SEISMIC ANALYSIS

AND DESIGN USING

THE ENDURANCE
TIME METHOD
SEISMIC ANALYSIS
AND DESIGN USING
THE ENDURANCE
TIME METHOD
Volume I

Concepts and Development

H.E. ESTEKANCHI AND H.A. VAFAI

MOMENTUM PRESS, LLC, NEW YORK

Seismic Analysis and Design Using the Endurance Time Method,
­Volume I: Concepts and Development

Copyright © Momentum Press®, LLC, 2018.

All rights reserved. No part of this publication may be reproduced, stored

in a retrieval system, or transmitted in any form or by any means—­
electronic, mechanical, photocopy, recording, or any other—except for
brief quotations, not to exceed 400 words, without the prior permission
of the publisher.

First published by Momentum Press®, LLC

222 East 46th Street, New York, NY 10017
www.momentumpress.net

ISBN-13: 978-1-94708-304-2 (print)

ISBN-13: 978-1-94708-305-9 (e-book)

Momentum Press Sustainable Structural Systems Collection

Collection ISSN: 2376-5119 (print)

Collection ISSN: 2376-5127 (electronic)

Cover and interior design by Exeter Premedia Services Private Ltd.,

Chennai, India

10 9 8 7 6 5 4 3 2 1

Printed in the United States of America

 Abstract

A new approach to seismic assessment of structures called endurance time

method (ETM) is developed. ETM is a dynamic analysis procedure in
which intensifying dynamic excitations are used as the loading function.
ETM provides many unique benefits in seismic assessment and design of
structures. ETM is a response history-based procedure. ETM consider-
ably reduces the computational effort needed in typical response history
analyses. Conceptual simplicity makes ETM a great tool for preliminary
response history analysis of almost any dynamic structural system. Most
important areas of application of ETM are in the fields of seismic design
optimization, value-based seismic design, and experimental studies. This
book is aimed to serve as a coherent source of information for students,
engineers, and researchers who want to familiarize themselves with the
concepts and put the concepts into practice.

KEYWORDS

endurance time method, dynamic structural analysis, earthquake engineer-

ing, seismic assessment, seismic design, structural design optimization,
value-based seismic design
 Contents

List of Figures xi
List of Tablesxvii
Foreword xix
Preface  xxi

1  Introduction to the Endurance Time Method  1

1.1 Introduction  1
1.2  The Endurance Time Concept  2
1.3  Dynamic Excitation  5
1.4  Analysis Method  8
1.5  Endurance Criteria  9
1.6  Application to Linear SDOF Systems  10
1.7  Application to MDOF Systems  14
1.8  Summary and Conclusions  18
2  Properties of the Endurance Time Excitation
Functions  21
2.1  Introduction  21
2.2  Generation of ET Acceleration Functions  22
2.3  Comparison of ET Response Spectrum with
Real Earthquake  24
2.4   Basic Ground Motion Properties  26
2.5   Frequency Content  27
2.6   Power Spectral Density  28
2.7   Other Ground Motion Parameters  30
2.8   Intensity Parameters  30
2.9   Period Parameters  33
2.10  Peak Velocity Acceleration Ratio  35
2.11  Cumulative Absolute Velocity  35
viii  •   Contents

2.12 Sustained Maximum Acceleration (SMA)

and Velocity (SMV) 36
2.13 Effective Design Acceleration (EDA) 36
2.14  Summary and Conclusions  36
3  Duration Properties of the Endurance Time Excitation
Functions  41
3.1 Introduction  41
3.2  Ground Motion Selection  42
3.3  ET Excitation Functions  43
3.4 Review of Definitions of Strong-Motion Duration 44
3.5 Comparison Between ET Accelerograms and
Real Ground Motions 47
3.6  Evaluation of Proposed Target Time 47
3.7  Summary and Conclusions 51
4  Generating ET Excitation Functions by Numerical
Optimization  55
4.1 Introduction  55
4.2 Generating ET Acceleration Functions 57
4.3 Nonlinear Least Squares Formulation 62
4.4 Improved ET Acceleration Functions 64
4.5 Optimization of Long Duration ET Acceleration
Functions  66
4.6 Comparison of ETAFs in the Analysis of SDOF
Systems71
4.7  Summary and Conclusions  73
5  Correlating Analysis Time with Significant
Intensity Indicators  77
5.1 Introduction  77
5.2  Endurance Time Method  78
5.3 Application of ET Method in Performance-Based Design 79
5.4 Correlation Between Time in ET Analysis and Return
Period82
5.5  Explanatory Case Study  89
5.6  Summary and Conclusions  92
6  ET Analysis of MDOFs  97
6.1 Introduction  97
6.2  Concept of ET Method  100
6.3 Specifications of Models, Ground Motions, and
Acceleration Functions  102
 Contents  •   ix

6.4 Comparison with Nonlinear Response History

Analysis106
6.5 Application of ET Method in Seismic Rehabilitation
of Buildings 119
6.6  Summary and Conclusions  121
Index 125
 List of Figures

Figure 1.1.  Hypothetical shaking table experiment. 3

Figure 1.2.  Damage curve against time for a typical structure
subjected to intensifying accelerogram. 4
Figure 1.3.  Input profile functions (a) Linear (b) Increasing rate
(c) Decreasing rate. 6
Figure 1.4.  A typical random accelerogram with PGA = 1. 6
Figure 1.5.  Filtered frequency content. 7
Figure 1.6.  Modified frequency content. 8
Figure 1.7.  Response of modified accelerogram compared to
codified value. 8
Figure 1.8.  Intensifying accelerogram acc1. 9
Figure 1.9.    Acceleration time history for ST05. 10
Figure 1.10. Maximum acceleration and average for ST05. 11
Figure 1.11. Displacement time history for ST05. 12
Figure 1.12. Maximum displacement and average for ST05. 12
Figure 1.13. Maximum average accelerations. 13
Figure 1.14.  Maximum average displacements. 13
Figure 1.15. Three-story steel moment frames. 15
Figure 1.16. Story drifts time history for frame f3 subjected to acc1. 15
Figure 1.17. Maximum story drifts for frame f3 subjected to acc1-3
accelerograms.16
Figure 1.18. Maximum story drifts for frame f3w subjected to acc1-3
accelerograms.17
Figure 1.19. Maximum average story drifts for frames f3 and f3w. 17
Figure 2.1. ETA20a03 acceleration function. 23
xii  •   List of Figures

Figure 2.2.  Average response spectra of ETA20a acceleration

functions for ξ = 5% at different time.
(a) Pseudo-Acceleration, (b) Displacement Response. 24
Figure 2.3.  Comparison of ET response spectra (t=10 sec), INBC
code, and Earthquakes response spectra. 25
Figure 2.4.  Velocity and displacement of ETA20a acceleration
functions, (a) velocity, (b) displacement. 26
Figure 2.5.  Comparison of frequency content between ET
acceleration functions (at t=10 sec) and ground motions. 27
Figure 2.6.  Power Spectral Density function for average of three
ET acceleration functions ETA20a01-03. 29
Figure 2.7.  Comparison of Power Spectral Density for average of
seven ground motions and average of ET acceleration
functions at t=10 sec. 29
Figure 2.8.  Comparison of input energy between ET acceleration
functions and ground motions; (a)EI (Input Energy)
at t=0.7 sec, (b) energy spectra (t=10 sec). 31
Figure 2.9.  Comparison of SED between ET acceleration functions
and ground motions. 31
Figure 2.10. Ia for ETA20a acceleration functions and ground
motions.32
Figure 2.11. A95 for ET acceleration functions and ground motions. 33
Figure 2.12. Ic for ET acceleration functions and Ground motions. 33
Figure 2.13. ASI for ET acceleration functions and ground motions. 33
Figure 2.14. VSI for ET acceleration functions and ground motions. 34
Figure 2.15. Tp of ET acceleration functions and ground motions. 34
Figure 2.16. Tm of ET acceleration functions and ground motions. 35
Figure 2.17. Vmax/amax for ET acceleration functions and ground
motions.35
Figure 2.18. CAV for ET acceleration functions and ground motions. 36
Figure 2.19. SMA for ET acceleration functions and ground motions. 37
Figure 2.20. SMV for ET acceleration functions and ground motions. 37
Figure 2.21. EDA for ET acceleration functions and ground motions. 37
Figure 3.1.  “Bracketed Duration” of an accelerogram. 45
Figure 3.2.  “Uniform Duration” of an accelerogram. 45
Figure 3.3.  “Significant Duration” of an accelerogram. 46
 List of Figures  •   xiii

Figure 3.4.  Definition of strong-motion duration by McCann and

Shah (1979). 46
Figure 3.5.  Procedure to determine the target time. 47
Figure 3.6.  Basic rules for peak-oriented hysteretic model. 48
Figure 3.7.  Revelation of correlation between the results of ET
records and real ground motions. 50
Figure 4.1.  General procedure of generating ET acceleration
functions.57
Figure 4.2.  ETA20d01 acceleration function. 60
Figure 4.3.  Response spectra of ETA20d01 at 5th, 10th, 15th,
and 20th seconds, (a) acceleration, (b) displacement. 60
Figure 4.4.  Response spectra of ETA20d01-03 at the target time
(the 10th second), (a) acceleration, (b) displacement. 61
Figure 4.5.  Comparison of convergence of least squares procedure
(this work) with unconstrained procedure in
optimization of ETA20d-TR01 acceleration function. 65
Figure 4.6.  Acceleration response spectra of ETA20d-TR01-03 at
the target time (the 10th second). 65
Figure 4.7.  Optimization procedure of 40 seconds ET acceleration
functions.67
Figure 4.8.  ETA40d01 acceleration function. 69
Figure 4.9.  Response spectra of ETA40d01 at 10th, 20th, 30th,
and 40th seconds, (a) acceleration, (b) displacement. 69
Figure 4.10. Acceleration response spectra of ETA40d01-03 at
the target time (the 20th second). 70
Figure 4.11. Comparison of three series of ET acceleration
functions.70
Figure 4.12. Acceleration response time—history of four SDOF
systems for the first series ETAFs. 71
Figure 4.13. Acceleration response time-history of four SDOF
systems for the second series ETAFs. 72
Figure 4.14. Acceleration response time-history of four SDOF
systems for the third series ETAFs. 72
Figure 5.1.    Typical ET accelerogram. 79
Figure 5.2.  ET performance curves for two steel moment frames. 80
Figure 5.3.  An example of target and performance curves [6]. 81
xiv  •   List of Figures

Figure 5.4.  Seismic hazard curve for Sa.83

Figure 5.5.  The ASCE41 response spectra for Tehran, for various
hazard levels. 84
Figure 5.6.  Correlating between seismic hazard and ET response
spectra using T0 and VT0 methods. 86
Figure 5.7.  Return period vs. time in ET analysis for different
fundamental periods (a) T0 method (b) VT0 method. 87
Figure 5.8.  Return period vs. fundamental period for various ET
analysis times (T0 and VT0 methods). 88
Figure 5.9.  Return period vs. structural period and ET analysis time
(Average method). 88
Figure 5.10. ETA20jn and hazard levels’ response spectra. 89
Figure 5.11. Performance curve for MF3S1 by ET analysis time. 91
Figure 5.12. Performance curve for MF3S1 by return period. 91
Figure 5.13. Performance curves for MF3S1 and MF3S2 frames. 92
Figure 6.1.  Schematics of the frames with seven stories and one
bay.103
Figure 6.2.  Total acceleration response spectra of ETA20f series
acceleration functions for ξ=5% at different time. 105
Figure 6.3.  (a) ETA20f02 acceleration function; (b) ETA20f02
velocity function; (c) ETA20f02 displacement function.106
Figure 6.4.  Comparison of smooth spectrum and ASCE-7 spectrum
with the average spectrum of the scaled records for
FM12B3RGW frame. 109
Figure 6.5.   ET maximum interstory drift ratio curves for
FM03B1RG frames with EPP material model and
values of teq of nonlinear response history analysis. 110
Figure 6.6.  Maximum interstory drift ratios of the accelerograms
and their average for FM03B1RG frames with EPP
material model. 111
Figure 6.7.  Interstory drift ratio response history of FM03B3RGW
frame for LPAND270. 112
Figure 6.8.  Interstory drift ratio response history of FM03B3RGW
frame for ETA20f02. 113
Figure 6.9.  ET maximum interstory drift ratio curves for
FM07B1RG frames with EPP material model and
values of teq of nonlinear response history analysis. 114
 List of Figures  •   xv

Figure 6.10. ET maximum interstory drift ratio curves for

FM07B3RG frames with EPP material model and
values of teq of nonlinear response history analysis. 114
Figure 6.11.  Comparison of average maximum interstory drift ratios
of the frames with EPP material model obtained by
nonlinear response history analysis and ET analysis. 116
Figure 6.12. Maximum interstory drift ratios of the accelerograms
and their average for FM03B1RGW frame with EPP
material model with and without considering P-∆
effects.116
Figure 6.13. ET maximum interstory drift ratio curves for
FM03B1RGW frame with EPP material model with
and without considering P-∆ effects. 117
Figure 6.14. ET maximum interstory drift ratio curves for
FM03B1RG frames with EPP and STL material
model.118
Figure 6.15. ET maximum interstory drift ratio curves for
FM03B3RGW frames with different locations for
dampers.120
 List of Tables

Table 2.1. Actual record events on soil condition C 25

Table 3.1. Characteristics of used ET accelerograms 44
Table 3.2. Target time for different series of ET accelerogram 48
Table 3.3. δ value for different series of ET accelerogram and
different strong-motion definitions 50
Table 3.4. Linear correlation factor for different series of ET
accelerogram and different strong-motion definitions 50
Table 3.5. Rankings of different series of ET accelerograms
considering different criteria  51
Table 4.1. Errors of acceleration responses of the first series ETAFs 61
Table 4.2. Errors of acceleration responses of the second series
ETAFs65
Table 4.3. Errors in different steps of optimization of ETA40d01
acceleration function 68
Table 4.4. Errors of acceleration responses of the third series
ETAFs70
Table 4.5. Acceleration responses errors for four SDOF systems
under three series of the ETAFs 73
Table 5.1. Equivalent time corresponding to each performance
level81
Table 5.2. MF3S1 frame basic properties 90
Table 6.1. Specifications of the frames 103
Table 6.2. Description of GM1 set of ground motions used in this
study105
Table 6.3. Scale factors of GM1 set for different frames 107
xviii  •   List of Tables

Table 6.4. Comparison between the results of nonlinear response

history analysis and ET analysis for different frames
with EPP material model 112
Table 6.5. Comparison between the results of nonlinear response
history analysis and ET analysis for different frames
with STL material model 119
Table 6.6. Comparison between the results of nonlinear response
history analysis for FM03B3RGW frames with different
locations for dampers 121
 Foreword

I am delighted to write the Foreword for this book, which captures the
recent advances on fundamentals of the “Endurance Time” methodology
and its application to earthquake resisting design of structures. Protec-
tion of buildings against earthquake is of critical importance for seismic
regions. In the conventional aseismic design, the structure is strengthened
to sustain the expected earthquake ground shaking without failure and/or
major damages. Observations of structural failures during earthquakes,
however, revealed that most destruction occurs due to gradual increase
of damages and weakening of the building due to the earthquake induced
loads. In this book a new innovative method for performance analysis
of buildings to earthquake strong motion is presented. In this approach,
the structure is subjected to a seismic excitation whose intensity increases
with time. The building damages are monitored until the collapse of struc-
ture, and the corresponding Endurance Time is evaluated. The Endurance
Time method provides a new realistic approach for aseismic design of
complex structural systems.
In my view, the authors of this book are the pioneering experts in
this area, and have been the developer of the Endurance Time method-
ology from its inception to formalizing its applications to the earthquake
resisting design of structures. I believe this book provides the readers
with an in-depth understanding of the Endurance Time methodology for
aseismic design of buildings. The challenge that has been presented to the
authors has been to introduce the fundamental aspect of the Endurance
Time concept and bring it all the way to the actual design of buildings
in a single text book. The authors can be proud that they have managed
to achieve such a task seamlessly. The description of the fundamentals
of Endurance Time methodology is broken down to easy-to-understand
exposition beginning from the first principals and building up toward a
thorough description of the method targeted for application to seismic
response analysis of structures and evaluation of the associated damages.
xx  •  Foreword

More importantly, the authors have provided useful best practice solution
methods that are accompanied by a practical approach in aseismic build-
ing designs. Whether readers view this text from either the fundamentals
of Endurance Time methodology or the earthquake engineering side, the
book has been written in a manner where readers will be guided through
a clear understanding in the comprehensive treatment of the structural
responses along the innovative approach of Endurance Time methodol-
ogy for designing buildings against earthquake. This book is certainly an
enthusiastic celebration of three basic elements–theory, modeling, and
practice–in handling a multitude of structural design issues under earth-
quake excitations.
—Goodarz Ahmadi
Clarkson University
Potsdam
March 3, 2017
 Preface

Endurance Time Method (ETM) is a relatively new approach to seismic

assessment of structures. This method has been developed into a versatile
tool in the field of earthquake engineering and its practical applications
are expected to highly expand in near future. ETM is a dynamic analysis
procedure in which intensifying dynamic excitations are used as the load-
ing function. ETM provides many unique benefits of its own among avail-
able seismic analysis procedures. First of all, it is a response history-based
­procedure. Thus, ETM inherits the capabilities of true dynamic procedures
in analysis of systems with complexities. These include issues arising
from nonlinear behavior and irregularity in the system that defies simpli-
fication beyond certain limits. Second, ET method considerably reduces
the computational effort needed in typical response history analyses. This
reduction is roughly in the order of 10 to 100 folds depending on the num-
ber of the ground motions and intensity levels intended in equivalent time
history analysis. Huge computational effort is still a factor in seismic anal-
ysis of large and/or complex dynamic systems that makes conventional
response history analysis of acceptably detailed models impractical. The
third important advantage of the method lies in the simplicity of its con-
cept. Conceptual simplicity makes it a great tool for preliminary response
history analysis of almost any dynamic structural system. By applying an
intensifying dynamic loading, many of the modeling issues can be readily
identified and corrected. The structural analyst is mostly relived of the
complexities involved in ground motion selection and scaling issues.
Some of the most important areas of application of ETM are in the
fields of Seismic Design Optimization, Value-Based Seismic Design, and
Experimental Studies. Seismic design optimization usually requires a
huge amount of computational effort and ETM can pave the way toward
practical use of response history-based analysis improving the reliability
of the design. Design optimization based on lifecycle cost evaluation and
more generally, value-based design optimization also involves too many
xxii  •   Preface

repetitive response history analyses and ETM can be useful in this regard.
Another area of application of ET method which is less explored till now
is in experimental investigations in shaking table labs. While the number
of analyses may be a less critical factor in computation, it is indeed among
the most critical factors in experimental studies and ETM can be very
useful in these studies.
This book is essentially a recompilation of selected research work
published through the years 2004 to 2016 in the Journal of Scientia Iranica
by the editors and their graduate students. This is the first book on ETM
and is aimed to serve as a coherent source of information for students,
engineers, and researchers who want to familiarize themselves with the
concepts and put the concepts into practice. The chapters of the book are
organized in a relatively independent manner. After reading Chapter 1,
the reader can essentially continue with any chapter that covers his or her
topic of interest. However, readers who want to develop a deep under-
standing of the method and its development are encouraged to cover all
chapters consequently.
We are indebted to Professor G. Ahmadi of the University of Clarkson
for reading the manuscript of the book and offering his scientific feedback.
We would like to thank Professor Mohammad Noori of the ­California Poly-
technic State University for ongoing encouragement. Lastly, we would
like to express our gratitude to Mr. Joel Stein, Ms. Charlene ­Kronstedt,
and Ms. Sheri Dean from Momentum Press for their professionalism and
support in the process of publication.
This work could not be materialized without the contribution of our
graduate students at Sharif University of Technology. The authors would
like to specially thank Dr. H T Riahi, Dr. M C Basim, Dr. Valamanesh,
Dr. A Nozari, Dr. A Mirzai, Dr. A Bazmuneh, Mr. Mashayekhi, and Mr.
Foyuzat who’s excellent contributions in the development of ET method
appears in this book. The authors would also like to thank the staff of the
Scientia Iranica Journal for their critical efforts and cooperation during
preparation of this work.
—H.E. Estekanchi and H.A. Vafai
 CHAPTER 1

Introduction to the
Endurance Time Method

1.1 INTRODUCTION

The basic objective of seismic design is to provide the structure with an

appropriate safety margin against failure when subjected to strong earth-
quakes (Estekanchi 1994).1 The common philosophy of most well-known
seismic design codes is to achieve the dual goal of keeping the nonstruc-
tural damage to a minimum in the case of service-level earthquakes and,
also, to prevent structural failure in the case of collapse-level earthquakes
(Moghaddam 2002; Newmark 1971).
Early observations of structural failure during earthquakes revealed
that most structure failures could be attributed to the weakness of
­structures in sustaining the imposed lateral loads and displacements. This
observation formed the basis of well-known earthquake design criteria
based on lateral load, also known as the static seismic design method. In
the static lateral load method, the structure is designed to resist a min-
imum lateral load specified by the code. In this way, minimum lateral
strength and stiffness is provided and lateral displacements are limited
­(Moghaddam 2002; Estekanchi, Vafai, and Shahbodaghkhan 2003).
Extensive research work in the field of earthquake engineering has
revealed many deficiencies and shortcomings in the traditional method of
static seismic analysis (Chandler and Lam 2001). According to the static
method of analysis, structures with higher lateral strength and stiffness
are superior to their less-stiff and less-strong counterparts. However,

1
Chapter Source: Estekanchi, H.E., A. Vafai, and M. Sadeghazar. 2004. “Endur-
ance Time Method for Seismic Analysis and Design of Structures.” Scientia
Iranica 11, no. 4, pp. 361–70.
2  •  THE ENDURANCE TIME METHOD

e­ xperimental and analytical investigation shows that this is not always

the case (Shortreed, Seible, and Benzoni 2002; Bertero and Bertero 2002;
Moehle and Elwood 2003). In fact, there are cases in which reducing
­lateral stiffness results in better seismic performance. The concept of
­seismic base isolation is one example (Moghaddam and Estekanchi 1999;
Estekanchi 1993).
These apparent shortcomings in traditional seismic design, along
with remarkable developments in the field of information technology and
the availability of vastly improved analytical tools, have led research-
ers and engineers to develop more rational and consistent methods for
earthquake engineering (Chopra and Geoel 1999; Chopra and Goel 2003;
Medhekar and Kennedy 2000; Gupta and Kunnath 2000). In this respect,
Performance-Based Seismic Engineering (PBSE) has gained increased
interest among practitioners in earthquake engineering field (Kelly and
­Chambers 2000; Judi, Davidson, and Fenwick 2003). Development of
these new methods and criteria should be mainly contributed to the amaz-
ing improvement in computational tools that have made possible the solu-
tion of sophisticated nonlinear models (Vafai, Estekanchi, and Ghadimi
2000). The use of static push over analysis is becoming standard practice
in structural engineering design offices and the application of nonlinear
time history analysis is also gaining popularity. Thanks to these recent
developments, it has become possible for the structural analyst to incor-
porate the most significant nonlinear material and geometric behavior into
the model and thus, perform a more realistic analysis of structural behav-
ior during earthquakes.
In recent years the development of new analysis tools has provided
the infrastructure for the development of new methods in structural engi-
neering. In this paper, a new method for seismic analysis and evaluation
of structures is introduced. The concept of the Endurance Time (ET) crite-
ria, which is quite intuitive, is first explained and prospective methods to
implement it are discussed. The criterion is then evaluated by applying the
concept to linear single and multi degree of freedom (SDOF and MDOF)
systems.

1.2 THE ENDURANCE TIME CONCEPT

The concept of ET method can be best explained by considering a hypo-

thetical experiment. Assume that three model buildings with unknown
seismic resistance characteristics are to be investigated with reference to
their performance against collapse in high-intensity earthquakes. Consider
 Introduction to the Endurance Time Method  •  3

that these models are put on the shaking table and fixed to it. The experi-
ment starts by subjecting the buildings to random vibration with gradually
increasing intensity. In the beginning (e.g., at t = 5 sec) the amplitude
of shaking is quite low so all three buildings vibrate but remain stable,
as shown in Figure 1.1a. As the amplitude of vibration is increased (say,
at t = 10 sec), a point is reached when one of the buildings collapses.
Assume that this happens to be Model Number 1, as in Figure 1.1b. As
time passes and the vibration amplitude is further increased (say, at t =
20 sec), the second structure fails. Assume this to be Building Number 3,
as in ­Figure 1.1c. Further, consider that Building Number 2 happens to be
the last building to fail in this hypothetical experiment.
Now, based on this experiment, it is concluded that Building
­Number 2, which endured longer, has the best performance, while Build-
ing ­Number 1, which failed soonest, performed the worst. Note that the
judgment is based on the ET, that is, the time during which each building
remained stable, without any reference to the building strength or stiffness
or other dynamic characteristics. If the goal was to evaluate the building’s
resistance to collapse, then, such experiments seem to show a direct and
relevant measure. This is the concept of the ET criterion in evaluating the
performance of buildings subjected to earthquakes.
In ET method, buildings are rated according to the time that they can
endure a standard calibrated intensifying accelerogram. Higher ET is to be
interpreted as a better performance. A minimum performance index can be

(a) t = 5.0 sec 1 2 3

(b) t = 10.0 sec 1 2 3

(c) t = 20.0 sec 1 2 3

Figure 1.1.  Hypothetical shaking table experiment.

4  •  THE ENDURANCE TIME METHOD

set to be used as the design criterion. The idea of the ET method is some-
what similar to the method used by cardiologists to evaluate the condition
of the heart, known as a stress test. In stress tests, the patient is asked to
walk on a treadmill with a variable slope and speed. The test starts with a
low-slope and low-speed condition. During the test, the slope and speed
are increased gradually, while the physical and biological condition of the
patient, such as blood pressure, heart beat rate, and so on, are monitored.
The test is commenced until signs of distress or abnormal conditions are
observed. The heart condition is then judged on the basis of the speed and
slope level that could be tolerated.
In the ET method, nearly the same concept is applied. The idea behind
ET method is roughly to put the structure on a ramp-like accelerogram and
see how far it can go. The structure is subjected to a standard accelerogram
with intensifying dynamic demand. The specified performance indexes are
monitored until they reach a predefined maximum value. The performance of
the structure is judged on the basis of the time when the damage limit index
is exceeded. This concept has been explained in Figure 1.2. Consider the
damage index curve (e.g., maximum drift, plastic energy, etc.) for a typical
structure subjected to an intensifying accelerogram, as shown in Figure 1.2.
If the limit value for the specified damage index is specified to be
1.00, then it can be concluded from Figure 1.2 that this structure has
endured the accelerogram up to about the 12th second. Moreover, consider
that the accelerogram has been calibrated and the design criterion is that
ET should be at least 10 seconds. As can be seen from this figure, the dam-
age value is about 0.82 at t = 10.0 seconds that is, below the value limit,
thus, one can conclude that the structure has met the design criteria. The
analysis should not necessarily be limited to a single damage criterion.
Various different damage indexes and structural criteria can be monitored
simultaneously in order to reach more conclusive results.

2.00
1.80
1.60
Damage index

1.40 Endurance limit

1.20
1.00
0.80
0.60
0.40
0.20
0.00
0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 18.0 20.0
Time (sec)

Figure 1.2.  Damage curve against time for a typical

structure subjected to intensifying accelerogram.
 Introduction to the Endurance Time Method  •  5

In order to achieve quantitative results, one needs to propose a method

for implementing the concept. Direct physical testing of actual structures
as described earlier is, of course, not a practical proposition for general
application considering time and cost. Analytical software that are capable
of modeling and predicting structural behavior up to the complete collapse
point are available, but are still considered to be at research level. In this
paper, some practical implementations of the ET method are introduced
by making some simplifying assumptions and by making use of already
available analytical tools. Modified and improved implementations should
be developed based on ongoing research. However, complete numerical
implementation of the method, which should, ideally, follow the nonlinear
response up to collapse level, should be left to the next generation of struc-
tural engineering software and computer hardware capable of performing
such analysis.
Basic implementation of the concept is based on three fundamental
bases, that is, dynamic input, structural model, and endurance criteria.
These will be discussed next.

1.3  DYNAMIC EXCITATION

The choice of an appropriate dynamic input is fundamental to the success-

ful implementation of the ET concept. The ideal input function is the one
that results in higher consistency and best correlation between the ET anal-
ysis results and the actual or well-known performance characteristics of
the structures subjected to earthquakes. Determination of the most appro-
priate and optimal dynamic input function should, in itself, be the subject
of extensive research work. However, in this research, a function based on
engineering judgment and some elementary calculations will be proposed
so that the concept and application of the ET method can be explained.
An important issue in determining dynamic input is the profile of
amplitude increase. As shown in Figure 1.3, the amplitude increase pro-
file can take various forms. After considering several possible alterna-
tives, the authors came to the conclusion that a linear profile, as shown in
­Figure 1.3a, is more suitable for the purpose of initial investigation. In this
­profile, maximum acceleration is directly proportional to time. Determina-
tion of an optimum profile that results in better correlation and consistent
results is under investigation.
Another important consideration in ET analysis is the specification of
the dynamic input function itself. The most basic form of dynamic input
is a simple harmonic. But this kind of input has obvious disadvantages
6  •  THE ENDURANCE TIME METHOD

(a) (b) (c)

Figure 1.3.  Input profile functions (a) Linear (b) Increasing rate (c) Decreasing
rate.

because of poor frequency content consisting of only one harmonic. The

structures that have natural vibration periods near the input frequency will
experience high dynamic magnification and, thus, will be too much penal-
ized. Considering this issue, a random vibration input with a frequency
content resembling that of a white noise has been used as the starting point
for generating of intensifying accelerograms. As will be explained later,
the frequency content is later modified to better correspond to what is
expected in a real earthquake. The accelerograms generated by this pro-
cedure seem to be good enough for investigation of ET concept and will
be applied in this study. Further research is required and is under way, in
order to propose optimized dynamic input so that the results of ET analy-
sis are best correlated with the known actual performance of structures in
earthquakes and shaking table experiments.

1.3.1  GENERATION OF INTENSIFYING ACCELEROGRAMS

The first generation of accelerograms to be used as dynamic input in the

ET method is produced using random numbers with a Gaussian distri-
bution of zero mean and a variance of unity (Clough and Penzien 1993).
A stationary random accelerogram is generated using δt = 0.01 and n = 211
= 2048 with PGA = 1 is shown in Figure 1.4. Duration of the accelerogram
is equal to δt × n = 20.48 seconds.

1.0
0.8
Acceleration (g)

0.6
0.4
0.2
0.0
−0.2
−0.4
−0.6
−0.8
−1.0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

Time (sec)

Figure 1.4.  A typical random accelerogram with PGA = 1.

 Introduction to the Endurance Time Method  •  7

The frequency content of the random accelerogram, that is statis-

tically similar to a white noise, is then modified, in order to resem-
ble actual earthquake accelerograms. For this purpose, filter functions
given by formulas (1) and (2) are applied to the random accelerograms.
Application of these filter functions is explained by Clough and Penzien
(1993).

ϖ
1 + 2iξ1
ϖ1
H 1(iϖ ) = (1)
 ϖ2  ϖ
1 − 2  + 2iξ1
 ϖ1  ϖ 1

2
ϖ 
 
H 2(iϖ ) = ϖ 2  (2)
 ϖ2  ϖ
1 − 2  + 2iξ 2
 ϖ2  ϖ2

Where ω1 = 2π/0.5, ξ1 = 0.2, ω2 = 2π/0.1, ξ2 = 0.2 are used. It should also

be noted that the frequencies higher than 100 Hz are filtered out. A sample
frequency content resulted from the accelerogram shown in Figure 1.4 is
depicted in Figure 1.5.
The frequency content is then further modified so as to make the
resulting response spectra compatible with typical seismic code Response
Accelerograms. Any given design response spectrum can be used for this
purpose. In this paper, the response spectra of the Iranian National Build-
ing Code (Standard 2800) (Estekanchi et al. 2003) has been used as a
sample. The resulting frequency content of a response spectra compatible
accelerogram, after several cycles of stepwise modification, is shown in
Figure 1.6.

1.4
1.2
Amplitude

1
0.8
0.6
0.4
0.2
0
0 5 10 15 20 25 30
Frequency (Hz)

Figure 1.5.  Filtered frequency content.

8  •  THE ENDURANCE TIME METHOD

4.5
4
3.5
Amplitude 3
2.5
2
1.5
1
0.5
0
0 5 10 15 20 25 30
Frequency (Hz)

Figure 1.6.  Modified frequency content.

3
Response acceleration (g)

2.5 Modified accelerogram

2 Code response spectra

1.5

0.5

0
0.0 1.0 2.0 3.0 4.0 5.0 6.0
Period (sec)

Figure 1.7.  Response of modified accelerogram

­compared to codified value.

Acceleration response, which is compared with the code response

spectrum curve, is depicted in Figure 1.7. The convergence is assumed to
be good enough for the purpose of explanation of the ET concept. Further
calibration of the accelerogram is under investigation.
In the next step, the acceleration values are adjusted for target values
of velocity and acceleration, which, in this case, are set to zero. Finally,
the acceleration values are multiplied by a profile function, which, in this
study, is a linear one starting from zero and reaching a value of 1.00 at
t = 10 seconds. Three accelerograms have been generated using the proce-
dure explained earlier. These accelerograms will be referred to hereafter as
acc1, acc2, and acc3. Accelerogram acc1 that is obtained from the original
accelerogram in Figure 1.4 is shown in Figure 1.8.

1.4 ANALYSIS METHOD

Collapse analysis of the structures is still considered a challenging task

for structural engineers involved in the numerical analysis of structures.
 Introduction to the Endurance Time Method  •  9

2.5
2.0

Acceleration (g)
1.5
1.0
0.5
0.0
−0.5
−1.0
−1.5
−2.0
−2.5
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
Time (sec)

Figure 1.8.  Intensifying accelerogram acc1.

There are a few programs that are capable of conducting such analysis with
reasonable assumptions. However, the experimental evidence to verify
­
the results of such analyses is still quite limited. The minimum modeling
requirement for collapse analysis is nonlinear material behavior, including
material degradation due to cyclic behavior, large strain, and fracture. Anal-
ysis should include the effect of large deformations, buckling, and collusion.
Furthermore, for the purpose of illustrating the ET method, the simplest
method of dynamic analysis is considered in this paper, that is, linear sys-
tems. The Newmark linear method has been applied for the analysis (Chopra
1995). It should be evident that, as far as the concept of the ET method
is considered, the procedure remains almost the same, even for complex
MDOF (multi degree of freedom) models, including nonlinear features.

1.5  ENDURANCE CRITERIA

The numerical definition of structural collapse and failure is not a straight-

forward procedure. An intuitive failure criterion can be defined as the
­displacement of the center of mass of the structure to a lower level from
its initial position. However, analytical implementation of such criteria is
still considered to be very demanding in terms of computational effort and
modeling complexities. The state of the art in this regard is to define the
structural damage in terms of damage indexes. Various damage indexes
have been defined and proposed by researchers in recent years. Seismic
codes that are based on performance-based design usually propose a cer-
tain damage index and set maximum acceptable values for it. In this way,
the structure is assumed to have collapsed when its damage index exceeds
specified code limits. The same simplified method shall be used for the
purpose of this study, as described in the next sections. ET is defined as
the time it takes for the specified damage criteria to reach its value limit as
the structure is subjected to intensifying accelerograms. For example, in
10  •  THE ENDURANCE TIME METHOD

the static method of earthquake resistant design, structures are required to

be designed according to a specified base shear, which is proportional to
certain peak ­lateral ground acceleration. Also, limits to maximum ­lateral
displacements and drift are usually specified. These can be considered
as the simplest forms of d­ amage indexes. In the discussion that follows,
basic parameters such as “max response acceleration,” “max displacement
response,” and “max interstory drift” have been used for the matter of
explanation of the idea. It will be up to the structural designer/analyst
to choose an applicable ­damage index that is more appropriate for the
structure under investigation. It should be evident from the discussion that
many different damage indexes can be applied in the procedure. ­Practical
implications of applying different d­ amage indexes for ET analysis are
under investigation.

1.6 APPLICATION TO LINEAR SDOF SYSTEMS

For linear SDOF systems, the most significant parameter that describes
the dynamic characteristics of the structure is its natural frequency of
vibration. Consider that three different linear structures with natural
periods of vibration equal to 0.1, 0.5, and 1.0 seconds are to be studied.
These structures will be hereafter referred to as ST01, ST05, and ST10.
ST01 roughly pertains to a stiff single-story masonry building, while
ST05 and ST10 have a period typical of three- and nine-story steel frame
buildings, respectively. Damping ratio will be assumed to be 0.05 of the
critical value, as commonly assumed in dynamic analysis. The accelera-
tion response for ST05, subjected to the three generated accelerograms, is
shown in ­Figure 1.9. As expected, peak acceleration increases with time
as the input acceleration is intensified.
In Figure 1.10, maximum acceleration as a function of time has been
plotted along with average value and a linear fitting curve. As can be
seen in this figure, in spite of the fact that the generated accelerograms

2.0
1.5 ST05_acc1 ST05_acc2
Acceleration (g)

ST05_acc3
1.0
0.5
0.0
−0.5
−1.0
−1.5
−2.0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
Time (sec)

Figure 1.9.  Acceleration time history for ST05.

 Introduction to the Endurance Time Method  •  11

2.5
ST05_acc1 ST05_acc2
2.0 ST05_acc3 Avr
Linear (Avr)
Acceleration (g)

1.5

1.0

0.5

0.0
0 2.5 5 7.5 10 12.5 15 17.5 20
Time (sec)

Figure 1.10.  Maximum acceleration and average for ST05.

are compatible with the same design response spectra and have a similar
­frequency content, the dynamic response can be significantly different at
specific time intervals. For example, consider that one wants to specify the
time at which the structure has experienced a maximum acceleration of
1g. From Figure 1.10, it can be seen that for acc1, this occurs at t equals to
about 11.5 seconds, while for acc2, this occurs at t = 16 seconds.
This could be expected considering the well-known characteristics of
time-history analyses. An averaging method can be used to achieve better
approximation. In this example, using the linear fitting curve, the time at
a = 1g can be seen to be about 14 seconds.
The displacement response of ST05 subjected to different accelero-
grams has been depicted in Figure 1.11. As can be seen in this figure, the
displacement response is also an increasing function of time, as expected.
The pulsating characteristic of the displacement response should be paid
due attention. These pulsations can result in the maximum response to
remain constant during a relatively long period of time, making it difficult
to interpret the result of analysis regarding the time corresponding to a cer-
tain response level. This problem can be avoided by using several acceler-
ograms, along with an appropriate averaging method. Even though, in this
paper, the number of accelerograms to be averaged has been set to three as
a practical minimum, it should be clear that the desired accuracy and con-
vergence can be achieved by considering a larger number of accelerograms.
Maximum and average accelerations are shown in Figure 1.12. Max-
imum displacement is considered to be a simple and effective damage
criterion. In multistory buildings, maximum displacement is usually
proportional to maximum drift, which is another significant response
­characteristic that can be related to building damage. Consider that the
maximum tolerable displacement for this building has been set to 4 cm.
12  •  THE ENDURANCE TIME METHOD

0.100
0.080
Displacement (m)

ST05_acc1 ST05_acc2 ST05_acc3

0.060
0.040
0.020
0.000
−0.020
−0.040
−0.060
−0.080
−0.100
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
Time (sec)

Figure 1.11.  Displacement time history for ST05.

0.1200
acc1 acc1
0.1000
acc3 Avr
Displacement (m)

0.0800 Linear (Avr)

0.0600

0.0400

0.0200

0.0000
0 2.5 5 7.5 10 12.5 15 17.5 20
Time (sec)

Figure 1.12.  Maximum displacement and average for ST05.

Based on Figure 1.12, one can conclude that the building can endure the
prescribed accelerogram up to t = 11 seconds.
A summary of the results for the maximum acceleration of ST01,
ST05, and ST10 are depicted in Figure 1.13. In general, the magnitude of
acceleration experienced by all three structures is nearly the same, with
that of ST05 being higher for almost the entire time range and that of ST10
being lower in the time interval from 7.5 to 12.5 seconds. These could
be expected considering the shape of the codified response spectrum and
the frequency content being amplified near to T0 = 0.5, according to the
assumed soil conditions.
Displacement responses have been depicted in Figure 1.14. As can be
seen in this figure, there is a marked distinction between the displacement
responses of sample structures. Significant insight on the structural per-
formance of these structures can be gained through studying the curves of
Figure 1.14. For example, consider that the maximum tolerable displace-
ment for all these structures was to be limited to 4cm. It can be concluded
from Figure 1.14 that the ET for ST10 is about 6 seconds, for ST05 about
11 seconds and for ST01 more that 20 seconds. Now, by specifying the
required endurance time of 10 seconds, it can be concluded that ST01 and
 Introduction to the Endurance Time Method  •  13

1.8
1.6 ST01 T=0.1sec
Acceleration (g) 1.4 ST05 T=0.5sec
1.2
ST10 T=1.0sec
1.0
0.8
0.6
0.4
0.2
0.0
0 2.5 5 7.5 10 12.5 15 17.5 20
Time (sec)

Figure 1.13.  Maximum average accelerations.

0.20
0.18 ST01 T=0.1 sec
0.16
ST05 T=0.5 sec
0.14
ST10 T=1.0 sec
Displacement (m)

0.12
0.10

0.08
0.06

0.04
0.02

0.00
0 2.5 5 7.5 10 12.5 15 17.5 20
Time (sec)

Figure 1.14.  Maximum average displacements.

ST05 are acceptable, that is, can endure the specified dynamic demand,
while ST10 is rejected.
As another example, consider that ST05 and ST10 are three- and nine-
story buildings with heights of 10 and 30 m, respectively, and the failure
criteria is set to be a maximum drift of 0.005. If one roughly, assumes
the vibration mode to be linear, maximum displacements will be 0.005 ×
10 = 0.05 m and 0.005 × 30 = 0.15 m, respectively. Now, by referring to
­Figure 1.14, it can be seen that ST05 reaches the value limit at about t = 15
seconds, while ST10 reaches its value limit at t = 15.5 seconds. Thus, the
ET for both structures is nearly the same and, if one sets the required ET to
10 seconds, then both structures will be considered acceptable.
14  •  THE ENDURANCE TIME METHOD

In the earlier discussion, the linear analysis of a single degree of free-

dom is considered for the purpose of describing the basic idea behind the
ET method. It should be clear from this discussion that the analysis can
be readily extended to nonlinear and multidegree of freedom systems. The
essence of the ET method lies in the definition of standard intensifying
accelerograms and appropriate damage criteria. It is interesting to note
that the concept of ET can also be readily applied in experimental dynamic
investigation of structures. In this case, the most realistic performance cri-
teria, that is, actual failure of the structure, can be considered the endur-
ance limit as measured against time. Another advantage of the ET method,
as compared to other dynamic analysis methods, is in its applicability to
experimental shaking table studies. In these cases, the cost and required
resources usually eliminate the possibility of conducting the experiment
several times with different levels of excitation. Using the concept of ET,
a single-run experiment can be used to obtain the desired results.

1.7 APPLICATION TO MDOF SYSTEMS

In this section, the application of ET method in dynamic analysis of

MDOF systems is explained by considering a three-story steel moment
frame example. A three-story steel frame, named as “f3” in Figure 1.15,
has been designed in accordance to conventional loading and analysis pro-
cedures appropriate for ordinary moment frames in high seismic zone.
For the matter of comparison, another frame with similar properties, but
utilizing steel sections with lower stiffness and strength, has also been
designed. This weak frame is named as f3w in Figure 1.15. We intend to
analyze these frames using ET concept and study the results. For multi-
story frames, interstory drift has proved to be a very significant and conve-
nient measure of building damage and performance criteria. We shall use
this criterion for the purpose of explanation of the method. It should be
clear that other criteria based on strain energy, nonlinear hysteresis behav-
ior, and so on can also be applied with no further complications. The only
limitation in the application of ET method will lie on the computational
capability of the analysis program used.
Frame f3 has been subjected to accelerogram acc1 and the time his-
tory of story drifts has been depicted in Figure 1.16. It can be seen from
this figure that the story drifts are nearly the same for most of the time his-
tory and also, they reach peak values at almost the same times. This is due
to the fact that in MDOF frame structures, Mode Number 1 is usually the
predominating deflection mode. Higher vibration modes have had much
less pronounced effect in peak values for interstory drifts.
 Introduction to the Endurance Time Method  •  15

6.00m 6.00m
IPE 500 IPE 400

HE180B

HE180B

HE160B

HE160B
IPE 500 IPE 400

3 @ 3.00m

3 @ 3.00m
HE200B

HE200B

HE180B

HE180B
IPE 500 IPE 400
HE220B

HE220B

HE200B

HE200B
Frame f3 Frame f3w

Figure 1.15.  Three-story steel moment frames.

0.020
Interstory drift (m)

0.015 Story1 drift

Story2 drift
0.010 Story3 drift
0.005
0.000
-0.005
-0.010
-0.015
-0.020
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
Time (sec)

Figure 1.16.  Story drifts time history for frame f3 subjected to acc1.

Frame f3 has been subjected to accelerograms acc2 and acc3 as well

and the maximum absolute values of drifts along with their average value
are depicted in Figure 1.17. As expected, the maximum drift increases
with time and by considering a limiting value for drift, the ET can be eval-
uated. Due to the randomness of input and resulting response, the peak
value can remain constant for a prolonged period of time. For example,
maximum drift has remained constant from t = 10 to t = 16 seconds for
f3 subjected to acc1 as can be seen in Figures 1.16 and 1.17. Considering
the irregular fluctuation in the drift values for the frame subjected to each
accelerogram, the average value should be used in order to avoid the inac-
curacy resulted from random nature of response.
A curve fitting method can also be used in order to better define the
ET value. In Figure 1.17, a linear trend line has been added for this pur-
pose. In linear systems subjected to linearly intensifying accelerograms,
we know that the response is roughly linear. This justifies the use of linear
trend line as the fitting curve. However, it should be clear that when apply-
ing ET method to nonlinear problems, a more appropriate fitting method,
such as spline or polynomial curves, should be used.
Now, considering a limiting value of 0.005 for interstory drift, as a
typically recommended maximum value, the limiting drift for frame f3
16  •  THE ENDURANCE TIME METHOD

0.0300
f3 MaxD1
f3 MaxD2
0.0250
f3 MaxD3
f3 Max average drift
0.0200
Interstory drift (m)

Linear (f3 Max average drift)

0.0150

0.0100

0.0050

0.0000
0 2.5 5 7.5 10 12.5 15 17.5 20
Time (sec)

Figure 1.17.  Maximum story drifts for frame f3 subjected to acc1-3

accelerograms.

with a story height of 3.0 m will be 0.005 × 3.0 = 0.015 m. Thus, we can
conclude from Figure 1.17 that frame f3 has endured our acc1-3 set of
accelerograms up to time t = 13.5 seconds. If we have standardized these
accelerograms for the specified site and had specified a minimum ET of
say 10 seconds, we would then conclude that frame f3 passes our seismic
design criteria. Generating a standard set of intensifying accelerograms to
be used as design criteria in ET method is under investigation and is the
subject of another study.
The result of the same analysis as applied to frame f3w, that is, the
weak version of f3 is summarized in Figure 1.18. The general trend for f3w
is the same as f3. It should be noted that the resulting drift values are gener-
ally higher for f3w. This could be expected considering the lower stiffness
of frame f3w. Considering the limiting value of 0.005 for interstory drifts
as earlier, it can be concluded that the ET for f3w is about 10.5 seconds.
Analysis results for frames f3 and f3w are compared in Figure 1.19.
The significant point in this figure is the ET for f3w being clearly lower
than that of f3. On the other hand, the weakness of f3w as compared to f3
shows up in its response to the intensifying accelerograms. Again, if we
consider that the accelerograms have been somehow standardized and a
minimum ET of 10.0 seconds has been specified to the building site, it can
be concluded that f3w passes our seismic design criteria. If, however, the
minimum ET was specified to be 12.0 seconds, f3 would pass the design
criteria while f3w would fail.
 Introduction to the Endurance Time Method  •  17

0.0300
f3w MaxD1
f3w MaxD2
0.0250
Interstory drift (m) f3w MaxD3
f3w Max average drift
0.0200 Linear (f3w Max average drift)

0.0150

0.0100

0.0050

0.0000
0 2.5 5 7.5 10 12.5 15 17.5 20
Time (sec)
Figure 1.18.  Maximum story drifts for frame f3w subjected to acc1-3
­accelerograms.

0.0300
f3 Max average drift
0.0250 f3w Max average drift
Linear (f3w Max average drift)
Interstory drift (m)

0.0200 Linear (f3 Max average drift)

0.0150

0.0100

0.0050

0.0000
0 2.5 5 7.5 10 12.5 15 17.5 20
Time (sec)

Figure 1.19.  Maximum average story drifts for frames f3 and f3w.

In case of the simple example considered earlier, one could already

predict the analysis result by considering the fact that f3w is less stiff that
f3. However, in case of more complex models, particularly the models
involving several sources of nonlinearity, it is not possible to predict the
ET and the advantage of one design over the other. For example, consider
that the effectiveness of various energy absorbing devices in reducing the
maximum drift of a structure is to be studied. These type of analyzes are
usually encountered when considering seismic retrofit studies. It should
be clear from the earlier discussion that by applying the ET method, a
conclusive analysis result can be readily achieved.
18  •  THE ENDURANCE TIME METHOD

1.8  SUMMARY AND CONCLUSIONS

In this paper, a new approach to performance-based earthquake analysis

and design has been introduced. In this method, the structure is subjected
to standard accelerograms that impose increasing dynamic demands on
structures with time. Damage indexes of interest are monitored through
time as the dynamic load intensity is increased. The time when collapse
level damage or other performance limits are reached is called the endur-
ance time of the structure. ET is defined as the length of the time interval
from start to failure point (or limit point), that is, the time at which the
damage index reaches its maximum tolerable value. Longer ET implies
better or improved performance. Standard accelerograms are calibrated
and a minimum ET is specified to be used as the design criteria. In this
way, the structure is considered acceptable or unacceptable by its ET being
higher or lower than the benchmark value.
The ET method provides a unified and objective approach to seis-
mic analysis and design of structures that can be applied in numerical
and experimental investigations regardless of model complexity. Also, the
method is readily applicable to complex nonlinear systems. The definition
and calibration of standard intensifying accelerograms is a key issue in
this method. A simple approach to generating required accelerograms has
been discussed. Three such accelerograms were applied in the simple case
of linear elastic single degree of freedom systems. Methods for generating
improved accelerograms to achieve better convergence are under investi-
gation. Also, practical implications in cases of nonlinear MDOF systems
are being studied.

REFERENCES

Bertero, R.D., and V.V. Bertero. 2002. “Performance-Based Seismic Engineering:

The Need for a Reliable Conceptual Comprehensive Approach.” Earthquake
Engineering & Structural Dynamics 31, no. 3, pp. 627–52.
Chandler, A.M., and N.T.K. Lam. 2001. “Performance Based Design in Earth-
quake Engineering: A Multidisciplinary Review.” Engineering Structures 23,
no. 12, pp. 1525–43.
Chopra, A.K. 1995. “Dynamics of Structures: Theory and Applications to Earth-
quake Engineering.” Prentice Hall International Series in Civil Engineering
and Engineering Mechanics. Englewood Cliffs, NJ: Prentice Hall.
Chopra, A.K., and R.K. Geoel. 1999. “Capacity-Demand-Diagram Methods for
Estimating Seismic Deformation of Inelastic Structures: SDF Systems.”
Report No. PEER-1999/02, Pacific Earthquake Engineering Research Center,
College of Engineering, University of California, Berkeley.
 Introduction to the Endurance Time Method  •  19

Chopra, A.K., and R.K. Goel. 2003. “A Modal Pushover Analysis Procedure to
Estimate Seismic Demands for Buildings: Summary and Evaluation.” Fifth
National Conference on Earthquake Engineering, 26–30, Istanbul, Turkey.
Clough, R.W., and J. Penzien. 1993. Dynamics of Structures. McGraw-Hill Inc.
Estekanchi, H.E. 1993. “A Review of History, Performance and Application of
Seismic Isolators (in Farsi).” Omran Magazine of Sharif, no. 10, pp. 42–44.
Estekanchi, H.E. 1994. “Basics of Earthquake Engineering (in Farsi).” Omran
Magazine of Sharif 15, pp. 26–29.
Estekanchi, H.E., A. Vafai, and P. Shahbodaghkhan. 2003. “Comparative Evalua-
tion of Braced Steel Frames Designed According to the First and Second Edi-
tion of the Iranian Code for Seismic Resistant Design of Buildings (in Farsi).”
Building Engineering and Housing Science 1, no. 1, pp. 11–16.
Gupta, B., and S.K. Kunnath. 2000. “Adaptive Spectra-Based Pushover Proce-
dure for Seismic Evaluation of Structures.” Earthquake Spectra 16, no. 2, pp.
367–91.
Judi, H.J., B.J. Davidson, and R.C. Fenwick. 2003. “Displacement Focused Seis-
mic Design Methods: A Comparative Study.” Proceedings of the 2003 Pacific
Conference on Earthquake Engineering, New Zealand Society for Earthquake
Engineering, 8 pages, Paper No. 108.
Kelly, T.E., and J.D. Chambers. 2000. “Analysis Procedures for Performance
Based Design.” 12th World Conference on Earthquake Engineering, New
Zealand Society for Earthquake Engineering, Upper Hutt, New Zealand,
Paper No. 2400.
Medhekar, M.S., and D.J.L. Kennedy. 2000. “Displacement-Based Seismic Design
of Buildings: Application.” Engineering Structures 22, no. 3, pp. 210–21.
Moehle, J.P., and K.J. Elwood. 2003. “Collapse Performance Prediction for RC
frame Structures.” Proceedings of the 2003 Pacific Conference on ­Earthquake
Engineering [electronic resource], New Zealand Society for Earthquake
­Engineering, [Wellington], 8 pages, Paper No. 154.
Moghaddam, H. 2002. Earthquake Engineering: Theory and Application. Iran:
Farahang Publications.
Moghaddam, H.A., and H.E. Estekanchi. 1999. “A Study of Off-Center Bracing
Systems.” Journal of Constructional Steel Research 51, no. 2, pp. 177–96.
Newmark, N.M. 1971. Fundamentals of Earthquake Engineering. Englewood
Cliffs: Prentice-Hall.
Shortreed, J.S., F. Seible, and G. Benzoni. 2002. “Simulation Issues with a Real-
time, Full-scale Seismic Testing System.” Special Issue: Controversial Issues
in Earthquake Engineering, Proceedings of First ROSE Seminar, Journal of
Earthquake Engineering 6, Special Issue 1, pp. 185–201.
Vafai, A., H.E. Estekanchi, and G. Ghadimi. 2000. “The Role of Information Tech-
nology in Construction Industry (in Farsi).” Iranian Journal of Engineering
Education 1, no. 4, pp. 23–37.
 Index

A B
A95 parameter, 32 base errors, 61, 62, 64, 67, 68
acceleration, 10, 12, 60 bilinear material model, 99
absolute, 30, 58 bracketed duration, 44, 45, 47, 49
amplitude, increase in, 78, 79
errors of responses, 61 C
filtered, 22 calculated errors, 60
functions, 11, 23, 26, 27, 56, 70, central period (CP), 22
117 characteristic Intensity, 32
ground, 45 collapse analysis of structures, 8
maximum average, 13 collapse prevention (CP), 77, 80,
random, 57 81, 92, 100
response, 8, 10, 58, 71–73 compressed matrix, 64, 66
reverse, 46 CP performance level, 77, 80, 92
spectrum, 43 cumulative absolute velocity
target, 57 (CAV), 35, 36
time history, 10 curve fitting method, 15
values, 8
Acceleration Spectrum Intensity D
(ASI), 32 damage, 12, 21
AISC-ASD building code, 89, 102 accumulated, 41
analysis method, 8, 9 criteria, 14
annual probability of exceedance, hysteretic energy dissipation, 49
78, 82, 83, 94 observation of buildings after
Applied Technology Council earthquakes, 22
(ATC-40) guidelines, 98 to nonstructural components, 110
Arias Intensity, 32 value, 4
ASCE Standard for Seismic damage index(es), 10
Rehabilitation of Existing curve, 4
Buildings, 82 forms of, 10
ASCE41, 82, 89, 93 of interest, 18
design spectrum, 78, 79, 89 Kunnath, 48
response spectrum, 83, 84, 92 levels, 18
126  •  Index

Park-Ang, 48 effective stiffness, 117

of structure, 49 elastic-perfectly-plastic (EPP)
damage limit index, 4 material model, 99
design spectrum, 22, 23, 26, 27, elastic-perfectly-plastic material,
29, 37, 56, 57, 78, 79, 89, 93, 99, 100, 102, 103, 110–118,
101, 106 120–122
displacement spectrum, 43 endurance criteria, 5, 9, 10
displacement time history, 12 endurance time (ET) method, 2–5,
ductility, 48, 49 12, 18, 42, 56, 77–79, 81, 98
dynamic excitation, 21, 43, 73, 74, acceleration functions, 21, 22,
98, 100 56, 102
generation of intensifying generation of, 57–62
accelerograms, 6–8 improved, 64–66
ideal input function, 5, 6 optimization of long duration,
dynamic time history, 55 66–70
real ground motions and,
E comparison between, 47
earthquake(s), 5, 32, 34, 77, 80, accelerograms
87, 101, 121, 122 characteristics of, 44
buildings observation after ETA20jn series, 89
strong, 22 production, methodology of,
collapse-level, 1 78, 79
complexity of ground motions, basic ground motion properties,
22 26, 27
destructive, structural failure in, definition of, 9, 21, 100
55 dynamic analysis procedure, 42
distance from site, 35 excitation functions, 43, 44
engineering, 1, 2, 42, 55, 73 fourier amplitude of, 27, 28
excitation, 58 frequency content, 27, 28
hazard, 100 generation of functions, 22–24
high-intensity, 2 PBSE, applicability of, 101
individual, 101 performance-based design,
induced deformation, 98 application in, 79–81
motions, 41, 42 power spectral density, 28–30
performance-based, 18 response spectrum with real earth
produce complicated loadings, 27 earthquake and, comparison
real, 6, 24–26, 99, 117 with, 24–26
records, 24 seismic rehabilitation of
resistance design, 10 buildings, application in,
of return period, 90 119–121
service-level, 1 time and return period,
effective design acceleration correlation between, 82–89
(EDA), 36, 37 energy parameter, for ground
effective peak acceleration, 47 motion, 30
 Index  •   127

energy spectrum, 30 period parameter, 33, 34

engineering demand parameters ground motion selection, 42, 43
(EDPs), 97
EPP and STL material model, 117, H
118 hazard level (or return period), 43,
EPP material model, 99, 110–118, 82–84, 90, 98, 100, 101
120–122 Hazard return period, 82
ET acceleration, 21
functions, 22, 23, 27–30, 33–38, I
67, 98, 99, 101, 102, 104–106, immediate occupancy (IO), 100
108–111, 113, 115, 122 Incremental Dynamic Analysis
properties of, 22 (IDA), 98, 108
ET curve, 110, 111, 113, 118 initial stiffness, 99, 102
ETA20d, 57, 60, 62, 63, 71 intensifying accelerograms, 3, 4,
ETA20d02, 61 6–9, 14–16, 18, 21, 36, 37, 42,
ETA20d-3, 56, 61 43, 78
ETA20d-TR, 68 intensity measures (IMs), 97
ETA20d-TR01, 65 intensity parameter, for ground
ETA20d-TR02, 65 motion, 30–33
ETA20d-TR03, 65 interstory drift, 14–17, 41, 79–81,
ETA20d-TR01-03, 64, 65, 71 90–92, 99, 102, 109–122
ETA20f, 99, 104–106, 109, 111, IO performance level, 80, 81, 100
113 Iranian National Building Code
(INBC), 7, 24–27, 56, 57, 102,
F 104
Federal Emergency Management
Agency (FEMA-356), 98 L
FEMA-356, 25, 81, 98 Life Cycle Cost An41, 79, 99,
FEMA 440, 104 109–122
FEMA 695, 43 life safety (LS), 100
first generation, 22
Fourier amplitude M
of ET acceleration, 27, 28 maximum average accelerations,
spectrum, 33 13
framework for PBEE, 77 maximum average displacements,
Frequency content, 6, 7, 11, 12, 13
27, 28, 30, 34, 108 maximum considered earthquake
(MCR) maps, 83
G max reponse acceleration, 10
Generation of intensifying mean period, 34
accelerograms, 6–8 modified frequency content, 8
ground motion parameters or multi degree of freedom (MDOF),
properties, 22, 25–29 2, 9, 18, 97
energy parameter, 30 applications to, 14–17
intensity parameter, 30–33 multiple hazard levels, 80
128  •  Index

N Pacific Earthquake Engineering

Newmark linear method, 9 Research Center (PEER), 89
nonlinear distributed plasticity, peak ground acceleration (PGA),
104 6, 22, 78, 80–82, 97, 105, 108
nonlinear dynamic analysis, 42 peak ground velocity (PGV), 108
nonlinear freedom system, 14 peak velocity acceleration ratio, 35
nonlinear hysteresis behavior, 14 PEER framework, 89
nonlinear inelastic phases, 98 performance-based earthquake
nonlinear least square, 62–64 engineering, 77
nonlinear material behavior, 2, 9 performance-based seismic design,
nonlinear range of behavior, 44 121
nonlinear region, 51 performance-based seismic
nonlinear response history engineering (PBSE), 2, 97, 100
analysis, 99, 100, 102, 106–119 performance criteria, 160
nonlinear response of SDOF performance curves, 14, 90, 91
system, 101 performance levels
nonlinear response spectra, 44 OP, IO, LS, CP, 77, 80, 81, 92,
nonlinear time history analysis, 2, 100
100, 121, 122 performance objectives, 77, 78, 80,
nonstructural damage, 1 92, 97, 100, 121
numerical analysis of structures, 8 performance-based design method,
9, 55, 78–81, 93
O period parameter, for ground
Objective function, 59, 60, 202 motion, 33, 34
OpenSees platform, 89, 90, 104 plastic rotation, 100
Opensees software, 90, 104 power spectral density (PSD) of
Optimization objective, 62 ET acceleration functions,
Optimization process, 22, 23, 28, 28–30
38, 44, 56, 59, 60 predominant period, 33, 34
aims and objectives of, 98, 101 time target, evaluation of, 47–51
of ETA20d-TR01 acceleration pseudo-acceleration, 24, 104
function, 65
of ETA40d01 acceleration Q
function, 68 quasi-Newton algorithm, 59, 62
least square, 64
multilevel, 66 R
objective function of, 62, 63 ramp-like accelerogram, 4
of 40 seconds ET acceleration resilience, 100
functions, 67 response history analyses, xxi,
unconstrained, 64 xxii, 74, 97, 99–102, 106–122
variable, 59 response spectra, 7, 11, 21, 23–25,
44, 56, 59–61, 64, 65, 68–70,
P 82–84, 86, 88, 89, 92, 93, 98,
P-D effects, 99, 100, 104, 111, 101, 104, 105
115–118, 121, 122 return period, 78, 80–93
 Index  •   129

root-mean-square of accelerogram, structural damage, 9, 32, 35, 36,

46 42, 55, 109
sustained maximum acceleration
S (SMA), 36, 37
scaling procedure, 109 sustained maximum velocity
second generation, 22, 23, 36, 44 (SMV), 36, 37
seismic acceleration, 22
seismic base isolation, concept T
of, 2 target curve, 78, 80–82, 89, 91, 92
seismic design, 16 target displacement spectrum, 43
codes, philosophy of, 1 target spectrum, 43
ET method tool, 73 target time, 23, 38, 48, 51, 64, 67,
objectives of, 1, 55 68, 70, 109
seismic hazard level, 88, 89 definition of, 42, 43, 49
seismic hazard return period, 77 determination of, 47
sensitivity analysis, 97 of ET accelerogram, 48
significant duration, 44–47 evaluation, 41
single degree freedom (SDOF), of functions, 57
2, 97 history response, 71
application to, 10–14 time history based procedure, 21
ETAFs comparison in analysis acceleration, 45, 46
of, 71–73 analysis, 2, 11, 14, 59, 63, 68,
specific energy density, 30, 31, 34 77, 78
spectral absolute seismic input based procedures, 21
energy, 22 calculation of, 58
spectral acceleration, 22, 24, 52, non-linear, 100
84, 108 responses, 71
standard deviation of the EDP, Trust Region algorithm, 63, 74
108 typical ET curve, 110
standard No. 2800 of INBC, 57 typical ET acceleration function,
story drifts time history, 15 23, 59
ST05 typical ET accelerogram, 79
acceleration time history for, 10
displacement time history for, 12 U
maximum acceleration and Uniform duration, 44, 45, 47
average for, 11
maximum displacement and V
average for, 12 Value based seismic design, xxi
STL material model, 104, Velocity Spectrum Intensity (VSI),
117–119, 122 32
strong-motion duration, 22, 41, 42, viscoelastic damper, 120
44–47, 51 viscous damper, 90, 91