation has been
performed. The three components of the El Centro earthquake have been used in the
study. The results of numerical analysis indicate that the earthquake-induced
collisions between two asymmetric steel structures may substantially influence
their behaviour leading to both increase and decrease in the response. The results
also indicate that torsional vibrations (due to eccentric pounding) play an important
role in the overall pounding-involved response of asymmetric steel buildings under
earthquake excitations.
Keywords Structural pounding • Earthquakes • Steel buildings • Asymmetric

structures • Numerical analysis
23.1 Introduction
Due to lack of space and high cost of land in the urban areas, buildings are often
constructed with very small in-between separation gap. That situation may result in
structural interactions during seismic excitations, known as the earthquake-induced
structural pounding. Collisions between adjacent, insufficiently separated buildings
have been repeatedly observed during moderate and strong ground motions (Kasai
B. Sołtysik (*) • R. Jankowski

Faculty of Civil and Environmental Engineering, Gdańsk University of Technology, ul.
Narutowicza 11/12, 80-233 Gdańsk, Poland

DOI 10.1007/978-3-319-14246-3_23
256 B. Sołtysik and R. Jankowski
and Maison 1997). For example, significant pounding damage on the parts of school
buildings was observed after the Athens earthquake of September 7, 1999
(Vasiliadis and Elenas 2002). Also, the San Fernando earthquake in 1971 caused
collisions between the main building of the Olive View Hospital and one of its
independently standing stairway towers resulting in its permanent tilting (Bertero
and Collins 1973). During the Kocaeli (Izmit) earthquake (17.08.1999), inadequate
separation gap between buildings resulted in serious damage at the places of
interactions.
The main reason of structural pounding between buildings is related to the
difference in mass or stiffness of colliding structures (Anagnostopoulos 1988;
Mahmoud et al. 2012; Mahmoud and Jankowski 2009). The difference in the
natural vibration periods of neighbouring structures results in their out-of-phase
vibrations. The phenomenon of earthquake-induced pounding between buildings
has recently been intensively studied using various structural models and different
models of collisions (see, for example, Anagnostopoulos and Karamaneas 2008;
Anagnostopoulos and Spiliopoulos 1992; Jankowski 2005, 2007; Karayannis and
Favvata 2005; Komodromos et al. 2007; Mahmoud et al. 2013; Mahmoud and
Jankowski 2011; Maison and Kasai 1992). However, the analyses for asymmetric
structures, which induce eccentric collisions, are very limited (see Leibovich
et al. 1996). Also, the scientific interest was mainly focused on pounding-involved
response of reinforced concrete buildings rather then on steel structures (Sołtysik
and Jankowski 2013). Meanwhile, because of the lower stiffness and damping
properties, steel structures usually experience vibrations with larger amplitudes
which make then more vulnerable to structural pounding.
Therefore, the aim of this paper is to show the results of the numerical analysis
dealing with structural pounding between two L-shaped asymmetric steel buildings
under seismic excitation. In the first part of the work, the dynamic properties of both
structures have been identified. Afterwards, the detailed dynamic analysis on
earthquake-induced collisions between asymmetric steel structures has been
conducted.
23.2 Numerical Model of Colliding Buildings
The research described in this paper has been focused on the pounding-involved
response of two asymmetric steel buildings with different number of storeys. The
3-storey building (modelled by 10560 shell elements), as well as the 4-storey
building (modelled by 14064 shell elements) with different dynamic properties
have been considered in the analysis. The Finite Element (FE) model of both
structures is presented in Fig. 23.1. All structural members have been modelled
by four-node quadrilateral elements. Steel columns have been rigidly fixed to the
ground and the soil-structure interaction has not been taken into account. Pounding
between structures has been modelled using six three-dimensional gap-friction
elements (two for each storey). These elements, placed between the corner nodes
23 Pounding of Asymmetric Steel Buildings 257
Fig. 23.1 FE model of two colliding asymmetric steel buildings
of the buildings at all storeys, assure frictional and gapping connection. When
contact is detected, the nodes become fixed in the longitudinal direction and friction
forces are imposed in the transverse and vertical directions. The initial separation
gap between structures equal to 2 cm has been considered. The friction coefficient
of 0.5 has been applied in the analysis. The numerical model has been implemented
using the commercial software MSC Marc 2008.
23.3 Modal Analysis
The modal analysis has been first conducted in order to identify the dynamic
properties of each asymmetric steel building. The results of the analysis showing
the first three natural vibration modes of both structures are presented in Fig. 23.2.
The corresponding natural frequencies for the modes of free vibrations are also
summarized in Table 23.1.
23.4 Dynamic Analysis
The detailed dynamic analysis, focused on the response of interacting structures

under seismic excitation, has been carried out in the second part of the study. The El
Centro earthquake (18.05.1940) records have been used in the numerical analysis.
The NS, EW and UD components of the ground motion have been applied in the
longitudinal (Y), transverse (X) and vertical (Z) direction, respectively. The exam-
ples of the results of the analysis are shown in Figs. 23.3, 23.4, 23.5, and 23.6. In
particular, Fig. 23.3 presents a comparison between the displacement time histories
in the longitudinal direction for the 3-storey building (node no. 31650 at the corner
of the third storey) with and without pounding (large separation gap); whereas
Fig. 23.4 shows the corresponding results for the 4-storey building (node
no. 27689 at the corner of the third storey). Furthermore, Figs. 23.5 and 23.6
Fig. 23.2 Modes of free vibrations for 3-storey and 4-storey asymmetric steel buildings. (a) 1st
modes of free vibrations (transverse direction). (b) 2nd modes of free vibrations (longitudinal
direction). (c) 3rd modes of free vibrations (torsional vibrations)
Table 23.1 Natural Transverse Longitudinal Torsional

frequencies for free vibration
3-storey building 5.526 Hz 3.002 Hz 3.309 Hz
modes
4-storey building 4.708 Hz 1.973 Hz 2.205 Hz
show the comparison between the pounding-involved and independent vibration

time histories of the rotation angle at the top floor of the 3-storey and 4-storey
building, respectively.
Figures 23.3 and 23.4 clearly show that structural pounding may substantially
influence the response of two asymmetric steel buildings in the longitudinal direc-
tion under earthquake excitation. It can be seen from Fig. 23.3 that the peak
response of the 3-storey structure increases due to collisions by 3.3 %. Also,
Fig. 23.4 indicates that collisions between two asymmetric steel structures lead to
the increase in the peak displacement of the 4-storey building, and this increase is as
large as 14.8 %.
0.03
0.02
Displacement [m]
0.01
-0.01
-0.02
3-storey building with no pounding
3-storey building with pounding 0.029 0.021
-0.03
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
Time [s]
Fig. 23.3 Displacement time histories in the longitudinal direction for pounding-involved and
independent vibration responses of 3-storey building
0.06
0.04
Displacement [m]
0.02
-0.02
-0.04 4-storey building with no pounding

4-storey building with pounding
0.051 0.059
-0.06
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
Time [s]
Fig. 23.4 Displacement time histories in the longitudinal direction for pounding-involved and
independent vibration responses of 4-storey building
Figures 23.5 and 23.6 indicate that torsional vibrations of both buildings are
important components of the overall structural responses, apart from the effect of
pounding. They are induced during the earthquake as the result of structural
eccentricity caused by asymmetric plan of the structures. The results indicate,
however, that structural pounding may significantly influence the torsional behav-
iour of both structures and, what is more, it may play both positive and negative
role. In the case of the 3-storey building, the increase in the peak value of the
rotation angle at the top floor is equal to 13.9 % (see Fig. 23.5). On the other hand,
Fig. 23.6 shows that the torsional response of the 4-storey building significantly
decreases due to pounding. The change in the peak value of the rotation angle at the
top floor of this structure during the earthquake is as large as 45.7 %.
0.2
0.15
Angle of rotation [cm/m]
0.1
0.05
-0.05
-0.1

3-storey building with pounding 0.180 0.158
-0.2
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
Time [s]
Fig. 23.5 Pounding-involved and independent vibration time histories of the rotation angle at the
top floor of the 3-storey building
0.4
0.3
0.191
Angle of rotation [cm/m]
0.2
0.1
-0.1
-0.2

4-storey building with pounding
0.350
-0.4
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
Time [s]
Fig. 23.6 Pounding-involved and independent vibration time histories of the rotation angle at the
top floor of the 4-storey building
The detailed numerical analysis, focused on pounding between two L-shaped asym-
metric steel buildings under seismic excitations, has been considered in this paper.
The results of the study clearly indicate that earthquake-induced pounding may
substantially influence the response of both asymmetric structures, especially in the
longitudinal and torsional directions. In the case of the analysis conducted, it resulted
in the considerable increase in the displacements of both structures. It also led to the
increase in the torsional response of 3-storey building and significant decrease in the
response 4-storey structure. The results of the study indicate that torsional vibrations
(due to eccentric pounding) play an important role in the overall pounding-involved
response of asymmetric steel buildings under earthquake excitations.
The results of the analysis focused on earthquake-induced pounding between the

examples of two elastic steel buildings, for specified gap size, have been presented
in this paper. Further studies are requires so as to extend our knowledge on the
phenomenon. Such studies should consider the effects of different gap size values,
inelastic responses as well as the influence of increasing the input excitation on the
structural behaviour.
References
Anagnostopoulos SA (1988) Pounding of buildings in series during earthquakes. Earthq Eng Struct
Dyn 16:443–456
Anagnostopoulos SA, Karamaneas CE (2008) Use of collision shear walls to minimize seismic
separation and to protect adjacent buildings from collapse due to earthquake-induced
pounding. Earthq Eng Struct Dyn 37:1371–1388
Anagnostopoulos SA, Spiliopoulos KV (1992) An investigation of earthquake induced pounding
between adjacent buildings. Earthq Eng Struct Dyn 21:289–302
Bertero VV, Collins RG (1973) Investigation of the failures of the Olive View stairtowers during
the San Fernando earthquake and their implications on seismic design. EERC Report
No. 73-26, Earthquake Engineering Research Center, University of California, Berkeley
Jankowski R (2005) Impact force spectrum for damage assessment of earthquake-induced struc-
tural pounding. Key Eng Mater 293–294:711–718
Jankowski R (2007) Assessment of damage due to earthquake-induced pounding between the main
building and the stairway tower. Key Eng Mater 347:339–344
Karayannis CG, Favvata MJ (2005) Earthquake-induced interaction between adjacent reinforced
concrete structures with non-equal heights. Earthq Eng Struct Dyn 34:1–20
Kasai K, Maison B (1997) Building pounding damage during the 1989 Loma Prieta earthquake.
Eng Struct 19:195–207
Komodromos P, Polycarpou PC, Papaloizou L, Phocas MC (2007) Response of seismically
isolated buildings considering poundings. Earthq Eng Struct Dyn 36:1605–1622
Leibovich E, Rutenberg A, Yankelevsky DZ (1996) On eccentric seismic pounding of symmetric
buildings. Earthq Eng Struct Dyn 25:219–233
Mahmoud S, Jankowski R (2009) Elastic and inelastic multi-storey buildings under earthquake
excitation with the effect of pounding. J Appl Sci 9:3250–3262
Mahmoud S, Jankowski R (2011) Modified linear viscoelastic model of earthquake-induced
structural pounding. Iran J Sci Technol 35 C1:51–62
Mahmoud S, Austrell PE, Jankowski R (2012) Simulation of the response of base-isolated
buildings under earthquake excitations considering soil flexibility. Earthq Eng Vibration
11:359–374
Mahmoud S, Abd-Elhamed A, Jankowski R (2013) Earthquake-induced pounding between equal
height multi-storey buildings considering soil-structure interaction. Bull Earthq Eng
11:1021–1048
Maison B, Kasai K (1992) Dynamics of pounding when two buildings collide. Earthq Eng Struct
Dyn 21:771–786
Sołtysik B, Jankowski R (2013) Non-linear strain rate analysis of earthquake-induced pounding
between steel buildings. Int J Earth Sci Eng 6:429–433
Vasiliadis L, Elenas A (2002) Performance of school buildings during the Athens earthquake of
September 7 1999. In: 12th European conference on earthquake engineering, paper ref.
264, Elsevier Science Ltd.
Chapter 24
Dynamic Analysis of Irregular Multistorey
Shear Wall Buildings Using
Continuous-Discrete Approach
Jacek Wdowicki, Elżbieta Wdowicka, and Zdzisław Pawlak
Abstract The paper presents an analysis of coupled flexural-torsional vibrations of

multi-storey shear wall buildings using continuous-discrete approach. A dynamic
model with masses in the form of rigid floor slabs has been adopted. The mass
matrix is generated exactly, according to real layout of walls, connecting beams and
floor slabs. The flexibility matrix is generated on the basis of the solution of the
governing differential equations for a three-dimensional continuous model. In order
to verify the presented method, for the first example the free vibration analysis of
non-planar, non-symmetrical coupled shear walls, previously analysed in the liter-
ature by SAP2000 program, has been chosen. In the second example, the possibility
of seismic location of the multistorey building, designed originally in Poznań,
stiffened by the asymmetrical system of coupled shear walls, was considered. The
analysis, in which a continuous-discrete approach and the response spectrum
technique were applied, was carried out by means of DAMB (Dynamic Analysis
of Multistorey Buildings) program. The presented method is efficient and can be
useful at the preliminary design stage when the solutions for a structure with many
various geometrical and physical arrangements are required.
Keywords Tall buildings • Coupled shear walls • Continuous connection method •

Flexural-torsional vibrations • Seismic analysis
J. Wdowicki (*) • E. Wdowicka

Polytechnic Faculty, Higher Vocational State School (PWSZ) in Kalisz,
ul. Nowy Świat 4, 62-800 Kalisz, Poland
Z. Pawlak
Institute of Structural Engineering, Poznań University of Technology,
ul. Piotrowo 5, 60-965 Poznań, Poland

DOI 10.1007/978-3-319-14246-3_24
264 J. Wdowicki et al.
24.1 Introduction
In multistorey buildings lateral loads that arise from effects of wind and earth-
quakes are often carried by a system of shear walls acting as vertical cantilevers.
Rutenberg (2013) reviews the literature on the seismic shear demand on reinforced
cantilever walls. Such walls are usually perforated by vertical bands of openings,
which are required for doors and windows, and they form a system of coupled shear
walls. There are basically two approaches for analysis of coupled shear walls:
discrete and continuous. In the continuous approach, which has been widely used
for shear walls being uniform along the height, the discrete set of connecting beams
is replaced by a continuous medium of equivalent properties.
The seismic design of multistorey buildings involves the dynamic analysis of
structure. For the dynamic analysis it is convenient to use a hybrid approach, based
on the analysis of an equivalent continuous medium and a discrete lumped mass
system (Wdowicki and Wdowicka 1991; Li and Choo 1996). In this approach the
continuous connection method is employed to find the structural flexibility matrix
but the structure mass matrix is found with the lumped mass assumption.
When multistorey buildings with non-coincident centres of mass and stiffness are
subjected to a ground motion due to earthquake they respond in coupled lateral and
torsional vibrations. Torsional coupling may induce significant amplification of inertia
force resulting from earthquakes and it is felt to be strongly influenced by building
asymmetry (Glück et al. 1979; De Stefano and Pintucchi 2008; Yiu et al. 2014).
The paper presents an analysis of coupled flexural-torsional vibrations of
multistorey shear wall buildings using continuous-discrete approach. The mass
matrix including flexural and torsional inertia is generated. To find the flexibility
matrix each lumped mass is loaded subsequently with a unit horizontal generalized
force and the corresponding horizontal displacement vector for the whole structure
is found by the continuous connection method. As a result of solving the
eigenproblem, which corresponds to free vibrations of the system, the natural
frequencies and mode shapes of vibration have been received.
To verify the presented method, the free vibration analysis of non-planar, non-
symmetric coupled shear walls, previously analysed in the literature by the discrete
method, has been presented. The results obtained by this method have been
compared with those available in the literature.
The paper presents also the results of the seismic analysis based on discrete-
continuous approach and the response spectrum technique. The subject of the
analysis is the multistorey building, designed originally in Poznań, which is stiff-
ened by the asymmetric system of coupled shear walls. The seismic analysis was
carried out using DAMB program, as a part of an Integrated System (Wdowicki
et al. 1995), using design spectrum for elastic analysis according to Eurocode
8 (EC8-1 2004).
24 Dynamic Analysis of Multistorey Buildings 265
24.2 Method of Analysis
The analysis is based on the following main assumptions:

1. The floor slabs are taken as diaphragms with infinite in-plane stiffness;
2. The out-of-plane stiffness of the floor slabs can be modelled by connecting
beams of appropriate stiffness spanning between shear walls;
3. Vlasov’s theory for thin walled beams of an open section is taken to be valid for
the individual shear walls;
4. The walls and beams are assumed to be linearly elastic;
5. The geometric and mechanical properties of the structure are constant through-
out the height of each segment (Wdowicki and Wdowicka 2012).
In our analysis the continuous connection method has been used in conjunction
with Vlasov’s theory of thin-walled beams. To simplify the analysis, the effect of
St. Venant’s torsion has been neglected. The dynamic model with masses in the
form of rigid floor slabs has been adopted since over a half of the building total mass
is usually concentrated on the floor levels. The coupled torsional-flexural vibrations
have been taken into consideration. For shear wall multistorey structure it is more
natural to determine the flexibility matrix D rather than the stiffness matrix K
(Clough and Penzien 1993). To find the flexibility matrix D each lumped mass is
loaded with a unit horizontal generalized force and the corresponding horizontal
displacement vector is found using the continuous connection method.
The vibration of a structure is described by the following relation:
D M €x þ D C x_ þ x ¼ D f ð24:1Þ
where: D – flexibility matrix, M – mass matrix, C – damping matrix, x – d-element

vector of generalised coordinates (d – number of dynamic degrees of freedom of the
considered structure), f – d-element vector of generalised excitation forces,
corresponding to the generalised coordinates.
The flexibility matrix D is generated on the basis of the exact solution of the
governing differential equations for three-dimensional continuous model of the
shear wall structure.
Also the mass matrix is generated exactly according to the real distribution of
walls, connecting beams and floor slabs, including flexural and torsional inertia.
Using mass properties of shear walls, connecting beam bands as well as floor slabs,
a quasi-diagonal mass matrix of a whole structure is created:
M ¼ diagðMk Þ, ðk ¼ 1, . . . , nk Þ ð24:2Þ
where: nk – number of storeys, Mk – the 3 3 symmetric matrix, which defines

inertia properties of k-th storey.
The seismic response of the structure is estimated using the response spectrum
technique (Chmielewski and Zembaty 1998). The involved stages of the analysis
are as follows:
1. Determination of natural frequencies and corresponding mode shapes;
2. Evaluation of modal participation factors of modal loading of the structure using
an appropriate design spectrum;
3. Estimation of response taking into account the contribution of all given modes
for various parameters of interest, using three methods: SRSS – the square root
of the sum of the squares, CQC – the complete quadratic combination or DSC –
the double sum combination (Maison et al. 1983).
On the basis of the elaborated algorithm the software in Object Pascal of Delphi
5 environment has been implemented and included in the DAMB (Dynamic
Analysis of Multistorey Buildings) program for the dynamic analysis of shear
wall tall buildings.
24.3 Free Vibration Analysis of Non-planar

Coupled Shear Walls
To verify the presented method, several examples have been solved. As the first
example the non-planar, non-symmetric coupled shear wall system, presented by
Aksogan et al. (2014), has been considered (Fig. 24.1). The total height of the shear
wall is 48 m and the storey height is 3 m. The mass density and the elasticity and
shear modules are as follows: ρ ¼ 2400 kg/m3, E ¼ 2.85 GPa and G ¼ 1.056 GPa.
The height of the connecting beams is 0.4 m. The thickness of the connecting beams
and walls is 0.4 m in the right part and 0.2 m in the left part of the structure. The
mass of the typical storey, including a connecting beam, equals to 104.3 103 kg
and the total mass of the building is 1669 103 kg. The mass of floor slabs has not
been considered in this example for comparison purposes.
It may be noted that in the present analysis the shear deformation of the walls has
been neglected due to the assumption in Vlasov’s theory. The same assumption has
been made in the analysis presented by Aksogan et al. (2014) and for comparison
purposes the shear deformation was neglected in SAP2000 application as well.
In Table 24.1 the first ten natural frequencies corresponding to each mode found
by the program DAMB are compared with the values given by Aksogan et al.
(2014), obtained by SAP2000 structural analysis program, using MacLeod’s frame
method, for the case without a stiffening beam.
The torsional modes are omitted in the results given by Aksogan et al. (2014),
but for the other frequencies a satisfactory match of results has been observed.
Figure 24.2 presents the first six mode shapes of the analysed non-planar shear
wall structure found by the presented method, using DAMB program.
2 2 2 2 3
10 3 13
2 12
4
9
8 1 1 11
4 14
3
7
6 5 15
4 2 2 3
Fig. 24.1 Plan view of the non-planar shear wall structure [m]
Table 24.1 Comparison of the first ten natural frequencies found by the present method (program
DAMB) and by SAP2000 program (Aksogan et al. 2014)
Natural frequencies [Hz]
Present method
Mode Predominant mode (program DAMB) SAP2000 % difference
1 First mode X 0.4748 0.50047 5.13
2 First mode Y 0.6656 0.66117 0.67
3 First torsional mode 1.0022 –
4 Second mode X 2.4180 2.46396 1.87
5 Second mode Y 4.1799 4.11839 1.49
6 Second torsional mode 6.2020 –
7 Third mode X 6.4394 6.47507 0.55
8 Third mode Y 11.729 11.45595 2.38
9 Fourth mode X 12.456 12.37532 0.65
10 Third torsional mode 17.265 –
F Y X XF Y F Y X
46,5 46,5 46,5
43,4 43,4 43,4
40,3 40,3 40,3
37,2 37,2 37,2
34,1 34,1 34,1
31,0 31,0 31,0
27,9 27,9 27,9
24,8 24,8 24,8
21,7 21,7 21,7
18,6 18,6 18,6
15,5 15,5 15,5
12,4 12,4 12,4
9,3 9,3 9,3
6,2 6,2 6,2
3,1 3,1 3,1
0,0 0,0 0,0
0 1 0 1 0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1
Functions: X, Y, F [f = 0.475 Hz T = 2.106 s] Functions: X, Y, F [f = 0.666 Hz T = 1.502 s] Functions: X, Y, F [f = 1.022 Hz T = 0.978 s]

max X = 1,000 z = 46,500 max Y = 1,000 z = 46,500 max X = 1,000 z = 46,500
min F = -0,089 z = 46,500 min X = -0,076 z = 46,500 min X = 0,000 z = 0,000
FY X XF Y YX F
46,5 46,5 46,5
43,4 43,4 43,4
40,3 40,3 40,3
37,2 37,2 37,2
34,1 34,1 34,1
31,0 31,0 31,0
27,9 27,9 27,9
24,8 24,8 24,8
21,7 21,7 21,7
18,6 18,6 18,6
15,5 15,5 15,5
12,4 12,4 12,4
9,3 9,3 9,3
6,2 6,2 6,2
3,1 3,1 3,1
0,0 0,0 0,0
0 1 0 1 0 1
Functions: X, Y, F [f = 2.418 Hz T = 0.414 s] Functions: X, Y, F [f = 4.180 Hz T = 0.239 s] Functions: X, Y, F [f = 6.202 Hz T = 0.161 s]

max X = 1,000 z = 46,500 max Y = 1,000 z = 46,500 max X = 1,000 z = 19,500
min X = -0,814 z = 22,500 min Y = -0,840 z = 22,500 min Y = -0,867 z = 46,500
Fig. 24.2 The first six mode shapes of the non-planar shear wall structure, obtained by the DAMB
program
24.4 Seismic Analysis of the Multi-storey Asymmetric

Shear Wall Building
In the second example, the seismic analysis of the multistorey building, designed
originally in Poznań, stiffened by the asymmetric system of coupled shear walls,
has been presented. After the static analysis, the possibility of seismic location of
the designed structure was considered. The investigated building is a multi-storey
reinforced concrete structure. The elevation of the building and shear walls in the
typical floor plan are shown in Figs. 24.3 and 24.4.
Fig. 24.3 The view of the analysed multistorey shear wall building
The building has 15 storeys above the ground level and 2 basement storeys. The
total height of the building is 53.5 m. In the analysed building, lateral loads that
arise as a result of winds and earthquakes are carried by the three-dimensional
system of coupled shear walls. The dimensions of the central core are 3.15 m 12 m.
The thickness of walls equals to 0.3 m. The height of connecting beams is 0.7 m.
The structural properties of shear walls and lintels are uniform along the building
height. The slabs are 0.25 m thick.
A diaphragm action of all floor slabs is taken into consideration as the effect of
the assumption of their in-plane infinite rigidity and negligible transverse one. The
Young’s modulus E ¼ 36 GPa and the shear modulus G ¼ 15.4 GPa are assumed for
concrete. All the elements are assumed to be fully fixed in foundations. The static
analysis was carried out on the basis of a variant of the continuous connection
method (Wdowicki and Wdowicka 1993).
Fig. 24.4 The plan of the shear walls in typical floor (from the pre-processor of DAMB program)
Dynamic solutions have been obtained by treating the structure as a lumped

parameter system with discrete masses in the form of rigid floor slabs, having
flexural and torsional inertia. The floor masses are determined according to
Eurocode 8. Complete masses resulting from the permanent load are considered,
whereas the masses associated with the variable-live load are reduced using the
factor ΨEi ¼ φ Ψ2i. Factor Ψ2i amounts to 0.3 in the case of an office building.
Factor φ is equal to 1.0 for the roof slab and 0.5 for the other slabs. The total mass
of the building is equal to 9 234 103 kg.
The analysis was made using DAMB program which gives a possibility to
perform the dynamic analysis of three-dimensional shear wall structures. As a
result of the first step the periods of vibration and the corresponding mode shapes
have been received. The three fundamental periods of vibration of the building
amount to 2.22 s, 1.60 s and 1.10 s. The first mode is predominantly torsional, the
second mode is predominantly translational in the X direction and the third mode is
predominantly translational in the Y direction. It may be noted that the analysed
building can be classified as torsionally flexible (Fajfar et al. 2005).
The seismic action is represented by the Type 1 elastic response spectrum,
recommended for high and moderate seismicity regions (Ms > 5.5), for subsoil
class C (EC8-1 2004). The reference peak ground acceleration amounts to
agR ¼ 0.25 g. The values of the periods (TB, TC, TD) and the soil factor S, which
describe the shape of the elastic response spectrum, amount to TB ¼ 0.2 s, TC ¼ 0.6 s,
TD ¼ 2.0 s and S ¼ 1.15. For the analysed building the importance category II,
according to Eurocode 8, the corresponding importance factor γ 1 ¼ 1.0 and the
reduction factor v ¼ 0.5 were assumed. The value of viscous damping ratio ξ ¼ 5 %
has been assumed in the analysis (Fajfar and Kreslin 2012).
Assuming that the specific provisions for all the structural elements are satisfied
to provide the appropriate amount of ductility, the ductility class DCM (medium
ductility) has been established. The value of behaviour factor q ¼ 2.0 has been
applied. In the analysis the design response spectrum according to Eurocode 8 has
been applied (EC8-1 2004).
Modal response spectrum analysis was performed for the ground excitation in
two horizontal directions, X (E-W) and Y (N-S), independently. The CQC rule for
the combination of different modes was used.
The results of the analysis based on damage limitation state have been presented.
It has been assumed that the analysed building has non-structural elements of brittle
materials attached to the structure. The requirement of damage limitation is accom-
plished by Eurocode 8 when interstorey drifts dr do not exceed the limited values,
obtained using the following equation:
dr 0:005 0:005
¼ ¼ 0:01 ð24:3Þ
h ν 0:5
where: dr – difference between the lateral displacements occurring at the top and
bottom of the storey, determined by the linear analysis based on the design response
spectrum, h – the storey height.
The results of the displacement analysis, based on the response spectrum tech-
nique, for (E-W) seismic wave direction, are shown in Figs. 24.5 and 24.6.
The ratio of the actual top displacement and the total height of the building
above the basement amounts to: 2.0 0.123/53.5 m ¼ 0.46 % and 2.0 0.0844/
53.5 m ¼ 0.32 %, for X and Y directions, respectively.
The obtained maximum value of interstorey drift dr/h was equal to 2.0 0.0176/
3.45 m ¼ 0.0102.
The horizontal displacements of the shear wall structure due to the seismic wave
parallel to X axis reveal the considerable influence of torsion. The torsional rigidity
of the structure should be increased by the additional walls.
Using the software DAMB, it is easy to obtain the graphs of normal stresses for
the entire cross-section of shear wall structure, for a selected three-dimensional
shear wall as well as for a single wall (Fig. 24.7).
It seems advisable to consider the application of the presented method in the
seismic analysis of the actual existing asymmetric, torsionally flexible shear wall
buildings.
Budynek apartamentowo-biurowo-handlowy z dwoma poziomami garazy podziemnych,

w Poznaniu przy ulicy Dabrowskiego 81, beton B-55
statyka oraz wymuszenia sejsmiczne
Data : C:\1-PODRECZNE\BW\DANE\AKTUALNE\309-5.BW8
Window1 Window2 Window3
Pattern:8
Vx Vy Fi
59,9 59,9 59,9
RANGE
55,9 55,9 55,9 All
51,9 51,9 51,9
UNITS
47,9 47,9 47,9
- length
43,9 43,9 43,9 [ metr ]
39,9 39,9 39,9 - angle
35,9 35,9 35,9 [radian]
31,9 31,9 31,9 - function n

(intensities
28,0 28,0 28,0 of shear forces
in beams & joints)
24,0 24,0 24,0
kiloniuton
20,0 20,0 20,0
metr
16,0 16,0 16,0
12,0 12,0 12,0
8,0 8,0 8,0
4,0 4,0 4,0
0,0 0,0 0,0

0 0,1 0 10 20 0 1 2 3 4 5 6 7
Function: Vx [metr] Function: Vy [mili-metr] Function: Fi [mili-radian]
max Vx = 0,123 z = 59,899 max Vy = 25,235 z = 59,899 max Fi = 7,615 z = 59,899

min Vx = 0,000 z = 0,000 min Vy = 0,000 z = 0,000 min Fi = 0,000 z = 0,000
Fig. 24.5 Horizontal displacements of shear wall structure due to (E-W) seismic wave direction
(parallel to X)
Budynek apartamentowo-biurowo-handlowy z dwoma poziomami garazy podziemnych,

w Poznaniu przy ulicy Dabrowskiego 81, beton B-55
statyka oraz wymuszenia sejsmiczne
Data : C:\1-PODRECZNE\BW\DANE\AKTUALNE\309-6.BW8
Max displacement: 0.2476 Point: 19(8.54,-14.14) Pattern: 8
Min displacement: 0.0394 Point: 25(-2.83,11.01) z = 53.500
Structure
Y Multiplier = 20
2.00 METR
non displaced
structure
displaced
structure
intermediate
levels
Fig. 24.6 View of the displacements of the shear wall structure due to (E-W) seismic wave
direction; displacements are 20 enlarged relative to the structure dimensions (from the post-
processor of DAMB program)
Fig. 24.7 Normal stresses at the base of the shear wall structure due to (N-S) seismic wave
direction (from the post-processor of DAMB program)
24.5 Conclusions
In this study the seismic analysis of the shear wall tall building has been carried out
using a continuous-discrete approach and the response spectrum technique. The
coupled flexural-torsional vibrations have been taken into account. The results
obtained by the abovementioned method have been compared with those obtained
using the SAP2000 structural analysis program, given in the literature, and a
satisfactory match has been observed. The proposed method is efficient and can
be useful, particularly at the preliminary design stage when quick checks with
different structural arrangements and dimensions are needed.
References
Aksogan O, Turkozer CD, Emsen R, Resatoglu R (2014) Dynamic analysis of non-planar coupled
shear walls with stiffeners using a continuous connection method. Thin-Walled Struct
82:95–104
CEN. Eurocode 8 (EC8-1) Design of structures for earthquake resistance. Part 1: General rules,
seismic actions and rules for buildings. EN 1998-1:2004, Comité Européen de Normalisation,
Brussels, Belgium
Chmielewski T, Zembaty M (1998) Podstawy dynamiki budowli. Arkady, Warszawa
Clough RW, Penzien J (1993) Dynamics of structures. McGraw-Hill, New York
De Stefano M, Pintucchi B (2008) A review of research on seismic behaviour of irregular building
structures since 2002. Bull Earthq Eng 6:285–308
Fajfar P, Kreslin M (2012) Introduction to the RC building example. Modeling and analysis of the
design example. In: Eurocode 8: seismic design of buildings. Worked examples. JRC Scientific
and Technical Reports, Eur 25204 EN- 2012, Chapter 2, pp 27–50, Ispra, Italy. Available from:
http://eurocodes.jrc.ec.europa.eu/doc/WS_335/report/EC8_Seismic_Design_of_Buildings-
Worked_examples.pdf
Fajfar P, Marusic D, Perus I (2005) Torsional effects in the pushover-based seismic analysis of
buildings. J Earthquake Eng 9(6):831–838
Glück J, Reinhorn A, Rutenberg A (1979) Dynamic torsional coupling in tall building structures.
Proc Inst Civ Eng Part 2 67:411–424
Li G-Q, Choo BS (1996) A continuous-discrete approach to the free vibration analysis of stiffened
pierced walls on flexible foundations. Int J Solids Struct 33(2):249–263
Maison B, Neuss C, Kasai K (1983) The comparative performance of seismic response spectrum
combination rules in building analysis. Earth Eng Struct Dyn 11(5):623–647
Rutenberg A (2013) Seismic shear forces on RC walls: review and bibliography. Bull Earthq Eng
11(5):1727–1751
Wdowicki J, Wdowicka E (1991) Integrated system for analysis of three-dimensional shear wall
structures. Comput Methods Civ Eng 1(3–4):53–60
Wdowicki J, Wdowicka E (1993) System of programs for analysis of three-dimensional shear wall
structures. Struct Des Tall Build 2(4):295–305
Wdowicki J, Wdowicka E (2012) Analysis of shear wall structures of variable cross section. Struct
Design Tall Spec Build 21:1–15
Wdowicki J, Wdowicka E, Błaszczyński T (1995) Integrated system for analysis of shear wall tall
buildings. In: Proceedings of the Fifth World Congress “Habitat and High- Rise: Tradition and
Innovation”, Council on Tall Buildings and Urban Habitat, Amsterdam, pp 1309–1324
Yiu C-F, Chan C-M, Huang M, Li G (2014) Evaluation of lateral-torsional coupling in earthquake
response of asymmetric multistory buildings. Struct Design Tall Spec Build 23:1007–1026
Chapter 25
Analysis of the Dynamic Response
of Masonry Buildings with Irregularities
of Localization of Bearing Elements
Due to Mining Shocks
Tadeusz Tatara and Filip Pachla
Abstract The study concerns the analysis of the impact of structural irregularities
in the masonry buildings on their dynamic characteristics and dynamic response to
kinematic loads. Several models of buildings deemed to be representative of this
class of structures in the areas covered by the influence of mining tremors were
analyzed. In areas Legnica-Głogow Copper District (LGCD) and the Upper Silesian
Coal Basin (USCB) are the most numerous masonry buildings with a height of 1–5
storeys. The first part of the study refers to the adoption of 3-D dynamic models of
buildings of different structural wall system. These were buildings with irregular,
symmetrical and bisymmetrical system load-bearing walls. The second part con-
cerns the dynamic characteristics of the building models and assessing the impact of
irregularities in the load-bearing walls on the natural frequencies. Subsequently,
using the response spectrum method and assuming different standard response
spectra, dynamic responses of building models were determined. The level of
dynamic principal and shear, depending on the adopted standard curve, describing
the response spectrum and irregularities in the structural walls, were also evaluated.
The results of numerical analyzes showed a significant influence of the structural
irregularities on the natural frequencies of the analyzed models of buildings. It also
showed the impact of the adopted standard response spectrum curve in the LGCD
on the level of calculated stresses.
Keywords Mine tremors • Masonry buildings • Irregular structures • Response

spectrum • Dynamic response • Eurocode 8
T. Tatara (*) • F. Pachla

Institute of Structural Mechanics, Cracow University of Technology, ul. Warszawska 24,
31-155 Cracow, Poland

DOI 10.1007/978-3-319-14246-3_25
276 T. Tatara and F. Pachla
25.1 Introduction
In Poland, Legnica-Głogow Copper District (LGCD) and the Upper Silesian Coal
Basin (USCB) most intensive mining tremors are characterized by a power of up to
1E10 J and intensity of surface vibrations that reach even 0.3 g (g – acceleration due
to gravity). Underground mining operations are conducted primarily in urban,
heavily populated areas, so the influence of surface mining related vibrations on
development and risk assessment is particularly important. Masonry buildings
constitute the largest part of the residential development in the above areas.
These are generally single-family or multi-family buildings (in lesser extent)
which have from one to five floors. Repairs and reconstructions impose the neces-
sity of structural changes in buildings. These changes often cause the appearance of
symmetry or even bisymmetry of the load-bearing walls. However, the bearing wall
systems of these buildings often represent abnormal structural systems as a result of
extension works such as building external staircases. In this study the method of
dynamic response spectrum Tatara (2012) was used. This method uses: (a) the
adopted models of buildings treated as a representative for the considered class of
objects, (b) standard response spectra characterizing the surface vibrations induced
by mining tremors. These spectra were also compared with the standard response
spectrum of EN 1998-1:2004 (2004).This comparison indicated the differences
between the vibration-type surface mining and seismic (deriving from earthquakes)
Tatara (2002, 2012). The numerical analysis includes simultaneous operations of
two mutually perpendicular loads (parallel to the transverse and longitudinal axes
of buildings’ models) as described by standard response spectra. The vertical
component has been omitted. Assessment of the impact of positioning the wall
bearing elements in buildings on the responses was made by comparing the results
of dynamic analysis of the models. The bases of this assessment were the additional
maximum principal stress and maximum shear stress.
25.2 Analyzed Buildings and Their Structural

Characteristics
All of the analyzed structures are masonry buildings of various height and design.
Buildings can be considered as representative of this class of structures, residential
single-family housing. It was built using the traditional method of unreinforced
masonry. Floors and staircases and flights of stairs are made as a monolithic
reinforced concrete. The first three are three-storey buildings; floor plans of their
repetitive floors are shown in Fig. 25.1a–c.
Building from Fig. 25.1a is characterized by the symmetry of the load-bearing
walls with respect to one axis, and the building from Fig. 25.2a is bisymmetrical.
25 Analysis of the Dynamic Response of Masonry Buildings with Irregularities. . . 277
Fig. 25.1 Floor plans of three storey buildings: no. 1 (a), no. 2 (b), no. 3 (c)
Fig. 25.2 Floor plans of five storey buildings: no. 4 (a), no. 5 (b)
There is no symmetry in building presented in Fig. 25.1c. Building from Fig. 25.1c
differs in relation to the building in Fig. 25.1a due to addition of external staircase to
one of the outer wall (gable wall). A characteristic feature is the lack of dilatation
between the outer staircase and gable wall.
Buildings No. 4 and No. 5 – see Fig. 25.2 – have the same distribution of the
bearing walls as buildings No. 3 and No. 1 respectively presented in Fig. 25.1a, c.
Building No. 4 has two staircases, including one outside, not militated from the
main part of the building – see Fig. 25.2a.
A five-storey building No. 6 has first three floors as building No. 4. The other two
floors are of a different arrangement of bearing walls – see Fig. 25.3. Access to
floors 1–3 is via an internal staircase – comp. Fig. 25.3a. Floors 4 and 5 can be
accessed only using an external staircase located at the gable wall and not militated
from the main part of the building – see Fig. 25.3b.
Fig. 25.3 Floor plans of five storey building no. 6
25.3 Assessment of Regularity of the Structural System

of the Analyzed Buildings
Using the information provided e.g. in EN 1998-1:2004 (2004) buildings analyzed

in the study can be characterized as structures in which:
• the ratio of height to width (length) is close to unity, which prevents the loss of
stability of the structure and, consequently, its overturning,
• height of floors is the same, approximate carrying capacity due to the horizontal
loads in two mutually perpendicular directions is similar,
• there is a uniform shape horizontally and vertically,
• there are stiffening rims and bearing walls on circuit of structure, which leads to
a reduction in the effects of torsion,
• there is a lack of cantilever elements,
• there are no floors at different levels within the same storey,
• statically indeterminate structural system increases resistance and the possibility
of transferring the loadings in case of damage to one or a group of bearing
elements.
The presence of the factors listed above in the analyzed buildings provides that
structures should be designed to carry additional dynamic loads caused by the
mining related surface vibrations. The analyzed buildings shown in Figs. 25.1a–c
and 25.2 are characterized by the regularity along the height. The center of gravity
and the center of the shear remain in the vertical line. These buildings fulfill the
conditions of regularity given in EN 1998-1:2004 (2004) – see Fig. 25.4.
Fig. 25.4 Criteria for regularity of buildings EN 1998-1:2004 (2004)
25.4 Analyzed Models of Buildings and Their Dynamic

Characteristics
The models take into account the different properties of building materials forming
the bearing system, loads due to the weight of plasters and finishing elements, light
warming of the roof and long-term part of the live load (40 %), according to polish
standard PN-85/B-02170:1985 (1985). Assumption of only long-term part of the
live load results from the fact that, during use of the building, occasionally this load
reaches the maximum value, which in combination with the occurrence of mining
shock is highly unlikely. In addition, the static load acting in accordance with the
direction of gravitational acceleration plays reductive factor on the tensile stresses
Fig. 25.5 Assumed geometric models of three storey buildings from Fig. 25.1a–c
caused by mining tremors, which leads to a safer estimate of the structure response
to dynamic loads. When determining the value of the Young’s modulus of the
cement – limestone wall results of secant modulus of elasticity of the wall given in
Kubica et al. (1999) and the standard PN-B-03002:1999 (1999) were used. The
value of Poisson’s ratio was adopted on the basis of the studies contained in Drobiec
et al. (2000). Parameters for other materials were adopted on the basis of building
standards PN-B-03002:1999 (1999), PN-B-03264:2002 (2002). In the analyzed
models following values were used: reinforced concrete floors of the characteristics
of materials – Young’s modulus E ¼ 29 GPa, Poisson’s ratio ν ¼ 0.2, mass density
ρm ¼ 2500 kg/m3, the walls of masonry cement mortar for masonry material
characteristics – Young’s modulus and E ¼ 2GPa, Poisson’s ratio ν ¼ 0.25, mass
density ρm ¼ 1800 kg/m3.
The models of the analyzed buildings take into account all the elements relevant
to the stiffness and mass of the buildings which include e.g. window and door
openings, stairways, lintels. In the considered models masonry elements are viewed
as homogeneous, treating such assumption as sufficiently accurate in terms of
precise engineering calculations Tatara (2002). Finishing elements are not modeled
using FEA for practical reasons. Models of elastic support building resulting from
the foundation on the flexible ground were assumed. An influence of ground was
taken into account by means of elastic mounting. Elastic support of the models was
achieved by springs localized at the level of the bottom of the foundation. The
stiffness of boundary elements was defined according to PN-80/B-03040:1980
(1980), as for substrate III category of medium stiffness, and the papers Sawinow
(1964), Lipiński (1985). The dynamic coefficients of the substrate for the ground in
a natural state of the residual amount up to 50 MPa/m. Adopted three storey
building models from Fig. 25.1a–c are shown in Fig. 25.5a–c, and five storey
building from Figs. 25.2 to 25.3 are presented in Fig. 25.6. These models will be
the subject of further computational analyzes.
Table 25.1 summarizes, for example, the first three natural frequencies calcu-
lated for each of the analyzed masonry building models. Calculated mode shapes
are complex. Therefore, determination of the sequence of natural frequency was
guided by the dominant relative displacements in that direction.
Fig. 25.6 Assumed geometric models of five storey buildings from Fig. 25.2 and 25.3
Table 25.1 First three calculated natural frequencies of analyzed models of masonry buildings
f1 [Hz] f2 [Hz] f3 [Hz]
Model no 1 7.08 7.57 11.22
Model no 2 7.34 7.73 10.92
Model no 3 6.93 8.04 10.51
Model no 4 3.70 4.42 6.14
Model no 5 3.81 4.07 6.54
Model no 6 3.73 4.46 6.18
In all models, first two calculated natural frequencies f1 and f2 correspond to the
transverse vibration and the frequency f3 to torsion vibration. In the case of building
models with added staircase to the gable wall, regardless of their height, the value
of the natural frequency f2 is almost 20 % higher than the frequency f1. This
demonstrates the significant effect of irregularities in the position of the bearing
walls on the stiffness of these buildings and consequently the value of the funda-
mental frequency of the vibrations in the transverse direction. In the case of other
buildings impact of irregularities does not exceed 7 %.
25.5 The Dynamic Response of the Analyzed Models

of Masonry Buildings
For further calculations, dynamic response spectrum method was used (RSA);
described in detail in Tatara (2002). This method uses a structural model and
corresponding computer software, which must implement the standard, relative
acceleration response spectrum. The standard spectra contain information about the
frequencies content of the recorded, mining related, surface vibrations. They may
be used in the calculation of the design of dynamic structures exposed to
kinematic load.
Fig. 25.7 Standard response spectra assumed as kinematic excitations
Figure 25.7 shows the standard spectra for the USCB and LGCD areas obtained by
analysis of multiple horizontal components of the surface vibration accelerations. A
comparison of the standard curves generated on the basis of investigation shows large
variations. Application of dynamic curves proposed in Zembaty (2011, 2012),
Zembaty et al. (2015), adapted from EN 1998-1:2004 (2004), to the analysis, can
lead to erroneous solutions due to the very wide range of oscillation periods, where a
constant maximum value of β is assumed. These curves are based on the EN 1998-
1:2004 (2004), which refers only to the natural phenomena of earthquakes.
The studies carried out for many years at the Institute of Structural Mechanics in
Cracow University of Technology related to construction of the standard response
spectra for mining areas showed differences in the shape of the spectral curves
corresponding to different regions of LGCD and USCB Ciesielski et al. (1996),
Czerwionka and Tatara (2007), Kowalski et al. (1997), Kuźniar et al. (2010, 2006),
Lipiński (1985), Tatara (2002, 2012). The results presented in these works indicated
a significant effect of the ground properties on the shape of these curves. In Tatara
(2002), there was a comparison made of standard curves for USCB and LGCD with
the curve corresponding to response spectrum of the first type from EN 1998-
1:2004 (2004). This comparison shows that in the case of a horizontal section of the
curve from EN 1998-1:2004 (2004) is described by much wider range of periods,
including much greater periods than in the case of mining tremors. Figure 25.8
shows the comparison of standard response spectra for USCB and LGCD area
resulting from previous research and analysis Ciesielski et al. (1996), Kowalski
et al. (1997), Tatara (2002) with type 2 spectral curves for different types of soil
according to EN 1998-1:2004 (2004). The comparison indicates that standard
curves correspond to different types of soil described in EN 1998-1:2004 (2004)
are significantly different compared to the standard curves obtained for areas USCB
and LGCD. It would be unreasonable to use these curves in the design of structures
to dynamic loads of mining origin. It is difficult to assume that in the LGCD
characteristics of the soil correspond to the rocky substrate (soil A) – for this type
of substrate spectral curve of EN 1998-1:2004 (2004) is closest to the response
Fig. 25.8 Comparison of USCB and LGCD response spectra with response spectra of type 2 from
EN 1998-1:2004 (2004) for different types of soils
curves from the LGCD and USCB – see Fig. 25.8. Similar comparative analysis of
the standard curves obtained according to EN 1998-1:2004 (2004) with standard
spectral curves from different mining areas of USCB (area KWK “Rydułtowy” and
KWK “Anna”) indicates a significant difference and demonstrates a significant
effect of local soil conditions on the shape of the standard response spectra
Tatara (2012).
These differences occur both for spectral curves drawn on the basis of horizontal
and vertical components of the ground surface vibrations. These differences relate
to the width of the horizontal portion of the spectrum (“plateau”) and the value
describing this part of curve. The horizontal part of the curve from EN 1998-1:2004
(2004) is described by much wider range of periods, including periods greater than
these occurring in the case of mining tremors.
Dynamic calculations were performed assuming: (a) models of buildings shown
in Chap. 4, (b) the ground acceleration ap ¼ 1 m/s2, (c) the kinematic force in the
form of a standard response spectrum applied in two mutually perpendicular
horizontal directions. Calculations were done according to the procedure described
in detail in Tatara (2012).
Tables 25.2 and 25.3 show for example calculated dynamic maximum principal
stress in particular load-bearing walls in each of the analyzed models of buildings
shown in Fig. 25.5b, c. Tables 25.4 and 25.5 present maximum dynamic shear
stress. Calculated stresses showed in Tables 25.2, 25.3, 25.4, and 25.5 refer to cases
without considering the impact of dead and live load.
From the analysis of exemplary calculations for LGCD area, listed in Tables 25.2
and 25.3, a significant influence of the type of the curve describing the standard
response spectrum used for the dynamic calculations can be noted. Considering
building in Fig. 25.1b (full symmetry), and taking into account the spectral curve
given in Kowalski et al. (1997), Tatara (2002, 2012), an increase in the value of
these stresses in the individual structural walls compared to the case of the use of
the curves representing soil type A, B and C Zembaty (2011, 2012), Zembaty
Table 25.2 Maximum dynamic principal stress in bearing walls for model from Fig. 25.5b
assuming ap ¼ 1 m/s2
Maximum dynamic principal stress in bearing walls
[MPa]
1 2 3 4 5 6 7
GZW Kowalski et al. (1997) 0.26 0.42 0.57 0.37 0.33 0.42 0.14
LGCD Kowalski et al. (1997) 0.33 0.54 0.72 0.47 0.42 0.53 0.18
LGCD ground type A Zembaty 0.21 0.34 0.46 0.30 0.27 0.34 0.11
et al. (2015)
LGCD ground type B Zembaty 0.26 0.43 0.57 0.37 0.33 0.42 0.14
et al. (2015)
LGCD ground type C Zembaty 0.26 0.43 0.58 0.38 0.34 0.42 0.14
et al. (2015)
Table 25.3 Maximum dynamic principal stress in bearing walls for model from Fig. 25.5c
assuming ap ¼ 1 m/s2
Maximum dynamic principal stress in bearing walls [MPa]
1 2 3 4 5 6 7 8
GZW Kowalski et al. (1997) 0.37 0.40 0.37 0.35 0.28 0.34 0.17 0.44
LGCD Kowalski et al. (1997) 0.44 0.48 0.46 0.42 0.33 0.41 0.22 0.53
LGCD ground type A Zembaty 0.30 0.32 0.30 0.29 0.22 0.27 0.14 0.36
et al. (2015)
LGCD ground type B Zembaty 0.37 0.40 0.38 0.36 0.28 0.34 0.17 0.45
et al. (2015)
LGCD ground type C Zembaty 0.36 0.39 0.37 0.35 0.27 0.33 0.18 0.44
et al. (2015)
Table 25.4 Maximum dynamic shear stress in bearing walls for model from Fig. 25.5b assuming
that ap ¼ 1 m/s2
Maximum dynamic shear stress in bearing walls [MPa]
1 2 3 4 5 6 7
GZW Kowalski et al. (1997) 0.106 0.130 0.176 0.069 0.096 0.106 0.085
LGCD Kowalski et al. (1997) 0.135 0.165 0.135 0.088 0.121 0.134 0.107
LGCD ground type A Zembaty 0.086 0.105 0.145 0.056 0.077 0.086 0.068
et al. (2015)
LGCD ground type B Zembaty 0.108 0.131 0.178 0.070 0.097 0.107 0.086
et al. (2015)
LGCD ground type C Zembaty 0.26 0.43 0.58 0.38 0.34 0.42 0.14
et al. (2015)
et al. (2015), reaches respectively, approximately 58, 27 and 22 %. In the case of the
irregular arrangement of the building bearing walls – see Fig. 25.1c obtained results
were larger of the average dynamic stress of about 50, 20 and 19 %. Practically in
all bearing walls of considered models – comp. Fig. 25.1b, c, calculated maximum
values of dynamic shear stress using spectral curves according to Kowalski
Table 25.5 Maximum dynamic shear stress in bearing walls for model from Fig. 25.5c assuming
that ap ¼ 1 m/s2
Maximum dynamic shear stress in bearing walls [MPa]
1 2 3 4 5 6 7 8
GZW Kowalski 0.115 0.118 0.132 0.086 0.107 0.077 0.081 0.083
et al. (1997)
LGCD Kowalski 0.139 0.143 0.160 0.106 0.131 0.094 0.103 0.100
et al. (1997)
LGCD ground type A 0.094 0.096 0.108 0.070 0.087 0.063 0.066 0.067
Zembaty et al. (2015)
LGCD ground type B 0.118 0.121 0.135 0.088 0.109 0.079 0.082 0.084
LGCD ground type C 0.115 0.117 0.132 0.086 0.107 0.077 0.084 0.082
et al. (1997), Tatara (2002, 2012) are larger than calculated from the curves of the
studies Zembaty (2011, 2012), Zembaty et al. (2015) – see Tables 25.4 and 25.5.
In the area of LGCD and USCB in most of the structural walls in the
bisymmetrical building higher calculated values of maximum principal stress
than in a building with an irregular distribution of load-bearing walls were
achieved. It is not possible to establish a clear trend indicating the impact of
irregular bearing walls on the maximum value of shear stress – see Tables 25.4
and 25.5.
25.6 Conclusions
The study presents series of dynamic analysis of masonry buildings of a height of

3 and 5 floors. Buildings vary in terms of an angle of the load-bearing walls and
show symmetry with respect to one or two axes. Models with an irregular locali-
zation of the building load-bearing walls are also analyzed. The results of numerical
analysis show a significant effect of the irregular load-bearing walls on the rigidity
of the system, hence on the value of the natural frequency. Buildings with irregu-
larities, such as considered in the presented study, obtained up to 20 % higher
natural frequency, compared to other models of buildings. The results of calcula-
tions show significant effect of the type of response spectrum curve taken into
consideration on the calculated maximum values of dynamic stresses, regardless of
the type of assumed bearing system of building. In both considered mining areas,
there was no effect of irregular distribution of load-bearing walls on the calculated
maximum dynamic values of principal and shear stresses.
Acknowledgments The study carried out as part of work funded by the Ministry of Science and
Higher Education (No L-4/124/2014/DS).
References
Ciesielski R, Kowalski W, Macia˛g E, Tatara T (1996) Response spectra of earthquakes, mining

tremors, and their use (Spektra odpowiedzi od trze˛sień ziemi i wstrza˛sow g orniczych oraz ich
zastosowanie). In: Proceedings of the 42nd conference of Committee of Civil Engineering of
The Polish Academy of Sciences and the Science Committee of the Polish Association of Civil
Engineers and Technicians, Krak ow-Krynica, 17–22 September 1996
Czerwionka L, Tatara T (2007), Standard response spectra from chosen mining regions at Upper
Silesian Coalfield (Wzorcowe spektra odpowiedzi z wybranych obszar ow GZW). Czasopismo
Techniczne z.6 seria Budownictwo z. 2-B:11–18
Drobiec Ł, Piekarczyk A, Kubica J (2000) Influence of the element shape and direction of the load
on the value of the Poisson ratio of brick walls (Wpływ kształtu elementu badawczego oraz
kierunku obcia˛żenia na wartość wsp
ołczynnika Poissona mur ow z cegły). In: Proceedings of
the 46th conference of Committee of Civil Engineering of The Polish Academy of Sciences
and the Science Committee of the Polish Association of Civil Engineers and Technicians,
Wrocław – Krynica, 17–22 September 2000
Kowalski W, Macia˛g E, Tatara T (1997) Influence of mining tremors on buildings and people
staying in them (Wpływ wstrza˛s ow g
orniczych na budynki i ludzi w nich przebywaja˛cych). In:
Kwiatek J (ed) Protection of buildings in mining areas (Ochrona obiekt ow budowlanych na
terenach gorniczych). GIG, Katowice, p 583
Kubica J, Drobiec Ł, Jasiński R (1999) Research of secant elasticity modulus of the brick walls
(Badania siecznego modułu spre˛żystości mur ow z cegły). In: Proceedings of the 45th confer-
ence of Committee of Civil Engineering of The Polish Academy of Sciences and the Science
Committee of the Polish Association of Civil Engineers and Technicians, Wrocław – Krynica,
13–18 September 1999
Kuźniar K, Macia˛g E, Tatara T (2006) Acceleration response spectra from mining tremors, First
European Conference on Earthquake Engineering and Seismology (a joint event of the 13th
ECEE & 30th General Assembly of the ESC), Geneva, Switzerland, 3–8 September 2006,
Paper Number: 665
Kuźniar K, Macia˛g E, Tatara T (2010) Prediction of response spectra of foundation vibrations of
buildings due to mining tremors using neural networks (Prognozowanie spektr ow odpowiedzi
drgań fundament ow budynk ow od wstrza˛s ow g orniczych z zastosowaniem sieci
neuronowych), Prace Naukowe GIG, G ornictwo i Środowisko, Kwartalnik nr 4/4/2010, GIG,
Katowice 2010:50–64
Lipiński J (1985) Foundations and supporting structures for the machines (Fundamenty i
konstrukcje wsporcze pod maszyny). Arkady, Warszawa
Sawinow OA (1964) The modern design of foundations for machines and their calculations
(Sowriemiennyje konstrukcii fundamientow pod masziny i ich rascziot). Strojizdat, Moskwa
Tatara T (2002) An influence of surface mining-related vibration on low-rise buildings (Działanie
drgań powierzchniowych wywołanych wstrza˛sami g orniczymi na niska˛ tradycyjna˛ zabudowe˛
mieszkalna˛), Zeszyty Naukowe Politechniki Krakowskiej, seria „Inżynieria La˛dowa”, nr 74, Krak
ow
Tatara T (2012) Dynamic resistance of buildings in mining tremors conditions (Odporność
dynamiczna obiekt ow budowlanych w warunkach wstrza˛s ow g orniczych). Wydawnictwo
Politechniki Krakowskiej, Krak ow
Zembaty Z (2011) Adaptation of Eurocode 8 for the calculation of structures on influence of
mining tremors (Adaptacja Eurokodu 8 do obliczeń budowli na wpływy wstrza˛s ow
gorniczych), Inżynieria i Budownictwo Nr 3/2011:161–164
Zembaty Z (2012) The velocity conception of evaluation for mining tremors intensity to design of
buildings exposed to mining tremors (Pre˛dkościowa koncepcja oceny intensywności
wstrza˛sow gorniczych dla celow projektowania budowli narażonych na wstrza˛sy g ornicze).
In: Kabiesz J (ed) Risks and technologies (Zagrożenia i technologie). GIG, Katowice, p 405
Zembaty Z, Kokot S, Bozzoni F, Scandella L, Lai C, Kuś J, Bobra P (2015) A system to mitigate
deep mine tremor effects in the design of civil infrastructure. Int J Rock Mech Min Sci,
Elsevier, 74 (2015), 81–90
Standards and Other Materials
EN 1998-1:2004 Eurocode 8: Design of structures for earthquake resistant Part 1: Genaral rules,
seismic actions and rules for buildings
PN-80/B-03040:1980 Foundations and supporting structures for the machines (Fundamenty i
konstrukcje wsporcze pod maszyny)
PN-85/B-02170:1985 Evaluation of the harmfulness of building vibrations due to ground motion
(Ocena szkodliwości drgań przekazywanych przez podłoże na budynki)
PN-B-03002:1999 Unreinforced masonry structures. Design and calculation (Konstrukcje murowe
niezbrojone – projektowanie i obliczanie)
PN-B-03264:2002 Concrete, reinforced concrete and prestressed structures. Design and calcula-
tion (Konstrukcje betonowe, żelbetowe i spre˛żone. Obliczenia statyczne i projektowanie)
Chapter 26
Numerical and Experimental Prediction
Methods for Assessment of Induced
Vibrations in Irregular Buildings
Jan Benčat
Abstract The prediction models for the ground–borne vibrations and irregular
buildings (IB) structure dynamic response due to traffic means have been intro-
duced. For the free–field dynamic response at the distance (e.g. from railway track)
and dynamic response of IB in the calculation procedures are described in this
paper. The numerical model of the soils is modelled as a viscoelastic half-space.
This model is used both for evaluation of the track–soil interaction forces as well as
for prediction of the ground–borne vibrations. The numerical results in time domain
are presented as the time histories damped amplitudes of the half-space vibration at
the distance. In presented frequency domain free–field response is calculated via
response spectra and frequency response function (FRF) of the viscoelastic soil
medium. In the next step this functions are applied for building structure dynamic
response calculation due to railway traffic (case study) via relevant computational
building structure model.
Keywords Microtremor • Vibrations • Prediction • FEM • Experiments
26.1 Introduction
The growing traffic volume, the higher population density and the diminishing
distance between the track and the structure can be considered to be responsible for
increasing vibration nuisance due to railway traffic. Therefore, the development and
validation of a numerical prediction model for traffic induced vibrations in build-
ings is treated in many works. Empirical models show a close relationship to a set of
experimental data but the application of the model is limited to similar conditions.
Also these models do not always provide insight in the influence of specific
parameters. Numerical models allow the influence of various parameters to be
J. Benčat (*)
University Research Centre, University of Žilina, Univerzitna 1,
SK 010 26 Žilina, Slovakia

DOI 10.1007/978-3-319-14246-3_26
290 J. Benčat
investigated but a validation of the model with experimental data required to verify
the underlying theoretical assumptions. Even though that validation is focuses on
traffic induced vibrations, the numerical prediction model can be generally appli-
cable to other types of vibration sources e.g. roadways Bencat J (1992, 2006),
Bencat J et al (2009). The dynamic train–track interaction is a coupled problem,
contrary to vehicle–road interaction problems, that requires the simultaneous solu-
tion of the equations of motion of the train and the track. The train–track interaction
forces due to the track unevenness are computed using a flexibility formulation. A
two–dimensional linear vehicle model with a limited number of DOF is coupled to
a linear elastic longitudinal invariant track model, which allows a solution of the
equations of motion in the frequency–wavenumber domain Bencat J (1992), The
transfer functions Bendat J and Piersol A (1993) between the track and the soil and
the computed interaction forces are used to compute the response at any arbitrary
point in the free field. Finally, the building structure dynamic response at the
distance calculation is performed using half – space output data (PSD, time history)
as the input data into the IB structure foundations, see Bencat J et al. (2009),
Fujikake T (1986), Ford R (1987), Grassie S and Cox S (1984), and Grundman H
et al. (1999).
26.2 Track Model Description: Numerical Approach
A numerical prediction model for ground–borne vibrations due to railway traffic on

ballasted track requires the modelling of several components, as indicated in
Fig. 26.1. This paper presents a numerical prediction model which calculates the
ground–borne vibration level due to railway traffic in two steps. The first step
determines the dynamic track–soil interaction forces using a detailed train model
and the dynamic behavior of the layered spring–damper system and the through–
soil coupling of the sleepers are accounted for the soil model (Broeck and Roeck
1999; Grassie and Cox 1984; Grundman et al. 1999; Knothe and Wu 1998;
Kotrasova 2009; Turek 1992).
The prediction of the ground–borne vibration level at the distance in the second
step is based on the viscous–elastics soil model (Bencat 2006; Ford 1987).
The vehicle car–body, the bogie and the wheelset are modelled as rigid bodies
connected by springs and dampers, Fig. 26.2e. The wheelset is connected to the rail
with a linearized Hertzian spring. The rail is modelled as a hinged Rayleigh beam
with rotational inertia. In the track model the rail is supported discretely by sleepers
modelled as rigid bodies with spring–damper systems representing the rail pads.
The sleeper is modelled as a short Rayleigh beam resting on flexural mass layer
supported by discretely Pasternak spring–damper systems representing the elastic
and attenuation characteristics of the railway ballast and substrate soils. As a result,
the model evaluates the track–soil interaction forces in terms of the spectral density
function which is often used as the statistical description of the rail roughness
Turek J (1992).
26 Prediction Methods for Vibrations Assessment 291
TRAIN
I6 CAR BODY
TRACK m5 BOGIE
RAIL WHEEL SET
RAIL PAD k2 b2 k2 b2
SLEEPER I8 I8
m9 m9
k1 b1 k1 b1 k1 b1 k1 b1 Q -e e Q
m 10 m7 m4 m1 y
TRACK - SIOL x
INTERACTION k1, b1 k1 b1 k1 b1 E, J(y), µ (y)
d
FORCES µ 2, 2, J 2
p
SOIL
h
k 2, b 2, c 2 BALLAST
SUBSTRATE
a a a a
L L l/2 l/2
Fig. 26.1 Components of the train–track–soil system
This calculations followed by a second step in which the spectral density of the
level of ground–borne vibrations is determined by FRF between track and
unbounded soil. Proposed prediction theoretical model for vertical track vibration
numerical program consists of three parts:
• model of vehicle
• model of train–track interaction
• model of track (sleepers/ballast and subsoil).
The frequency characteristics method (input–output) was used for calculation
modelled feedback linear dynamic system train–track–soil system parameters,
(program Interaction). Final products of the numerical calculations are: vehicle,
rail, sleeper, railpads and ballast frequency response functions (also sleeper deflec-
tion and bending moment in time domain) using spectral density functions (SDF) of
the rail roughness used by railway operators or experimentally measured in situ for
case study.
An important example of non–linear behavior is the wheel–rail contact but also
the rail pads and the suspension of the train can deflect in a non–linear manner.
Nevertheless the results presented in this study are limited to linear analyses. Also
in this model is accepted symmetrical dynamic response of the sleepers to longitu-
dinal axis of the track (rail roughness coherence function for left and right rail is
equals to 1).
26.2.1 Track Model
The track model commonly found in the literature represents the rail as infinite
Timoshenko, Euler or Rayleigh beam, see also Grassie S and Cox S (1984) and
Grundman H et al. (1999), on a continuous uniform support, Fig. 26.2a. The beam is
c b l
292
a f(t)
e
E 1 I1 rail E2 I2 (y)
− −
Q1 = − Q1 = − A2 (y)
y2 2 2 ρ
k1, b1 railpads
sleepers 1 23 45 N I y
h
ρb
k2, b2, c ballast k2 (y)
Q2
c2 (y)
b2
c d L L
L L
= ( )
= ( )
E 1, I1 x
E1, I1, μ1,ρ1 x
0 0
k1, b1 k 1, b 1
v1 sleepers v1 sleeper
ms
a a a a a a a a layer a a a a a a a
I6
e m5
f x v
z
k2 b2 k2 b2
g
I8 I3 ζ
m9 m2 WQ
Mw F0 z z =ξ + v + η
k1 b1 k1 b1 k1 b1 k1 b1 ξ z Q
+ W Q
ξ(x) m10 m7 m4 m1 v Q Q
kH z4 p=4 z3 3 z2 2 z1 1 Wv
Q
η Q Q
Wη
v1 2,4 m 6,6 m 2,4 m
Fig. 26.2 The railway components model for numerical analysis

J. Benčat
taken as uniform flexural rigidity EI, rail mass per unit mr and distributed sleeper
mass ms. The railpad stiffness and viscous damping constant per unit length are
taken to be k1 and b1 and the corresponding parameters for ballast are k2 and b2
respectively. An harmonic point force p(t) ¼ Pcosωt is assumed to run at constant
velocity v along the rail. The FRF of the track excited to a harmonic force for
proposed prediction model are discussed in this section. The track model consists of
two parts: (i) model of sleeper with ballast and subballast, Fig. 26.2b and (ii) track
model (rail supported discretely by sleepers modelled as rigid bodies with spring–
damper systems representing the rail pads), Fig. 26.2c, d.
Finally, the solution for FRF of the linear dynamic system model e.g. in which
on input are rail roughness ξ and on output are wheel forces Q enable calculations of
the interaction matrix FRF according to scheme as shown Fig. 26.2g. The dynamic
displacement of the wheel z is defined by z ¼ ξ þ v þ η, where ξ represents rail
roughness, v – rail vertical deflection and η – wheel and rail contact deformation in
contact location.
The track irregularities are great source of the track and vehicles dynamic
excitations. Such excitation arises from discrete irregularities such as wheel flats
and rail joints as well as periodic irregularities such as corrugation of the railhead. It
is assumed that excitation of the track arises from a wheel passing over a sinusoidal
irregularity on the rail head (Fig. 26.2f). The stochastic theory analysis enables to
define irregularities by PSD function by Sξ ðΩÞ ¼ AΩa , where A, a are empirical
(experimental) constants. The distance x is used as the independent variable to
define ξ(x).
26.2.2 Calculation of Track Component Arbitrary FRF
The frequency response function of arbitrary dynamic system part is calculated by

rule of FRF summing as follows W QI ¼ W Q1 I
:W Q1 Q2 Qi Qn
Q2 :W Q3 . . . . . . ::W Qiþ1 . . . ::W Q :
The same way is used for summing of the arbitrary FRF matrix, e.g. consider a
dynamic system with defined input rail roughness ξ producing output sleeper
deflection v2, than the FRF is given by equation
W ξv2 ¼ W ξQ :W QQ1 :W Q
v2 :
1
26.2.3 The Response of Track Resting on a Continuous

Rail Supports
An advantage of the continuous track model (Fig. 26.2c, d) is that enables for
arbitrary track variables to calculate Turek J (1992): FRF, input spectra and
standard deviations of the track dynamic response parameters – dynamic forces
Q, Q1, Q2, dynamic deflections v1, v2, wheel center dynamic displacement z and
294 J. Benčat
Fig. 26.3 The vertical [mm/kN]

track receptance of v1 0
-1
-2
-3
-4
-5
1 10 100 1000 [Hz]
Fig. 26.4 The input PSD: [(m.s-2)2/Hz ]

Gv́€2 v́€2 ð f Þ 2.00E-06
1.60E-06
1.20E-06
8.00E-07
4.00E-07
0.00E+00
0 50 100 150 200 250 [Hz]
dynamic bending moment in rail and sleeper. As an example on Fig. 26.3 is plot of
numerical calculations results for vertical track receptance of rail deflection v1 due
to wheel contact forces Q.
Parameters Used in Calculation:

Sleeper SB8: k2 ¼ 49.40 Mpa; b2 ¼ 0.023 Mpa.s; c2 ¼ 0; h ¼ 0.45 m; a ¼ 0.55;
ρb ¼ 0.0017 kg.m3.
Rail R65: L ¼ 12.12 m; k1 ¼ 217,000 Mpa; b1 ¼ 0.037 Mpa.s; E ¼ 210,000 Mpa;
I1 ¼ 3.6 105 m4; mr ¼ 65 kg.m1.
Vehicle: SKODA E 699; 10 degree of freedom; 4 axles; kH ¼ 1.5 109 [N/m]; mass of
axle: 1,250 kg; mass of bogie: 4750 kg; mass of body casing: 23.500 kg; bogie inertia
moment: 5.5 103 kg m2; bogie casing inertia moment: 5.105 kgm2;
k1 ¼ 7266.105 Nm1 spring stiffness: k2 ¼ 9.5 106 Nm1; k1 ¼ 7.266 105 Nm1;
spring damping: b1 ¼ 7.37 104 Nm1; b2 ¼ 3.68 104 Nm1; axle base: 2.8 m;
boggie base: 10.3 m (Fig. 26.4). From the proposed model it is possible to calculate
also sleepers dynamic deflection, velocities and acceleration power spectral densi-
ties (PSD) as the input spectra or PSD into the ballast and roadbed, see Fig. 26.4.
26.3 Prediction Models for Ground Vibration from

Railways
26.3.1 The Analytic–Experimental Approach
The analytic–experimental approach proposes the test and the theory data combi-
nation to calculate the prediction level of ground vibration (suitable in practical
cases). In this process as an input signal can be used accelerations spectra
(or spectral densities) derived from experimental data bank of authorized railway
category with corresponding rail profile or accelerations spectrum Sw€ w€ ð f Þ measured
at nearest ground point to the track for individual case study. The frequency
response function (or transfer function) of the ground can be derived via experi-
mental impulse seismic method (ISM) Bencat J (1992, 2006) or cross–hole test data,
SASW method, from which elastic and attenuation parameters of the ground can be
obtained, too. The measuring output response acceleration spectrum (RAS) at the
distance Sẅẅ( f ) due to input accelerations spectrum Sw€ w€ ð f Þ the FRF– H( f ) can be
derived, see also Bencat J et al. (2009). In accordance with Bendat J and Piersol A
(1993) then RAS is given by the formula: Sw€́ w€́ ð f Þ ¼ Hð f ÞS €€́ ð f Þ:
ẃ w
26.3.2 In: Situ Soil Dynamic Parameter Tests Experimental

Tests at Nearby Building IBM Region
To calculate prediction vibration level and dynamic response of the projected new
building in the projected new railway line area it was needs to know the building
site soils dynamic parameters, soil FRF and project building parameters. Therefore
the in situ ISM tests in the IBM Data Centre building site were performed. The
building site is situated in the same area in which the new Trans European Network
(TEN–T) line is projected, too. After the both structures erection the distance
between by them will be approximately 20 m. Hence the prediction of building
vibration level and response spectra due to operating trains were required.
The building site is situated on level ground (sandy loam 3.5 m and gravel sand
12.0 m). This permits the ground to be modelled as a damped, viscoelastic half
space. The viscoelastic model of soil simulation using the complex modulus
conception E* ¼ E(1 + δE) and G* ¼ G(1 + δG) respectively, offers a very good
approach to the actual soil behavior (E, G and δEδG are real and imaginary
components of complex modulus). The Raleigh’s and shear waves propagation vr
and vs in half space in this form for the case study were analyzed in Bencat J (1992).
The experimental tests procedure for the purpose of the evaluation of elastic and
attenuation soil parameters is described in Bencat J et al. (2009). The IBM building
site layout, accelerometers and impact loading positions (Ii) during the experimen-
tal tests are shown in Fig. 26.5. The ISM results are as follows:
• vr ¼ 145.10 ms–1; δG ¼ 0.117; E0 ¼ 109.20 MPa; G0 ¼ 41.10 MPa,
• the ISM test No. 5 spectral analysis results example are plotted on Fig. 26.6.
296 J. Benčat
8.00m
0.10m
1.10m 13.90m 3.30m 1.70m 1.50m1.50m
(ALLIANZ)
IBM DC
B2 B1
I2 I3 I4
1.10m
I1
17.30m
19.00m
22.00m
10.50m
Betónové oplotenie
Fig. 26.5 Accelerometers and impacts (I1 – 4) position and projected building site layout
BK2z – PSD – G22(f)

[(m/s^2)^2/Hz]
[(m/s^2)^2/Hz]
51.52 Hz
2.5E-05 45.35 Hz 60.33 Hz 27.95 Hz
0.035
2E-05 0.03 50.76 Hz 56.53 Hz
0.025
1.5E-05 0.02
80.11 Hz
1E-05 29.33 Hz 0.015
25.44 Hz 0.01
5.E-06
0.005
20 30 40 50 60 70 80 [Hz] 20 30 40 50 60 70 80 [Hz]
BK1z – PSD – G11 (f) BK1z, BK2z – CP SD G 12 (f)
[(m/s^2)^2/Hz]
2E-06
50.71 Hz 56.47 Hz
2E-06 5.E-06
60.42 Hz
1.E-06
4.E-06 46.91 Hz
1.E-06
3.E-06
8E-07 25.44 Hz
2.E-06
5.E-07
3E-07 1.E-06 27.47 Hz
20 30 40 50 60 70 80 [Hz] 20 30 40 50 60 70 80 [Hz]
BK2z – PSD – G22(f) BK1z, BK2z – Transfer function –H 12 (f)
Fig. 26.6 The ISM5 test spectral analysis results at B1 and B2 points
The calculations includes data: λR ¼ 9.2 m, (Raleigh’s wave length);

ρ ¼ 1950 kgm–3, (soil mass density); α ¼ 0.0398 m–1, (the attenuation coefficient
obtained by standard deviations σ 0, σ y of displacement amplitude vibration at the
distance l0, ly from source of excitation using the displacement power spectral
ð0Þ ðyÞ
densities Gii and Gkk ).
26.3.3 Vibration Propagation Process Experimental

Spectral Analysis
An experimental study of ground vibration transmission from a railway was carried

out in the same region as the impact tests, adjacent to the ŽSR railway line
Bratislava – Vienna, track No. 1 (No. 2) in the town district Bratislava Trn avka.
The object of the experimental measurements was to find: spectral characteristics
of the vibration components of the track near region soils by the acceleration power
spectral densities Gii( f ), Gkk( f ) and Gik( f ). The pickups positions are shown in
Fig. 26.7. The roadbed and ground accelerations of the vibrations were recorded
using the portable computer with relevant software and hardware facilities. The test
experimental procedure in details is described in Bencat J et al. (2009). As an
example of the train induced vibrations accelerations spectral analysis results (PSD)
in the ground at measured point BK1 are plotted on Fig. 26.8.
Fig. 26.7 The pickups and impacts positions in track region

298 J. Benčat
[m/s^2] [(m/s^2)^2/Hz]
0.3
8.5E -06 43.11 Hz 54.94 Hz
0.2 39.13 Hz
17.02 Hz 60.00 Hz
8 E -06
0.1
7.5E -06
0 32.06 Hz
-0.1 7 E -06
66.30 Hz
-0.2 6.5 E -06
-0.3
5 10 15 20 25 [t] 10 20 30 40 50 60 70 80 [Hz]
BK1y – ẅ(t) BK1y – PSD – Gyy(f)
[m/s^2] [(m/s^2)^2/Hz]
0.15
44.04 Hz
5.E-06
0.1
55.05 Hz
4.E-06
0.05
59.02 Hz
0 3.E-06 38.07 Hz
-0.05 2.E-06 22.02 Hz
63.75 Hz
-0.1
1.E-06
-0.15
5 10 15 20 25 [t] 10 20 30 40 50 60 70 80 [Hz]
BK1x – ẅ(t) BK1x – PSD – Gxx(f)
[m/s^2] [(m/s^2)^2/Hz]
39.14 Hz
0.2
4.E-06
27.98 Hz
18.05 Hz
0.1 3.E-06
45.11 Hz
0 2.E-06
55.05 Hz
1.E-06
-0.1
5 10 15 20 25 [t] 10 20 30 40 50 60 70 80 [Hz]
BK1z – ẅ(t) BK1z – PSD – Gzz(f)
Fig. 26.8 The acceleration time histories and PSD at point BK1
26.4 Dynamic Response of the Building Structure

Prediction
To calculate the prediction vibration level and the structure dynamic response of
projected new IB structure situated at nearest area of the projected new railway line
it was needs to know the building site soils dynamic parameters, site geological
medium FRF and also the representative input accelerations spectra due to trains
Fig. 26.9 FEM model

render
P6
K6
P5
K5
P4
K4
P3
K3
P2
K2 z
y
K1≡P1 x
comparable to expected real train spectra due to trains in future traffic. For
calculation expected structure dynamic response it was used: (i) PSD – Gii( f ) of
ground acceleration at the track nearest region (Fig. 26.7) with similar geological
medium data as the site geological medium input spectra (see Fig. 26.8), (ii) The
halfspace transfer function Hik( f ) of the building site geological medium obtained
by ISM tests (see Fig. 26.6) and (iii) project of building structure.
The response PSD – Gkk ( f ) of the halfspace point at the projected structure
foundations location was calculated and plotted on Fig. 26.11. In the next step these
spectra were used as the input spectra for expected building structure dynamic
response calculation due to train. The dynamic response of the building structure
numerical calculation was carried out by the Visual Fea program package. The
calculated values of natural frequencies and natural modes were used in the next
step of the IB structure dynamic response calculations. The structure FEM model
render and two natural modes examples are shown in Figs. 26.9 and 26.10.
In the next step the building structure dynamic response calculations were
performed in frequency domain. The spectral analysis was divided into two parts:
low frequency band (0–10 Hz) and higher frequency band (10–130 Hz). The low
frequency spectral analysis gives the basic building natural frequencies vibration
range which enables to predict possible resonance effects of structure vibration due
to traffic. Because of the vibration sensitive technologies installing in the IBM DC
building the power spectral densities are determined in the structure relevant points
(over columns and in the middle of beam spans) for the structure dynamic response
considering in the range of frequency band 10–130 Hz. The higher frequency
spectral analysis band is mainly required for vibration level assessment on moni-
tored frequency according to the IBM Corporate Standard C–S1–9711–002, 1990–
03 requirements. The accelerations PSD, displacements extreme and rms values
were numerically calculated at the selected render points (K1. . .K6, P1. . .P6) in the
three vibration directions x, y, z. The results of the IBM DC building structure
300 J. Benčat
Fig. 26.10 Structure FEM natural modes examples
-6 -6
9.10 40,59 [Hz] 8,1.10
-6
52,00 [Hz] 57,68 [Hz]
-6 8,89.10 9,29.10
-6
-6 48,52 [Hz] 8,38.10
-6
8.10 [(m/s2)2/Hz] [(m/s2)2/Hz] 7,2.10 5,41.10

-6 [(m/s2)2/Hz]
-6 -6 [(m/s2)2/Hz]
7.10 6,3.10
-6 -6
6.10 31,80 [Hz] 5,4.10 26,18 [Hz]
-6
-6 4,26.10
[(m/s ) /Hz]
2,45.10
[(m/s ) /Hz]
-6
[(m/s2)2/Hz] -6 [(m/s2)2/Hz]
5.10 4,5.10
2 2
2 2
-6 -6 35,33 [Hz]
4.10 57,31 [Hz] 3,6.10 1,84.10
-6
26,10 [Hz] 2,45.10-6
-6 [(m/s2)2/Hz] 61,40 [Hz]
-6 2,46.10 [(m/s2)2/Hz] -6 -6
3.10 2,7.10 3,63.10
[(m/s2)2/Hz] 22,10 [Hz] [(m/s2)2/Hz]
-6
-6 -6 1,26.10
2.10 1,8.10 [(m/s2)2/Hz]
-6 -6
1.10 0,9.10
0 0
[Hz] 20 40 60 80 100 120 [Hz] 20 40 60 80 100 120
x – direction y – direction
-5
14.10 51,94 [Hz]
-4
1,42.10
-5 [(m/s2)2/Hz]
12.10
32,04 [Hz]
-5 35,40 [Hz]
3,78.10 -5
-5 [(m/s2)2/Hz] 2,68.10
10.10 [(m/s2)2/Hz]
[(m/s ) /Hz]
-5
8.10 26,37 [Hz]
2 2
-5
5,46.10 67,08 [Hz]
[(m/s2)2/Hz]
48,52 [Hz] 2,42.10
-5
-5 -5
6.10 2,22.10 [(m/s2)2/Hz]
[(m/s2)2/Hz]
-5
4.10
-5
2.10
0
[Hz] 20 40 60 80 100 120
z – direction
Fig. 26.11 The input acceleration PSD – G11( f ) to foundation at point K1/10 – 130 Hz
x y
-7
4,0.10
72,27 [Hz] 67,08 [Hz]
-7 -7 51,94 [Hz] 53,65 [Hz] 9,67.10
-7
3,5.10 3,99.10 -7 4,56.10
-7
[(m/s2)2/Hz] 3,11.10 [(m/s2)2/Hz]
-7 [(m/s2)2/Hz]
8,0.10 [(m/s2)2/Hz]
-7 42,48 [Hz]
3,0.10 40,52 [Hz]
1,56.10
-7
-8
9,38.10 [(m/s2)2/Hz]
-7 [(m/s2)2/Hz] 48,52 [Hz]
2,5.10 -7
1,71.10
-7
6,0.10
[(m/s ) /Hz]
74,28 [Hz]
[(m/s ) /Hz]
62,93 [Hz] -7 [(m/s2)2/Hz]

-7 -7
1,79.10
2 2
2 2
2,0.10 1,39.10 [(m/s2)2/Hz]

38,76 [Hz] [(m/s2)2/Hz]
-8
5,09.10 -7 70,35 [Hz]
1,5.10
-7
[(m/s2)2/Hz] 4,0.10 42,54 [Hz] -7
2,17.10
-7
2,57.10 [(m/s2)2/Hz]
-7 35,43 [Hz] [(m/s2)2/Hz]
1,0.10 8,57.10
-8
-7
[(m/s2)2/Hz] 2,0.10 72,21 [Hz]
-8 -7
5.10 2,15.10
[(m/s2)2/Hz]
0 0
[Hz] 20 40 60 80 100 120 [Hz] 20 40 60 80 100 120
z
-7
2,1.10
-7
49,61 [Hz]
-7
1,8.10 2,07.10
[(m/s2)2/Hz]
-7
51,94 [Hz]
-7
1,5.10 1,87.10
[(m/s2)2/Hz]
[(m/s ) /Hz]
-7
1,2.10
2 2
-7
9,0.10
42,48 [Hz]
-7 -8
6,0.10 3,00.10
[(m/s2)2/Hz]
-7
3,0.10
0
[Hz] 20 40 60 80 100 120
Fig. 26.12 The response acceleration PSD – G66( f ) of roof at point K6/10 – 130 Hz
dynamic response calculations are presented in detail Bencat J et al. (2009). The
adequate input load acceleration PSD – G11( f ) into the foundation structure (render
point K1) for x, y and z vibration directions are plotted in the Fig. 26.11 and the
response acceleration PSD – G66( f ) of the building roof structure (render point K6)
is plotted in the Fig. 26.12.
26.5 Conclusions
Based on the results of this paper the following conclusions can be drawn:
(i) The numerical prediction model can account for many parameters of the train–
track–soil interaction problem. Final products of the numerical calculations
are vehicle, rail, sleeper, rail pads and ballast frequency response functions
using spectral density functions of the rail roughness. To predict the level of
ground vibration in the vicinity of railways it needs to calculate the response
spectrum at distance point on the ground surface Sww( f ) via the (FRF) –
Hik( f ) of the ground by a method involving integral transform.
(ii) The numeric–experimental approach process proposes the test and the theory
data combination to calculate the prediction level of ground vibration. This
approach is suitable for practical tasks and case studies (e.g. law studies,
predictions).
302 J. Benčat
(iii) The frequency response function – Hik( f ) of the ground for the case study was
derived via experimental impulse seismic method (ISM) test data, from which
elastic and attenuation parameters of the ground were obtained, too.
(iv) The calculation results of the predicted IBM Data Centre building dynamic
response using the relevant input experimental data are introduced, too. The
relevant calculated data values following from spectral and amplitude analysis
of the predicted IB dynamic response (spectral picks limit, vibration levels,
etc.) were compared with relevant standards prescription values and criteria
(IBM Corporate Standard C–S1, STN EN 1998 – 1/NA/Z1 (2010) Slovak
Standard STN 73 0032 (2013), etc.). From this comparisons it follows that all
relevant standards prescription values and criteria in future will be fulfilled.
Acknowledgement We kindly acknowledge the project “Research Centre of University of

Zilina”- ITMS 26220220183, supported by European regional development fund and Slovak
state budget.
References
Benčat J (1992) Microtremor due to traffic. In: Research report A – 4 – 92/b, UTC Žilina, 1992
(in Slovak)
Benčat J (2006) Microtremor from railway traffic. In: Proceedings of the 8th international
conference on Computational Structures Technology, Gran Canaria
Benčat J et al (2009) Studies of the TEN–T railway traffic effects on IBM Data Centrum – ST
Building in Bratislava. In: Report PC 16/SvF/2009, University of Žilina (in Slovak)
Bendat J, Piersol A (1993) Engineering applications of correlation and spectral analysis. Wiley,
New York
Ford R (1987) The production of ground vibrations by railway trains. J Sound Vib 116(3):585–589
Fujikake T (1986) A prediction method for the propagation of ground vibration from railway
trains. J Sound Vib 111(2):357–360
Grassie S, Cox S (1984) The dynamic response of railway track with flexible sleepers to high
frequency vertical excitation. Proc Inst Mech Eng 198(2):117–124
Grundmann H et al (1999) The response of a layered half–space to traffic loads moving along its
surface. Arch Appl Mech 69:55–67
Knothe K, Wu Y (1998) Receptance behaviour of railway track and subgrade. Arch Appl Mech
68:457–470
Kotrasová K (2009) Influence of category of sub-soil on liquid storage circular tanks during
earthquake. In: Proceedings of 12th international scientific conference on structural mechanics.
Brno University of Technology. In: Selected Scientific Papers, Journal of Civil Engineering,
TU Košice (Slovakia), 3(1). ISSN 1336-9024.
Slovak Standard (2013) Calculation of building structures and foundations loaded by dynamic
effect of machines (in Slovak). STN 73 0032. Slovak Institute of Standards, SUTN, Bratislava,
2010
Slovak National Annex to Eurocode 8 (2010) Design of structures for earthquake resistance. Part
1: General rules, seismic actions and rules for buildings (in Slovak). STN EN 1998 – 1/NA/Z1.
Slovak Institute of Standards, SUTN, Bratislava
Turek J (1992) The interaction model of the vehicle–track system. In: Research report 245/92 VUŽ
Prague
Van den Broeck, P, De Roeck G (1999) The vertical receptance of track including soil–structure
interaction. In: Proceedings of the 4th European conference on structural dynamic: “Eurodyn
99”, Prague, pp 849–853
Chapter 27
Stability Analysis of Żelazny Most Tailings
Dam Loaded by Mining-Induced
Earthquakes
Waldemar Świdziński, Aleksandra Korzec, and Kinga Woźniczko
Abstract The paper presents the stability analysis of Żelazny Most tailings dam
subjected to mining-induced paraseismic events. The tailings dam as well as
foundation have complex geotechnical structure. The foundation of the dam is
built of series of tertiary and quaternary formations affected in the past by three
glaciations whereas the complexity of tailings structure results from the method of
its deposition within the pond. Simplified dynamic displacement analysis of the
dam is performed based on Newmark’s approach using GeoStudio2007 software.
The paper highlights the necessity of proper domain and time discretization,
especially in the case of complex geological structure. Dynamic response of the
tailings dam was obtained for the equivalent linear one-phase soil model. Stability
calculations were performed for Coulomb-Mohr model and Stress History And
Normalized Soil Engineering Properties (SHANSEP). The applied input signals
were selected by Arias intensity criterion from huge set of signals recorded on the
ground surface caused by mining tremors. The peak ground acceleration value was
scaled to the values predicted by hazard analysis. The results of the calculations
carried out for three selected acceleration signals have revealed no permanent
deformations.
Keywords Earth structure • Dynamic response • Dynamic properties • Stability

analysis • Newmark’s method
W. Świdziński • A. Korzec (*)

Institute of Hydro-Engineering Polish Academy of Sciences, ul. Kościerska 7,
80-328 Gdańsk-Oliwa, Poland
K. Woźniczko
DHV HYDROPROJEKT Sp. z o.o., ul. Dubois 7, 00-182 Warszawa, Poland

DOI 10.1007/978-3-319-14246-3_27
304 W. Świdziński et al.
27.1 Introduction
Żelazny Most tailings dam is the only place to store huge amount of post-flotation
mineral material coming from copper production in KGHM Polska Miedź S.A. Thus
it is a key component of the production process. The depository is built by upstream
method. The tailings-water mixture is discharged inside the depository surrounded by
embankments. It needs to be continuously raised to accommodate subsequent batches
of tailings. At present the maximum height of the depository is 66 m (East Dam).
Other dams are lower due to natural morphology of the terrain on which they are
founded. The development of the depository is based on observational method which
needs well developed monitoring system to observe the response of the dams and
foundation against increasing loadings. The monitoring system consist of thousands
of various measuring points in which different measuring devices are installed (e.g.:
benchmarks, vibrating-wire piezometers, deep inclinometers, seismic stations),
Jamiolkowski et al. (2010). Due to extreme geotechnical complexity of both tailings
and foundation, numerous laboratory and filed tests have been done.
The region of the Żelazny Most depository is characterized by very low natural
seismicity Guterch (2009), however due to mining activity it is exposed to mining-
induced tremors with close epicentral distances Lasocki (2011), Lasocki
et al. (2012a) and Zembaty (2004), Fig. 27.1. Accelerometers that record three
components of earthquake vibrations, are located in six cross-sections on the slope
and at the toe of the dam, Fig. 27.1. Seismic hazard analysis has been carried out by
Fig. 27.1 Regions of epicentres near the Żelazny Most depository, based on Lasocki (2011) and
Lasocki et al. (2012a)
27 Tailings Dam Affected by Paraseismic Events 305
Lasocki et al. based on collected records the seismic energy of which was greater than
1∙107 J. For the mining activity plan in the period 2012–2042, the forecast of peak
horizontal ground acceleration value (PHA) for one of the West Dam cross-sections
was estimated to be 0.17 g, Lasocki et al. (2012b). This values with 5 % of
exceedance over a period of 30 years has been assumed as design acceleration.
High risk due to social, economic and environmental consequences of failure requires
reliable dynamic stability analysis of the dams loaded by paraseismic events.
According to Eurocode 8 - EC8 (2003), pseudo-static stability analysis may be
adopted for PGA lower than 0.08 g. Otherwise pseudo-dynamic or full dynamic
analyses should be applied. One of the most popular pseudo-dynamic approaches is
Newmark’s concept which assess the dam stability in terms of permanent displace-
ments produced during shaking, Newmark (1965). The original concept considers
rigid mass sliding along surface when rigid perfectly plastic strength criterion is
reached. The yield acceleration value that causes relative motion corresponds to
factor of safety equal to one. The relative downward displacement occurs as long as
relative velocity is non zero. Nowadays the modified Newmark’s approach based
on stress state derived from finite element method is being used.
In the paper the application of modified Newmark’s approach to assess the
stability of Żelazny Most tailings dam subjected to paraseismic loadings is
presented. Simplified dynamic displacement analysis has been performed using
GeoStudio2007 software reported in GEO-SLOPE (2010a, b). Dynamic response of
the tailings dam is calculated based on the equivalent linear one-phase soil model.
The attention was focused on the determination of differentiated material stiffness
in order to reflect the soil structure complexity and also taking into account its stress
and strain dependency. The paper highlights the necessity of proper domain and
time discretization, as well as the numerical efficiency, which is not trivial issue in
the case of complex geological structure.
27.2 Geotechnical Characterization of Żelazny Most

Tailings Dam
One of typical cross-sections of the Żelazny Most depository dams is shown in

Fig. 27.2. It consists of starter dam (2) and embankment (1), which create some kind
of external shell for inside deposited tailings and they are made in the form of
construction fill. Additionally, in order to improve the overall stability, in some
parts of the depository the loading berms have been constructed (5). Hydraulically
transported tailings are discharged from the top of the dam crest creating a beach
that becomes the foundation for future embankments and deposited tailings. In the
process of the gravitational flow the segregation of the tailings occurs, due to which
the coarser materials (3) deposit close to embankment whereas the finer ones
(4) flow towards the pond. Due to free water infiltration extensive mass of tailings
if fully saturated, (saturated materials below the phreatic line have been marked as
Fig. 27.2 Typical cross-section of ZM tailings depository
20 , 30 , 40 ). The results of Cross-Hole tests have revealed large variability of tailings

stiffness within the distance of 240 m from the embankment. Thus three additional
zones (4a, 4b, 4c) have been distinguished for numerical simulations. The Żelazny
Most tailings pond is founded on a very complex soil structure consisting of
Tertiary (Tr) and Quaternary (Q) clays and sands affected by the three glaciations.
Dynamic calculations have been performed for equivalent linear model which
takes into account hysteretic soil behaviour during cyclic loading and soil stiffness
dependence on shear strain level. The hysteretic loop allows for the determination
of a tangent shear modulus (equivalent modulus) and a damping ratio. The formula
given by Ishibashi and Zhang (1993) was adopted in Quake/W to model normalized
shear modulus degradation curves G/Gmax depending on confining pressure ( p0 ) as
well as plasticity index (PI) thus it is applicable for both cohesive and cohesionless
soils. G/Gmax curves were fitted based on experimental TX shearing test results for
Tertiary clays and tailings, which is shown in Fig. 27.3a.
It is also well known that soil stiffness depends on stress state. Thus the
maximum shear modulus Gmax depending on vertical effective stress is considered
by a modified version of the formula given by Seed and Idriss for non cohesive soils
and a function proposed by Hardin for cohesive soils, Ishihara (2003). Respective
curves for Tertiary clays and tailings (3, 4) were approximated based on numerous
series of shear wave velocity measurements for undisturbed soils samples, made in
TX with piezoelectric bender elements, at different level of confining stress,
Fig. 27.3b.
Due to the lack of direct measurements of internal soil damping (according to the
recommendation of EC8), constant damping ratio of 3 % has been assumed for
foundation soils and 5 % for tailings. The Poisson ratio for dynamic calculations
was defined as 0.495 for all soils under phreatic line which is the closest allowable
approach to the constant volume condition in computational model.
Strength of the majority of soils in stability analysis has been modelled by
Coulomb-Mohr yield condition, excluding fine tailings under phreatic line (40 ) for
which linear variation of strength with mean effective stress was assumed
(SHANSEP model). Angle of internal friction of the dam’s materials is equal to
34 and of foundation soils is varying from 17 to 34 .
Fig. 27.3 (a) Normalized shear modulus against shear strain for tailings compared to laboratory
tests. (b) Maximum shear modulus dependence on effective vertical stress
27.3 Dynamic Response of the ZM Dam
The 2D plane strain dynamic response of the dam was carried out based on finite
element software Quake/W and Wilson-θ time integration method (GeoStudio2007
package). The size of the model was tested to minimize the boundary impact and
wave reflections. Suggested geometry of the model dimensions is 5:1 (horizontal to
vertical), Świdziński and Korzec (2013). The discrete model consist of 9571 trian-
gular elements with 3 Gauss points. The mesh scheme is consistent with dynamic
material properties and one period of shear wave described by at least five nodes. The
conditions ensure that the distance between nodes is lower than the distance that the
wave can travel in one single time step. Proposed approach gives almost constant
value of Courant number in whole model. Due to memory and time restrictions only
selected dynamic results are saved. Saving frequency must fulfil the Nyquist’s
criterion. After many tests the dynamic calculation times step was established as
0.013 s and saving time step as 0.052 s, Świdziński and Korzec (2013).
Fig. 27.4 Response spectra of selected signals compared to elastic spectrum proposed in EC8
The first stage in the stability analysis using FEM is to establish the initial static
equilibrium stress state within the dam mass before earthquake. In this stage the
initial shear modulus dependent on vertical effective stress is set for each element.
For the sake of simplicity the stage construction and overconsolidation was
neglected. Subsequently, the simplified one-phase dynamic calculations are done
using equivalent linear procedure. Due to these assumptions, only elastic dynamic
strains are expected and the pore pressure is not analysed. The model was loaded by
three uniform horizontal acceleration time-histories that results from deconvolution
procedure, Kramer (1996) and Dulińska (2012). Deconvolution allows to gain
recorded free surface motions at the toe with designed PHA. The linear chirp signal
which covers interested frequency range with the same amplitude was used to
calculate the transmittance function. Accordance to EC8 recommendations, three
accelerograms with waveforms representing near, medium, far field conditions
have been applied. Peak acceleration ratio of these groups of signals has been
assumed to be 1.0:0.5:0.3. An Arias intensity criterion has been used to select one
signal from each group. Response spectra of selected accelerograms are quite
compatible with type 2 of elastic response spectrum Se given in EC8 for the ground
type C, Fig. 27.4. To improve calculation efficiency, the important part of signal
(denoted by ta) was extracted using period criterion that covers an increase of Arias
intensity from 1 to 95 %. Non-standard lower limit has been applied for the
numerical stability purpose. The baseline correction and band-pass filtration were
applied for the selected signals to ensure zero dynamic displacement at the end of
vibration, reducing noise and high frequencies with very low amplitude, Boore and
Bommer (2005).
There are two main steps of Newmark safety analyses using Slope/W. At first, an
average acceleration history is determined based on shear forces mobilized along
slip surface (from both static and dynamic FEM analysis), divided by the mass of
sliding block. Next, the factor of safety is determined as the ratio of the available
static resistance shear force to mobilized shear force along a slip surface. Knowing
time-history of both the average acceleration and the factor of safety, yield accel-
eration ay is determined. For the instability time periods (when the average
acceleration is greater than ay) the positive relative average accelerations of the
block arise. By double integration of the latter time-history, the relative velocity
and then relative displacement time-history are determined. Around one hundred
various slip surfaces were examined covering all admissible failure mechanisms;
from shallow to deep surface, passing through the soft clays, and the boundary
between saturated and unsaturated tailings.
27.4 Selected Calculation Results
Calculations of permanent displacements based on modified Newmark’s approach

have revealed that for predicted future mining induced seismic activity no perma-
nent displacements occur. In order to verify and validate the model some addi-
tional calculations for scaled acceleration signals (design PHA multiplied by 3),
applied to the base, have been carried out. Many various tests were performed and
only selected results are presented. The main conclusions from validity tests and
parametric study are as follows:
1. higher horizontal accelerations occur at the top part of the dam, however higher
vertical accelerations are located in the middle part of the dam body;
2. higher initial shear modulus for the defined soils stronger response and higher
wave reflection effect;
3. the permanent displacement for rigid base case should be calculated only for
strong input signal duration;
4. the deconvolution procedure showed that due to horizontal loading applied to the
base, vertical acceleration component appears which is mostly caused by differ-
entiated elevation of model (the dam);
5. spatial variability of tailings should be considered, giving 30 % larger permanent
deformation and different failure mechanism, Fig. 27.5.
Fig. 27.5 The impact of tailings stiffness on the permanent displacements (G4 – constant stiffness
of tailings; finl_G4var, fin4_G4var – various stiffness of tailings depending on its distance from
the dam crest)
27.5 Conclusions
In the paper the assessment of the Żelazny Most tailings dam stability subjected to
mining-induced paraseismic loading has been presented. Due to the fact that
predicted peak ground accelerations for the mining plan up to 2042 are larger
than given by EC8 allowable limit for application of pseudo-static approach, the
assessment was carried out by simplified dynamic Newmark’s analysis.
The Newmark approach coupled with FEM was chosen for stability analysis.
The adopted methodology assumes one-phase equivalent linear shearing behaviour
and it is correct for predicted PHA level and low cyclic mobility of saturated
tailings. In the numerical analysis made by GeoStudio software the attention was
focused on how to reflect the material complexity of dam and foundation. Such
complexity required a determination of many material properties, which was based
on the interpretation of large series of laboratory test results. The correctness and
the efficiency of discrete model was proved by series of numerical tests. The proper
simulation of the design motion was achieved by deconvolution procedure using
linear chirp signal. The work should be continued to model compliant base case and
check its influence to dynamic response.
For predicted design peak ground acceleration derived for the mining plan
period from 2012 to 2042 no permanent displacements should occur.
Acknowledgments The authors express their thankfulness to the Management of KGHM for
publishing permission. The authors are also grateful for useful advices given by prof. S. Lasocki
and Geoteko specialists (especially by dr inż. P. Sorbjan).
References
Boore DM, Bommer JJ (2005) Processing of strong-motion accelerograms: needs, options and
consequences. Soil Dyn Earthq Eng 25:93–115
Dulińska J (2012) Ziemne budowle hydrotechniczne na terenach sejsmicznych i parasejsmicznych
w Polsce. Wybrane aspekty modelowania i obliczeń. Wydawnictwo PK, Krak ow
Eurocode 8 (2003) Design of structures for earthquake resistance. Part 5: Foundations, retaining
structures and geotechnical aspects. Final Draft. CEN
GEO-SLOPE International Ltd. (2010a) Dynamic modeling with QUAKE/W 2007. An Engineer-
ing Methodology
GEO-SLOPE International Ltd. (2010b) Stability modeling with SLOPE/W 2007 version. An
Engineering Methodology
Guterch B (2009) Sejsmiczność Polski w świetle danych historycznych. Przegla˛d Geologiczny
57(6):513–520
Ishibashi I, Zhang X (1993) Unified dynamic shear moduli and damping ratios of sand and clay.
Soils Found 33(1):182–191
Ishihara K (2003) Soil behaviour in earthquake geotechnics. Clarendon Press, Oxford
Jamiolkowski M, Carrier WD, Chandler RJ, Hoeh K, Swierczynski W, Wolski W (2010) The
geotechnical problems of the second world largest copper tailings pond at Zelazny Most,
Poland. Geotech Eng J SEAGS AGSSEA 41(1):1–15
Kramer SL (1996) Geotechnical earthquake engineering. Prentice-Hall, Upper Saddle River
Lasocki S (2011) Wyb or rejestracji sejsmometrycznych zgodnie z zaleceniami Zespołu Ekspertow

Mie˛dzynarodowych. IGF PAN, Warszawa
Lasocki S, Orlecka-Sikora B, Mirek J, Zarzycka E, Kozłowska M (2012a) Nadz or nad pomiarami
oddziaływania sejsmicznego eksploatacji g orniczej w zakresie sieci obserwacyjnej zap or
OUOW „Żelazny Most” oraz stacji monitoruja˛cych zachodnie przedpole OUOW w 2012
roku. IGF PAN, Krak ow
Lasocki S, Popiołek E, Zorychta A, Orlecka-Sikora B, Sopata P, Stoch T, Urban P (2012b)
Szczegołowa prognoza oddziaływania wstrza˛s ow indukowanych działalnościa˛ g ornicza˛ i
deformacjami powierzchni terenu na OUOW „Żelazny Most” z uwzgle˛dnieniem jego
rozbudowy oraz rozwoju wydobycia złoża rudy miedzi do roku 2042. IGF PAN, Warszawa
Newmark NM (1965) Effects of earthquakes on dams and embankments. Geotechnique
15(2):139–160
Świdziński W, Korzec A (2013) Opracowanie procedury projektowania zap or Obiektu Unieszko-
dliwiania Odpadow Wydobywczych (OUOW) Żelazny Most do wyższych rze˛dnych metoda˛
elementow skończonych oraz uproszczona˛ metoda˛ Newmarka z uwzgle˛dnieniem obcia˛żeń
parasejsmicznych oraz zaleceń Eurokodu 8. IBW PAN, Gdańsk
Zembaty Z (2004) Rockburst induced ground motion – a comparative study. Soil Dyn Earthq Eng
24:11–23
Part III
Seismic Control and Monitoring of
Irregular Structures
Chapter 28
Optimal Drift and Acceleration Control of 3D
Irregular Buildings by Means of Multiple
Tuned Mass Dampers
Yael Daniel and Oren Lavan
Abstract This chapter presents a formal optimization methodology for the seismic
design of multiple tuned-mass-dampers (MTMDs) for the multi-modal control of
3D irregular buildings. The total weight of all TMDs is minimized while both inter-
story drifts and total accelerations are constrained to allowable values so as to lead
to a performance-based-design. The results reveal that, with the right design,
MTMDs can mitigate both structural and nonstructural earthquake damage.
Hence, they can potentially present a multi-hazard strategy to mitigate both winds
and earthquakes.
Keywords Irregular structures • Multiple tuned mass dampers • Multi modal •

Control • Seismic design of tall buildings • Acceleration control
28.1 Introduction
The main criterion in seismic design under strong ground motions (life safety) has
long been limiting the amount of casualties following a severe earthquake. How-
ever, the financial consequences of recent ground motion (e.g. Northridge 1994;
Kobe 1995; Christchurch 2011) led to the notion that financial criteria should also
be considered. This understanding inspired the development of the performance-
based design (PBD) philosophy, where limiting damage under less severe ground
motions is also considered (see e.g. Fajfar and Krawinkler 1997, or Priestley 2000,
for elaboration on the PBD concept). This also motivated retrofitting of existing
structures to limit damage following ground motions of various levels.
Y. Daniel
AJS Consulting Engineers, Haifa, Israel
O. Lavan (*)
Faculty of Civil and Environmental Engineering, Technion – Israel Institute of Technology,
Haifa, Israel

DOI 10.1007/978-3-319-14246-3_28
316 Y. Daniel and O. Lavan
One of the advanced means for retrofitting such structures relies on tuned-mass
dampers (TMD) (e.g. Den-Hartog 1940; Soong and Dargush 1997). Such devices
can efficiently reduce the response of a linear system to a harmonic loading over a
specific narrow band of frequencies. Therefore, the application to wind design is
straight-forward (e.g. McNamara 1977; Luft 1979; Wiesner 1979). Under seismic
events however, elastic behaviour is often experienced by tall buildings as well as
buildings equipped with energy dissipation systems (Kasai et al. 2012). Here,
limiting absolute acceleration, in addition to the traditional limitations on inter-
story drifts, is of much importance (Kasai et al. 2012). As several modes often
contribute to these responses, multiple TMDs (MTMDs), tuned to various frequen-
cies have an advantage over a single TMD controlling only one mode of vibration
(Clark 1988; Lavan and Daniel 2013; Daniel and Lavan 2014, 2015). This makes
such devices attractive for multi-hazard mitigation of winds and earthquake
hazards.
This chapter presents the main derivations and findings of Daniel and Lavan
(2014). An optimization problem formulation suitable for the optimal design of
such MTMDs in 3D irregular structures is first presented. The total mass of
MTMDs is minimized, while local responses of interest of the peripheral frames
are constrained to allowable values, each one individually. Hence, excessive local
damage, which usually characterizes irregular structures, may be prevented, and
elastic response could be ensured. Similarly, absolute accelerations at the perime-
ters of each floor are also constrained to allowable values, allowing the reduction of
nonstructural damage levels. In turn, the derivation of a methodology that allows a
formal optimal design for this PBD problem is presented. This methodology
requires a relatively small computational effort, without predefining the amount
of added devices, their locations, modes to be dampened, or sizes. An example of a
design for retrofitting of an eight story asymmetric structure using the proposed
methodology is presented, and its characteristics are thoroughly discussed.
28.2 Problem Formulation
28.2.1 Equations of Motion and Their Solution Using

Lyapunov’s Equation
The equations of motion of a general 3D linear asymmetric building were formu-

lated and represented in state-space notation (see e.g. Soong 1990). Furthermore, a
stochastic description of the ground motion hazard was adopted. This enabled a
very efficient computation process by using Lyapunov’s equation (see
e.g. Kwakernaak and Sivan 1972), together with a more realistic representation of
the ground motion input, by filtering a white noise input (Daniel and Lavan 2014).
28 Optimal Control of Buildings Using MTMDs 317
28.2.2 Performance Measures
Inter-story drifts are well known to describe damage to structural elements, and
therefore by limiting their value, the damage to these important elements can be
effectively reduced (e.g. Williams and Sexsmith 1995; FEMA 356 2000; Charmpis
et al. 2012). Inter-story drifts are a damage measure to some types of nonstructural
components as well (e.g. partition walls). It is also important to note that by
constraining the level of drifts, the structure can be brought to behave linearly,
thus eliminating the problems associated with structural yielding and detuning of
TMDs. When considering damage to other types of nonstructural elements
(e.g. piping systems, air-conditioning systems, sensitive equipment), absolute
accelerations are an important measure. Thus, these two parameters were chosen
herein to describe the various aspects of seismic structural performance
enhancement.
The measure of cost of TMDs is taken here as the amount of added mass. As
more mass is added to the structure, the retrofit becomes more expensive and thus
less cost-effective. Thus, this is taken as the objective function to be minimized.
28.2.3 Optimization Problem
In the problem to follow, TMDs are potentially allocated at each peripheral location
(i.e. at the four edges) of each floor. At each location, multiple TMDs could be
assigned, each tuned to a different frequency, aimed to control a different mode of
the structure. The optimization problem is formulated so as to minimize the total
amount of added mass in all TMDs while constraining root mean-square responses
of the structure. These responses are measured at all peripheral locations of all
floors, as they are the largest expected within story limits. These locations, that are
also the potential locations for the TMDs, are shown in Fig. 28.1, for the story n, as:
(xpyl)n, (xpyr)n, (xpxt)n and (xpxb)n, and are the peripheral coordinates in the “y”, “y”,
“x” and “x” directions, at the left, right, top and bottom edges of floor n,
Fig. 28.1 Definition of (xpxt)n

dynamic DOFs and
peripheral coordinates of
the nth floor (Daniel and (dy)n
Lavan 2014)
(xpyl)n
(xpyr)n
(dθ)n (dx)n
(xpxb)n
respectively. The responses at these locations can be found by using the transfor-
mation matrix, T, from the DOFs of the original structure ((dx)n, (dy)n and (dθ)n in
Fig. 28.1) to peripheral coordinates.
The optimization problem is thus formulated as:
all all
X frequencies
locations X
min J ¼ ðmTMD Þl, f
mTMD
l f ð28:1Þ
s:t:
RMS r p l
1:0 8l ¼ 1, 2, . . . , N locations
rRMS
all, l
where (mTMD)l,f is the mass of the TMD located at peripheral location l tuned to
frequency f, RMS((rp)l) is the root mean-square of the response of interest at each
location, l, (the lth term of RMS(rp) (such reference to a component of a vector or a
matrix, (·)l, will be used throughout this chapter), rRMS
all;l is the allowable root mean-
square response at the location l, and Nlocations is the number of locations to be
constrained (¼4Nfloors where Nfloors is the number of floors). Generally, the design
variables in problems such as the one presented in Eq. (28.1) can be the masses of
the TMDs, their stiffnesses and damping coefficients. Optimal frequencies (stiff-
nesses) and damping ratios (damping coefficients) for SDOF systems as a function
of their mass were proposed by Den Hartog (1940). These could be applied to each
mode separately as was proposed by Lavan and Daniel (2013). These are also
adopted herein, thus, the masses of TMDs remain the only design variables.
28.3 Solution Scheme
A first order optimization method was adopted to solve to optimization problem.

The required gradients were efficiently derived analytically. All constraints were
first normalized by their allowable values and combined to a single constraint on
their maximum value. The adjoint method was then utilized to efficiently derive its
gradient. For further details of these derivations the reader is referred to Daniel and
Lavan (2014).
28.4 Example
The following eight-story asymmetric setback reinforced concrete (RC) frame

structure (Fig. 28.2) introduced by Tso and Yao (1994) is retrofitted using
MTMDs. This structure was also adopted by Lavan and Daniel (2013) and
y
5
6.0
6
θ
x
6.0 ag
7
6.0
8
1 2 3 4
3.5
3.5
3.5
3.5
3.5
3.5
3.5
3.5
9.0 6.0 9.0
Fig. 28.2 Eight-story setback structure (Daniel and Lavan 2014)
retrofitted using MTMDs to limit total accelerations. Here, only the “y” direction
component of the excitation is considered. A uniform distributed mass of 0.75 ton/
m2 was taken. The column dimensions are 0.5 m by 0.5 m for frames 1 and 2 and
0.7 m by 0.7 m for frames 3 and 4. The beams are 0.4 m wide and 0.6 m tall. Five
percent Rayleigh damping for the first and second modes was used. The design
variables are the locations and properties of the individual tuned mass dampers. The
dampers are to potentially be located on the peripheral frames, where they are most
effective, and as the excitation is in the “y” direction only, dampers will be assigned
only to the peripheral frames 1 (lower 4 floors), 3 (upper 4 floors) and 4, to dampen
frequencies of modes which involve “y” and “θ”. The response is analysed under a
Clough-Penzien filtered Kanai-Tajimi power-spectral density with parameters
suited to match the SE 10/50 ground motion ensemble.
This example is solved for two constraints simultaneously: a constraint on RMS
absolute accelerations and RMS inter-story drifts. Herein, it is desired to reduce the
maximal responses by 45 % (i.e. the allowable RMS acceleration is 55 % of the max
RMS acceleration of all locations at the bare structure’s response, and the allowable
RMS inter-story drift is 55 % of the max RMS inter-story drift of all locations at the
bare structure’s response).
Table 28.1 Natural frequencies of the structure in the coupled y,θ direction
Mode # Angular frequency (rad/s) Mode # Angular frequency (rad/s)
2 7.36 9 43.48
3 10.37 11 57.41
5 17.88 12 67.27
6 22.61 13 71.95
8 35.96 15 94.66
Table 28.2 Final properties of added TMDs

Frame Floor Mode to dampen Mass (ton) Stiffness (kN/m) Damping ratio
1 4 3 8.33 841.46 0.1041
3 8 2 122.44 4490.50 0.2120
3 8 3 15.05 1520.62 0.1041
3 8 5 3.68 1140.16 0.0759
4 8 2 18.99 696.47 0.2120
The first few natural frequencies and modes, associated with the “y” translation
(and “θ” rotation) of the building, are shown in Table 28.1.
A 45 % reduction of both the absolute acceleration and the peripheral peak total
inter-story drift in the bare structure is desired. 160 TMDs were added, as a first
guess. These are comprised of 10 dampers each tuned to a different natural
frequency (related to the “y” and “θ” modes) at each of the 16 peripheral locations
of frames 1, upper 4 floors of frame 3, and frame 4. The initial frequencies and
damping ratios were computed based on the initial masses of TMDs as per Den
Hartog (1940). Using the Active Set algorithm for optimization, the mass con-
verged to the final stiffness using 117 function/gradient evaluations. The final
solution attained is given in Table 28.2.
Finally, an analysis of the retrofitted structure yields the peripheral RMS inter-
story drifts and accelerations shown in Fig. 28.3. Also shown in Fig. 28.3 is the total
amount of mass at each floor. As can be seen, in this case both constraints
(i.e. accelerations and drifts) are active at the same time.
The convergence of the constraint as well as the performance index can be seen
in Fig. 28.4.
28.5 Conclusions
This chapter presented a formal optimization methodology for the seismic

retrofitting of 3D irregular structures. The proposed formulation minimizes the
total mass of added MTMDs while constrains inter-story drifts and total accelera-
tions to allowable values, thus limiting both structural and non-structural damage
a b
8 8 8 8 8 8
7 7 7 7 7 7 dRMS
all
6 6 6 6 6 6
floor number
floor number
floor number
floor number
floor number
floor number
5 5 5 5 5 5
aRMS
all
aRMS
4 4 4 4 4 all 4
dRMS
all
3 3 3 3 3 3
2 2 2 2 2 2
1 1 1 1 1 1
0 50 100 150 0 10 20 0 0.05 0 50 100 150 0 10 20 0 0.05
sum of RMS acceleration RMS drift sum of RMS acceleration RMS drift
masses (ton) masses (ton)
Fig. 28.3 Peripheral RMS accelerations and inter-story drifts of structure with final TMDs
(continuous or dashed) and sum of added masses (dots) (a) frame 1 (floors 1–4) and 3 (floors 5–
8) and (b) frame 4 (Daniel and Lavan 2014)
a
300
total added mass
250
200
150
100
0 20 40 60 80 100 120
function evaluation
b 4
Constraint value
0
0 20 40 60 80 100 120
function evaluation
Fig. 28.4 Convergence of (a) performance index and (b) objective function, constraining absolute
accelerations and inter-story drifts (Daniel and Lavan 2014)
levels. As responses of interest are constrained, the methodology could be used in

the context of Performance-Based Design. It is believed by the authors that with the
formulation of the optimization problem and the tools derived for its solution to
assist in the design, a door is opened for a wider use of MTMDs for the seismic
retrofitting of buildings.
References
Charmpis DC, Komodromos P, Phocas C (2012) Optimized earthquake response of multi-storey

buildings with seismic isolation at various elevations. Earthq Eng Struct Dyn 41(15):
2289–2310
Clark AJ (1988) Multiple passive tuned mass dampers for reducing earthquake induced building
motion. In: Proceedings of the 9th world conference on earthquake engineering, Tokyo-Kyoto,
Japan, vol V, pp 779–784
Daniel Y, Lavan O (2014) Gradient based optimal seismic retrofitting of 3D irregular buildings
using multiple tuned mass dampers. Comput Struct 139:84–97
Daniel Y, Lavan O (2015) Optimality criteria based seismic design of multiple tuned-mass-
dampers for the control of 3D irregular buildings. Earthq Struct 8(1):77–100
Den-Hartog JP (1940) Mechanical vibrations, 2nd edn. McGraw-Hill Book Company, New York
Fajfar P, Krawinkler H (eds) (1997) Seismic design methodologies for the next generation
of codes. Balkema, Rotterdam
FEMA 356 (2000) Pre-standard and commentary for the seismic rehabilitation buildings.
Federal Emergency Management Agency, Washington, DC
Kasai K, Pu W, Wada A (2012) Response of passively-controlled tall buildings in Tokyo during
2011 Great East Japan Earthquake. In: Proceedings of the 15th world conference on earthquake
engineering, Auckland, New Zealand, 2000
Kwakernaak H, Sivan R (1972) Linear optimal control systems. Wiley, New York
Lavan O, Daniel Y (2013) Full resources utilization seismic design of irregular structures using
multiple tuned mass dampers. Struct Multidiscip Optim 43(3):517–532
Luft RW (1979) Optimum tuned mass dampers for buildings. ASCE J Struct Div 105:2766–2772
McNamara RJ (1977) Tuned mass dampers for buildings. ASCE J Struct Div 103:1785–1798
Priestley MJN (2000) Performance based seismic design. In: Proceedings of the 12th world
conference on earthquake engineering, Lisbon, Portugal, 2012. Paper no. 2831
Soong TT (1990) Active structural control: theory and practice. Longman Scientific & Technical,
Harlow
Soong TT, Dargush GF (1997) Passive energy dissipation systems in structural engineering.
Wiley, Chichester
Tso WK, Yao S (1994) Seismic load distribution in buildings with eccentric setback. Can J Civ
Eng 21:50–62
Wiesner KB (1979) Tuned mass dampers to reduce building wind motion. ASCE Convention and
Exposition, Boston, pp 1–21
Williams MS, Sexsmith RG (1995) Seismic damage indices for concrete structures: a state-of-the-
art review. Earthq Spectra 11(2):319–349
Chapter 29
Improved Seismic Performance of RCC
Building Irregular in Plan with Water Tank
as Passive TMD
Suraj N. Khante and Rutuja S. Meshram
Abstract The choices of building shapes and structural systems have significant
effect on their seismic performance. Structures that have non-coincident centers of
mass and stiffness are referred to as plan-irregular structures. Such structures could
be highly vulnerable to earthquake damage due to torsional response. Regular
buildings result in a fairly uniform distribution of seismic forces throughout its
components, whereas irregular buildings result in highly indeterminate distribution
of forces making the analysis prediction more complicated. This paper deals with
analytical investigation of feasibility of implementing water tank as passive TMD
on plan irregular building using SAP2000. The water tank was installed at the
terrace level of L-shaped building. The response of building to earthquake data,
namely, El-Centro 1940 and Bhuj 2001 were studied. The responses of the building
with tank provided at three different locations of building were studied under five
conditions, namely, tank empty, 0.25 h, 0.50 h, 0.75 h and full tank, where, h is the
height of water level. It is concluded that water tank at top roof level in partially
filled condition mitigate response of irregular buildings.
Keywords Plan irregularity • Water tank as Passive TMD • Nonlinear analysis •

SAP2000
29.1 Introduction
To protect civil structures from significant damage, the response reduction during
severe earthquakes has become an important topic in structural engineering. A
building is said to be irregular when it lacks symmetry and discontinuity in
geometry, mass or load resisting elements. There are two types of irregularities
namely, Horizontal irregularities refers to asymmetrical plan shapes (L, T, U and F)
S.N. Khante (*) • R.S. Meshram

Applied Mechanics Department, Government College of Engineering, Amravati, Maharashtra,
India

DOI 10.1007/978-3-319-14246-3_29
324 S.N. Khante and R.S. Meshram
or discontinuities in horizontal resisting elements and Vertical irregularities refer-

ring to sudden change of strength, stiffness, geometry and mass of a structure in
vertical direction.
An ideal multi-storey building designed to resist lateral loads due to earthquake
would consist of only symmetric distribution of mass and stiffness in plan at every
storey and a uniform distribution along height of the building. Such a building
would respond only laterally and is considered as torsionally balanced building. But
it is very difficult to achieve such a condition because of restrictions such as
architectural requirements and functional needs. The issue of mitigating the
response of such irregular structures due to seismic loads has drawn the interest
of many researchers in recent years.
Tuned mass dampers (TMD) have been widely used for the vibration control in
civil engineering structures. The concept of vibration control, using a mass damper,
dates back to the year 1909, when Frahm invented a vibration control device called
a dynamic vibration absorber. Through intensive research and development in
recent years, the TMD has been accepted as an effective vibration control device
for both new and existing structures. On similar lines TLD (Tuned Liquid Dampers)
have also proven themselves in the form of successful passive vibration mitigating
system. It is found to be a simple, effective, inexpensive and reliable means for
suppressing undesirable vibrations of structures caused by seismic excitations. It is
attached to a structure in order to reduce the dynamic response of the structure.
Since the water storage tanks are built-in component of buildings and mostly
these are constructed on the top roof level, hence they add dead load on the
structure. During earthquakes, this extra mass can be employed as damper to take
over the surplus energy transmitted to the structure.
Lin et al. (1994) studied vibration control effectiveness of passive tuned mass
dampers (PTMDs) for irregular buildings. The results verified that PTMDs are
capable of reducing building response effectively. Jaiswal (2004) employed Simple
Tuned Mass Damper to Control Seismic Response of Elevated Tanks. Mehboob
et al. (2013) carried out numerical investigation of water tank as TLD, installed at
the roof level of multi-storey building using ANSYS. Investigation was carried out
by locating water tank off the centre of mass of building and for different water
level conditions. Gulve and Murnal (2013) conducted study on implementation of
water tank as passive tuned mass damper. Three multi-storey regular structures,
three, seven and ten were taken for the study. Hemalatha and Jaya (2008) focused
on the feasibility of providing passive TMD on building when subjected to earth-
quake. The researchers had only varied water levels for analysis.
The studies available in literature mainly concentrate on regular buildings. More
so ever two mass model of water tank is missing in the structural modelling of water
tank. The present study bridges this gap. This study is an attempt to investigate
effect of different earthquake excitation on irregular building with varying height
on maximum displacement. Deliberately two earthquake spectra are employed for
excitation of building because of their typical characteristic signature. As the
purpose of study is to establish effectiveness of existing water tanks as TLD, the
parameters viz. frequency ratio, mass ratio which affect the performance of TLD
were not optimized. The water tanks were provided on bay of 3 m 3 m.
29 Building with Water Tank‐TMD: Seismic Response 325
29.2 Basic Principle
It was thought, modeling of water tank as two mass model will lead to more realistic
dynamic response of the structure of which water tank is integral part. Two mass
model for elevated tank was proposed by Housner (1963) which is more appropriate
and is being commonly used in most of the international codes including IS 1893
(Part-II) (Liquid Retaining Tanks 2000). The pressure generated within the fluid
due to the dynamic motion of the tank can be separated into impulsive and
convective parts. For representing these two masses and in order to include the
effect of their hydrodynamic pressure in analysis two-mass model is adopted for
tanks. In spring mass model convective mass (mc) is attached to the tank wall by the
spring having stiffness (Kc), where as impulsive mass (mi) is rigidly attached to
tank wall. For elevated tanks two-mass model is considered, which consist of two
degrees of freedom system. The two- mass model is shown in Fig. 29.1.
29.3 Building Description
The models of building considered for the analysis are G + 4 (M5), G + 7 (M8) and
G + 12 (M13) storeys RCC structures. The buildings are irregular in plan as shown
in Fig. 29.2. The building has bay width of 3 m in X and Y direction with 3 m storey
height. Slab is modeled as rigid diaphragm. Tuned mass damper in the form of
water tank is installed at three different locations at the top of the building.
a
mc
Kc
b
mi + ms mi + ms
Ks Ks mc
+ Kc
Fig. 29.1 Two mass model

Fig. 29.2 Plan and location of water tank
Table 29.1 Details of the model

Model name No. of floors Dimension of column (m) Dimension of beam (B x D) in m
M5 5 0.30 0.45 0.30 0.40
M8 8 0.40 0.40 0.30 0.40
M13 13 0.45 0.45 0.30 0.45
Non-linear time history analysis is carried out in SAP2000 software (Structural

Analysis Program SAP 2000) using El Centro and Bhuj Earthquake records.
CM (Centre of mass) is location where total mass of the system can be considered
to be located. CR (Centre of rigidity) is the stiffness centroid within a floor
diaphragm plan. When CM do not coincide with CR then eccentricity is created
in structure i.e. distances between CM and CR. CM and CR are calculated by using
ETABS software.
The scope of this study includes investigation of vibrational response of a 3D
model of an irregular building provided with a water tank on the top roof level. Two
random ground acceleration cases corresponding to past earthquake time-histories
were applied to explore the dynamic response of buildings.
Plan irregular buildings i.e. L-shaped buildings were considered for the present
study. Location of center of mass (CM) and center of rigidity (CR) along with
different locations of water tank are shown in Fig. 29.2a, b. Table 29.1 shows their
detail.
The material properties used for the analysis are Grade of concrete – M25,
Poisson’s ratio – 0.16 and Density of concrete – 2500 kg/m3. The horizontal input
ground motions, in the form of acceleration time-history are summarized in the
Table 29.2.
Table 29.2 Time history record summary

Country/record Station/year Magnitude PGA (g)
California/Imperial Valley El-Centro (1940) 6.9 0.32
Bhuj Ahmadabad (2001) 7.0 0.47
Fig. 29.3 Locations of water tank (a) Water Tank at Location I (b) Water Tank at Location II (c)
Water Tank at Location III
Fig. 29.4 Acceleration value for El-Centro earthquake and Bhuj earthquake. (a) El-Centro
Earthquake (b) Bhuj Earthquake
In the present work the water tank was placed at three different locations. The
tank had a plan dimension of 3 3 m and 2 m height for all the models and was
placed over 1 m high columns. The beam-column supports for the tank were
rectangular concrete sections and the walls and roof were modeled as concrete.
Models of the tank with buildings are shown in Fig. 29.3.
Acceleration values for El-Centro and Bhuj earthquakes are shown in Fig. 29.4.
29.4 Results and Discussion
The location of water tank was changed and for each case of tank location, all three
models were investigated for El-Centro and Bhuj earthquake data. The extreme
recorded values of roof displacement are plotted and shown in Figs. 29.5, 29.6, and
29.7. From Fig. 29.5(a) (i) displacement of building M5 with empty tank at location
I is lowest and displacement of building with 3/4th filled tank is adverse. The
maximum displacement reduction of building with empty tank as compared to
building without tank is 7.8 % for location II it is 9.4 % for El-Centro.
Generally it is observed that for buildings without water tank (Figs. 29.6(a)–(i),
(a)-(ii), (b)-(i), (b)-(ii) and 29.7(a)-(i), (a)-(ii), (b)-(i), (b)-(ii)), with empty water
tank (Figs. 29.6(a)-(iii) and 29.7(a)–(iii), (b)–(iii)) and with full water tank
(Fig. 29.5(a)-(ii), (a)-(iii), (b)-(ii), (b)-(iii)) maximum displacement is more. On
the contrary partially filled tank was found to be effective in mitigating response.
29.5 Conclusions
After the numerical investigation of water tank as TMD, for three different building
models (difference in terms of height of building) irregular in plan, subjected to
seismic excitation, following conclusions can be made:
• Water tank at top can be designed to serve as TMD provided the parameters i.e.,
plan location of water tank; water level and mass ratio are tuned properly.
• TMD (tank + water) located near the centre of mass and centre of rigidity of an
irregular building shows maximum response reduction as compared to TMD
(tank + water) located away from it. As the distance between water tank and
centre of mass building increases the response of the building also increases.
• In higher building, maximum percentage reduction in response of plan irregular
building is noted.
a b
Maximum Displacement (mm) for Maximum Displacement (mm) for
water tank location I water tank location I
80 80
w/o water w/o water
70 tank 70 tank
Empty 60 Empty
60
Displacement (mm)
Displacement (mm)
50 0.25h 50
0.25H
40 40
0.50h 0.50H
30 30
20 0.75h 20 0.75H
10 10
Full
0 Full
0
0 1 2 3 4 5 6 0 1 2 3 4 5 6
No. of Storey No. of Storey
(i) (i)

water tank location II water tank location II
80 w/o water 80 w/o water
tank tank
70 70
Empty Empty
60 60
Displacement (mm)
Displacement (mm)
50 50 0.25H
0.25h
40
40
30 0.50H
30 0.50h
20
20 0.75H
0.75h 10
10
0 Full
0 Full 0 1 2 3 4 5 6
0 1 2 3 4 5 6
(ii) (ii)

water tank location III water tank location III
80 w/o water 80 w/o water
tank 70 tank
70
Empty 60 Empty
60
Displacement (mm)
Displacement (mm)
50 50
0.25h 0.25h
40 40
30 0.50h 30 0.50h
20 20
0.75h 0.75h
10 10
0 Full 0 Full
0 1 2 3 4 5 6 0 1 2 3 4 5 6
(iii) (iii)
Fig. 29.5 Displacement of five storey building for El-Centro and Bhuj earthquake. (a) El-Centro
a b
Maximum Displacement (mm) with Maximum Displacement (mm) with
water tank at location I water tank at location I
120
w/o water 120 w/o water
100 tank
tank
Empty 100
Empty
Displacement (mm)
80
Displacement (mm)
80
0.25h
60 0.25h
60
0.50h
40 0.50h
40
0.75h
20 0.75h
20
0 Full
0 Full
0 5 10
0 5 10
(i) (i)

water tank at location II water tank at location II
120 w/o water

120 w/o water
tank tank
100
100
Empty Empty
Displacement (mm)
80
Displacement (mm)
80
0.25h 0.25h
60
60
0.50h 40
0.50h
40
0.75h 20 0.75h
20
0 Full
0 Full
0 5 10
0 5 10
(ii) (ii)

water tank at location III water tank at location III
120
140 w/o water w/o water
tank 100 tank
120 Empty
Empty
Displacement (mm)
100 80
Displacement (mm)
0.25h 0.25h
80 60
60 0.50h 0.50h
40
40
0.75h 20
0.75h
20
Full 0 Full
0
0 5 10 0 5 10
(iii) (iii)
Fig. 29.6 Displacement of eight storey building for El-Centro and Bhuj earthquake. (a) El-Centro
a b
water tank at location I water tank at location I
180 w/o water
160 w/o water
160 tank
140 tank
140 Empty Empty
120
Displacement (mm)
Displacement (mm)
120
100
100
0.25h 0.25h
80
80 0.50h
0.50h 60
60
40
40 0.75h
0.75h 20
20 Full
0
0 Full 0 5 10 15
0 3 6 9 12 15
No. of Storey
No. of Storey
(i) (i)

water tank at location II water tank at location II
180 160
w/o water w/o water
160 tank 140 tank
140 Empty Empty
120
Displacement (mm)
120
Displacement (mm)
0.25h 100
100 0.25h
80 80
0.50h
60 60 0.50h
40 0.75h 40
20 0.75h
20
0 Full
0 Full
0 5 10 15
0 5 10 15
(ii) (ii)
Maximum Displacement in mm for Maximum Displacement (mm) with

water tank at location III water tank at location III
160
180 w/o water
160 w/o water 140 tank
tank Empty
Displacement in mm
140 120
Displacement (mm)
120
Empty 100
100 0.25h
80 80
60 0.25h 60 0.50h
40
40 0.75h
20 0.50h
0 20
Full
1 3 5 7 9 11 13 0
0.75h
0 5 10 15
No. of Stories
No. of Storey
(iii) (iii)
Fig. 29.7 Displacement of thirteen storey building for El-Centro and Bhuj earthquake. (a)
El-Centro Earthquake (b) Bhuj Earthquake
References
Gulve TS, Murnal P (2013) Feasibility of implementing water tank as passive tuned mass damper.
Int J Innov Technol Explor Eng (IJITEE) 3(3):12–19, ISSN: 2278-3075
Hemlatha G, Jaya KP (2008) Water tank as passive TMD for seismically excited structures. Asian
J Civ Eng (Build Hous) 9(4):349–366
Housner GW (1963) The dynamic behavior of water tanks. Bull Seismol Soc Am 53(2):381–387
IS: 1893 (Part-II, Liquid Retaining Tanks) (2000) Criteria for earthquake resistant design of
structures. Bureau of Indian Standards, New Delhi
Jaiswal OR (2004) Simple tuned mass damper to control seismic response of elevated tanks. In:
Proceedings of the 13th world conference on earthquake engineering, Vancouver
Lin CC, Hu C-M, Wang JF, Hu R-Y (1994) Vibration control effectiveness of passive tuned mass
damper. J Chin Inst Eng 17(3):367–376
Mehboob SS, Khan QZ, Tahir F, Ahmad MJ (2013) Investigation of water tank as TLD for
vibration control of frame structures under seismic excitation. Life Sci J 10(7s):1182–1189,
ISSN: 1097-8135
Chapter 30
Behaviour of Asymmetric Structure
with Base Isolation Made of Polymeric
Bearings
Tomasz Falborski and Robert Jankowski
Abstract Earthquake-induced ground motions are the most severe and

unpredictable threats to the structures all around the world. Seismic excitations
cause a lot of damage in a wide variety of ways, leaving thousands of casualties in
their wake. Due to randomness of earthquake occurrence, lack of visible causes and
their power of destructiveness, structural engineers need to develop new technical
solutions and protection systems against earthquake forces and their devastating
effects in order to minimize loss of life and property damage. Accordingly, design-
ing seismic-resistant structures became an issue of great importance in many
seismically active regions of the world. This concerns especially designing asym-
metric structures, since their response under earthquake excitations is much more
complex and often difficult to be predicted comparing with the symmetric ones. The
present paper reports the results obtained from detailed finite-element analysis on a
dynamic response of a base-isolated asymmetric multi-storey steel building under
various seismic excitations. The numerical investigation aims to verify the effec-
tiveness of Polymeric Bearings in suppressing vibrations of asymmetric structures
under strong earthquakes.
Keywords Polymeric bearings • Seismic isolation • Asymmetric structures •

Earthquakes • Numerical analysis
30.1 Base Isolation
Earthquakes produce large-magnitude forces of short duration that must be resisted

by a structure without causing collapse and preferably without any significant
damage to the structural members. It has been gradually being recognized that
various methods of structural control are probably the most effective means of
T. Falborski (*) • R. Jankowski

Faculty of Civil and Environmental Engineering, Gdańsk University of Technology, ul.
Narutowicza 11/12, 80-233 Gdańsk, Poland

DOI 10.1007/978-3-319-14246-3_30
334 T. Falborski and R. Jankowski
protecting structures during seismic excitations (Buckle 2000). They are not only
employed for mitigating earthquake forces, but are also equally useful in control-
ling undesirable structural vibrations produced due to wind loads and other minor
dynamic excitations (Mayes and Naeim 2001).
Base isolation is considered to be one of the most popular passive structural
control techniques of protecting structures against strong ground motions (Naeim
and Kelly 1999). Base isolators, like Lead-Rubber Bearings (LRBs), High-
Damping Rubber Bearings (HDRBs), and Friction Pendulum Systems (FPSs) are
frequently used in practice in many earthquake-prone countries all around the
world.
Base isolation systems work by decoupling the building from the horizontal
components of the earthquake ground motion by interposing a layer with low
horizontal stiffness between the structure and its foundation (Kelly 1993; Skinner
1993). The philosophy behind the concept of base isolation is to lengthen the period
of vibration of the protected structure, so as to reduce the base shear induced by the
earthquake, while providing additional damping to the system (Jankowski 2003;
Mahmoud et al. 2012; Robinson 1982). This is why most seismic design codes
suggest the use of base isolation systems that have the dual function of period
elongation (period shift effect) and energy dissipation (increasing damping effect).
30.2 Numerical Model
Numerical simulation of the seismic behaviour of a multi-storey steel building, both

fixed-base and base-isolated, under various seismic excitations has been performed
using the Finite Element (FE) software MSC Marc® 2008. The FE model of the
fixed-base asymmetric 4-storey steel structure building is composed of six steel
rigid frames made of HEB 300 sections and 25 cm thick reinforced C25/30 concrete
slabs (Fig. 30.1). In order to increase structural stiffness of the building, additional
bracings in two sidewall vertical planes have been used. The spacing between the
columns in the plane of steel frames is 6.0 m. The distance between the steel frames
in longitudinal direction is 4.2 m. The height of each storey is 3.6 m. Accordingly,
the analyzed structure is 12.6 m long, 12.0 m wide and its total height is 14.4 m.
Steel frames as well as the reinforced concrete slabs have been modelled using
4-node shell elements (Zienkiewicz and Taylor 2002) available in MSC Marc®
2008. In order to generate bracings in vertical planes, 2-node truss elements have
been employed. Both columns and girders are made of steel with the following
elastic parameters: Young’s modulus Es ¼ 210 GPa and Poisson’s ratio νs ¼ 0.3.
The material is isotropic and its mass density is ρs ¼ 7850 kg/m3, which corre-
sponds to standard steel. Floor and roof slabs are made of concrete with the
following elastic parameters: Young’s modulus Ec ¼ 31 GPa and Poisson’s ratio
νc ¼ 0.2, which is typical for C25/30 concrete. The material is also isotropic and its
mass density is ρc ¼ 2400 kg/m3.
30 Asymmetric Base‐Isolated Structure with PB 335
Fig. 30.1 FE model of a

fixed-base asymmetric
4-storey steel building
30.3 Polymeric Bearings
As indicated by Fig. 30.2, the analyzed 4-storey building has been decoupled from
the shaking ground with the Polymeric Bearings, which can be considered a
proposal of a new seismic isolation system. It consists of rectangular-shaped blocks
(600 600 400 mm) made of a specially prepared flexible polymeric material
whose chemical composition includes certain additives so as to improve its
damping potential. The basic mechanical properties of this material have been
already determined in experimental studies and the results have been presented in
previous publications (Falborski 2012; Falborski et al. 2012a, b, c). Hysteresis
loops observed during the cycling testing have confirmed its relatively high
damping and energy-dissipation properties, which are extremely desirable for
materials used for seismic isolation devices. Moreover, shaking table experimental
study, using the prototype of the Polymeric Bearings supporting a 1.20 m high
single-storey and a 2.30 m high two-storey steel structure models, has also been
performed and the results have proven the effectiveness of this base isolation
system in suppressing structural vibrations during dynamic excitations (Falborski
and Jankowski 2012, 2013).
Rubber-like materials behave as hyperelastic, incompressible and, from a mac-
roscopic point of view, homogenous and isotropic solids. To analyze rubber
bearings, different analytical methods are available. In the present study, Polymeric
Bearings have been modelled using 8-node hexahedral solid elements available in
Fig. 30.2 FE model of a

base-isolated asymmetric
4-storey steel building
MSC Marc® 2008. The 5-parameter Mooney-Rivlin material has been adopted to
simulate the behaviour of seismic isolation bearings. This hyperelastic model pro-
vides the best fit with the previously obtained experimental results. Besides,
Mooney-Rivlin material model is counted among the most widely adopted consti-
tutive relationships in the stress analysis of polymers (Finney 2001). The following
material constants have been calculated based on the uniaxial tension and com-
pression test data: C10 ¼ 0.889 MPa, C01 ¼ 0.246 MPa, C20 ¼ 0.155 MPa,
C11 ¼ 0.094 MPa, C30 ¼ 0.011 MPa and bulk modulus G ¼ 2.5 GPa.
30.4 Modal Analysis
The first sequence in the numerical investigation has been focused on determination
of the dynamic properties of the 4-storey building, both fixed-base and base-
isolated. The results of the modal analysis are briefly summarized in Table 30.1.
The first dynamic mode of the base-isolated structure involves deformation only in
Polymeric Bearings, while the superstructure remains basically rigid, as indicated
by Fig. 30.3. The fundamental period corresponding to the first mode of the base-
isolated building has been shifted away and it has been calculated to be T ¼ 1.306 s
(f ¼ 0.766 Hz).
Table 30.1 Results of the modal analysis

Vibration mode Natural frequency (Hz) Period of vibration (s)
Fixed-base structure Transverse 1.594 0.627
Longitudinal 2.132 0.469
Torsional 2.736 0.365
Base-isolated structure First mode 0.766 1.306
Fig. 30.3 Vibration mode

of a base-isolated
asymmetric 4-storey steel
building
30.5 Dynamic Analysis
In the second stage of the numerical investigation, the dynamic transient analysis
has been performed to investigate the response of the 4-storey building, both fixed-
base and base-isolated, during a seismic excitation. Both structures, with and
without base isolation system, have been subjected to the El Centro earthquake of
1940 (NS component, PGA ¼ 3.070 m/s2, EW component, PGA ¼ 2.107 m/s2) and
the Northridge earthquake of 1994 (Santa Monica station, NS component,
PGA ¼ 3.628 m/s2, EW component, PGA ¼ 8.664 m/s2). The NS and EW compo-
nents of the ground motions have been applied along the Y and X direction,
respectively. The first 20 s of the accelerograms, with a time step of 0.01 s, have
been employed in this study.
The peak values of the seismic response of the analyzed building are briefly
summarized in Table 30.2. Additionally, the time-acceleration histories recorded
along the Y direction at the top of the structure, both fixed-base and base-isolated,
are presented in Figs. 30.4, 30.5, 30.6, and 30.7.
Table 30.2 Results of the dynamic analysis

Peak acceleration at the top (m/s2)
Base-isolated Fixed-base Reduction
Dynamic excitation structure structure (%)
El Centro earthquake of 1940 4.02 12.74 68.45
Northridge earthquake of 6.50 11.38 42.88
1994
15
fixed-base building
10
Acceleration at top (m/s2)
-5
-10
-15
0 2 4 6 8 10 12 14 16 18 20
Time (s)
Fig. 30.4 Time-acceleration history plot for the fixed-base building during the 1940 El Centro
earthquake (Y direction)
30.6 Results and Conclusions
The effectiveness of Polymeric Bearings in suppressing vibrations of asymmetric

building under seismic excitations has been verified in this paper. The reduction in
lateral response has been measured by comparing the peak accelerations of a
4-storey asymmetric steel building, both fixed-base and base-isolated. The peak
lateral accelerations during the El Centro earthquake of 1940, and the Northridge
earthquake of 1994 have been reduced by over 68 % and 42 %, respectively. The
results obtained from the numerical analysis on seismic response of the analyzed
multi-storey steel structure building indicate that Polymeric Bearings can be suc-
cessfully used to reduce damage of structures during seismic excitations.
Nevertheless, further experimental studies and numerical analyses (including the
nonlinear ones – see, for example, Mahmoud and Jankowski 2009; Jankowski
15
base-isolated building
10
-5
-10
-15
0 2 4 6 8 10 12 14 16 18 20
Time (s)
Fig. 30.5 Time-acceleration history plot for the base-isolated building during the 1940 El Centro
15
fixed-base building
10
-5
-10
-15
0 2 4 6 8 10 12 14 16 18 20
Time (s)
Fig. 30.6 Time-acceleration history plot for the fixed-base building during the 1994 Northridge
15
base-isolated building
10
-5
-10
-15
0 2 4 6 8 10 12 14 16 18 20
Time (s)
Fig. 30.7 Time-acceleration history plot for the base-isolated building during the 1994
Northridge earthquake (Y direction)
2007) are required to fully verify the efficiency of the Polymeric Bearings as a new
base isolation system for asymmetric buildings.
References
Buckle IG (2000) Passive control of structures for seismic loads. In: Proceedings of the 12th world
conference on earthquake engineering, Auckland, 30 January–4 February 2000
Falborski T (2012) Experimental determination of basic mechanical properties of elastomeric
polymer. In: Fijało C, Fijało P (ed) Advances in chemical and mechanical engineering, Gdańsk
University of Technology Publishers, Gdańsk, Poland, pp 147–150
Falborski T, Jankowski R (2012) Shaking table experimental study on the base isolation system
made of polymer bearings, In: Proceedings of the 15th world conference on earthquake
engineering, Lisbon, 24–28 September 2012
Falborski T, Jankowski R (2013) Polymeric Bearings – a new base isolation system to reduce
structural damage during earthquakes. Key Eng Mater 569–570:143–150
Falborski T, Jankowski R, Kwiecień A (2012a) Experimental study on polymer mass used to repair
damaged structures. Key Eng Mater 488–489:347–350
Falborski T, Kwiecień A, Strankowski M, Piszczyk Ł, Jankowski R (2012b) The influence
of temperature on properties of the polymer flexible joint used for strengthening historical
masonries. In: Jasieńko J (ed) Structural analysis of historical constructions. DWE,
Switzerland, pp 816–822
Falborski T, Strankowski M, Piszczyk Ł, Jankowski R, Kwiecień A (2012c) Experimental
examination of an elastomeric polymer. Technol Art 3:98–101
Finney RH (2001) The finite element method. In: Gent AN (ed) Engineering with rubber: how to
design rubber components. Hanser, Berlin, pp 257–305
Jankowski R (2003) Nonlinear rate dependent model of high damping rubber bearing. Bull Earthq
Eng 1:397–403
Jankowski R (2007) Assessment of damage due to earthquake-induced pounding between the main
building and the stairway tower. Key Eng Mater 347:339–344
Kelly JM (1993) Earthquake-resistant design with rubber. Springer, New York
Mahmoud S, Jankowski R (2009) Elastic and inelastic multi-storey buildings under earthquake
excitation with the effect of pounding. J Appl Sci 9:3250–3262
Mahmoud S, Austrell P-E, Jankowski R (2012) Simulation of the response of base-isolated
buildings under earthquake excitations considering soil flexibility. Earthq Eng Eng Vib
11:59–374
Mayes RL, Naeim F (2001) Design of structures with seismic isolation. In: Naeim F (ed) The
seismic design handbook, 2nd edn. Kluwer, Boston, pp 723–755
Naeim F, Kelly JM (1999) Design of seismic isolated structures: from theory to practice. Wiley,
New York
Robinson WH (1982) Lead-rubber hysteretic bearings suitable for protecting structures during
earthquakes. Earthq Eng Struct Dyn 10:593–604
Skinner RI (1993) An introduction to seismic isolation. Wiley, New York
Zienkiewicz OC, Taylor RL (2002) The finite element method. Butterworth-Heinemann, Oxford
Index
A Code provision, 137, 231, 235

Accidental eccentricity(ies), 26, 92, 150, 157 Coefficient of variation (CoV), 149–154, 156,
AFORS. See Autonomous fibre-optic rotational 157, 161–163, 165, 168
seismograph Column failure, 184
Angular motion, 73 Concrete mechanical properties, 150
ASPU. See Autonomous signal processing unit Concrete strength, 139, 150, 151, 153, 156,
Asymmetric-plan building, 4, 8, 9, 11, 123–134 161, 163–165, 168
Asymmetric system, 194, 197, 198, 206, 209, Confidence factor (CF), 137, 150, 162, 186
212, 219, 244, 250–252, 264, 268 Continuous monitoring, 50, 53
Autonomous fibre-optic rotational seismograph Corner displacement amplification, 243–253
(AFORS), 50–54, 56, 60 Corrective eccentricities, 194, 198–201, 216,
Autonomous signal processing unit (ASPU), 217, 222
52, 53, 60 Coupled shear wall, 264, 266, 268, 269
B D
Bare structure, 319, 320 Damping, 5, 20, 37, 38, 43, 66, 71, 89, 93, 118,
Base isolated structure, 336–338 125, 171, 187, 197, 209, 219, 235, 244,
Bidirectional ground motion, 194 247, 248, 251, 256, 265, 271, 293, 294,
Body waves, 36 306, 318, 320, 334, 335
Braced frame, 9, 125, 195, 196, 206, 218, 222 Deformation, 4–7, 29, 50, 81, 125, 130, 150,
Brittle failure, 142 161, 185, 244, 266, 293, 309, 336
Deformed shape, 124, 134
Dense array, 26, 27, 31, 33
C Design variables, 318, 319
Capacity spectrum method, 234, 235 Displacement response, 234, 244, 246
Center of mass, 5, 114, 227, 234–236, 245, Ductility, 173, 176, 178, 179, 186, 271
250, 326 Ductility demand, 176, 178, 179, 186
Center of resistance, 172, 178, 235 Dynamic analysis, 89, 113, 138, 140, 141,
Centre of rigidity, 194, 206, 207, 217, 143, 145, 172, 188, 197–198, 209, 219,
326, 328 236, 256, 257, 264, 266, 270, 276, 285,
Center of strength, 227 337, 338

DOI 10.1007/978-3-319-14246-3
344 Index
E Inelastic
Earthquake behaviour, 172
excitation, 178, 258, 260, 324 deformation, 6, 7, 172, 178, 213
induced torsion, 260 range, 217, 229
intensity, 227, 234 response, 112, 117, 130, 151, 201, 213, 215,
resistant design, 13, 100 216, 261
Elastic torsional response, 213
analysis, 140, 145, 172–174, 182–186, 216, Infilled frame, 95
217, 264 Infill panel, 90, 139
base shear, 198 In-plan distribution, 168, 199, 206, 217, 219
displacement, 235 Inter story drift, 50, 51, 60, 317, 319–321
parameter, 334 Intrinsic variability, 135
response, 20, 113, 118, 145, 197, 198, 270, Irregular structure, 112, 136, 193, 225, 226, 228,
308, 316 229, 231, 234, 235, 241, 316, 320, 324
response spectrum, 94, 138, 151, 183, 186,
187, 195, 197, 198, 209, 270, 308
stiffness, 140 L
Equivalent single degree of freedom (SDOF), Lateral torsional coupling, 194, 197, 205–213
94, 95, 100, 130 Lateral torsional response, 244, 252
Eurocode 2, 92 Life safety, 315
Eurocode 8, 36, 40, 47, 89, 90, 125, 136, 150, Limit state, 90, 94, 95, 136, 138, 140, 141, 145,
160, 172, 175, 209, 216, 226, 228, 234, 150, 151, 153–155, 157, 160–168, 184,
236, 240, 264, 270, 271, 305 188, 190, 191, 215, 230, 231
L shaped building, 10, 256, 260, 326
F
FEMA, 125, 317 M
Fibre-Optic System for Rotational Events and Masonry building, 276, 280–285
Phenomena Monitoring (FOSREM), Maximum dynamic displacement, 198, 209, 219
49–62 Maximum rotational response, 246, 251, 252
Finite difference method, 28, 30, 31, 33 Mining shock, 275–285
Flexural and shear capacity, 160 Modal damping, 67
FORS Telemetric Server, 53 Modal properties, 128
Friction dampers, 306 Modal pushover analysis (MPA), 124,
129–132, 233
Mode superposition, 37, 38, 40, 66
G Modulus of elasticity, 116, 280
Gravity load, 125, 162, 181, 186, 187, 191, 197 Multi-hazard mitigation, 316
Ground motion, 4–8, 10, 11, 13, 14, 17, 20, 22, Multi modal control
26, 29, 33, 35, 36, 93, 96, 99, 113, 117, Multiple tuned mass dampers (MTMDs),
121, 124, 127, 128, 130, 138–141, 159, 315–321
161, 164, 172, 175, 176, 179, 182, 183, Multi story buildings, 4, 8–10, 112, 118
197, 209, 215, 219, 227, 234, 235, 247,
251, 255, 257, 264, 315, 316, 319, 326,
334, 337 N
Ground rotational acceleration, 28 Nonlinear
analysis, 3–11, 89
dynamic analysis, 90, 139–141, 144, 182,
H 187, 191, 194, 197, 199, 209, 213,
High rise building, 4 219, 222
response history analysis, 4, 7, 8
static analysis, 5, 138, 143, 145, 194,
I 211–213, 222
Incremental dynamic analysis, 91 static method, 193, 194, 197, 199, 201, 213,
Induced vibration, 289, 297 215, 216, 219, 221, 222
Index 345
O 171, 176, 179, 182, 186, 188–191, 194,

Optical gyro configuration 197, 199, 205–213, 216, 219, 221, 222,
225, 227, 236, 266, 337, 338
Seismostruct, 151, 161
P Serviceability limit state, 140, 145, 157
Performance SESA method, 234–237, 240, 241
based design, 244, 321 Setback structure, 117, 118, 319
evaluation, 136, 143, 145, 150, 157, 163 Shear
measures, 317 capacity, 93
Period shifting effect, 244, 246 demand, 264
Permalloy particles, 53 forces, 43, 45, 128, 131, 133, 184, 308
Plan asymmetric buildings, 216 modulus, 269, 306–309
Plastic strength, 181, 182, 184, 186, 190
deformation, 207 wall building, 264, 269, 271
hinge, 92, 95, 138, 187 wave, 28, 36, 138, 247, 295, 306, 307
Polymeric bearings, 335–336, 338, 340 Spatial seismic effect, 36
Pounding, 255–261 Spectral analysis, 234, 295–297, 299
Pushover analysis, 17, 90, 94, 124, 134, 140, Spectral density(ies), 290, 291, 295–297, 299,
145, 198, 209, 210, 217 301, 319
Spectrum compatible accelerogram, 228, 229,
236
R Stiffness distribution, 73, 112, 172
Radius of gyration, 113, 115–117, 152, 153, Story moment, 8
195, 198, 206, 218, 246 Strength eccentricity, 152, 172, 194, 195, 198,
Rayleigh damping, 93, 319 206, 208, 217
Reinforced concrete, 71, 89, 151, 160, 172, Structural
256, 268, 276, 280, 318, 334 control, 333, 334
Rigidity, 112, 115, 121, 194, 195, 198, 201, engineering, 50, 55, 323
206, 207, 217, 269, 271, 285, 293, 326 health monitoring, 75
eccentricity, 194, 195, 198, 206, 207, 217 irregularities, 138
Rocking excitation, 36, 43, 46 Surface rocking, 36
Roof displacement, 4–7, 10, 11, 17, 20, 21, 90, Survivability Limit State, 236
94, 96, 199, 217, 328 Symmetric structure, 227, 235
Rotational component, 26, 35, 36, 50, 109
Rotational motion, 26, 50, 60, 62
Rotation rate amplitude, 68–70 T
Rotation rate sensor, 66, 73, 75 Tailings dam, 304–306, 310
Time history analyses, 20, 22, 118
Torsion, 90, 115, 128, 231, 233–235, 265, 271,
S 278, 281
Sagnac effect, 50–52, 58, 62 balanced system, 107, 115, 194, 206,
Sagnac phase shift, 51, 52 217, 218
Seismic capacity, 226, 236
demand, 90, 91, 93, 95, 124, 131, 150, 154, flexible system, 222
157, 159 index, 17, 100, 102, 104, 108, 109
intensity, 138, 151, 153, 172, 178, 179, 227, response, 133, 136, 145, 157, 198, 211, 213,
228, 232, 234, 236, 237, 240, 241 217, 222, 244, 259, 260
response, 14, 22, 41, 43, 100, 117–118, 124, sensitivity, 240
128, 136, 139, 140, 145, 153, 156, 168, stiff systems, 207
346 Index
Translational component, 26, 27 217, 219, 244–247, 256–260, 264, 265,

Translational period, 26, 228, 236 270, 273, 276, 278, 281–283, 289–291,
Trusses, 186 295–299, 301, 302, 304, 308, 316, 324,
334, 335, 337, 338
mode, 120, 152, 257, 258, 337
U Viscoelastic damper, 295
Ultimate limit state, 145, 153, 157
Unsymmetric plan buildings
W
Water tank, 324–328
V Weakened column, 178
Vertical irregularity, 324
Vibration, 5, 9, 37, 66–68, 70, 73–75, 83, 93,
100, 115, 117, 119, 120, 124, 125, 145, Y
151, 152, 186, 187, 195, 197, 209, 216, Yield moment, 92, 176

Seismic Behaviour and Design of Irregular and Complex Civil Structures II

Uploaded by

Copyright:

Available Formats

Seismic Behaviour and Design of Irregular and Complex Civil Structures II

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Seismic Behaviour and Design of Irregular and Complex Civil Structures II

Uploaded by

Copyright:

Available Formats

What is the book about?

What is the book about?

What topics are covered in the book?

What topics are covered in the book?

Geotechnical, Geological and Earthquake Engineering

Seismic Behaviour and

ISSN 1573-6059 ISSN 1872-4671 (electronic)

Library of Congress Control Number: 2013930359

Springer Cham Heidelberg New York Dordrecht London

Printed on acid-free paper

Springer International Publishing AG Switzerland is part of Springer ScienceþBusiness Media

Structural regularity was a logical preference for traditional civil engineering

Opole, Poland Zbigniew Zembaty

Part I Ground Motion, Rotations and Wave

6 Application of Rotation Rate Sensors in Measuring Beam

Part II Seismic Analysis and Design of Irregular

17 Application of Nonlinear Static Method with Corrective

Part III Seismic Control and Monitoring of Irregular Structures

Asimina M. Athanatopoulou Department of Civil Engineering, Aristotle

Giada Gasparini Department DICAM, University of Bologna, Bologna, Italy

Valentine Mariani Department of Architecture (DiDA), University of Florence,

Marco Tanganelli Department of Architecture (DiDA), University of Florence,

Juan C. Reyes, Erol Kalkan, and Andrea C. Ria~

Abstract In performance assessment and design verification of complex structural

J.C. Reyes (*) • A.C. Ria~

© Springer International Publishing Switzerland 2016 3

of expected median engineering demand parameters (EDPs)] and efficiency

Keywords Scaling seismic records • Modal pushover based scaling • Nonlinear

Performance-based procedures for evaluating existing buildings and proposed

1.2 Modal Pushover-Based Scaling (MPS) Procedure

1.2.1 Target Roof Displacement and Pushover Analyses

€ n ðtÞ þ 2ξn ωn D_ n ðtÞ þ Fsn ¼ €ug ðtÞ ! Dn

Determine D ^ n as the median of the Dn values. Calculate roof displacement in

1.2.2 Scaling Phase

(d) Estimate error: ε ¼ ur u^ r

1.2.3 Selection Phase

7. Select the first k records with the lower values of:

1.3 Point of Comparison (Benchmark)

The benchmark or comparison point of an EDP is defined in this study as the

1.4 Study Cases for One Component RHAs

19-story building 52-story building

1.5 Study Cases for Two Components RHAs

L05 L10 L15

T05 T10 T15

corner c1 x-girder y-girder center of mass center of stiffness

Fig. 1.2 Plan views of the nine multi-story asymmetric-plan buildings

buildings were considered: single-story and multi-story buildings. The single-story

Kalkan E, Chopra AK (2011) Modal-pushover-based ground-motion scaling procedure. J Struct

Keywords Asymmetric building • Bi-directional ground motion • Geometric mean

In designing new buildings for earthquake resistance or when conducting seismic

© Springer International Publishing Switzerland 2016 13

asymmetric buildings this procedure may be inadequate because the most

2.2 Description of the Procedure

2.2.1 Definition of the “Reduced Response Spectrum”

γ ¼ p SAζ ðT Þ=p SAξ ðT Þ ¼ const: ð2:1Þ

If the reduced response spectrum p SAU ðT Þ is defined as the geometric mean

Note that in general, the geometric mean spectrum of orthogonal components

2.2.2 Outline of Simplified Procedure

An outline of the proposed procedure to predict the “average” peak response is