See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/261551429

Structural Seismic Design Optimization and Earthquake Engineering:

Formulations and Applications

Chapter · May 2012

DOI: 10.4018/978-1-4666-1640-0.ch006

CITATIONS READS
0 1,217

1 author:

Abbas Moustafa
Minia University
54 PUBLICATIONS   833 CITATIONS   

SEE PROFILE

All content following this page was uploaded by Abbas Moustafa on 12 April 2014.

The user has requested enhancement of the downloaded file.

Structural Seismic
Design Optimization and
Earthquake Engineering:
Formulations and Applications
Vagelis Plevris
School of Pedagogical and Technological Education,
Greece National Technical University of Athens, Greece

Chara Ch. Mitropoulou

Institute of Structural Analysis and Seismic Research,
National Technical University of Athens, Greece

Nikos D. Lagaros
Institute of Structural Analysis and Seismic Research,
National Technical University of Athens, Greece Hellenic Army, Greece
Managing Director: Lindsay Johnston
Senior Editorial Director: Heather A. Probst
Book Production Manager: Sean Woznicki
Development Manager: Joel Gamon
Development Editor: Myla Harty
Acquisitions Editor: Erika Gallagher
Typesetter: Nicole Sparano
Cover Design: Nick Newcomer, Lisandro Gonzalez

Published in the United States of America by

Engineering Science Reference (an imprint of IGI Global)
701 E. Chocolate Avenue
Hershey PA 17033
Tel: 717-533-8845
Fax: 717-533-8661
E-mail: cust@igi-global.com
Web site: http://www.igi-global.com

Copyright © 2012 by IGI Global. All rights reserved. No part of this publication may be reproduced, stored or distributed in
any form or by any means, electronic or mechanical, including photocopying, without written permission from the publisher.
Product or company names used in this set are for identification purposes only. Inclusion of the names of the products or
companies does not indicate a claim of ownership by IGI Global of the trademark or registered trademark.

Library of Congress Cataloging-in-Publication Data

Structural seismic design optimization and earthquake engineering: formulations and applications / Vagelis Plevris ... [et al.],
editors.
p. cm.
Includes bibliographical references and index.
Summary: “This book focuses on the research around earthquake engineering, in particular, the field of implementation of
optimization algorithms in earthquake engineering problems, including simulation issues for the accurate prediction of the
seismic response of structures, design optimization procedures, soft computing applications, and other important advance-
ments in seismic analysis and design”-- Provided by publisher.
ISBN 978-1-4666-1640-0 (hardcover) -- ISBN 978-1-4666-1641-7 (ebook) -- ISBN 978-1-4666-1642-4 (print & perpetual
access) 1. Earthquake engineering. 2. Structural optimization. I. Plevris, Vagelis, 1976-
TA654.6.S8673 2012
624.1’762--dc23
2012003307

British Cataloguing in Publication Data

A Cataloguing in Publication record for this book is available from the British Library.

All work contributed to this book is new, previously-unpublished material. The views expressed in this book are those of the
authors, but not necessarily of the publisher.
128

Chapter 6
Damage Assessment of
Inelastic Structures under
Simulated Critical Earthquakes
Abbas Moustafa
Minia University, Egypt

ABSTRACT
Damage of structures can be significantly reduced through robust prediction of possible future earth-
quakes that can occur during the life-time of the structure and through accurate modeling of the nonlinear
behavior of the structure under seismic loads. Modern seismic codes specify natural records and arti-
ficially generated ground accelerations as input to the nonlinear time-history analysis of the structure.
The advantage of using natural records is the inclusion of all important characteristics of the ground
motion (fault properties, path effects and local soil condition) in the design input. This option requires
selecting and scaling a set of proper accelerograms from the available records. However, the site under
consideration may have limited or scarce earthquake data. In such case, numerically simulated ground
motions can be employed as input to the dynamic analysis of the structure. This chapter deals with
the damage assessment of inelastic structures under numerically simulated critical earthquakes using
nonlinear optimization, inelastic time-history analysis, and damage indices.

1. INTRODUCTION tion techniques. The critical excitation method

relies on the high uncertainty associated with
Damage indices describe the state of the structural the occurrence of the earthquake phenomenon
damage and correlate well with actual damage and on the safety requirements of important and
displayed during earthquakes. The critical ground lifeline structures. The earthquake input is taken to
motion for a give structure is estimated by solving maximize the damage index of the structure while
an inverse nonlinear dynamic problem in time satisfying predefined constraints that are quanti-
domain using constrained nonlinear optimiza- fied from the earthquake data available at the site.
Numerical illustrations for damage assessment of
DOI: 10.4018/978-1-4666-1640-0.ch006

Copyright © 2012, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.
Damage Assessment of Inelastic Structures

one-storey and two-storey plane frame structures the structure under possible future earthquakes
under possible future earthquakes are presented. while maintaining optimal use of the construc-
Earthquakes continue to claim thousands of tion material. The design objectives in current
lives and to damage structures every year (Comar- seismic building codes are to ensure life safety
tin et al, 2004). Each earthquake brings out new and to prevent damage of the structure in minor
surprises and lessons with it. In fact, the unexpected and moderate frequent earthquakes, and to control
loss of lives and the severe damage of infrastruc- local and global damage (prevent total collapse)
tures and buildings during past strong earthquakes and reduce life loss in a rare major earthquake.
(e.g., 1994 Northridge, 1995 Kobe, 2010 Haiti and This can be achieved through: (1) robust predic-
the most recent 2011 Tohoku earthquakes) have tion of expected future strong ground motions
raised significant concern and questions on life at the site, (2) accurate modeling of the material
safety and performance of engineering structures behavior under seismic loads, and (3) optimal
under possible future earthquakes. The occurrence distribution of the construction material.
of strong earthquakes in densely populated regions, Early works on seismic design have dealt
especially in developing countries with vulnerable with the specification of earthquake loads using
building stock and fragile infrastructure, could the response spectrum method, the time history
lead to catastrophic consequences. A notable of the ground acceleration or using the theory of
example is the 2010 Haiti earthquake that killed random vibrations. The nonlinear time-history
250,000 people and left a long-term suffering for analysis method is recognized as the most accurate
the residents of this developing country (USGS/ tool for dynamic analysis of structures (Pinho,
EERI 2010). On the other hand, the severe dam- 2007). Many researchers have also established
age caused by the 2011 Tohoku earthquake and deterministic and probabilistic hazard spectra
associated tsunami in Japan has raised significant for the site (Reiter, 1990, McGuire, 1995). The
challenges to one of the most developed countries development of mathematical models to describe
as well (Takewaki et al, 2011). Hence, the assess- the hysteretic nonlinear behavior of the structure
ment of seismic performance of structures under during earthquakes has also been pursued in
strong ground motions is an important problem in several studies (e.g., Takeda et al, 1970, Otani,
earthquake engineering. Structures need to resist 1981, Akiyama, 1985). New design concepts
unknown future earthquakes which adds more and methods, such as energy-, performance- and
complexity to the problem (Moustafa 2011, 2009, displacement-based design, base-isolation and
Moustafa & Takewaki 2010a, Abbas & Manohar structural control have been recently developed
2007, Takewaki 2002a, 2007). The consideration (Priestely et al, 2007, Takewaki, 2009, Fardis,
of the earthquake inherent uncertainty, the vari- 2010). Similarly, the optimal design of the struc-
ability in the structure parameters and modeling tures under earthquake loads has been investigated
the nonlinear behavior of the structure is essential in several studies (Fardis, 2010, Elishakoff &
for the accurate prediction of the actual response Ohsaki, 2010, Plevris, 2009, Haldar, 2006, Liang,
of the structure. Earthquake uncertainties include 2005). The evaluation of the current procedures
time, location, magnitude, duration, frequency and new practical procedures for ground motion
content and amplitude, referred to as aleatory selection and modification are provided in the
uncertainties. recent special issue on earthquake ground motion
The earthquake-resistant design of structures selection and modification for nonlinear dynamic
has been an active area of research for many analysis of structures (Kalkan & Luco, 2011). The
decades (e.g., Penelis & Kappos 1997). The struc- two edited books by Papadrakakis et al (2009)
tural engineer aims to ensure safe performance of and Tsompanakis et al. (2008) and the doctoral

129
 Damage Assessment of Inelastic Structures

thesis by Plevris (2009) present the state-of-the-art and other related subjects has been extensively
on advances and applications of optimal seismic investigated by several researchers, especially
design of structures considering uncertainties. at Stanford University (e.g., Baker & Cornell,
The definition of the worst (also known as 2006, Moustafa et al, 2010, Baker, 2011, Baker et
critical) ground motion represents a major chal- al, 2011, Haselton et al, 2011, Buratti et al 2011,
lenge in earthquake-resistance design of structures. Bommer & Acevedo 2004). If the number of the
This is because of the high uncertainty involved available records is small, records from other
in the occurrence of the earthquake phenomenon sites with similar soil condition or artificially
compared to the relatively low variability in the simulated ground motions could be employed.
structure’s properties. Strasser & Bommer (2009) The critical excitation method provides another
pose an important question on whether we have alternative in case of scarce, inhomogeneous or
seen the worst ground motion yet, and the answer limited earthquake data at the site. This method
is not. They identify the worst ground motions as has been used to assess the structure’s response
those having large amplitude. It may be empha- under mathematically simulated earthquake inputs
sized that the worst ground motion for a structure representing possible worst future earthquakes
may not be the worst input for a different structure. (e.g., Moustafa, 2011, 2009, Takewaki, 2002a,
For example, the 2002 earthquake of magnitude 2007). The method relies on the high uncertainty
7.9 occurred along Alaska’s Denali fault killed associated with the occurrence of the earthquake
no one and did a little serious damage (Worth phenomenon, associated characteristics and also
2005), while the 1995 Kobe earthquake of 6.9 on the safety requirements of important and lifeline
magnitude killed 5,100 people and caused billion structures (nuclear plants, storage tanks, industrial
of dollars in structural damage. Early studies on installations, etc.). The critical earthquake input
defining severity of strong ground motion and for a given structure is computed by minimizing
earthquake capability to create large damage have the structure’s performance while satisfying pre-
focused on the earthquake intensity, peak values defined constraints observed in real earthquake
of ground acceleration, velocity and displacement records. The structural performance may be de-
(PGA, PGV, PGD), effective PGV, etc. (Housner, scribed in terms of the structure’s response or in
1970, Housner & Jennings, 1976). Near-field terms of reliability measures or damage indices
ground motion with pulse-like characteristic is a (Abbas & Manohar, 2007, Moustafa, 2009). The
phenomenon representing one scenario of sever- optimum design of the structure under varying
ity in the near-field region. Other scenarios of critical earthquake loads has also been studied
earthquake severity include repeated occurrence of (e.g., Fujita et al, 2010, Saikat & Manohar, 2005,
ground motion in sequences (Elnashai et al, 1998, Takewaki, 2002b). Several practical applications
Moustafa & Takewaki 2009). Deep soft soils can have evolved from the concept of critical excita-
also amplify earthquake amplitudes and modify tions. This includes design of structures to critical
frequency content. Secondary causes include also excitations, deriving critical response spectra for
the travel path effects. In reality, some of these the site, estimating critical cross power spectral
causes could exist together. density functions of multi-point and spatially-
To carry out nonlinear time-history analysis, a varying ground motions, and reliability analysis of
set of suitable accelerograms need to be selected structures to partially specified earthquake loads.
from available records (see, e.g., PEER, 2011). In The method of critical excitations has also been
this context, the criteria based on which records employed in identifying resonant accelerations and
are selected and scaled represents an interest- in selecting critical recorded accelerograms based
ing subject. A notable effort in this direction on the notion of the entropy principle (Moustafa,

130
Damage Assessment of Inelastic Structures

2009, 2010). Comprehensive reviews on these 2.1 Energy Dissipated by Inelastic

aspects can be found in Moustafa (2011). Structures
This chapter deals with the damage assess-
ment for inelastic structures under worst future The energy dissipated by an N multi-degree-of-
earthquakes using the critical excitations method. freedom (MDOF) structure under the ground
The novelty of this research is in combining acceleration x(t )can be computed by integrating
damage indices, for the first time, with nonlinear the equations of motion as follows (Zahrah &
optimization and nonlinear time-history analysis Hall, 1984, Akiyama, 1985, Uang & Bertero,
for assessing the structural performance under 1990, Takewaki, 2004, Kalkan & Kunnath, 2008):
possible future ground motions. The use of dam-
age indices provides a quantitative measure for t
1 N
∫ X  (τ )d τ =
∑ mi xi2 (t )
T
damage and necessary repair for the structure. E K (t ) = (τ )MX
0
2 i =1
Bilinear and elastic-plastic force-displacement (1a)
relationships are taken to model the material
nonlinearity, and thus the present work is lim- t N t
 T (τ )CX
 (τ )d τ =
ited to non-deteriorating structures. Numerical E D (t ) = ∫ X ∑ ∫ xi (τ )fDi (τ )d τ
examples for one-storey and two-storey plane 0 i =1 0
(1b)
frames without irregularities are provided. Future
practical applications of the proposed methodol- t
N
ogy in seismic analysis and design of structures E H (t ) = ∑ ∫ xi (τ )fsi (τ )d τ − Es (t ) (1c)
are also discussed. i =1 0

2. DAMAGE ASSESSMENT where, M, C, are the mass and damping matrices

OF INELASTIC STRUCTURES of the structure, respectively, fsi (t )is the ith hys-
UNDER EARTHQUAKE LOADS teretic restoring force, X(t) is the structure dis-
placement vector and dot indicates differentiation
Damage assessment for structures is generally
with respect to time. The quantities EK(t), ED(t),
based on the nonlinear response quantities under
Es(t) and EH(t) represent the kinetic, damping,
earthquake loads (see Table 1). The bilinear and
strain and hysteretic energies, respectively
the elastic-plastic models are shown in Figure 1.
(Moustafa, 2009). For viscous damping models,
The evaluation of the structural damage is usu-
the damping energy reduces to
ally carried out using damage indices which are N N t
quantified in terms of the structure response and ∑ ∑ ∫ c x (τ )x (τ )d τ .
ij i j
the associated absorbed energy. Therefore, the i =1 j =1 0

quantification of damage indices is carried out after Note that equations (1) provide the relative
performing nonlinear time-history analysis for the energy terms. Note also that, by the end of the
structure. The nonlinear time-history analysis for earthquake duration the kinetic and elastic strain
the structure is performed by solving the equations energies diminish. Thus, the earthquake input
of motions using numerical integration schemes energy to the structure is dissipated by hysteretic
(Moustafa, 2009, Hart & Wong 2000). and damping energies. The next section demon-
strates the use of the structure’s response and the
hysteretic energy in developing damage indices.

131
 Damage Assessment of Inelastic Structures

Table 1. Response descriptors for inelastic buildings under earthquake ground motion

Response parameter Definition

Maximum ductility
x (t )
µmax = max | |
0≤t ≤t
f xy
Number of yield reversals Number of times velocity changes sign
∆x p,i
∆x p,i = max | |
Maximum normalized plastic deformation range 0≤t ≤t f xy
N | ∆x
p ,i |
µac . =∑ +1
Normalized cumulative ductility i =1 xy
x (t f )
µres =| |
Residual (permanent) ductility xy
tf
1
EI = ∫ E (t )dt
Normalized earthquake input energy fy x y 0 I
t
1 f
E = ∫ E (t )dt
Normalized total hysteretic energy dissipated H fy x y 0 H
EH
rE =
Ratio of total hysteretic energy to input energy EI
1  dE (t ) 
PI ,max = max  I 
Maximum rate of normalized input energy fy x y 0≤t ≤t f  dt 
1  dE (t ) 
PD ,max = max  D 
Maximum rate of normalized damping energy
fy x y 0≤t ≤t f  dt 
1  dE (t ) 
PH ,max = max  H 
fy x y 0≤t ≤t f  dt 
Maximum rate of normalized hysteretic energy

2.2 Damage Measures for during the ground shaking. This measure does not
Inelastic Structures incorporate information on how the earthquake
input energy is imparted on the structure nor how
The literature on damage measures for structures this energy is dissipated. Earthquake damage oc-
under earthquake loads is vast (e.g., Cosenza et curs due to the maximum deformation or ductil-
al, 1993, Ghobarah et al, 1999). Damage indices ity and the hysteretic energy dissipated by the
are quantified in terms of a single or a combina- structure. Therefore, the definition of structural
tion of structural response parameters. Table 1 damage in terms of the ductility is inadequate.
summarizes several damage measures that are The last three measures indicate the rate of the
based on a single response parameter (Powell & earthquake input energy to the structure (i.e., how
Allahabadi 1988, Cosenza et al, 1993). The first fast the input energy E I is imparted by the earth-
measure represents the ultimate ductility produced

132
Damage Assessment of Inelastic Structures

Figure 1. (a) Force-displacement relation for nonlinear materials

quake and how fast it gets dissipated). Damage E H / ( fy x y )

DI H = (4)
indices can be estimated by comparing the re- µu − 1
sponse parameters demanded by the earthquake
with the structural capacities. Powell & Alla-
A robust damage measure should include not
habadi (1988) proposed a damage index in terms
only the maximum response but also the effect
of the ultimate ductility (capacity) µu and the
of repeated cyclic loading. Park and co-workers
maximum ductility attained during ground shak- developed a simple damage index, given as (Park
ing µmax : et al, 1985, Park & Ang 1985, Park et al, 1987):

x max − x y µmax − 1 x max E µ EH

DI µ = = (2) DI PA = + β H = max + β
xu − xy µu − 1 xu fy x u µu fy x y µu
(5)
However DI µ does not include effects from
hysteretic energy dissipation. Additionally, this Here, x max , E H are the maximum absolute
damage index may not be zero for undamaged displacement and the dissipated hysteretic energy
structures. A damage index that overcomes this excluding elastic energy. x u is the ultimate defor-
problem has been proposed by Khashaee (2004): mation capacity under monotonic loading and β
is a positive constant that weights the effect of
−1
1 − µmax cyclic loading on structural damage. Note that if
DI K = (3) β = 0, the contribution to DIPA from cyclic load-
1 − µu−1
ing is omitted.
The state of the structure damage is defined
Fajfar (1992) and Cosenza et al., (1993) quan- as: (a) repairable damage, when DI PA < 0.40 ,
tified damage based on the structure hysteretic
( b ) d a m a g e d b e y o n d r e p a i r, w h e n
energy E H :
0.40 ≤ DI PA < 1.0 , and (c) total or complete

133
 Damage Assessment of Inelastic Structures

collapse, when DI PA ≥ 1.0 . These criteria are tion which are selected to span satisfactory the
based on calibration of DIPA against experimen- frequency range of xg (t ) . In constructing critical
tal results and field observations in earthquakes seismic inputs, the envelope function is taken to
(Park et al., 1987). Note that Eq (5) reveals that be known. The information on energy E, peak
both maximum ductility and hysteretic energy ground acceleration (PGA) M1, peak ground
dissipation contribute to the structural damage velocity (PGV) M2, peak ground displacement
during earthquakes. Eq. (5) expresses damage as (PGD) M3, upper bound Fourier amplitude spec-
a linear combination of the damage caused by tra (UBFAS) M 4 (ω), and lower bound Fourier
excessive deformation and that contributed by amplitude spectra (LBFAS) M 5 (ω) are also taken
repeated cyclic loading effect. Note also that the to be available which enables defining the fol-
quantities x max , E H depend on the loading history lowing nonlinear constraints (Abbas & Manohar,
while the quantities β, x u , fy are independent of 2002, Abbas, 2006):
the loading history and are determined from ex-
perimental tests. It should also be emphasized ∞
1
2
that Eqs (2-5) can be used to estimate damage for  x2 (t ) dt  ≤ E
∫ g 
a member in a structure which defines the local  0 
damage. To estimate the global damage of the max | xg (t ) |≤ M 1
0<t <∞
structure, a weighted sum of the local damage
max | xg (t ) |≤ M 2 (7)
indices need to be estimated (Park et al, 1987). 0<t <∞

In this chapter, Eq (5) is adopted in quantifying max | x g (t ) |≤ M 3

0<t <∞
the structural damage. The next section develops
M 5 (ω) ≤| X g (ω) |≤ M 4 (ω)
the mathematical modeling of critical future
earthquake loads.
Here, X g (ω) is the Fourier transform of xg (t )
. Note that the constraint on the earthquake en-
3. DERIVATION OF WORST ergy is related to the Arias intensity (Arias 1970).
FUTURE EARTHQUAKE LOADS The spectra constraints aim to replicate the fre-
quency content and amplitude observed in past
The worst future ground acceleration is repre- recorded accelerograms on the future earthquake.
sented as a product of a Fourier series and an The ground velocity and displacement are obtained
envelope function in Box 1. from Eq. (6) and seen in Box 2:
Here, A0 is a scaling constant and the param- Making use of the conditions x g (0) = 0 and
eters α1 , α2 impart the transient trend to xg (t ) . lim xg (t ) → 0 (Shinozuka & Henry, 1965), the
t →∞
Ri and ϕi are 2N f unknown amplitudes and phase
constants in the above equation can be shown to
angles, respectively and ωi , i = 1, 2, ..., N f are be given as (Abbas & Manohar, 2002, Abbas,
the frequencies presented in the ground accelera- 2006):

Box 1. ­

Nf Nf

xg (t ) = e(t )∑ Ri cos(ωit − ϕi ) = A0 [exp(−α1t ) − exp(−α2t )]∑ Ri cos(ωit − ϕi ) (6)

i =1 i =1

134
Damage Assessment of Inelastic Structures

Box 2. ­

Nf t Nf t

xg (t ) = ∑ ∫ Rie(τ ) cos(ωi τ − ϕi )d τ + C 1 ; x g (t ) = ∑ ∫ Rie(τ )(t − τ ) cos(ωi τ − ϕi )d τ + C 1t + C 2

i =1 0 i =1 0
(8)

Finally, the problem of deriving critical future

∞
Nf
earthquake loads on inelastic structures can be
C 2 = 0;C 1 = −∑ ∫ Rie(τ ) cos(ωi τ − ϕi )d τ
i =1
posed as determining the optimization variables
0
(9) y = {R1 , R2 , ..., RN f , ϕ1 , ϕ2 , ..., ϕN f }t such that the
damage index DIPA is maximized subjected to
The constraints of Eq (7) can be expressed in the constraints of Eq (10). The solution to this
terms of the variables Ri , ϕi , i = 1, 2, ..., N f (see nonlinear constrained optimization problem is
Box 3) tackled by using the sequential quadratic program-
Here i = −1 . To quantify the constraints ming method (Arora, 2004). The following con-
quantities E, M1, M2, M3, M 4 (ω), and M 5 (ω) it vergence criteria are adopted:
is assumed that a set of Nr earthquake records
| fj − fj −1 |≤ ε1 ; | yi , j − yi , j −1 |≤ ε2 (12)
denoted by vgi (t ), i = 1, 2, ..., N r are available for
the site under consideration or from other sites
with similar geological soil conditions. The values Herein, fj is the objective function at the jth
of energy, PGA, PGV and PGD are obtained for iteration, yi,j is the ith optimization variable at the
each of these records. The highest of these values jth iteration and ε1 , ε2 are small quantities to be
across all records define E, M1, M2 and M3. The specified. The structure inelastic deformation is
available records are further normalized such that estimated using the Newmark β-method which is
the Arias intensity of each record is set to unity built as a subroutine inside the optimization pro-
∞
gram. The details of the optimization procedures
(i.e.,[ ∫ vgi2 (t )dt ]1/2 = 1, Arias, 1970), and are involved in the computation of the critical earth-
0
quake and the associated damage index are shown
denoted by {vgi }iN=1r . The bounds M 4 (ω) and
in Figure 2. Further details can also be found in
M 5 (ω) are obtained as: Abbas (2006).
It may be emphasized that the quantities µ(t )
M 4 (ω) = E max | Vgi (ω) |; M 5 (ω) = E min | Vgi (ω) | and E H (t ) do not reach their respective maxima
1≤i ≤N r 1≤i ≤N r

(11) at the same time. Therefore, the optimization is

performed at discrete points of time and the op-
Here Vgi (ω), i = 1, 2, ..., N r denotes the Fou- timal solution
rier transform of the ith normalized accelerogram
vgi (t ) . The bound M 4 (ω) has been considered y * = [R1* , R2* , ..., RN* f , ϕ1* , ϕ2* , ..., ϕN* f ]t
earlier (Shinozuka, 1970, Takewaki, 2001, 2002).
The lower bound was considered by Moustafa is the one that produces the maximum DIPA across
(2002) and Abbas & Manohar (2002). all time points. The critical earthquake loads are
characterized in terms of the critical accelerations

135
 Damage Assessment of Inelastic Structures

Box 3. ­

1
 Nf Nf ∞ 2
A2  ≤E
 0 ∑ ∑ Rm Rn ∫ [exp(−α1t ) − exp(−α2t )]2
cos(ωm t − ϕm ) cos(ωn t − ϕn ) dt 
 m =1 n =1 0 
Nf

max | A0 [exp(−α1t ) − exp(−α2t )]∑ Rn cos(ωnt − ϕn ) |≤ M 1

0<t <∞
n =1
Nf t

max | A0 ∑ ∫ Rn [exp(−α1τ ) − exp(−α2 τ )] cos(ωn τ − ϕn )d τ −

0<t <∞
n =1 0
Nf ∞
(10)
A0 ∑ ∫ Rn [exp(−α1τ ) − exp(−α2 τ )] cos(ωn τ − ϕn )d τ |≤ M 2
n =1 0
Nf t

max | A0 ∑ ∫ Rn [exp(−α1τ ) − exp(−α2 τ )](t − τ ) cos(ωn τ − ϕn )d τ −

0<t <∞
n =1 0
Nf ∞

A0t ∑ ∫ Rn [exp(−α1τ ) − exp(−α2 τ )] cos(ωn τ − ϕn )d τ |≤ M 3

n =1 0
Nf ∞

M 5 (ω) ≤| A0 ∑ ∫ Rn {exp[−α1τ ] − exp[−α2 τ ]} cos(ωn τ − ϕn ) exp[−i ωτ ]d τ |≤ M 4 (ω)

n =1 0

and associated damage indices, inelastic deforma- also been developed in the existing literature,
tions and energy dissipated by the structure. The mainly by seismologists (see, e.g., Brune, 1970,
next section provides numerical illustrations for Hanks & McGuire, 1981, Boore, 1983, Queck et
the formulation developed in this section. al, 1990). In using these models, one needs to
In the numerical analysis, the constraints input values for a host of parameters and the suc-
quantities E, M1, M2, M3, M 4 (ω), and M 5 (ω) cess of the model depends on how realistically
are estimated using past recorded earthquake data. this is done. It is possible to formulate the critical
This approach is considered to be consistent with earthquake models based on the latter class of
the aspirations of the ground motion models that models in which one can aim to optimize the
are commonly used by engineers, which, aim to parameters of the model so as to realize the least
replicate the gross features of recorded motions, favorable conditions. Note that the class of admis-
such as, amplitude, frequency content, nonsta- sible functions, in the determination of critical
tionarity trend, local soil amplification effects, excitations, in this case, becomes further con-
and duration. It is of interest to note in this context strained by the choice that one makes for the
that, predictive or physical models for ground physical model. The approach adopted in this
motions, which take into account several details, study, in this sense, is nonparametric in nature. A
such as, fault dimension, fault orientation, rupture comparison of results based on this approach with
velocity, magnitude of earthquake, attenuation, those from ‘model-based’ approaches is of inter-
stress drop, density of the intervening medium, est; however, these questions are not considered
local soil condition and epicentral distance, have in the present study.

136
Damage Assessment of Inelastic Structures

Figure 2. Flowchart for deriving critical earth-

parameters are changed later to study their influ-
quake loads
ence on the estimated worst earthquake loads and
the associated damage. The yield displacement is
taken as 0.10 m and the structure is taken to start
from rest. The objective function is adopted as
the Park and Ang damage index DIPA given by
Eq (5). The parameters of the Newmark β-method
are taken as δ = 1 / 2; α = 1 / 6 and ∆t = 0.005
s.

4.1.2 Quantification of Constraints

A set of 20 earthquake records is used to quan-

tify the constraint bounds E, M1, M2, M3, M 4 (ω)
and M 5 (ω) (COSMOS, 2005). Table 2 provides
information on these records. Based on numerical
analysis of these records, the constraints were
computed as E = 4.17 m/s1.5, M1 = 4.63 m/s2
(0.47 g), M2 = 0.60 m/s and M3 = 0.15 m and the
average dominant frequency was about 1.65 Hz.
The envelope parameters were taken as A0 = 2.17,
α1 = 0.13, and α2 = 0.50. The convergence limits
ε1 , ε2 were taken as 10-6 and the convergence
criterion on the secant stiffness is taken as 10-3
N/m. The frequency content for xg (t ) is taken as
(0.1-25) Hz. Additionally, in distributing the
frequencies ωi , i = 1, 2, ..., N f in the interval
(0.1, 25) , it was found advantageous to select
some of these ωi to coincide with the natural
frequency of the elastic structure and also to place
relatively more points within the modal half-
power bandwidth.
The constraint scenarios considered in deriv-
4. NUMERICAL EXAMPLES ing the worst earthquake inputs are listed in Table
3. The constrained nonlinear optimization problem
4.1 Bilinear Inelastic Frame Structure is solved using the sequential quadratic optimiza-
tion algorithm ‘fmincon’ of the Matlab optimiza-
A SDOF frame structure with mass 9× 103 kg, tion toolbox (Caleman et al, 1999). In the nu-
initial stiffness k1 = 1.49× 105 N/m and viscous merical calculations, alternative initial starting
damping of 0.03 damping ratio is considered solutions, within the feasible region, were exam-
(initial natural frequency = 4.07 rad/s). The strain ined and were found to yield the same optimal
hardening ratio is taken equal to 0.05. These solution. To select the number of frequency terms
N f a parametric study was carried out and N f =

137
 Damage Assessment of Inelastic Structures

51 was found to give satisfactory results. Figure the total energy and PGA, the worst input is nar-
3 depicts the influence of Nf on the convergence row band (highly resonant) and the structure
of the objective function for constraints scenarios deformation is conservative (see Figure 4 and
1 and 4 (see Table 3). Table 4). Furthermore, most of the power of the
Fourier amplitude is concentrated at a frequency
4.1.3 Numerical Results and close to the natural frequency of the elastic struc-
Discussions ture. This amplitude gets shifted towards a
higher frequency when the strain hardening ratio
The numerical results obtained are presented in increases. The Fourier amplitudes at other frequen-
Figures 4, 5, 6, 7, 8 and 9 and Table 4. Figure 4 cies are low and uniformly distributed. This result
shows results for constraint scenario 1 and similar is substantially different from that for the elastic
results for case 4 are shown in Figure 5. Each of structure where all power of the acceleration
these figures shows the Fourier amplitude spec- amplitude is concentrated around ω0 with no
trum of the worst ground acceleration, the inelastic amplitude at other frequencies (Abbas & Manohar,
deformation, the hysteretic force and the energy 2002). Additional constraints on the Fourier am-
dissipated by the structure. Figure 6 shows the plitude spectra (see Table 3) force the Fourier
time history of the ground acceleration. Based amplitude of the worst acceleration to get distrib-
on extensive analysis of the numerical results, uted across other frequencies. The critical accel-
the major observations are summarized below. eration possesses a dominant frequency that is
The frequency content and Fourier amplitude close to the average dominant frequency observed
of the worst earthquake are strongly dependent in past records (see Figure 5). The realism of the
on the constraints imposed (see Table 3). If avail- earthquake input is also evident from the maximum
able information on earthquake data is limited to damage index it produces. For instance, the dam-

Table 2. Information on past recorded ground motion records for a firm soil site

Earthquake date Magnitude Epic. Dist. Comp- PGA PGV PGD Energy* Site
(km) onent (m/s2) (m/s) (m) (m/s1.5)
Mamoth lakes 6.2 1.5 W 4.02 0.21 0.05 3.73 Convict Greek
05.25.1980 S 3.92 0.23 0.05 4.01
Loma prieta 7.0 9.7 W 3.91 0.31 0.07 3.82 Capitola
10.18.1989 S 4.63 0.36 0.11 2.61
Morgan hill 6.1 4.5 S60E 3.06 0.40 0.07 2.33 Halls valley
04.24.1984 S30W 1.53 0.30 0.02 1.64
San Fernando 6.6 27.6 N69W 3.09 0.17 0.04 2.07 Castaic old ridge
02.09.1971 N21E 2.66 0.28 0.10 2.47
Parkfield 5.0 9.1 W 2.88 0.44 0.01 1.33 Parkfield fault
12.20.1994 S 3.80 0.10 0.01 1.74
Caolinga 6.5 30.1 W 2.83 0.26 0.10 2.67 Cantua creek
05.02.1983 N 2.20 0.26 0.10 2.14
Northridge 6.7 5.9 S74E 3.81 0.60 0.12 4.17 Canoga park
01.17.1994 S16W 3.43 0.34 0.09 3.50
Cape Mendocino 7.0 5.4 W 3.25 0.45 0.15 2.44 Petrolia general
04.25.1992 S 2.89 0.24 0.08 2.31
Westmorland 5.0 6.6 E 4.35 0.33 0.11 3.26 Westmorland fire
04.26.1981 S 3.54 0.44 0.15 3.25
Imperial valley 6.4 17.4 S45W 2.68 0.22 0.10 2.30 Calexico fire
10.15.1979 N45W 1.98 0.19 0.15 2.14
∞

* E = [ ∫ vg2 (t ) dt ]1/2 (Arias, 1970).

0

138
Damage Assessment of Inelastic Structures

Table 3. Nomenclature of constraint scenarios

examined by comparing maximum response from
considered
these accelerations with those from past recorded
Case Constraints imposed ground motions. Thus, the maximum ductility
1 Energy and PGA
factor of the structure from the worst earthquake
2 Energy, PGA, PGV and PGD is about 3.9 (case 1) and 2.6 (case 4) times that
3 Energy, PGA and UBFAS
4 Energy, PGA, UBFAS and LBFAS
from the Kobe earthquake and is 2.7 (case 1) and
1.5 (case 4) times that from the San Fernando
earthquake.
To examine the effect of the strain harden-
age index for case 4 is 0.37 which is substan-
ing ratio on the design earthquake acceleration
tially smaller than 1.15 for case 1 (Table 4). The
computed, limited studies were carried out. The
constraints on PGV and PGD were not found to
value of α was changed and the critical input
be significant in producing realistic critical inputs
was determined by solving a new optimization
compared to the constraints on UBFAS and LB-
problem. Namely, α was taken as 0.20, 0.10, 0.05,
FAS. Also, the realism of the critical acceleration
and 0.01. The strain hardening ratio was not seen
for case 4 can be examined by comparing the
to significantly influence the frequency content
Fourier amplitude spectra and frequency content
of the critical earthquake input. It was observed,
of the worst acceleration (Figures 4, 5) with the
however, that the inelastic structure with lower
Fourier amplitude spectra of past recorded earth-
values of α yields more frequently compared to the
quakes (Figure 7). Note that, while the constraint
same system with higher α values. Accordingly,
scenario 1 leads to pulse-like ground motion, such
the cumulative hysteretic energy dissipated was
scenario is observable in past recorded earthquakes
observed to decrease for higher values of α (Fig-
(e.g., 1971 San Fernando, 1985 Mexico, and 1995
ure 8(a)). This feature is particularly remarkable
Kobe earthquakes). Resonant or pulse-like earth-
at the end of the earthquake duration. It was also
quakes are also observable in near-field ground
observed that the results on critical earthquake
motion with directivity focusing, known as for-
accelerations for bilinear inelastic structure with
ward- and backward-directivity ground motion
α = 0.01 are close to those for the elastic-plastic
(Housner & Hudson, 1958, Kalkan & Kunnath,
structure (Abbas, 2006).
2006, He & Agrawal, 2008, Moustafa 2008). The
realism of worst earthquake loads can be also

Figure 3. Convergence of objective function in terms of frequency terms Nf (a) Case 1 (b) Case 4

139
 Damage Assessment of Inelastic Structures

Figure 4. Optimal earthquake input and associated structural responses for case 1 (a) Fourier ampli-
tude of the ground acceleration (b) Normalized inelastic deformation (c) Hysteretic restoring force (d)
Dissipated energy

To investigate the influence of the damping compared with the same structure with lower
ratio on the computed worst earthquake load, damping ratio (see Figure 8(b)). The damage
limited studies were carried out. The damping index also reduces when the damping ratio in-
ratio was changed, namely, 0.01, 0.03 and 0.05, creases.
while all other parameters were kept unchanged. To assess the structure safety, Eq. (5) was used
The critical earthquake is computed by solving a to estimate the damage index of the structure
new optimization problem for each case. The ef- subjected to the critical earthquake load. The ef-
fect of the change in η0 was seen to be similar to fect of the parameter β on the damage index is
that due to α. In other words, the value of the examined first. Based on experimental tests, it
damping ratio was not seen to significantly influ- was reported that β ranges between 0.05 and 0.20
ence the frequency content of the earthquake with an average value of 0.15 as suggested by
acceleration. It was observed, however, that the Park et al, (1987). Figure 9(a) shows the influence
ductility ratio and the maximum inelastic defor- of β on the damage index. To study the effect of
mation of the structure decrease for higher damp- the initial natural frequency of the structure on
ing ratios. Thus, the ductility ratio decreases to the damage index, the structure stiffness was
2.43 when the damping ratio is taken as 0.05 while varied while keeping all other parameters un-
the ductility ratio increases to 2.89 when the changed and the critical earthquake was com-
damping ratio reduces to 0.01. It was also observed puted for each case. Subsequently, the value of
that the inelastic structure with higher damping DIPA was calculated for each case. In the nu-
ratio dissipates more energy through damping merical calculations β was taken as 0.15 and

140
Damage Assessment of Inelastic Structures

Figure 5. Optimal earthquake input and associated structural responses for case 4 (a) Fourier ampli-
tude of the ground acceleration (b) Normalized inelastic deformation (c) Hysteretic restoring force (d)
Dissipated energy

x max , µmax are taken as 0.10 m and 2.64, respec- of the structure is expected. The value of DIPA
tively. for the structure with ω0 greater than about 1.70
The value of µu was taken as 6 in Figures. 9(a) Hz is less than 0.40 and thus the structure does
and 8 in Figure 9(b). It was found that the damage not experience total damage but repairable dam-
index for the structure with initial natural fre- age. This observation is consistent since the site
quency smaller than 1.65 is higher than 0.40 and dominant frequency is around 1.65 Hz and since
thus either total collapse or damage beyond repair the Fourier amplitude of the ground acceleration

Figure 6. Optimal earthquake acceleration and velocity (a) Case 1 (b) Case 4

141
 Damage Assessment of Inelastic Structures

Figure 7. Fourier amplitude of recorded earthquakes (a) San Fernando 1971 (b) Hyogoken-Nanbu 1995

Figure 8. (a) Effect of strain hardening ratio on dissipated yield energy (b) Effect of damping on dis-
sipated damping energy

Figure 9. (a) Effect of the value of β on the damage index (b) damage spectra for inelastic SDOF build-
ings

142
Damage Assessment of Inelastic Structures

Table 4. Response parameters for alternative constraint scenarios (α = 0.05, ζ = 0.03)

Case xmax (m) μmax xp (m) Nrv* DIPA Damage status

1 0.47 4.65 0.07 60 1.15 Total collapse
2 0.45 4.53 0.06 54 0.97 Damaged beyond repair
3 0.41 4.14 0.07 49 0.72 Damaged beyond repair
4 0.26 2.64 0.05 44 0.37 Repairable damage
* Nrv = number of yield reversals (see Table 1)

is seen to be located in the stiff side of the initial x y = Lεy / cos θ = 0.0381 m.
frequency of the inelastic structure.
Thus, brace 1 yields when | x 1 | = 0.0381
4.2 Inelastic Two-Story m and brace 2 yields when | x 2 − x 1 | = 0.0381
Frame Structure m. The objective function is taken as the weight-
ed damage indices in braces 1 and 2. In the nu-
A two-story braced building frame is considered merical analysis, the parameters of the Newmark
to demonstrate the formulation developed in this β-method were taken as δ = 1 / 2; α = 1 / 6 and
chapter for MDOF inelastic structures (Moustafa the time step ∆t = 0.005 s.
2009). The material behavior of the braces is The results of this example are shown in Fig-
taken as bilinear ( k2 = γk1 ) as shown in Figure ures 10 and 11. In general, the feature observed
1(a). The floor masses are taken as for the future earthquakes in the previous example
m1 = m 2 = 1.75 × 105 Ns2/m, the cross-section- was also observed in this example. However, the
al areas of the braces are A1 = A2 = 6.45 × 10−4 inelastic deformation and the associated damage
m2, the Young’s modulus = 2.59 × 1011 N/m2, and were seen to depend on the two vibration modes.
the strain hardening ratio = 0.10 (i.e., ratio of the Thus, the maximum ductility ratio μ for case 1 is
post-yield stiffness to the pre-yield stiffness). 4.34 while that produced from constraint case 4
When both braces are behaving elastically, the is 2.27. Similarly, the maximum response reduces
stiffness matrix Ks = Kel , if brace 1 yields from 0.15 m to 0.08 m when the constraints on
Ks = K1 , if brace 2 yields Ks = K2 and if both UBFAS and LBFAS are brought in. The earthquake
braces yield Ks = K12 . These matrices are given input energy to the inelastic system is mainly dis-
sipated by yielding and nonlinear damping of the
in Moustafa (2009). The structure is assumed to
structure. The hysteretic and damping energies are
start from rest. The first two natural frequencies
significantly higher than the recoverable strain and
of the elastic structure were computed as
kinetic energy. The kinetic and recoverable strain
ω1 = 6.18 rad/s and ω2 = 16.18 rad/s. A Rayleigh
energies are small and diminish near the end of the
proportional damping C = aM + bKs with a =
ground shaking. The energy dissipated by yield-
0.2683, b = 0.0027 is adopted. These values are ing is significantly higher than that dissipated by
selected such that the damping ratio in the first damping. The weighted damage index for case 1
two modes is 0.03. This implies that the damping was about 0.96 implying total collapse while for
forces in braces are nonlinear hysteretic functions case 4 the damage index was about 0.35 implying
of the deformed shape of the structure. Let the repairable damage.
yield strain of the braces εy = 0.002 for both It may be noted that the numerical illustrations
tension and compression. The braces will yield of the formulation developed in this paper were
at a relative displacement demonstrated for simple structures with bilinear

143
 Damage Assessment of Inelastic Structures

Figure 10. Critical acceleration xg (t ) for inelastic structure for case 1 (a) Time history (b) Fourier
amplitude spectrum

Figure 11. Critical acceleration xg (t ) for inelastic structure for case 2 (a) Time history (b) Fourier
amplitude spectrum

and elastic-plastic force-deformation laws. The This includes: (1) the weak cumulative component
application of the proposed method to more com- for practical cases given the typical dominance
plex structures and the use of more detailed of the peak displacement term over the accumu-
degradation models (e.g., trilinear degradation, lated energy term, (2) the use of a linear combina-
Takeda and Clough models) need to be investi- tion of deformation and energy in spite of the
gated. Additionally, in this paper Park and Ang obvious nonlinearity of the problem and the inter-
damage index has been used to assess the structure dependence of the two quantities, and (3) the lack
performance. It may be emphasized that this dam- of considering the loading sequence effect in the
age index has some limitations (Mehanny & cumulative energy term. Furthermore, when EH
Deierlein, 2000, Bozorgenia & Bertero 2003). = 0 (elastic behavior), the value of DIPA should

144
Damage Assessment of Inelastic Structures

be zero. However, the value of DIPA computed is of substantial importance in deriving critical
from Eq (5) will be greater than zero. Similarly, earthquake loads for inelastic structures. This is
when the system reaches its maximum mono- because damage indices imply that the structure
tonic deformation, while DIPA should be 1.0, is damaged by a combination of repeated stress
however, Eq (5) leads to DIPA greater than 1.0. reversals and high stress excursions. This also
Chai et al (1995) proposed modification to DIPA facilitates assessing the safety of the structure by
to correct for the second drawback only. The study, providing a quantitative measure on the neces-
also, examined experimentally the implication of sary repair.
the energy-based linear damage model of DIPA. In this chapter, the worst earthquake load is
Despite the drawbacks of DIPA, it has been exten- derived based on available information using
sively used by many researchers, mainly due to inverse nonlinear dynamic analysis, optimization
its simplicity and the extensive calibration against techniques and damage indices. It was seen that if
experimentally observed seismic structural dam- available information is limited to the energy and
age during earthquakes (mainly for reinforced PGA, the resulting earthquake is highly resonant
concrete structures). Bozorgenia & Bertero (2003) and produces conservative damage. When extra
proposed two improved damage indices that information on the Fourier amplitude spectra is
overcome some of the drawbacks associated with available, more realistic earthquake loads (in
DIPA. terms of frequency content, amplitude, inelastic
In this chapter, worst earthquakes that maxi- deformations and damage indices produce) are
mize the structure’s damage were obtained using obtained. The influences of the strain hardening
deterministic methods. Critical earthquakes can and damping ratios on the estimated design loads
be formulated using stochastic processes, random were studied. Critical damage spectra for the site
vibration theory and reliability analysis which were also established. These spectra provide upper
provides a powerful alternative to the methodol- bounds on the structural damage and necessary
ogy developed here (see, e.g., Abbas & Manohar, repair under possible future earthquakes. The
2005, 2007). formulation developed in this chapter was dem-
onstrated for inelastic frame structures modeled
with bilinear and elastic-plastic force-deformation
5. CONCLUDING REMARKS laws. In other words, non-deteriorating structures
are only considered. Future extension of the
A methodology for assessing damage in structures present research requires the use of nonlinear
under critical future earthquake loads is developed degradation models that facilitate the development
in this chapter. The novelty of this research lies of plastic hinges in the structure. In this case the
in combining damage indices, nonlinear opti- computations will increase considerably due to the
mization and nonlinear time-history analysis in complexity in estimating the structural response.
assessing the structural performance under future Finally, it may be emphasized that in the present
earthquakes. Damage descriptors are introduced work, the structural properties have been kept
in deriving the worst earthquake ground motion. unchangeable. It is possible to apply the proposed
The structural damage is quantified in terms of methodology for optimal design of the structure
Park and Ang damage indices. As is well known, under future earthquakes. Herein, an initial guess
damage indices describe the damage state of the for the dimensions of the structure’s members
structure and correlate well with the actual damage needs to be assumed and an iterative procedure
displayed during earthquakes. The quantification has to be carried out leading to the optimal de-
of the structure’s damage using damage indices sign of the structure, the system-dependent worst

145
 Damage Assessment of Inelastic Structures

earthquake and the associated damage (Takewaki, Baker, J. W. (2011). Conditional mean spectrum:
2002b, 2007, Saikat & Manohar, 2005, Fujita et Tool for ground motion selection. Journal of Struc-
al, 2010). tural Engineering, 137(3), 322–331. doi:10.1061/
(ASCE)ST.1943-541X.0000215
Baker, J. W., & Cornell, C. A. (2006). Spectral
ACKNOWLEDGMENT
shape, epsilon and record selection. Earthquake
Engineering & Structural Dynamics, 35, 1077–
This research work is partly supported by research
1095. doi:10.1002/eqe.571
funds from the Japanese Society for the Promotion
of Science (JSPS) under Grant No. JSPS-P-08073. Baker, J. W., Lin, T., Shahi, S. K., & Jayaram, N.
The support is gratefully acknowledged. (2011). New ground motion selection procedures
and selected motions for the PEER transporta-
tion research program. PEER Technical Report
REFERENCES 2011/03. 106p.

Abbas, A. M. (2006). Critical seismic load inputs Bommer, J. J., & Acevedo, A. B. (2004). The use of
for simple inelastic structures. Journal of Sound real earthquake accelerograms as input to dynamic
and Vibration, 296, 949–967. doi:10.1016/j. analysis. Journal of Earthquake Engineering, 1,
jsv.2006.03.021 43–91. doi:10.1080/13632460409350521

Abbas, A. M., & Manohar, C. S. (2002). Inves- Boore, D. M. (1983). Stochastic simulation of
tigations into critical earthquake load models high-frequency ground motions based on seis-
within deterministic and probabilistic frameworks. mological models of the radiated spectra. Bul-
Earthquake Engineering & Structural Dynamics, letin of the Seismological Society of America, 73,
31, 813–832. doi:10.1002/eqe.124 1865–1894.

Abbas, A. M., & Manohar, C. S. (2005). Reliabil- Bozorgenia, Y., & Bertero, V. V. (2003). Dam-
ity-based critical earthquake load models. Part 2: age spectra: Characteristics and applications to
Nonlinear structures. Journal of Sound and Vibra- seismic risk reduction. Journal of Structural
tion, 287, 883–900. doi:10.1016/j.jsv.2004.12.003 Engineering, 129(4), 1330–1340. doi:10.1061/
(ASCE)0733-9445(2003)129:10(1330)
Abbas, A. M., & Manohar, C. S. (2007).
Reliability-based vector nonstationary random Bozorgnia, Y., & Bertero, V. V. (Eds.). (2004).
critical earthquake excitations for parametrically Earthquake engineering. New York, NY: CRC
excited systems. Structural Safety, 29, 32–48. Press. doi:10.1201/9780203486245
doi:10.1016/j.strusafe.2005.11.003 Brune, J. N. (1970). Tectonic stress and the spectra
Akiyama, H. (1985). Earthquake-resistant limit- of seismic shear waves from earthquakes. Jour-
state design for buildings. Tokyo, Japan: Univer- nal of Geophysical Research, 75, 4997–5009.
sity of Tokyo Press. doi:10.1029/JB075i026p04997

Arias, A. (1970). A measure of earthquake inten- Buratti, N., Stafford, P. J., & Bommer, J. J. (2011).
sity: Seismic design of nuclear power plants (pp. Earthquake accelerogram selection and scaling
438–468). Cambridge, MA: MIT Press. procedures for estimating the distribution of drift
response. Journal of Structural Engineering,
Arora, J. S. (2004). Introduction to optimum de- 137(3), 345–357. doi:10.1061/(ASCE)ST.1943-
sign. San Diego, CA: Elsevier Academic Press. 541X.0000217

146
Damage Assessment of Inelastic Structures

Caleman, T., Branch, M. A., & Grace, A. (1999). Elishakoff, I., & Ohsaki, M. (2010). Optimiza-
Optimization toolbox for the use with Matlab. tion and anti-optimization of structures under
User’s guide. USA: The MATH WORKS Inc. uncertainty. Singapore: Imperial College Press.
doi:10.1142/9781848164789
Chai, Y. H., Romstad, K. M., & Bird, S. M.
(1995). Energy-based linear damage model for Fajfar, P. (1992). Equivalent ductility factors, tak-
high-intensity seismic loading. Journal of Struc- ing into account low-cyclic fatigue. Earthquake
tural Engineering, 121(5), 857–864. doi:10.1061/ Engineering & Structural Dynamics, 21, 837–848.
(ASCE)0733-9445(1995)121:5(857) doi:10.1002/eqe.4290211001
Choi, H., & Kim, J. (2006). Energy-based design Fajfar, P., & Krawinklere, H. (1997). Seismic
of buckling-restrained braced frames using hys- design methodologies for the next generation of
teretic energy spectrum. Engineering Structures, codes. Rotterdam, The Netherlands: Balkema.
28, 304–311. doi:10.1016/j.engstruct.2005.08.008
Fardis, M. N. (Ed.). (2010). Advances in perfor-
Comartin, C., Brvez, S., Naeim, F., Greene, mance-based earthquake engineering. New York,
M., Blondet, M., & Cherry, S. (2004). A chal- NY: Springer.
lenge to the earthquake engineering profession-
Fujita, K., Moustafa, A., & Takewaki, I. (2010).
als. Earthquake Spectra, 20(4), 1049–1056.
Optimal placement of viscoelastic dampers and
doi:10.1193/1.1809130
supporting members under variable critical excita-
Conte, J. P., & Peng, B. F. (1997). Fully nonstation- tions. Earthquakes and Structures, 1(1), 43–67.
ary analytical earthquake ground motion model.
Ghobara, A., Abou-Elfath, H., & Biddah, A.
Journal of Engineering Mechanics, 123, 15–24.
(1999). Response-based damage assessment of
doi:10.1061/(ASCE)0733-9399(1997)123:1(15)
structures. Earthquake Engineering & Structural
Cosenza, C., Manfredi, G., & Ramasco, R. (1993). Dynamics, 28, 79–104. doi:10.1002/(SICI)1096-
The use of damage functionals in earthquake engi- 9845(199901)28:1<79::AID-EQE805>3.0.CO;2-
neering: A comparison between different methods. J
Earthquake Engineering & Structural Dynamics,
Goel, R. K. (1997). Seismic response of asym-
22, 855–868. doi:10.1002/eqe.4290221003
metric systems: Energy-based approach.
COSMOS. (2005). Consortium organizations for Journal of Structural Engineering, 123(11),
strong-motion observation systems. Retrieved 1444–1453. doi:10.1061/(ASCE)0733-
from http://db.cosmos-eq.org/scripts/default.plx 9445(1997)123:11(1444)
Decanini, L. D., & Mollaioli, F. (2001). An Haldar, A. (Ed.). (2006). Recent developments
energy-based methodology for the assessment of in reliability-based civil engineering. Singa-
seismic demand. Soil Dynamics and Earthquake pore: World Scientific Publishing Co. Pte. Ltd.
Engineering, 21, 113–137. doi:10.1016/S0267- doi:10.1142/9789812707222
7261(00)00102-0
Hanks, T. G., & McGuire, R. K. (1981). The char-
Drenick, R. F. (1977). The critical excitation of acter of high frequency ground motions based on
nonlinear systems. Journal of Applied Mechanics, seismic shear waves. Bulletin of the Seismological
18, 333–336. doi:10.1115/1.3424047 Society of America, 71, 2071–2095.

147
 Damage Assessment of Inelastic Structures

Hart, G. C., & Wong, K. (2000). Structural dy- Khashaee, P. (2004). Damage-based seismic
namics for structural engineers. New York, NY: design of structures. Earthquake Spectra, 21(2),
John Wiley & Sons. 371–387. doi:10.1193/1.1896366
Haselton, C., Baker, J. W., Liel, A. B., & Deier- Kiureghian, A. D., & Crempien, J. (1989). An
lein, G. G. (2011). Accounting for ground mo- evolutionary model for earthquake ground motion.
tion spectral shape characteristics in structural Structural Safety, 6, 235–246. doi:10.1016/0167-
collapse assessment through an adjustment for 4730(89)90024-6
epsilon. Journal of Structural Engineering,
Liang, Q. Q. (2005). Performance-based optimi-
137(3), 332–344. doi:10.1061/(ASCE)ST.1943-
zation of structures. New York, NY: Spon Press.
541X.0000103
doi:10.4324/9780203334713
He, W. L., & Agrawal, A. K. (2008). Ana-
Mahin, S. A., & Bertero, V. V. (1981). An evalu-
lytical model of ground motion pulses for the
ation of inelastic seismic design spectra. Journal
design and assessment of seismic protective
of the Structural Division, 107(ST9), 1777–1795.
systems. Journal of Structural Engineering,
134(7), 1177–1188. doi:10.1061/(ASCE)0733- McGuire, R. K. (1995). Probabilistic seismic
9445(2008)134:7(1177) hazard analysis and design earthquake: Closing
the loop. Bulletin of the Seismological Society
Housner, G. W., & Hudson, D. E. (1958). The
of America, 85, 1275–1284.
Port Hueneme earthquake of March 18, 1957.
Bulletin of the Seismological Society of America, Mehanny, S. S., & Deierlein, G. G. (2000). Model-
48, 163–168. ing of assessment of seismic performance of com-
posite frames with reinforced concrete columns
Iyengar, R. N. (1972). Worst inputs and a bound
and steel beams. Report No. 135, The John Blume
on the highest peak statistics of a class of non-
Earthquake Research Center, Stanford University.
linear systems. Journal of Sound and Vibration,
25, 29–37. doi:10.1016/0022-460X(72)90593-7 Moustafa, A. (2002). Deterministic/reliability-
based critical earthquake load models for linear/
Kalkan, E., & Kunnath, S. K. (2006). Effects of
nonlinear engineering structures. Doctoral disser-
fling step and forward directivity on seismic re-
tation, Department of Civil Engineering, Indian
sponse of buildings. Earthquake Spectra, 22(2),
Institute of Science, Bangalore, India.
367–390. doi:10.1193/1.2192560
Moustafa, A. (2009). Critical earthquake load
Kalkan, E., & Kunnath, S. K. (2008). Rel-
inputs for multi-degree-of-freedom inelastic
evance of absolute and relative energy con-
structures. Journal of Sound and Vibration, 325,
tent in seismic evaluation of structures. Ad-
532–544. doi:10.1016/j.jsv.2009.03.022
vances in Structural Engineering, 11(1), 1–18.
doi:10.1260/136943308784069469 Moustafa, A. (2009). Discussion of the effect of
energy concentration of earthquake ground mo-
Kalkan, E., & Luco, N. (Eds.). (2011). Earthquake
tions on the nonlinear response of RC structures.
ground motion selection and modification for
Soil Dynamics and Earthquake Engineering, 29,
nonlinear dynamic analysis of structures. Jour-
1181–1183. doi:10.1016/j.soildyn.2009.02.004
nal of Structural Engineering, 137(3), 277–467.
doi:10.1061/(ASCE)ST.1943-541X.0000355

148
Damage Assessment of Inelastic Structures

Moustafa, A. (2009). Discussion of a new ap- Moustafa, A., & Takewaki, I. (2010). Charac-
proach of selecting real input ground motions terization and modeling of near-fault pulse-like
for seismic design: The most unfavourable real strong ground motion via damage-based critical
seismic design ground motions. Earthquake En- excitation method. Structural Engineering &
gineering & Structural Dynamics, 38, 1143–1149. Mechanics, 34(6), 755–778.
doi:10.1002/eqe.885
Moustafa, A., & Takewaki, I. (2011). Response of
Moustafa, A. (2010). Identification of resonant nonlinear SDOF structures to random accelera-
earthquake ground motion. Indian Academy of tion sequences. Engineering Structures, 33(4),
Sciences, 35(3), 355–371. 1251–1258. doi:10.1016/j.engstruct.2011.01.002
Moustafa, A. (2010). Discussion of analytical Moustafa, A., Takewaki, I., & Wijeyewickrema,
model of ground motion pulses for the design and A. K. (2010). Selection of earthquake records
assessment of seismic protective systems. Jour- for time-history analysis of structures. Joint 7th
nal of Structural Engineering, 137(1), 229–230. International Conference on Urban Earthquake
doi:10.1061/(ASCE)ST.1943-541X.134 Engineering and 5th International Conference on
Earthquake Engineering, 3-5 March, Tokyo, Japan.
Moustafa, A. (2010). Closure to discussion of
critical earthquake load inputs for multi-degree- Moustafa, A., Ueno, K., & Takewaki, I. (2010).
of-freedom inelastic structures. Journal of Sound Critical earthquake loads for SDOF inelastic
and Vibration, 330, 356–360. doi:10.1016/j. structures considering evolution of seismic waves.
jsv.2010.09.002 Earthquakes and Structures, 1(2), 147–162.
Moustafa, A. (2011). Damage-based design earth- Nakashima, M., Saburi, K., & Tsuji, B. (1996).
quake loads for SDOF inelastic structures. Jour- Energy input and dissipation behavior of structures
nal of Structural Engineering, 137(3), 456–467. with hysteretic dampers. Earthquake Engineering
doi:10.1061/(ASCE)ST.1943-541X.0000074 & Structural Dynamics, 25, 483–496. doi:10.1002/
(SICI)1096-9845(199605)25:5<483::AID-
Moustafa, A., & Mahadevan, S. (2010). Probabilis-
EQE564>3.0.CO;2-K
tic critical earthquake excitations using earthquake
response spectra. [Building and Housing]. Asian Newmark, N. M., & Hall, W. J. (1982). Earthquake
Journal of Civil Engineering, 11(3), 295–319. spectra and design. Monograph, Earthquake
Engineering Research Institute. EERI.
Moustafa, A., & Takewaki, I. (2010). Modeling
critical strong ground motion sequences on inelas- Otani, S. (1981). Hysteretic models of reinforced
tic structures. Advances in Structural Engineering, concrete for earthquake response analysis. Journal
13(4), 665–679. doi:10.1260/1369-4332.13.4.665 of the Faculty of Engineering. University of Tokyo,
36(2), 407–441.
Moustafa, A., & Takewaki, I. (2010). Determin-
istic and probabilistic representation of near- Pacific Earthquake Research (PEER) Center.
field pulse-like ground motion. Soil Dynamics (2011). Retrieved from http://peer.berkeley.edu/
and Earthquake Engineering, 30, 412–422. peer_ground_motion_database
doi:10.1016/j.soildyn.2009.12.013

149
 Damage Assessment of Inelastic Structures

Papadrakakis, M., Charmpis, D. C., Lagaros, N. D., Priestely, M. J. N., Calvi, G. M., & Kowalsky, M.
& Tsompanakis, Y. (Eds.). (2009). Computational J. (2007). Displacement-based seismic design of
structural dynamics and earthquake engineering. structures. Pavia, Italy: IUSS Press.
London, UK: Taylor & Francis.
Quek, S. T., Teo, Y. P., & Balendra, T. (1990).
Park, Y. J., & Ang, A. H.-S. (1985). Mecha- Non-stationary structural response with evolu-
nistic seismic damage model for reinforced tionary spectra using seismological input model.
concrete. Journal of Structural Engineering, Earthquake Engineering & Structural Dynamics,
111(4), 722–739. doi:10.1061/(ASCE)0733- 19, 275–288. doi:10.1002/eqe.4290190210
9445(1985)111:4(722)
Reiter, L. (1990). Earthquake hazard analysis.
Park, Y. J., Ang, A. H.-S., & Wen, Y. K. (1985). New York, NY: Columbia University Press.
Seismic damage analysis of reinforced concrete
Riddell, F. (1995). Inelastic design spectra ac-
buildings. Journal of Structural Engineering,
counting for soil conditions. Earthquake Engi-
111(4), 740–757. doi:10.1061/(ASCE)0733-
neering & Structural Dynamics, 24, 1491–1510.
9445(1985)111:4(740)
doi:10.1002/eqe.4290241106
Park, Y. J., Ang, A. H.-S., & Wen, Y. K.
Saikat, S., & Manohar, C. S. (2005). Inverse
(1987). Damage-limiting aseismic design of
reliability based structural design for system de-
buildings. Earthquake Spectra, 3(1), 1–26.
pendent critical earthquake loads. Probabilistic
doi:10.1193/1.1585416
Engineering Mechanics, 20, 19–31. doi:10.1016/j.
Pecker, A. (Ed.). (2007). Advanced earthquake probengmech.2004.03.006
engineering analysis. Udine, Italy: Springer.
Sarkar, A. (2003). Linear stochastic dynamical
doi:10.1007/978-3-211-74214-3
system under uncertain load: Inverse reliability
Penelis, G. G., & Kappos, A. J. (1997). Earth- analysis. Journal of Engineering Mechanics,
quake-resistant concrete structures. London, UK: 129(6), 665–671. doi:10.1061/(ASCE)0733-
E & FN SPON. 9399(2003)129:6(665)
Philippacopoulos, A. J., & Wang, P. C. (1984). SEAOC. Vision Committee. (2002). Performance
Seismic inputs for nonlinear structures. Jour- based seismic design engineering. Sacramento,
nal of Engineering Mechanics, 110, 828–836. CA: Structural Engineers Association of California
doi:10.1061/(ASCE)0733-9399(1984)110:5(828) (SEAOC) Report.
Pinho, R. (2007). Non linear dynamic analysis of Shinozuka, M. (1970). Maximum structural
structures subjected to seismic action. In Pecker, A. response to seismic excitations. Journal of En-
(Ed.), Advanced earthquake engineering analysis. gineering Mechanics, 96, 729–738.
New York, NY: Springer Wien. doi:10.1007/978-
Shinozuka, M., & Henry, L. (1965). Random vibra-
3-211-74214-3_5
tion of a beam column. Journal of Engineering
Plevris, V. (2009). Innovative computational Mechanics, 91, 123–143.
techniques for the optimum structural design
Strasser, F. O., & Bommer, J. J. (2009). Review:
considering uncertainties. Ph. D. dissertation,
Strong ground motions – Have we seen the worst?
Institute of Structural Analysis and Seismic Re-
Bulletin of the Seismological Society of America,
search, National Technical University of Athens,
99(5), 2613–2637. doi:10.1785/0120080300
June 2009.

150
 Damage Assessment of Inelastic Structures

Takeda, T., Sozen, M. A., & Nielsen, N. (1970). Tsompanakis, Y., Lagaros, N. D., & Papadrakakis,
Reinforced concrete response to simulated M. (Eds.). (2008). Structural design optimization
earthquakes. Journal of the Structural Division, considering uncertainties. London, UK: Taylor
96(ST12), 2557–2573. & Francis.
Takewaki, I. (2001). Probabilistic critical ex- Uang, C.-M., & Bertero, V. V. (1990). Evalua-
citation for MDOF elastic-plastic structures tion of seismic energy in structures. Earthquake
on compliant ground. Earthquake Engineer- Engineering & Structural Dynamics, 19, 77–90.
ing & Structural Dynamics, 30, 1345–1360. doi:10.1002/eqe.4290190108
doi:10.1002/eqe.66
USGS/EERI Advance Reconnaissance Team.
Takewaki, I. (2002a). Seismic critical excita- (2010). The Mw 7.0 Haiti earthquake of January
tion method for robust design: A review. Jour- 12, 2010.
nal of Structural Engineering, 128, 665–672.
Westermo, B. D. (1985). The critical excita-
doi:10.1061/(ASCE)0733-9445(2002)128:5(665)
tion and response of simple dynamic systems.
Takewaki, I. (2002b). Robust building stiff- Journal of Sound and Vibration, 100, 233–242.
ness design for variable critical excitations. doi:10.1016/0022-460X(85)90417-1
Journal of Structural Engineering, 128(12),
Wong, K. K. F., & Yong, R. (2002). Earthquake
1565–1574. doi:10.1061/(ASCE)0733-
response and energy evaluation of inelastic
9445(2002)128:12(1565)
structures. Journal of Engineering Mechanics,
Takewaki, I. (2004). Bound of earthquake input 128(3), 308–317. doi:10.1061/(ASCE)0733-
energy. Journal of Structural Engineering, 9399(2002)128:3(308)
130, 1289–1297. doi:10.1061/(ASCE)0733-
Yamaguchi, H., & El-Abd, A. (2003). Effect of
9445(2004)130:9(1289)
energy input characteristics on hysteretic damper
Takewaki, I. (2007). Critical excitation methods efficiency. Earthquake Engineering & Structural
in earthquake engineering. Elsevier Science. Dynamics, 32, 827–843. doi:10.1002/eqe.250
Takewaki, I. (2009). Building control with pas- Zahrah, T. F., & Hall, W. J. (1984). Earthquake en-
sive dampers. Singapore: John Wiley & Sons. ergy absorption in sdof structures. Journal of Struc-
doi:10.1002/9780470824931 tural Engineering, 110, 1757–1772. doi:10.1061/
(ASCE)0733-9445(1984)110:8(1757)
Takewaki, I., Murakami, S., Fujita, K., Yoshi-
tomi, S., & Tsuji, M. (2011). The 2011 off the
Pacific coast of Tohoku earthquake and response
of high-rise buildings under long-period ground
motions. Soil Dynamics and Earthquake En-
gineering, 31(11), 1511–1528. doi:10.1016/j.
soildyn.2011.06.001

151

View publication stats