Department of Civil and Environmental Engineering

Stanford University

EFFECTS OF NEAR-FAULT GROUND MOTIONS

ON FRAME STRUCTURES

by
Babak Alavi
and
Helmut Krawinkler

Report No. 138

February 2001
The John A. Blume Earthquake Engineering Center was established to promote
research and education in earthquake engineering. Through its activities our
understanding of earthquakes and their effects on mankind’s facilities and structures
is improving. The Center conducts research, provides instruction, publishes reports
and articles, conducts seminar and conferences, and provides financial support for
students. The Center is named for Dr. John A. Blume, a well-known consulting
engineer and Stanford alumnus.

Address:

The John A. Blume Earthquake Engineering Center

Department of Civil and Environmental Engineering
Stanford University
Stanford CA 94305-4020

(650) 723-4150
(650) 725-9755 (fax)
earthquake @ce. [Link]
[Link]

©2001 The John A. Blume Earthquake Engineering Center

EFFECTS OF NEAR-FAULT GROUND MOTIONS
ON FRAME STRUCTURES

by
Babak Alavi
and
Helmut Krawinkler

The John A. Blume Earthquake Engineering Center

Department of Civil and Environmental Engineering
Stanford University
Stanford, CA 94305-4020

A report on research sponsored by

CUREe-Kajima Phase III

National Science Foundation Grant CMS-9812478
California SMIP Contract No. 1097-601

Report No. 138

February 2001
 ABSTRACT

Near-fault ground motions have caused much damage in the vicinity of seismic sources
during recent earthquakes. These ground motions come in large varieties and impose
high demands on structures compared to “ordinary” ground motions. Recordings suggest
that near-fault ground motions are characterized by a large high-energy pulse. This
impulsive motion, which is particular to the “forward” direction, is mostly oriented in a
direction perpendicular to the fault, causing the fault-normal component of the motion to
be more severe than the fault-parallel component. This study is intended to evaluate and
quantify salient response attributes of near-fault ground motions and to investigate design
guidelines that explicitly account for near-fault effects.

In this study the elastic and inelastic response of SDOF (single degree of freedom)
systems and MDOF (multi degree of freedom) frame structures to near-fault and pulse-
type ground motions is investigated. Generic frame models are utilized to represent
MDOF structures. The stiffness and strength of the models are tuned to a story shear
distribution based on the SRSS (square root of sum of squares) combination of modal
responses. The extent to which these models represent code-compliant structures is
evaluated by comparing the dynamic response of the generic frames with that of steel
structure models. Near-fault ground motions are represented by equivalent pulses, which
have a comparable effect on structural response but whose characteristics are defined by
a small number of parameters. The inelastic dynamic response to both near-fault records
and basic pulses demonstrates that structures with a fundamental period greater than the
pulse period respond differently than shorter period structures. For the former, early
yielding occurs in higher stories but the high ductility demands migrate to the bottom
stories as the ground motion becomes stronger. For the latter, the maximum demand
always occurs in the bottom stories.

Abstract
i
Models are proposed that relate the parameters of the equivalent pulse to magnitude and
distance by means of regression analysis. A preliminary design methodology is
developed based on the equivalent pulse concept, including a procedure that provides an
estimate of the base shear strength required to limit story ductility ratios to specific target
values. Alternative story shear strength distributions are introduced that can improve the
distribution of ductility demands over the height for long period frames. Strengthening of
frames with walls that are either fixed or hinged at the base is studied, and it is shown
that strengthening with hinged walls can provide effective protection against near-fault
effects at all performance levels.

Abstract
ii
 ACKNOWLEDGEMENTS

This report is a reproduction of the senior author’s Ph.D. dissertation. The dissertation is
the result of an extensive study on the effects of near-fault ground motions on frame
structures. This study has been supported by a grant from the CUREe/Kajima Research
Program, by the National Science Foundation through Grant CMS-9812478 of the US-
Japan Cooperative Research Program in Urban Hazard Mitigation, and by the California
Department of Conservation as a SMIP 1997 Data Interpretation Project (Department of
Conservation Contract No. 1097-601). This support is gratefully acknowledged. The
constructive collaboration of Dr. Paul Somerville on the ground motion aspects of this
research is much appreciated.

The authors express their appreciation to Professors Allin Cornell, Eduardo Miranda, and
Greg Deierlein, who provided constructive feedback on the manuscript. Several
researchers of the John A. Blume Earthquake Engineering Center have contributed to this
work. Sincere thanks are due to all of them, especially to Nicolas Luco for his
collaboration and constructive feedback, and to Dr. Hjortur Thrainsson for sharing his
extensive knowledge of earthquake engineering with the senior author.

Acknowledgements
iii
 TABLE OF CONTENTS

Chapter 1. Introduction
1.1. Statement of Problem ....................................................................................... 1
1.2. Historical Perspective ....................................................................................... 3
1.3. Objectives and Scope ....................................................................................... 4

Chapter 2. Near-Fault Ground Motions Used in this Study

2.1. Ground Motion Records ................................................................................... 8
2.1.1. Directivity Effects ............................................................................. 9
2.1.2. Ground Motion Components ........................................................... 10
2.2. Elastic Spectra of Near-Fault Ground Motions .............................................. 10
2.2.1. Comparison with Ordinary Ground Motions .................................. 12

Chapter 3. SDOF and MDOF Systems Used in this Study

3.1. SDOF Systems ............................................................................................... 23
3.2. MDOF Systems .............................................................................................. 23
3.2.1. Properties of Generic Structure ....................................................... 23
3.2.2. Design Load Pattern ........................................................................ 25
3.2.3. P-Delta Effects ................................................................................ 26

Chapter 4. Response of Structures to Near-Fault Ground Motions

4.1. Elastic Response of MDOF Structures ........................................................... 31
4.1.1. Elastic Base Shear Demands ........................................................... 31
4.1.2. Elastic Shear Force Distribution Over Height of Structure ............. 32
4.1.3. Elastic Roof Displacement Demands .............................................. 33
4.2. Ductility Demands for Inelastic Structures .................................................... 34
4.2.1. SDOF Systems ................................................................................ 34

Table of Contents
iv
 4.2.2. MDOF Systems ............................................................................... 37
4.3. Displacement Demands for Inelastic Structures ............................................ 42
4.3.1. SDOF Inelastic Displacement Demands ......................................... 42
4.3.2. MDOF Inelastic Roof Displacement Demands ............................... 43

Chapter 5. Pulse-Type Seismic Input

5.1. Historical Perspective ..................................................................................... 68
5.2. Basic Pulse Shapes Used in this Study ........................................................... 69
5.2.1. Elastic Response Spectra ................................................................. 71
5.3. Other Pulse Shapes ......................................................................................... 72
5.3.1. Triangular Pulses ............................................................................. 72
5.3.2. Pulse Histories with Different Duration .......................................... 73

Chapter 6. Response of Structures to Pulse-Type Seismic Input

6.1. Elastic Response of MDOF Structures to Pulse-Type Input .......................... 86
6.1.1. Deflected Shapes of Structure ......................................................... 87
6.1.2. Maximum Elastic Base Shear Force ............................................... 88
6.1.3. Distribution of Elastic Story Shear Over Height ............................. 89
6.1.4. Maximum Elastic Roof Displacement ............................................ 90
6.2. Ductility Demands for Inelastic Structures .................................................... 90
6.2.1. SDOF Systems ................................................................................ 91
6.2.2. MDOF Systems ............................................................................... 94
6.3. Displacement Demands for Inelastic Structures .......................................... 100
6.3.1. SDOF Systems .............................................................................. 101
6.3.2. MDOF Systems ............................................................................. 102
6.4. Investigation of Other Pulse Shapes ............................................................. 104
6.4.1. Triangular Pulses ........................................................................... 104
6.4.2. Pulse Input Motions with Different Duration ............................... 105
6.5. Sensitivity of Inelastic Demands to Number of Stories ............................... 106
6.5.1. Generic 3-Story and 9-story Structures ......................................... 106
6.5.2. Ductility Demands ......................................................................... 107
6.5.3. Inelastic Displacement Demands .................................................. 109
6.6. Sensitivity of Inelastic Demands to Plastic Hinge Locations ...................... 111
6.6.1. MDOF System Investigated in this Study ..................................... 111
6.6.2. Story Ductility Demands Over Height .......................................... 111
6.6.3. Maximum Story Ductility Demands ............................................. 112

Table of Contents
v
 6.6.4. Base Shear Strength Demands for Target Ductility ...................... 113

Chapter 7. Representation of Near-Fault Ground Motions by Equivalent Pulses

7.1. Matching of Near-Fault Ground Motions to Equivalent Pulses ................... 156
7.1.1. Parameters of Equivalent Pulses ................................................... 156
7.1.2. Procedure for Matching ................................................................. 157
7.2. Evaluation of Equivalent Pulses ................................................................... 161
7.2.1. Comparison of SDOF Response Time Histories ........................... 161
7.2.2. Comparison of Story Ductility Demands ...................................... 163
7.2.3. Sensitivity to Pulse Type and Period ............................................. 163
7.3. Equivalent Pulse for Rotated Components .................................................. 165
7.4 Estimation of Structure Response to Near-Fault Ground Motions ............... 166

Chapter 8. Study of Models of Steel Structures

8.1. Models of Steel Structures Used in this Study ............................................. 188
8.2. Inelastic Static Analysis and Calibration of Generic Structures .................. 189
8.3. Inelastic Dynamic Analysis .......................................................................... 192
8.3.1. Roof Displacement ........................................................................ 192
8.3.2. Story Drift Angles ......................................................................... 193

Chapter 9. Design Considerations for Near-Fault Ground Motions

9.1. Relationships between Equivalent Pulse Parameters and Earthquake
Magnitude and Distance ............................................................................... 208
9.1.1. Pulse Period ................................................................................... 209
9.1.2. Pulse Intensity ............................................................................... 211
9.2. Base Shear Strength Demands for Targeted Maximum Ductilities ............. 213
9.3. Effect of Story Shear Strength Distribution on Ductility Demands ............. 215
9.3.1. Story Shear Strength Distribution for Uniform Story Ductility .... 216
9.3.2. Strengthening Schemes Based on Base Shear Strength and
Story Shear Strength Distribution 217
9.4. Strengthening of Frames with Walls ............................................................ 221
9.4.1. Dual Systems Investigated in this Study ....................................... 221
9.4.2. Demands for Dual Systems with Elastic Walls ............................. 222
9.4.3. Demands for Inelastic Walls ......................................................... 228

Table of Contents
vi
Chapter 10. Near-Fault Ground Motions from Moderate Earthquakes
10.1. Ground Motion Records Used in this Study .............................................. 263
10.2. Elastic Response Spectra ............................................................................ 264
10.2.1. Comparison with Ordinary Ground Motions .............................. 265
10.3. Story Ductility Demands Over Height ....................................................... 268

Chapter 11. Summary and Conclusions ................................................................... 280

Appendix A. Time History Traces and Elastic Spectra of Near-Fault

Ground Motions .................................................................................. 287

List of References ........................................................................................................ 298

Table of Contents
vii
 CHAPTER 1

INTRODUCTION

1.1. Statement of Problem

Near-fault ground motions have caused much damage in the vicinity of seismic sources
during recent earthquakes (Northridge 1994, Kobe 1995, and Taiwan 1999). There is
evidence indicating that ground shaking near a fault rupture may be characterized by a
short-duration impulsive motion that exposes structures to high input energy at the
beginning of the record. This pulse-type motion is particular to the “forward” direction,
where the fault rupture propagates towards the site at a velocity close to the shear wave
velocity, causing most of the seismic energy to arrive at the site within a short time
(Singh, 1985). The radiation pattern of the shear dislocation of the fault causes the pulse
to be mostly oriented perpendicular to the fault, i.e., the fault-normal component of the
motion is more severe than the fault-parallel component (Somerville, 1998). [A summary
of the near-fault seismological phenomenon is given in Somerville et al. (1997b).] The
near-fault phenomenon requires consideration in the design process for structures that are
located in the near-fault region, which is usually assumed to extend about 10 to 15 km
from the seismic source (1996 SEAOC Blue Book).

Aside from directivity effects, near-fault ground motions are more severe than “ordinary”
ground motions recorded during the same event and under similar site conditions because
proximity to the seismic source does not allow considerable attenuation of ground
motion. Furthermore, modified attenuation relationships (Somerville et al., (1997b)),
which incorporate directivity effects, suggest that for given magnitude and distance
values the pulse-type characteristics of near-fault ground motion in the forward-
directivity region may lead to significantly larger elastic spectral values compared to
those without directivity effects. To put the severity of pulse-type near-fault ground
motions in perspective, Fig. 1.1 compares velocity response spectra of near-fault and

Chapter 1 Introduction
1
ordinary design ground motions. The solid line (denoted as 15-D*) represents the mean
velocity spectrum of a set of ordinary ground motions whose individual spectra resemble
the UBC’97 soil type SD spectrum. The other lines correspond to the velocity spectra of
individual near-fault ground motions with forward directivity from various events. The
figure illustrates significant variations in the response of SDOF systems to near-fault
ground motions. It also indicates that near-fault ground motions impose seismic demands
on structures that may be several times those imposed by the mean of design level
“ordinary” ground motions. However, it should be noted that individual spectra for the
15-D* ground motions also may be much higher than the mean spectrum shown, and
exhibit variations around the mean.

The response of MDOF structures to near-fault ground motions also demonstrates special
characteristics. Figure 1.2 compares the story ductility demands for a 2-second 20-story
MDOF structure (for model description see Chapter 3) subjected to near-fault and
ordinary ground motions. The base shear strength of this structure is 15 percent of its
weight. The heavy solid line represents the mean story ductility demands for the same set
of ordinary ground motions as shown in Fig. 1.1. The peculiarity of the MDOF response
to near-fault records is again prevalent. Unlike for the ordinary ground motions, the
distribution of the demands over the height of the structure is highly non-uniform for the
near-fault records. The severity of near-fault ground motions leads to ductility demands
that are significantly larger than those for the mean of the ordinary records that represent
UBC design ground motions. Even though individual 15-D* ground motions may cause
ductility demands larger than the mean demand illustrated, on average the demands for
the near-fault records are much larger that those for the ordinary records and follow a less
uniform distribution.

The special response characteristics of near-fault ground motions deserve much scrutiny.
The development (or improvement) of design guidelines for structures close to a seismic
source requires a thorough understanding of near-fault response phenomena. Recent
seismic codes, e.g. the 1997 Uniform Building Code, have incorporated near-fault effects
by introducing source type and distance dependent near-fault factors to the customary
design spectrum. However, it is believed that these factors are not sufficient to solve the
problem consistently, because they pay little attention to the physical response
characteristics of near-fault ground motions. It may also be necessary to modify the
design shear strength distribution over the height of the structure. Moreover, the
emerging concepts of performance-based seismic design require a quantitative

Chapter 1 Introduction
2
understanding of response at different performance levels, ranging from nearly elastic
behavior to highly inelastic behavior associated with incipient collapse.

1.2. Historical Perspective

The first strong seismological evidence for the near-fault phenomenon was reported by
Benioff (1955) in his explanation of the intensity patterns observed in the 1952 Kern
County, California, earthquake. He showed that the propagation of fault rupture as a
moving source could lead to different types of ground motions at opposite ends of the
ruptured area, “with larger intensities of higher frequency in the direction of propagation
and smaller intensities and lower frequencies at the opposite end” (Singh, 1985). He also
demonstrated the kinematics of the moving radiation source along a straight line and its
effects on wave amplitudes and shapes.

The peculiar structural response to the large pulse of motion in the vicinity of seismic
source, a.k.a. “fling”, was pointed out by Mahin et al. (1976) and Bertero et al. (1978)
after the 1971 San Fernando earthquake. They noted that the building of the Olive View
Medical Center suffered extensive damage caused by a severe pulse, which they
identified as characteristic of near-fault ground motions. They concluded that the
damage was the result of only a few large displacement excursions rather than of a large
number of oscillations as in ordinary ground motions. They also concluded that short
period structures designed to code requirements could experience very large ductility
demands when subjected to near-fault ground motions, and thus special design
precautions should be taken for structures located near active faults.

After the 1979 Imperial Valley earthquake, Anderson and Bertero (1987) reported the
sensitivity of inelastic near-fault response to structure strength, as well as to the
fundamental period of the structure with respect to the period of a pulse contained in the
near-fault record. They emphasized the importance of directivity effects associated with
the direction of rupture propagation. Their investigations of MDOF structures
demonstrated that increased ductility demands in bottom stories, where axial loads are
largest, could lead to significant P-delta effects. They also suggested that the shape of the
design spectrum in the long-period range be modified for structures exposed to pulse-type
ground motions. Near-fault effects received much recognition as a result of tremendous
losses due to structural damage in populated urban areas in the 1994 Northridge and 1995
Hyogo-ken Nanbu (Kobe) earthquakes.

Chapter 1 Introduction
3
Hall et al. (1995) employed wave propagation theory to study the response of a
continuous shear building to pulse-type ground motions. They, too, warned about the
damaging effects of near-fault ground motions and the inadequacy of current code
provisions to address the problem effectively. Iwan (1997) utilized a similar elastic shear
building to obtain the “drift spectrum” (maximum story drift plotted vs. structure period)
as a measure of seismic demand for MDOF structures subjected to near-fault ground
motions with pulse-type characteristics. He showed that even for elastic structures near-
fault effects cannot be accounted for simply by multiplying the code base shear
coefficient by a near-fault factor that is constant beyond a relatively short period (as in
UBC’97).

Several studies have been aimed at mitigating the near-fault problem by improving the
performance of structures that are exposed to near-fault ground motions. Hall et al.
(1995), and Makris and Chang (2000), among others, studied the efficiency of base
isolation with various dissipative mechanisms to protect structures from pulse-type and
near-fault ground motions. Although there are some promising results, large
displacement demands imposed by severe pulses of near-fault ground motions pose many
difficulties. Anderson et al. (1999) evaluated the performance of several tall R/C and
steel frames strengthened by single or coupled shear walls (for R/C frames) and lateral
bracing systems (for steel frames). They concluded that when long period structures are
subjected to severe pulse-type ground motions, conventional retrofit strategies, such as
increasing the stiffness and/or strength of the system by adding shear walls, are not
efficient. The reason is that increasing the stiffness shortens the period of the system,
moving it into a range of higher spectral accelerations. They, however, argued for the use
of energy dissipation devices, particularly viscous dampers, as a more effective technique
to provide protection against near-fault effects.

1.3. Objectives and Scope

This study attempts a systematic evaluation of the elastic and inelastic response of SDOF
systems and MDOF frame structures subjected to near-fault ground motions. The global
objective is to acquire quantitative knowledge on near-fault ground motion effects. The
results of this study are intended to identify salient response characteristics, to describe
near-fault ground motions by simple equivalent pulses, and to utilize the pulse response
characteristics to define behavior attributes of structures when subjected to near-fault
ground motions. The ultimate goal is to develop design guidelines or strengthening

Chapter 1 Introduction
4
schemes that provide more consistent protection for structures located in near-fault
regions.

A set of recorded near-fault ground motions is utilized in the response investigations. To

enlarge the relatively small size of the recorded ground motions, a set of simulated near-
fault records is also used. The ground motions are introduced in Chapter 2, which also
addresses the effect of directivity and rotation of components of near-fault motions. In
order to derive general rather than specific information, generic rather than particular
structures are used in the response evaluations. Chapter 3 presents a description of the
generic frame structures used in this study and the assumptions made in their design.

Chapter 4 focuses on the elastic and inelastic response of structures to near-fault ground
motions. Salient near-fault response characteristics and their differences from the
characteristics of ordinary ground motions are identified. Global and story drift demands
of the generic structures are investigated through a comprehensive parametric study that
describes the variation of seismic demands with structure parameters such as fundamental
period and base shear strength. The pulse-type properties of near-fault ground motions
provide motivation for representing these ground motions by a small number of simple
pulses, which can significantly facilitate the process of response prediction and design.
Such simple pulse shapes and their spectral properties are discussed in Chapter 5.
Chapter 6 addresses the elastic and inelastic demands of structures subjected to the
simple pulses using an extensive parametric study that takes into account the effects of
structure and pulse parameters. Since near-fault ground motions tend to impose large
displacement demands on frame structures, giving rise to second-order demand
amplification, P-delta effects are also addressed in this study. The issue of representing
near-fault ground motions by equivalent pulses is pursued in Chapter 7.

In Chapter 8 particular steel structure models are employed for verification and
calibration purposes, and to assess the extent to which the results obtained from the
generic structures can be generalized. Design implications for near-fault ground motions
are presented in Chapter 9, which summarizes the results of a study that relates the design
base shear to the magnitude of the event and closest distance from the site to the source.
Improved distributions of the design story shear strength over the height of the structure
are also investigated. To provide improved protection against near-fault ground motions,
techniques are evaluated in which frame structures are strengthened by walls that are
either fixed or hinged at the base.

Chapter 1 Introduction
5
Chapter 10 is concerned with the study of near-fault effects in moderate earthquakes.
Even though collapse safety is not a matter of concern for these earthquakes, damage
control is an issue. The SDOF and MDOF response of structures to a set of ground
motions recorded during events with magnitude 6.1 and smaller is investigated. The
spectral properties of the ground motions are compared with those of ordinary records
that form the basis for current code guidelines.

Many fundamental characteristics of near-fault ground motions and their effects on frame
structures have been identified and quantified in this study. But it is recognized that the
near-fault problem is very complex, and that more work is needed before a
comprehensive understanding of all important aspects of the problem will be
accomplished. This work attempts to address the most important issues concerning near-
fault ground motions and their response attributes in order to form a foundation on which
to base future research and development of design guidelines.

Chapter 1 Introduction
6
 Elastic SDOF Velocity Demands
15-D* vs. Recorded Near-Fault, ξ== 2%
800
LP89lgpc LP89lex
EZ92erzi LN92lucr
NR94rrs NR94sylm
KB95kobj KB95tato
600
15-D* (mean)
Sv (cm / sec)

400

200

0
0 0.5 1 1.5 2 2.5 3 3.5 4
T (sec)

Figure 1.1 Velocity Response Spectra of Near-fault and Ordinary Ground Motions

Story Ductility Demands

15-D* vs. Recorded Near-Fault, T = 2.0 sec, γ = 0.15
1
LP89lgpc
LP89lex
EZ92erzi
0.8 LN92lucr
NR94rrs
NR94sylm
Relative Height

0.6 KB95kobj
KB95tato
15-D* (mean)

0.4

0.2

0
0 2 4 6 8 10 12
Story Ductility Ratio, µi = δmax,i / δy,i

Figure 1.2 Story Ductility Demands for Near-fault and Ordinary Ground Motions

Chapter 1 Introduction
7
 CHAPTER 2

NEAR-FAULT GROUND MOTIONS

USED IN THIS STUDY

2.1. Ground Motion Records

A set of 22 recorded and 18 simulated near-fault ground motion records is utilized in this
study. The designation and basic properties of the recorded ground motions are listed in
Table 2.1. The first 10 ground motions in the table were assembled by Somerville for the
SAC Steel Project (Somerville et al., 1997a). The other 12 near-fault records listed in the
table were provided by Somerville for the CDMG Strong Motion Instrumentation
Program (Somerville, 1998). The ground motions are either recorded on soil or have
been modified to NEHRP soil type D conditions. They cover a moment magnitude (Mw)
range from 6.2 to 7.4 and a distance (R) range from 0.0 to 10.0 km. Rupture distance is
used in this study, which is defined as the closest distance from the site to the fault
rupture plane. A complete set of the ground time history traces is presented in Appendix
A (Fig. A.1) for the fault-normal component of the recorded motions with forward
directivity.

To enlarge the size of the ground motion set with forward directivity, a simulated record
set is also utilized in the investigations, which covers systematic ranges of magnitude
(6.5, 7.0, and 7.5) and distance (3, 5, and 10 km). These records, which were generated
for a project sponsored by the CDMG Strong Motion Instrumentation Program
(Somerville, 1998), simulate motions for two stations (f6 and f8) in the forward direction
of a seismic source with a strike-slip faulting mechanism. The simulated ground motions
were generated using the procedure described in Somerville et al. (1996), and modified
from rock to soil conditions (see Figs. A.3 and A.4 for elastic spectra).

Chapter 2 Near-Fault Ground Motions Used …

8
In all near-fault time histories there should be static displacements due to the static
dislocation field of the earthquake. However, most recording systems do not adequately
record the permanent displacements, which are filtered out of the recordings in the course
of processing. Somerville has not attempted to retain the static displacement field in any
of the time histories, with the exception of the Lucerne recording of the 1992 Landers
earthquake (LN92lucr). This time history has been modified by Graves (1996) compared
to the version of Iwan and Chen (1994) to include geodetically defined static
displacements. MacRae et al. (1998) have shown that the effect of baseline correction on
SDOF response to near-fault ground motions is small.

2.1.1. Directivity Effects

The record set includes recordings with both forward and backward rupture directivity. If
the rupture propagates towards the site, the recording at the site will show forward-
directivity effects. Since the propagation occurs at a velocity that is close to the shear
wave velocity, most of the seismic energy from the rupture arrives at the site in a large
pulse of motion at the beginning of the record (Somerville et al., 1997b). This large pulse
is mostly oriented in the fault-normal direction on account of the radiation pattern of
shear dislocation on the fault. Figure 2.1 illustrates ground time history traces for the
fault-normal component of a near-fault ground motion (LN92lucr) that was recorded in
the forward-directivity region during the 1992 Landers earthquake (Wald and Heaton,
1994). The large pulse of motion is clearly observed in the velocity and displacement
time histories.

If the rupture propagates away from the site, the recording at the site will show
backward-directivity effects. Records with backward directivity exhibit long-duration
motions that have low amplitudes at long periods (Somerville et al., 1997b). Figure 2.2
presents the time histories for a ground motion (LN92josh) that was recorded in the
backward-directivity region of the Landers earthquake. As can be seen, this record does
not show the pulse-type characteristics of the type observed from records with forward
directivity. Instead, the seismic energy arriving at the site is scattered throughout a long-
duration ground motion. It is also observed that the maximum ground acceleration,
velocity, and displacement of this backward-directivity record are significantly smaller
than their corresponding values of the forward-directivity record LN92lucr, even though
LN92josh is recorded at a station that is closer to the epicenter of the Landers earthquake.

Chapter 2 Near-Fault Ground Motions Used …

9
This study focuses only on the response characteristics of near-fault ground motions with
forward directivity.

2.1.2. Ground Motion Components

Figure 2.3 illustrates ground velocity and displacement traces for the fault-normal and
fault-parallel components of the near-fault record NR94rrs. This ground motion, which
was recorded in the forward-directivity region of the 1994 Northridge earthquake, shows
a large pulse of motion in the fault-normal trace within the time range from 2 to 3 sec.
As pointed out earlier, the fault-normal component of the motion is much more severe
than the fault-parallel component due to the radiation pattern of shear dislocation.
Therefore, the orientation of the structure with respect to the fault direction may
determine the severity of the ground motion that the structure will experience in the near-
fault region of a fault rupture.

In order to obtain a better understanding of the orientation effect, Fig. 2.4 shows ground
velocity and displacement time histories for two rotated components of the ground
motion under consideration. These components, which are rotated by 45 degrees with
respect to the fault direction, are obtained by combining the fault-normal and fault-
parallel time histories. It can be seen that the two rotated components also exhibit pulse-
type characteristics. The time history trace of one of the rotated components is very
similar to that of the fault-normal component (Fig. 2.4(a)). Thus, it appears that pulse-
type characteristics are not limited to the fault-normal direction. The study of the time
history traces also suggests that the rotated components are relatively severe. The
severity of the rotated components is further addressed using spectral values.

2.2. Elastic Spectra of Near-Fault Ground Motions

Figure 2.5 illustrates acceleration (elastic strength demand), velocity, and displacement
spectra of the near-fault ground motion NR94rrs, whose ground time histories were
illustrated earlier. Each graph includes the spectra for the fault-normal, fault-parallel, and
the two 45° rotated components of this ground motion. All spectra are computed for 2%
damping. The figure clearly shows the large differences between the fault-normal and
fault-parallel components. These results as well as the elastic spectra of other near-fault
ground motions with forward directivity (see Appendix A, Fig. A.2) emphasize that the
fault-normal component is much more severe than the fault-parallel component. When

Chapter 2 Near-Fault Ground Motions Used …

10
these two components are rotated by 45 degrees, the difference in the spectra becomes
smaller, but one of the two rotated components still will impose demands close to (and
sometimes even higher than) those associated with the fault-normal component. This
pattern is consistent for all of the near-fault records with forward directivity studied here
(Fig. A.2). Thus, when a 3-D structure composed of frames in two perpendicular
directions is subjected to a near-fault ground motion, frames in one of these two
directions will always be exposed to excitations with an intensity level close to that of the
fault-normal component. This provides sufficient justification for focusing in this study
on the fault-normal component of near-fault ground motions.

The magnitude and distance dependence of spectral values can be observed from the
elastic spectra of the simulated ground motions, for which magnitude and distance values
vary systematically. Elastic spectra for the fault-normal component of all simulated
ground motions used in this study are illustrated in Appendix A (Figs. A.3 and A.4).

An important observation from the spectra is the existence of a predominant peak in the
fault-normal velocity spectrum of most of the near-fault records. However, some of the
records used in this study have more than one clear velocity peak. Later in Chapter 7, it
is shown that identifying the predominant peak of the velocity response spectrum is the
key to estimating the period of the pulse contained in the near-fault record.

As pointed out earlier, some of the ground motions used in this study are originally
recorded on rock and have been analytically converted into soil motions by Somerville
(1998). Figure 2.6 shows the elastic response spectra of the fault-normal component of
the near-fault record KB95kobj, which has been modified from rock to soil conditions.
The figure compares the elastic spectra of the ground motion before and after the
modification is made. As the figure indicates, the spectral values of the original and
converted records are almost identical in the period range T < 0.7 sec. However, at
longer periods the spectral values of the converted soil motion vary between 1.6 and 1.9
times those of the original rock ground motion.

This is a simplified treatment of the complicated issue of soil effects, and may lead to
questionable implications for the response of structures to the converted soil motion. For
example, as Fig. 2.6 indicates, the elastic spectral values are very close for the rock and
converted soil motions at periods around 0.7 sec., whereas inelastic SDOF demands (e.g.
ductility) for a system with T ≅ 0.7 sec. subjected to the converted soil motion will be

Chapter 2 Near-Fault Ground Motions Used …

11
much larger than the corresponding demands for the same system subjected to the
original rock motion. The reason is that the effective period of the inelastic system
elongates and moves into the period range T > 0.7 sec., in which the spectral values for
the converted record are much larger than those for the original record. The large
difference in inelastic demands is a direct consequence of the scheme employed to
account for soil effects. Moreover, the large ground velocities in near-fault ground
motions may cause nonlinear soil response, which is not considered in this simplified soil
modification method.

2.2.1. Comparison with Ordinary Ground Motions

Since near-fault ground motions are recorded close to the seismic source, the ground
shaking has very little time to attenuate. Thus, these ground motions are more severe
than the ground motions recorded far from the rupture in the same event, even without
accounting for directivity effects. Based on an empirical analysis of near-fault data,
Somerville et al. (1997b) developed modifications to empirical attenuation models to
account for the effect of rupture directivity on strong motion amplitudes in the near-fault
region. In their directivity model, the amplitude modification factor depends on two
geometrical parameters: 1) the angle between the rupture propagation direction and the
direction of waves traveling from the fault to the site, and 2) the fraction of the rupture
plane that lies between the hypocenter and the site. They conclude that forward
directivity effects cause larger spectral response at periods longer than 0.6 sec. For
example, for strike-slip faulting, maximum directivity conditions amplify the average
spectral value at T = 2.0 sec. by a factor of 1.8, which can be attributed to the pulse-type
nature of near-fault ground motions with forward directivity.

To put the severity of near-fault ground motions in perspective, a reference set of 15

“ordinary” records is utilized for comparison purposes. These records, which were used
in past studies (Seneviratna and Krawinkler, 1997), are scaled in a way such that the
spectrum of each individual record matches the UBC’97 soil type SD spectrum with a
minimum error, using discrete periods in the range from 0.6 to 4.0 seconds (constant
velocity range). The mean acceleration response spectrum of the 15 scaled records,
referred to as 15-D* (mean), is shown in Fig. 2.7 together with the UBC soil type SD
spectrum (Z = 0.4) without the near-fault factor. Thus, on average, these 15-D* time
histories reasonably represent the UBC design spectrum, which corresponds to a 10/50
(10% in 50 years) seismic hazard level.

Chapter 2 Near-Fault Ground Motions Used …

12
Figure 2.8 illustrates the mean velocity and displacement spectra of the 15-D* records
superimposed on the velocity and displacement spectra of several of the recorded near-
fault ground motions with forward directivity. This figure is presented for two reasons:
first, to illustrate the great variations in the response spectra that have to be expected from
near-fault ground motions, and second, to put the severity of near-fault ground motions in
perspective with present design ground motions. Maximum values of spectral velocities
and displacements of the near-fault records are several times those of the mean of the
design ground motions. This indicates that near-fault records can impose very large
demands that need to be considered in the design process. The response of MDOF
structures to the near-fault ground motions represented by these spectra is discussed in
Chapter 4.

Chapter 2 Near-Fault Ground Motions Used …

13
 Table 2.1 Designation and Basic Properties of Recorded Near-Fault Ground Motions
Used in this Study

Designation Earthquake Station Directivity Mw R (km)

TB78tab Tabas, 1978 Tabas backward 7.4 1.2
LP89lgpc Loma Prieta, 1989 Los Gatos forward 7.0 3.5
LP89lex Loma Prieta, 1989 Lexington forward 7.0 6.3
CM92petr Mendocino, 1992 Petrolia backward 7.1 8.5
EZ92erzi Erzincan, 1992 Erzincan forward 6.7 2.0
LN92lucr Landers, 1992 Lucerne forward 7.3 1.1
NR94rrs Nothridge, 1994 Rinaldi forward 6.7 7.5
NR94sylm Nothridge, 1994 Olive View forward 6.7 6.4
KB95kobj Kobe, 1995 JMA forward 6.9 0.6
KB95tato Kobe, 1995 Takatori forward 6.9 1.5
IV79ar06 Imperial Valley, 1979 Array 6 forward 6.5 1.2
IV79bond Imperial Valley, 1979 Bond's Corn backward 6.5 2.4
IV79melo Imperial Valley, 1979 Meloland forward 6.5 0.0
KB95kpi1 Kobe, 1995 Port Island forward 6.9 3.7
LN92josh Landers, 1992 Joshua Tree backward 7.3 7.4
LP89corr Loma Prieta, 1989 Corralitos backward 7.0 3.4
MH84andd Morgan Hill, 1984 Anderson D forward 6.2 4.5
MH84cyld Morgan Hill, 1984 Coyote L D forward 6.2 0.1
MH84hall Morgan Hill, 1984 Halls Valley backward 6.2 2.4
NR94newh Nothridge, 1994 Newhall forward 6.7 7.1
NR94nord Nothridge, 1994 Arleta backward 6.7 9.2
NR94spva Nothridge, 1994 Sepulveda forward 6.7 8.9

Chapter 2 Near-Fault Ground Motions Used …

14
 Ground Acceleration Time History
LN92lucr, Fault-Normal
800
max = 700

400

ag (cm / sec2) 0

-400

-800
0 5 10 15 20 25 30 35 40
Time (sec)

Ground Velocity Time History

LN92lucr, Fault-Normal
150
max = 136

100

50
vg (cm / sec)

-50

-100

-150
0 5 10 15 20 25 30 35 40
Time (sec)

Ground Displacement Time History

LN92lucr, Fault-Normal
300
max = 230

200

100
ug (cm)

-100

-200

-300
0 5 10 15 20 25 30 35 40
Time (sec)

Figure 2.1 Ground Acceleration, Velocity, and Displacement Time Histories of Fault-
Normal Component of Record LN92lucr with Forward Directivity

Chapter 2 Near-Fault Ground Motions Used …

15
 Ground Acceleration Time History
LN92josh, Fault-Normal
300
max = 272

200

ag (cm / sec2)
100

-100

-200

-300
0 5 10 15 20 25 30 35 40
Time (sec)

Ground Velocity Time History

LN92josh, Fault-Normal
50
max = 42.2

25
vg (cm / sec)

-25

-50
0 5 10 15 20 25 30 35 40
Time (sec)

Ground Displacement Time History

LN92josh, Fault-Normal
20
max = 14.9

10
ug (cm)

-10

-20
0 5 10 15 20 25 30 35 40
Time (sec)

Figure 2.2 Ground Acceleration, Velocity, and Displacement Time Histories of Fault-
Normal Component of Record LN92josh with Backward Directivity

Chapter 2 Near-Fault Ground Motions Used …

16
 NR94rrs, Fault-Normal
200
max = 174

100

vg (cm / sec)
0

-100

-200
40
max = 38.3

20
ug (cm)

-20

-40
0 2 4 6 8 10 12 14
Time (sec)

(a) Fault-Normal Component

NR94rrs, Fault-Parallel
200
max = 60.9

100
vg (cm / sec)

-100

-200
40
max = 17.3

20
ug (cm)

-20

-40
0 2 4 6 8 10 12 14
Time (sec)

(b) Fault-Parallel Component

Figure 2.3 Velocity and Displacement Time Histories of the Fault-Normal and Fault-
Parallel Components of Record NR94rrs

Chapter 2 Near-Fault Ground Motions Used …

17
 NR94rrs, 0.707(FN+FP)
200
max = 127

100

vg (cm / sec)
0

-100

-200
40
max = 31.9

20
ug (cm)

-20

-40
0 2 4 6 8 10 12 14
Time (sec)

(a) 0.707(FN+FP) Component

NR94rrs, 0.707(FN-FP)
200
max = 125

100
vg (cm / sec)

-100

-200
40
max = 23.4

20
ug (cm)

-20

-40
0 2 4 6 8 10 12 14
Time (sec)

(b) 0.707(FN-FP) Component

Figure 2.4 Velocity and Displacement Time Histories of 45° Rotated Components of
Record NR94rrs

Chapter 2 Near-Fault Ground Motions Used …

18
 Elastic SDOF Strength Demands
NR94rrs, ξ = 2%
3
Fault-Normal
2.5 Fault-Parallel
0.707(FN+FP)
0.707(FN-FP)
2

Sa / g 1.5

0.5

0
0 0.5 1 1.5 2 2.5 3 3.5 4
T (sec)

Elastic SDOF Velocity Demands

NR94rrs, ξ = 2%
400
Fault-Normal
Fault-Parallel
0.707(FN+FP)
300
0.707(FN-FP)
Sv (cm / sec)

200

100

0
0 0.5 1 1.5 2 2.5 3 3.5 4
T (sec)

Elastic SDOF Displacement Demands

NR94rrs, ξ = 2%
100
Fault-Normal
Fault-Parallel
80 0.707(FN+FP)
0.707(FN-FP)
60
Sd (cm)

40

20

0
0 0.5 1 1.5 2 2.5 3 3.5 4
T (sec)

Figure 2.5 Acceleration (Elastic Strength Demand), Velocity, and Displacement Spectra
for Various Components of Record NR94rrs

Chapter 2 Near-Fault Ground Motions Used …

19
 Elastic SDOF Strength Demands
KB95kobj, Fault-Normal, ξ = 2%
5
Rock

4 Soil

Sa / g
2

0
0 0.5 1 1.5 2 2.5 3 3.5 4
T (sec)

Elastic SDOF Velocity Demands

KB95kobj, Fault-Normal, ξ = 2%
800
Rock

Soil
600
Sv (cm / sec)

400

200

0
0 0.5 1 1.5 2 2.5 3 3.5 4
T (sec)

Elastic SDOF Displacement Demands

KB95kobj, Fault-Normal, ξ = 2%
120
Rock
100
Soil

80
Sd (cm)

60

40

20

0
0 0.5 1 1.5 2 2.5 3 3.5 4
T (sec)

Figure 2.6 Comparison of Elastic Spectra of Original (Rock) and Converted (Soil)
Motions for Fault-Normal Component of Record KB95kobj

Chapter 2 Near-Fault Ground Motions Used …

20
 Elastic SDOF Strength Demand Spectra
15-D* Records and UBC 97 Soil Type SD, ξ = 5%
2
UBC 97

15-D* (mean)
1.6

1.2
Sa (g)

0.8

0.4

0
0 0.5 1 1.5 2 2.5 3 3.5 4
T (sec)

Figure 2.7 Mean Acceleration (Elastic Strength Demand) Spectrum of Reference Set of
Records (15-D*) Superimposed on UBC’97 Soil Type SD Spectrum

Chapter 2 Near-Fault Ground Motions Used …

21
 Elastic SDOF Velocity Demands
15-D* vs. Recorded Near-Fault, ξ== 2%
800
LP89lgpc LP89lex
EZ92erzi LN92lucr
NR94rrs NR94sylm
KB95kobj KB95tato
600
15-D* (mean)
Sv (cm / sec)

400

200

0
0 0.5 1 1.5 2 2.5 3 3.5 4
T (sec)

(a) Velocity Spectra

Elastic SDOF Displacement Demands

15-D* vs. Recorded Near-Fault, ξ== 2%
250
LP89lgpc LP89lex
EZ92erzi LN92lucr
NR94rrs NR94sylm
200
KB95kobj KB95tato
15-D* (mean)

150
Sd (cm)

100

50

0
0 0.5 1 1.5 2 2.5 3 3.5 4
T (sec)

(b) Displacement Spectra

Figure 2.8 Velocity and Displacement Response Spectra of Near-Fault Ground Motions
and Reference Ordinary Ground Motions

Chapter 2 Near-Fault Ground Motions Used …

22
 CHAPTER 3

SDOF AND MDOF SYSTEMS USED IN THIS STUDY

3.1. SDOF Systems

Fundamental studies are carried out with elastic and inelastic SDOF systems in order to
capture basic response characteristics that differentiate near-fault ground motions from
“ordinary” ground motions. The elastic period T of the SDOF system is varied at closely
spaced intervals to provide accurate spectral information within the range of interest. For
recorded ground motions the period range is between 0 and 4.0 seconds, and for pulse-
type ground motions the primary range of interest for T/Tp is between 0 and 3.0, where Tp
is the period of the pulse. In most of the studies, a damping ratio of ξ = 2% is used rather
than the more customary value of 5%. The reason is that the focus of the study is on steel
frame structures for which 5% damping is difficult to justify.

Inelastic SDOF systems are typically defined by non-degrading bilinear hysteresis rules.
The yield strength is denoted as Fy, and the strain-hardening ratio is represented by α. A
value of α = 0.03 is used to model hardening that is representative of typical steel frame
structures.

3.2. MDOF Systems

3.2.1. Properties of Generic Structure

One of the main objectives of this study is to quantify the seismic demands of multistory
frame structures subjected to near-fault ground motions and simple pulses. To achieve
this goal, a generic 2-dimensional frame structure is used whose strength and stiffness
properties are tuned to specific requirements in order to facilitate interpretation and
generalization of response results. In this generic structure, the fundamental elastic

Chapter 3 MDOF and SDOF Systems Used …

23
period T is a variable, but the number of stories is kept constant at 20. It was considered
impractical to vary the number of stories because of the emphasis on pulse loading,
which is characterized by a pulse period Tp rather than a specific numerical value of T
that can be associated with a specific number of stories. However, the sensitivity of
response to the number of stories will be studied in Section 6.5.

In its physical configuration, the generic structure is a single-bay moment-resisting frame

whose story strengths and stiffnesses are tuned to specific requirements that are discussed
in the next section. Inelastic deformations are permitted only at the ends of the beam in
each story and at the base of the columns. Thus, the basic plastic hinge mechanism under
lateral loads involves all stories, with no individual story mechanism allowed. This
mechanism, which is illustrated in Fig. 3.1, represents structures that comply with the
“weak beam – strong column” provisions of current seismic codes. Structures that can
form story mechanisms will be studied in Section 6.6.

Response to near-fault ground motions is affected by many variables. This study is

intended to shed light on near-fault response characteristics in general terms, while
keeping the results useful for practical purposes. This requires that the assumptions made
in the design of the generic structure be realistic enough to keep the results practical, but
not too specific to compromise generality. The following assumptions are made in the
design of the generic model structure:

•= Floor mass is the same in every story and at the roof level.
•= Story height is the same in every story.
•= Bay width is twice the story height.
•= Beam and column moments of inertia are the same in each story.
•= Only flexural deformations are considered.
•= The variation of the moment of inertia over the height is tuned such that the later
defined SRSS lateral load pattern results in a straight-line deflected shape of the
structure.
•= The beam bending strength in each story is tuned such that under the SRSS lateral
load pattern simultaneous yielding occurs in all stories.
•= The effect of gravity load moments on plastic hinge formation is not considered.
•= A bilinear non-degrading hysteresis model with a 3% strain-hardening ratio is used at
all plastic hinge locations.

Chapter 3 MDOF and SDOF Systems Used …

24
Some of the assumptions will be revisited in the later chapters. In Chapter 8 the
representativeness of the generic structure will be evaluated using models of multi-story
steel frames.

For time history analyses, Rayleigh damping is used to obtain a damping ratio of 2% at
the first mode period T and at 0.1T. All MDOF structural analyses in this study are
performed using the DRAIN-2DX computer program (Prakash et al., 1993).

Caution should be exercised in the interpretation of results because the assumptions made
here cannot represent the properties of all real structures. For example, the story stiffness
and strength of multi-story frame structures typically do not change in every story, which
may cause a concentration of demands where there is a discontinuity in structure
properties.

3.2.2. Design Load Pattern

In order to establish story stiffness and strength properties, a design lateral load pattern
and base shear strength are required. The base shear yield strength is varied according to
specific objectives of the analysis and is discussed later. Given the base shear yield
strength, the individual story shear yield strengths are tuned to the story shear forces
obtained from the design load pattern. As a result, all stories will yield simultaneously if
the lateral loads follow the design load pattern. Thus, global and story “shear force-drift”
relationships obtained from a pushover analysis with the design load pattern will mimic
the bilinear shape corresponding to the SDOF systems summarized in Section 3.1.

In previous studies by Nassar and Krawinkler (1991), and Seneviratna and Krawinkler
(1997), the UBC seismic load pattern was used for stiffness and strength design of
generic models. In this study it was decided to utilize a load pattern that is based on
dynamic properties rather than code assumptions. A load pattern was selected for this
purpose that is based on story shear forces obtained from the SRSS modal superposition
method. The SRSS analysis requires the selection of a design spectrum. It is assumed
that the design spectrum follows a 1/T shape for acceleration (or constant velocity) at all
modal periods that contribute significantly to the SRSS combination. This assumption,
together with the requirement that the deflected shape under the design load pattern
should be a straight line, results in the story shear force and design load patterns
illustrated in Figs. 3.2 and 3.3. It should be noted that the spectral shape used in the

Chapter 3 MDOF and SDOF Systems Used …

25
SRSS combination may not be suitable for short structures. However, to be consistent,
the same spectral shape is used regardless of the period. Short structures will be
investigated in Section 6.5.

Since the story shear forces obtained from the SRSS combination depend on relative
story stiffnesses, an iterative procedure is required to tune the element stiffnesses so that
a straight-line deflected shape is obtained under the SRSS load pattern. Basic dynamic
properties of the generic structure (period ratios, effective masses, and modal
participation factors for modes normalized to unity at the roof level) that fulfill the
stiffness design requirements are listed in Table 3.1.

3.2.3. P-Delta Effects

It is expected that dynamic P-delta effects will be of major concern for structures
subjected to the large displacement pulses of near-fault ground motions, particularly if
inelastic interstory drifts become large and lead to ratcheting of the seismic response
(Gupta and Krawinkler, 2000).

To simulate P-delta effects, identical gravity loads are assigned to each story. This
implies that axial column forces due to gravity loads increase linearly from the top to the
bottom of the frame. The magnitude of the story gravity load is determined so that in the
first story the elastic second-order interstory drift is 10% of the first-order interstory drift
under the SRSS lateral loads. This is equivalent to a “stability coefficient” of 10% in the
first story, i.e., θ1 = (P.∆1)/(V.h1) = 0.1, where P is the total vertical gravity load, ∆1 is the
elastic first-story drift caused by the base shear V, and h1 denotes the height of the first
story. In the elastic range the consequence of incorporating P-delta effects is a 10%
reduction in elastic stiffness in the first story, and a smaller reduction in higher stories.

In this study the stability coefficient θ1 = 0.1 is used for the generic structure regardless
of the period. In reality this value will be too large for short period structures with few
stories, resulting in overestimating P-delta effects. Nevertheless, the value of θ1 is kept
constant because the period of the structure is best described relative to a pulse period,
which varies depending on the characteristics of the near-fault ground motion.

In typical US practice, steel structures consist of perimeter moment resisting frames and
interior gravity frames with simple connections. It is recognized that these gravity

Chapter 3 MDOF and SDOF Systems Used …

26
frames, which are not incorporated in this study, contribute to the lateral stiffness and
strength of the system, and may significantly reduce P-delta effects (Gupta and
Krawinkler, 2000).

The effect of P-delta on inelastic behavior is illustrated in Fig. 3.4, which shows (a) base
shear versus roof displacement, and (b) base shear versus first story displacement
diagrams obtained from a pushover analysis. Results without and with consideration of
P-delta effects are presented. The following observations can be made:

•= If P-delta effects are neglected (without P-delta), the global and interstory strain-
hardening stiffnesses are 3.7% and 3.6% of the elastic stiffness, respectively.
These values are different from the 3% strain hardening assumed at plastic hinge
locations, because the columns remain elastic after the beam plastic hinges have
formed, and contribute to the stiffness in the post-elastic range.

•= Incorporating P-delta effects decreases the elastic stiffness by 10%, and decreases
the strain-hardening ratio from +3.7% to -14.4% for the global response, and from
+3.6% to -2.8% for the first story response. The large effect on the global
response is due to the cumulative nature of the global displacement response
(summation of all story drifts). The fact that the decrease of post-elastic stiffness
in the first story is less than 10% of the elastic stiffness is attributed to the change
in the deflected shape of the structure once a mechanism has formed.

Chapter 3 MDOF and SDOF Systems Used …

27
 Table 3.1 Basic Elastic Dynamic Properties of Generic Structures

Mode # Ti / T1 Effective Mass, % Participation Factor

1 1.000 78.5 1.37
2 0.371 10.9 0.59
3 0.225 4.1 0.37
4 0.159 2.1 0.26
5 0.121 1.3 0.18
6 0.096 0.8 0.14
7 0.078 0.6 0.11
8 0.065 0.4 0.10
9 0.055 0.3 0.09
10 0.047 0.2 0.07

2h

Figure 3.1 Plastic Hinge Mechanism for Generic Frame Structure under Lateral Loads

Chapter 3 MDOF and SDOF Systems Used …

28
 SRSS Story Shear Force Pattern
Generic 20-Story Frame Structure
1

0.8
Relative Height

0.6

0.4

0.2

0
0 0.2 0.4 0.6 0.8 1 1.2
Vi / Vbase

Figure 3.2 Story Shear Force Pattern Based on SRSS Combination

SRSS Lateral Load Pattern

Generic 20-Story Frame Structure
1

0.8
Relative Height

0.6

0.4

0.2

0
0 0.2 0.4 0.6 0.8 1 1.2
Fi / Froof

Figure 3.3 Lateral Load Pattern Based on SRSS Story Shear Forces

Chapter 3 MDOF and SDOF Systems Used …

29
 Roof Displacement vs. Base Shear
Generic 20-Story Structure, SRSS Load Pattern, αelement = 3%
1.2

without P-delta
α = 3.7%
1 with P-delta

0.8
α = -14.4%
Vbase / Vbase,y

0.6

0.4

0.2

0
0 0.5 1 1.5 2 2.5
δroof=/=δroof,y

(a) Roof

First Story Drift vs. Base Shear

Generic 20-Story Structure, SRSS Load Pattern, αelement = 3%
1.2

without P-delta
α = 3.6%
1 with P-delta

α = -2.8%

0.8
Vbase / Vbase,y

0.6

0.4

0.2

0
0 0.5 1 1.5 2 2.5
δ1=/=δ1,y

(b) First Story

Figure 3.4 Global and First Story Pushover Results with and without P-Delta

Chapter 3 MDOF and SDOF Systems Used …

30
 CHAPTER 4

RESPONSE OF STRUCTURES TO NEAR-FAULT

GROUND MOTIONS

This part of the study is devoted to evaluating and quantifying the elastic and inelastic
response of SDOF and MDOF structures to near-fault ground motions. Attempts are
made to characterize important response characteristics of near-fault records. The near-
fault ground motions introduced in Chapter 2 and the structure models introduced in
Chapter 3 are utilized in the response evaluations. In most of this chapter, two near-fault
records NR94rrs and KB95kobj are used for illustration, but the observed behavior
patterns hold also for other near-fault ground motions with forward directivity.
Structures with various base shear strength levels are investigated to identify near-fault
behavior patterns at different performance levels. The reference ground motion set
introduced in Section 2.2.1 is used to emphasize major differences in the inelastic
response of MDOF structures subjected to near-fault and ordinary ground motions.

4.1. Elastic Response of MDOF Structures

4.1.1. Elastic Base Shear Demands

Examples of elastic base shear demands for the generic structures subjected to the fault-
normal component of near-fault ground motions are presented in Fig. 4.1. Each graph
compares the maximum MDOF base shear forces for structures with various fundamental
periods with the corresponding elastic SDOF strength demand spectrum.

There is a rather close agreement between the MDOF and SDOF results. The general
observation is that the MDOF base shear demand is smaller than the first mode SDOF
strength demand if higher-mode effects are not significant, and is larger than the SDOF
demand if higher-mode effects are significant (large peaks in spectrum at second and/or

Chapter 4 Response of Structures to Near-Fault …

31
third mode periods). The design implication is that if the elastic design spectrum
incorporates the effects of near-fault ground motions, elastic MDOF base shear demands
follow patterns similar to those for ordinary ground motions (see Seneviratna and
Krawinkler, 1997).

4.1.2. Elastic Shear Force Distribution Over Height of Structure

Figures 4.2 and 4.3 compare the SRSS story shear strength distribution, which was used
in the design of the generic structure (denoted as “Design”), with the elastic story shear
force distributions obtained from (a) time history analyses, and (b) SRSS modal
combinations for ground motions NR94rrs and KB95kobj, using MDOF systems with
various fundamental periods T. The following observations can be made from these
figures:

•= The results of the time history analyses indicate that for structures with a
fundamental period ≤ 1.0 sec., the story shear force distributions are smooth and
relatively close to the design distribution.

•= The distributions of the story shear forces over the height for long period systems
differ significantly from the design distribution and exhibit the effect of a wave
traveling up the structure. It appears that the traveling wave effect dominates the
MDOF response of structures whose fundamental period is longer than a
particular value that depends on the properties of the pulse contained in the near-
fault ground motion. Several researchers, e.g. Hall et al. (1995) and Iwan (1997),
have studied the traveling wave in structures subjected to near-fault and pulse-
type ground motions by means of elastic wave propagation theory rather than time
history dynamic analysis, which was used in this study. Later in Section 6.1.2
elastic demands obtained using these two methods will be compared for pulse-
type ground motions.

•= The SRSS modal superposition technique can capture only partially the traveling
wave effect. For short period structures, the distribution of story shear forces
obtained from the SRSS analysis is close to the corresponding distribution
obtained from the time history analysis, whereas for long period structures larger
differences can be seen. The reason is that in long period structures the wave

Chapter 4 Response of Structures to Near-Fault …

32
 traveling up the structure gives rise to higher-mode effects, which are not taken
into account accurately by the SRSS modal combination.

The significant deviation of the elastic story shear force distribution obtained from a time
history analysis from that of the design distribution indicates that presently employed
design story shear strength distributions will lead to early yielding of upper stories in
structures with a long fundamental period. The reason is a traveling wave effect caused
by the pulse-type nature of near-fault ground motions. The effect of this traveling wave
on inelastic demands for MDOF structures is investigated in Section 4.2.2.

4.1.3. Elastic Roof Displacement Demands

Figure 4.4 illustrates ratios of the elastic MDOF roof displacement demand to the first-
mode spectral displacement, δroof,max/Sd, for structure with various periods T, subjected to
the fault-normal component of typical near-fault records. Each graph includes two
curves, one for MDOF systems in which P-delta effects are neglected, and the other for
systems with P-delta effects. When P-delta effects are considered, the fundamental
period of the structure slightly elongates because the secondary effects reduce the
effective stiffness of the structure. This is manifested in Fig. 4.4 by a small shift of the
periods to the right. The first mode participation factor (PF1) is 1.37, which is equivalent
to δroof,max/Sd when only the first mode of the structure is taken into account. Different
patterns of deviation from this reference value for different ground motions emphasize
the uniqueness of the near-fault records. The following observations can be made from
the figure:

•= The general pattern is that the ratio oscillates about the predicted value of PF1 for
relatively short periods, and usually exceeds the predicted value by a large
amount at long periods. This observation does not comply with the results of the
study performed by Seneviratna and Krawinkler (1997) for ordinary ground
motions. This indicates that higher-mode effects are more significant in long
period MDOF systems subjected to near-fault ground motions. The reason lies in
the spectral shape of near-fault ground motions. As Fig. 4.4 indicates, for a given
record the δroof,max/Sd ratio is largest for a structure whose second mode period (T2
= 0.37 T [see Table 3.1]) corresponds to a large peak in the elastic acceleration
spectrum. Since this large peak for the record LN92lucr is at a period longer than

Chapter 4 Response of Structures to Near-Fault …

33
 4.0 seconds, the δroof,max/Sd ratio does not exceed 1.5 within the period range
investigated.

•= Figure 4.4 shows that if the period elongation due to secondary effects is
accounted for in the computation of the first mode spectral displacement (Sd), the
ratio of δroof,max/Sd will be close to the corresponding ratio obtained when P-delta
effects are neglected.

4.2. Ductility Demands for Inelastic Structures

In this part of the study ductility demands for inelastic SDOF and MDOF systems
subjected to near-fault ground motions are investigated. For MDOF systems the base
shear yield strength is quantified by a base shear coefficient, γ, defined as:

Vy Vy
γ= = (4.1)
mg W

where Vy is the base shear strength, g is the acceleration of gravity, and W and m are the
seismically effective weight and mass of the structure, respectively. Once the base shear
strength is defined, the distribution of strength over the height follows an SRSS story
shear strength pattern obtained from a constant velocity spectrum, which was discussed
in Section 3.2.2. The yield strength of SDOF systems is defined using the same
coefficient, γ, by substituting the SDOF yield strength, Fy, for the base shear strength, Vy,
in Eq. 4.1.

4.2.1. SDOF Systems

Displacement Response Time History:

Figure 4.5 illustrates inelastic displacement response time histories for SDOF systems
with various fundamental periods subjected to near-fault ground motions LP89lex and
NR94rrs. In each case the strength of the SDOF system is selected such that a ductility
ratio (µ = umax/uy) of 6 is obtained. The displacement time history values are normalized
by the yield displacement uy for the corresponding structure period and input ground
motion. The figure clearly demonstrates that the response to the near-fault ground
motions is one-sided, with no more than two large inelastic excursions followed by small

Chapter 4 Response of Structures to Near-Fault …

34
elastic cycles. These pulse-type response characteristics differentiate near-fault ground
motions from ordinary ground motions. These observations invite the conjecture that the
response of structures to near-fault records can be replicated using pulse shapes as the
input motion. Identifying such pulse shapes is one of the main objectives of this study.

Constant Ductility Strength Demand Spectra:

Figure 4.6 shows examples of elastic (µ = 1) and constant ductility inelastic strength
demand spectra for the fault-normal component of near-fault records NR94rrs and
KB95kobj. As discussed in Section 3.1, the results are computed for a non-degrading
bilinear skeleton curve with a strain-hardening ratio of 3%. The inelastic spectra are
presented for target ductility ratios µ = 2, 3, 4, 6, and 8. Similar to observations made in
past studies (Nassar and Krawinkler, 1991, and Rahnama and Krawinkler, 1993), the
humps of the elastic spectra diminish and even disappear at large ductility ratios. At the
same time, the smaller peaks and valleys of the inelastic spectra shift to lower periods,
which can be rationalized by the fact that the “effective period” of the structure elongates
when the ductility increases.

In order to evaluate the sensitivity of the inelastic strength demands to the hysteresis
model for near-fault ground motions, inelastic SDOF strength demand spectra are
computed also using the modified Clough model (see Rahnama and Krawinkler, 1993,
for hysteresis rules). Figure 4.7 illustrates ratios of the strength demands obtained from
the modified Clough model to the corresponding demands obtained from the bilinear
model for target ductility ratios µ = 2, 3, 4, 6, and 8. The figure shows the ratios for
structures with various periods T subjected to records NR94rrs and KB95kobj, which
represent the near-fault ground motions with forward directivity introduced in Chapter 2.
The following observations can be made:

•= For the ground motion NR94rrs and a given target ductility, the ratios are very
close to 1.0 except for a narrow range of period in which the strength demands
become more sensitive to the hysteresis model. This period range depends on the
target ductility ratio. Most of the sensitivity is observed in the moderate to short
period rage (T < 2.0 sec.).

•= For the ground motion KB95kobj, there is a large fluctuation of the ratios, which
is not limited to narrow ranges of period. The larger ratios, especially for T > 1.0

Chapter 4 Response of Structures to Near-Fault …

35
 sec., indicate that the strength demands for KB95kobj are more sensitive to the
hysteresis model compared to NR94rrs. Later in Chapter 7 it will be shown that
KB95kobj contains a pulse with more cycles than the pulse contained by
NR94rrs.

The conclusion is that particularly for ground motions that contain pulses with multiple
cycles, sensitivity to the hysteresis model may not be negligible.

Ductility Demands for Various Strength Levels and Periods:

Figures 4.8 and 4.9 present ductility demands of SDOF systems, subjected to near-fault
records NR94rrs and KB95kobj, versus (a) the normalized yield strength γ, and (b) the
strength reduction factor defined as R = Fy,e/Fy(µ), where Fy,e is the elastic strength
demand and Fy(µ) is the inelastic strength demand corresponding to the ductility µ. In
order to evaluate the effect of the structure period, the variation of the ductility demands
with the yield strength or the strength reduction factor is presented for selected period
values T = 0.5, 1.0, 1.5, 2.0, 2.5, 3.0 and 4.0 seconds. The following observations are
made:

•= In many cases the ductility demand, µ, varies almost linearly with the strength
reduction factor, R, particularly in the large ductility range.

•= For very short period structures (T = 0.5 sec.) µ is larger than R at all ductility
values larger than 1.0, which indicates that the inelastic displacement is larger
than the elastic displacement (δin/δel = µ/R).

•= For a given ductility ratio, R becomes larger than µ beyond a certain period
(between 0.5 and 1.0 sec. for the two records investigated here). Later in Chapter
6 it will be shown that this period depends on the period of the pulse contained in
the near-fault ground motion.

•= At very long periods, µ approaches R, which indicates that the inelastic and
elastic displacements are close, both approaching the maximum ground
displacement. Figure 4.9(b) indicates that for the record KB95kobj the equal-
displacement condition will occur at periods longer than 4.0 sec.

Chapter 4 Response of Structures to Near-Fault …

36
The patterns observed previously for R-µ-T relationships are not specific to near-fault
ground motions. However, MacRae et al. (1998) have shown that for a given ductility
ratio, the mean R value obtained for the fault-normal component of near-fault records
(without giving consideration to the pulse period) is smaller than that obtained for
ordinary records in the period range T < 1.7 sec. This shows that in this period range
limiting ductility to a target ratio requires more strength when the structure is subjected to
a near-fault record compared to an ordinary record with the same elastic spectral value.
This finding highlights the impact of directivity effects on the inelastic response of
structures subjected to near-fault ground motions.

4.2.2. MDOF Systems

Story ductility demands are used here as the basic performance parameter for MDOF
structures. The story ductility ratio is defined as the maximum story drift normalized by
the story yield drift, i.e., µi = δmax,i/δy,i. The story yield drift is obtained from a static
pushover analysis under the SRSS lateral load pattern. Distributions of story ductility
demands over the height of the structure are studied to evaluate the story response
characteristics of MDOF frame structures subjected to near-fault ground motions.
Maximum story ductility demands (maximum of all stories) are also presented, which can
be directly compared with the SDOF ductility demands discussed in the previous section.
The story ductility demands are utilized also to evaluate the importance of dynamic P-
delta effects.

Story Ductility Demands Over Height:

Strength dependent distributions of story ductility demands over the height for structures
subjected to near-fault records NR94rrs and KB95kobj are shown in Figs. 4.10 and 4.11.
Each figure illustrates the variation of ductility distribution as the base shear strength
changes, covering a range from elastic behavior to large ductility demands. This
variation is shown for structures with two different fundamental periods, i.e., (a) T = 0.5
sec. (shorter than the effective pulse period), and (b) T = 2.0 sec. (longer than the
effective pulse period). For long period systems (Figs. 4.10(b) and 4.11(b)) the
consistent observations are:

(1) the occurrence of maximum ductility demands in upper stories for relatively
strong structures,

Chapter 4 Response of Structures to Near-Fault …

37
 (2) the consequent stabilization (no further growth) of the story ductility demand in
the upper portion, and
(3) the migration of demands toward the base as the structure becomes weaker.

However, these phenomena are not observed for short period structures (Figs. 4.10(a) and
4.11(a)). For short period structures, the maximum ductility demands occur close to the
base even at high strength values, indicating that the traveling wave effect, which causes
early yielding in upper stories, only occurs in structures with a long period (longer than
the period of the effective pulse).

The reason for the early inelastic behavior in the top portion of long period strong
structures is that, as Figs. 4.2(a) and 4.3(a) indicate, the traveling wave effect in long
period systems causes the elastic shear forces in upper stories to be the first to reach the
story shear capacities (which follow the SRSS distribution). This leads to premature
yielding in the upper stories and translates into significant ductility demands in the top
portion of long period structures. On the other hand, in short period structures, the shear
forces in lower stories exceed the provided capacities first, resulting in large ductility
demands in the bottom portion of the structure. The results presented here are obtained
for structures whose relative story shear strengths follow the SRSS shear force
distribution for a 1/T-type design spectrum. However, a strength design according to this
shear force distribution may not be the best choice. Desirable story shear strength
distributions are discussed in Chapter 9.

Story ductility demands obtained from near-fault records need to be put in perspective
with the demands obtained from ordinary ground motions. For this purpose, Fig. 4.12
illustrates the distributions of story ductility demands over the height of structures
subjected to the near-fault records whose velocity and displacement response spectra are
shown in Fig. 2.8. The demands are computed for MDOF systems with a fundamental
period T = 2.0 sec. and base shear strength coefficients of γ = 0.4 and γ = 0.15, which
represent a relatively strong and a relatively weak structure, respectively. For
comparison purposes, the mean story ductility demands obtained from the reference
record set 15-D* (see Section 2.2.1) are superimposed. The following observations are
made:

•= For most of the near-fault records the maximum story ductility demand occurs in
the upper portion of the structure when the structure is strong (large γ), whereas a

Chapter 4 Response of Structures to Near-Fault …

38
 migration of ductility demands toward the base takes place when the structure
becomes weaker (small γ) or the ground motion becomes more severe. This
migration of ductility demands and the consequent concentration of demands at
the base, which occur for structures whose first mode period is longer than the
period of the pulse contained in the ground motion, are basic phenomena that
characterize near-fault ground motions with pulse-type characteristics.

•= In the mean, an SRSS-based story shear strength design results in a relatively

uniform ductility distribution for ordinary ground motions (15-D*). When the
same design is subjected to the near-fault ground motions, the pattern observed
for the individual near-fault records is maintained also in the mean, i.e., high
ductility demands occur in the upper stories for strong structures and the
maximum demand migrates to the bottom story for weak structures.

•= The severity of the near-fault records, as indicated by Fig. 2.8, translates into
larger story ductility demands for these records compared to the mean demands
for the ordinary ground motions scaled to the UBC design spectrum (15-D*).

Maximum Story Ductility Demands:

A comprehensive assessment of maximum story ductility demands (maximum of all

stories) of MDOF systems subjected to near-fault records NR94rrs and KB95kobj can be
obtained from the γ-µmax curves presented for various periods in Figs. 4.13 and 4.14. In
each figure the top diagram is for structures in which P-delta effects are neglected, and
the bottom diagram pertains to structures in which P-delta effects of the magnitude
summarized in Section 3.2.3 are accounted for. These plots can also be compared with
the corresponding plots presented for SDOF systems (Figs. 4.8(a) and 4.9(a)).

Many of the γ-µmax curves for long period systems include a range with a very steep slope
(small changes in ductility for large changes in strength), which often indicates a
migration of ductility demands from the upper stories to the base. These ranges become
more noticeable in the 1/γ-µmax diagrams, which are shown in Figs. 4.15 and 4.16 for the
same near-fault records. These graphs clearly show a steep slope for long period
structures around a ductility of 4, which is even vertical (no increase in ductility demand
with a decrease in strength) in some cases. In this range of strength the maximum
ductility in the upper stories stabilizes and grows no further as the strength is reduced,

Chapter 4 Response of Structures to Near-Fault …

39
whereas the ductility demands at the base increase and finally exceed the upper story
demands.

The ground motion intensity (or the structure strength) level corresponding to the
migration of ductility is an important level for the effect of P-delta on the maximum
ductility demand. A comparison of parts (a) and (b) of Figs. 4.13 and 4.14, and also 4.15
and 4.16, reveals that for long period structures, in which the migration phenomenon is
observed, P-delta effects are insignificant when the structure is sufficiently strong such
that the maximum story ductility occurs in upper stories. However, once the base shear
strength is reduced to a level at which the maximum demand occurs at the base, P-delta
effects become significant as evidenced by the large difference in slope between the 1/γ-
µmax curves without and with P-delta effects. The reason is that the cumulative gravity
load is largest in the first story.

The maximum ductility demands without and with P-delta effects are compared directly
in Fig. 4.17 for NR94rrs and selected periods in both the γ-µmax and 1/γ-µmax domains. It
can be seen that in the latter domain P-delta effects decrease the slope of the curves,
which means larger story drifts for a given base shear strength. This decrease is more
significant for weak structures.

Figures 4.13 to 4.17 demonstrate that P-delta effects are large for short period structures
even when the structure is relatively strong (provided that the stability coefficient of 0.1,
which is used here to evaluate P-delta effects, applies to both long and short period
structures). The reason is that for structures with a short period and a story shear strength
distribution based on the SRSS lateral load pattern, the migration phenomenon, which is
the consequence of the traveling wave effect, does not occur, and the lower stories sustain
the largest demands even when the structure is strong. Whenever large displacement
demands occur at the bottom of the structure, where gravity loads are largest, significant
P-delta effects should be expected. The large effect of P-delta even leads to the
instability of the structure with T = 0.5 sec in this case. It should be noted, however, that
the stability coefficient θ1 = 0.1 may be too large for short structures, resulting in
overestimating P-delta effects.

Near-fault and ordinary ground motions are again compared in Fig. 4.18, which shows
the maximum story ductility demand versus the base shear strength of the structure for T
= 1.0 and 2.0 sec. The following observations are made:

Chapter 4 Response of Structures to Near-Fault …

40
 •= For T = 2.0 sec. several of the curves exhibit a range with a very steep slope
around a maximum ductility demand of about 3 to 4, which is the range in which
the maximum story ductility demand migrates from the upper portion of the
structure to the base. This phenomenon is evident for ground motions whose
effective pulse period is shorter than the structure period of 2.0 seconds.

•= This phenomenon is not observed for short period structures (T = 1.0 sec.)
because the pulse period of all ground motions is ≥ 1 sec. and the maximum
ductility occurs near the bottom of the structure at all strength levels (i.e., no
migration of ductility demands takes place).

•= The large ductility demands for the near-fault records compared to the mean
demands for the15-D* ground motions are evident.

Base Shear Strength Demands for a Target Ductility:

The MDOF γ-µmax diagrams presented previously can be used to estimate ductility
demands of a structure subjected to a near-fault record for a given strength level.
Inversely, they can be used to determine the base shear strength required to limit the
maximum story ductility to a target value. For a given fundamental period T, such base
shear strength demands can be obtained from straight-line interpolations of the data
points presented in the γ-µmax graphs (e.g., Figs. 4.13 and 4.14), in order to compute γ for
a target µmax value. Figure 4.19 illustrates examples of the so obtained base shear
strength demand spectra for the fault-normal component of near-fault records NR94rrs
and KB95kobj and target story ductility ratios µmax ranging from 1 to 8.

If the design objective is to limit the maximum story ductility to a specific target value,
graphs such as those presented in Fig. 4.19 can be used to obtain the required base shear
strength for structures with various first mode periods. These graphs are directly
comparable with the SDOF constant ductility strength demand spectra (Fig. 4.6). If
SDOF and MDOF systems responded identically to near-fault ground motions, Figs. 4.6
and 4.19 would look the same. However, a comparison of the corresponding diagrams in
these two figures shows that, for a given ductility ratio, long period MDOF structures
require significantly higher strength than their SDOF counterparts on account of large
higher-mode effects. On the other hand, at short periods the strength demands for MDOF
systems are close to (and sometimes smaller than) the corresponding demands for SDOF

Chapter 4 Response of Structures to Near-Fault …

41
systems with the same period. Again, the dividing line between “short” and “long”
periods is the period of the pulse contained in the ground motion (1.0 sec. for the
NR94rrs record and 0.9 sec. for the KB95kobj record).

Strength Demands for Rotated Components:

The elastic response spectra of 45° rotated components of near-fault ground motions were
evaluated in Section 2.2. It was shown that the elastic response for one of the rotated
components is comparable to the response associated with the fault-normal component.
In this part of the study the inelastic response of MDOF structures to the rotated
components of near-fault records is investigated. MDOF base shear strength demand
spectra similar to those presented in Fig. 4.19 can be computed for the rotated
components by employing the interpolation scheme discussed earlier. If the strength
demand obtained from a rotated component is divided by the corresponding value
obtained from the fault-normal component for the same T and µ values, the resulting
ratios represent the required base shear strength for the rotated component as a fraction of
the strength required for the fault-normal component. Examples of such ratios are
illustrated in Figs. 4.20 and 4.21 for the 45° rotated components of near-fault ground
motions NR94rrs and KB95kobj.

The results indicate that one of the 45° components exposes the structure to strength
demands almost as high as those obtained for the fault-normal component. The ratios
larger than unity for one of the rotated components of KB95kobj (Fig. 4.21(b)) highlight
the intensity of this component. One could argue for a strength reduction factor of about
0.8 for the 45° rotated component, but this argument does not apply consistently. In view
of the many uncertainties and unknowns involved in quantifying near-fault effects, it is
argued that the emphasis on the fault-normal component as a representative component is
justified.

4.3. Displacement Demands for Inelastic Structures

4.3.1. SDOF Inelastic Displacement Demands

The elastic displacement spectra of near-fault ground motions were addressed in Chapter
2. This section focuses on a comparison between elastic and inelastic displacement
demands for SDOF structures subjected to near-fault records. Examples of this

Chapter 4 Response of Structures to Near-Fault …

42
comparison are provided in Figure 4.22, which presents ratios of inelastic to elastic
spectral displacement demands (δin/δel = µ/R) for the fault-normal component of ground
motions NR94rrs and KB95kobj. In each case the strength of the inelastic system is
selected such that target ductility ratios of µ = 2, 3, 4, 6, and 8 are obtained. The ratios
follow patterns observed for ordinary ground motions: in the short period range the
inelastic spectral displacements are larger than the elastic ones, and increase with a
reduction in strength, whereas the reverse is usually observed in the intermediate period
range; in the long period range the inelastic and elastic displacements are close because
they both approach the maximum ground displacement.

Baez and Miranda (2000) investigated the inelastic to elastic displacement ratios for sets
of ordinary and near-fault ground motions, and showed that in the period range from 0.1
to 1.3 sec., the ratios obtained for near-fault records are larger than the ratios obtained for
ordinary ground motions, whereas the opposite is true at periods longer than 1.8 sec.
They also concluded that the ratios are larger for the fault-normal component of near-
fault ground motions compared to the fault-parallel component in the period range from
0.1 to 1.3 sec.

4.3.2. MDOF Inelastic Roof Displacement Demands

Information on elastic roof displacement demands for near-fault ground motions was
provided in Section 4.1.3. This section addresses the relationship between elastic and
inelastic roof displacement demands without P-delta effects. Figure 4.23 illustrates the
inelastic roof displacement of MDOF systems with various base shear strength
coefficients γ, normalized by their corresponding elastic roof displacement, for ground
motions NR94rrs and KB95kobj.

The relative values of inelastic to elastic roof displacements generally follow patterns
observed for inelastic to elastic spectral displacements (Fig. 4.22). The inelastic roof
displacement demands are larger than the elastic ones in the short period range, and
become much larger with decreasing period and base shear strength. On the other hand,
in the longer period range the reverse is observed, i.e., the inelastic roof displacement
demands are smaller than the elastic ones and decrease when the structure becomes
weaker.

Chapter 4 Response of Structures to Near-Fault …

43
Summary:

The major observation that summarizes the investigations presented in this chapter is that
the response of structures to near-fault ground motions with forward directivity has
peculiar characteristics that set such motions apart from ordinary ground motions. The
response clearly shows pulse-type characteristics that are specific to individual ground
motions and strongly depend on the structure period and strength. The effect of the
structure period on the response has to be put in perspective with the effective period of
the pulse contained in the near-fault ground motion. The identification of the pulse
period will be discussed in Chapter 7. In the next chapter, attempts are made to define
simple pulse motions whose response attributes are similar to those of near-fault records.

Chapter 4 Response of Structures to Near-Fault …

44
 Elastic Base Shear Strength Demands Elastic Base Shear Strength Demands
LP89lex, ξ = 2%, without P-∆
∆ LN92lucr, ξ = 2%, without P-∆
∆
4 1
MDOF MDOF

SDOF 0.8 SDOF

3

0.6
Vy,e / W

0.4

1
0.2

0 0
0 0.5 1 1.5 2 2.5 3 3.5 4 0 0.5 1 1.5 2 2.5 3 3.5 4
T (sec) T (sec)

(a) LP89lex (b) LN92lucr

Elastic Base Shear Strength Demands Elastic Base Shear Strength Demands
NR94rrs, ξ = 2%, without P-∆
∆ NR94sylm, ξ = 2%, without P-∆
∆
3 2
MDOF MDOF
2.5
SDOF SDOF
1.5
2
Vy,e / W

1.5 1

1
0.5
0.5

0 0
0 0.5 1 1.5 2 2.5 3 3.5 4 0 0.5 1 1.5 2 2.5 3 3.5 4
T (sec) T (sec)

(c) NR94rrs (d) NR94sylm

Elastic Base Shear Strength Demands Elastic Base Shear Strength Demands
KB95kobj, ξ = 2%, without P-∆
∆ KB95tato, ξ = 2%, without P-∆
∆
5 4
MDOF MDOF

4 SDOF SDOF
3

3
Vy,e / W

1
1

0 0
0 0.5 1 1.5 2 2.5 3 3.5 4 0 0.5 1 1.5 2 2.5 3 3.5 4
T (sec) T (sec)

(e) KB95kobj (f) KB95tato

Figure 4.1 Elastic MDOF Base Shear and SDOF Strength Demands for Fault-Normal
Component of Near-Fault Ground Motions

Chapter 4 Response of Structures to Near-Fault …

45
 Normalized Elastic Story Shear Demands
∆, Time History
NR94rrs, without P-∆
1

0.8
Relative Height

0.6

Design
0.4 T = 0.5 sec
T = 1.0 sec
T = 1.5 sec
T = 2.0 sec
0.2 T = 2.5 sec
T = 3.0 sec
T = 4.0 sec
0
0 0.2 0.4 0.6 0.8 1 1.2
Vi,max / Vbase,max

(a) Time History

Normalized Elastic Story Shear Demands

∆, SRSS
NR94rrs, without P-∆
1

0.8
Relative Height

0.6

Design
0.4 T = 0.5 sec
T = 1.0 sec
T = 1.5 sec
T = 2.0 sec
0.2 T = 2.5 sec
T = 3.0 sec
T = 4.0 sec
0
0 0.2 0.4 0.6 0.8 1 1.2
Vi,max / Vbase,max

(b) SRSS

Figure 4.2 Normalized Elastic Story Shear Demands Obtained from Time History and
SRSS Analyses for Record NR94rrs

Chapter 4 Response of Structures to Near-Fault …

46
 Normalized Elastic Story Shear Demands
∆, Time History
KB95kobj, without P-∆
1

0.8
Relative Height

0.6

Design
0.4 T = 0.5 sec
T = 1.0 sec
T = 1.5 sec
T = 2.0 sec
0.2 T = 2.5 sec
T = 3.0 sec
T = 4.0 sec
0
0 0.2 0.4 0.6 0.8 1 1.2
Vi,max / Vbase,max

(a) Time History

Normalized Elastic Story Shear Demands

∆, SRSS
KB95kobj, without P-∆
1

0.8
Relative Height

0.6

Design
0.4 T = 0.5 sec
T = 1.0 sec
T = 1.5 sec
T = 2.0 sec
0.2 T = 2.5 sec
T = 3.0 sec
T = 4.0 sec
0
0 0.2 0.4 0.6 0.8 1 1.2
Vi,max / Vbase,max

(b) SRSS

Figure 4.3 Normalized Elastic Story Shear Demands Obtained from Time History and
SRSS Analyses for Record KB95kobj

Chapter 4 Response of Structures to Near-Fault …

47
 Elastic Roof Displacement Demands Elastic Roof Displacement Demands
LP89lex, ξ = 2% LN92lucr, ξ = 2%
2 2
without P-delta without P-delta

1.8 with P-delta 1.8 with P-delta

δ roof,max / Sd

1.6 1.6

1.4 1.4

1.2 1.2

1 1
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5
T (sec) T (sec)

(a) LP89lex (b) LN92lucr

Elastic Roof Displacement Demands Elastic Roof Displacement Demands

NR94rrs, ξ = 2% NR94sylm, ξ = 2%
2 2
without P-delta without P-delta

1.8 with P-delta 1.8 with P-delta

δ roof,max / Sd

1.6 1.6

1.4 1.4

1.2 1.2

1 1
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5
T (sec) T (sec)

(c) NR94rrs (d) NR94sylm

Elastic Roof Displacement Demands Elastic Roof Displacement Demands

KB95kobj, ξ = 2% KB95tato, ξ = 2%

2 2
without P-delta without P-delta

1.8 with P-delta 1.8 with P-delta

δ roof,max / Sd

1.6 1.6

1.4 1.4

1.2 1.2

1 1
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5
T (sec) T (sec)

(e) KB95kobj (f) KB95tato

Figure 4.4 Ratio of Elastic MDOF Roof Displacement Demand to First Mode Spectral
Displacement for Near-Fault Records

Chapter 4 Response of Structures to Near-Fault …

48
 Normalized Inelastic SDOF Displacement Time History
LP89lex, µ = 6, α = 3%, ξ = 2%
6
T = 0.25 sec
T = 0.50 sec
4
T = 1.00 sec
T = 2.00 sec
2
u / uy

-2

-4

-6
0 1 2 3 4 5 6
t (sec)

(a) Record LP89lex

Normalized Inelastic SDOF Displacement Time History

NR94rrs, µ = 6, α = 3%, ξ = 2%
6
T = 0.25 sec
T = 0.50 sec
4
T = 1.00 sec
T = 2.00 sec
2
u / uy

-2

-4

-6
0 1 2 3 4 5
t (sec)

(b) Record NR94rrs

Figure 4.5 Normalized Inelastic SDOF Displacement Time Histories for Various Periods
and µ = 6

Chapter 4 Response of Structures to Near-Fault …

49
 SDOF Strength Demands for Constant Ductility
NR94rrs, Bilinear, α = 3%, ξ = 2%
3
µ===1
µ===2
2.5 µ===3
µ===4
µ===6
2 µ===8
γ = Fy / W

1.5

0.5

0
0 0.5 1 1.5 2 2.5 3 3.5 4
T (sec)

(a) Record NR94rrs

SDOF Strength Demands for Constant Ductility

KB95kobj, Bilinear, α = 3%, ξ = 2%
5
µ===1
µ===2
µ===3
4
µ===4
µ===6
µ===8
3
γ = Fy / W

0
0 0.5 1 1.5 2 2.5 3 3.5 4
T (sec)

(b) Record KB95kobj

Figure 4.6 Elastic and Inelastic SDOF Strength Demand Spectra for Constant Ductility
Ratios for Near-Fault Ground Motions

Chapter 4 Response of Structures to Near-Fault …

50
 SDOF Strength Demands for Constant Ductility
NR94rrs, Clough / Bilinear, α = 3%, ξ = 2%
2
µ===2
µ===3
µ===4
1.5 µ===6
µ===8
γ Clough / γ Bi

0.5

0
0 0.5 1 1.5 2 2.5 3 3.5 4
T (sec)

(a) Record NR94rrs

SDOF Strength Demands for Constant Ductility

KB95kobj, Clough / Bilinear, α = 3%, ξ = 2%
2
µ===2
µ===3
µ===4
1.5 µ===6
µ===8
γ Clough / γ Bi

0.5

0
0 0.5 1 1.5 2 2.5 3 3.5 4
T (sec)

(b) Record KB95kobj

Figure 4.7 Comparison of Inelastic SDOF Strength Demand Spectra for Constant
Ductility for Bilinear and Modified Clough Hysteresis Models

Chapter 4 Response of Structures to Near-Fault …

51
 SDOF Ductility Demands
NR94rrs, Bilinear, α = 3%, ξ = 2%
1
T = 0.5 sec
T = 1.0 sec
T = 1.5 sec
0.8 T = 2.0 sec
T = 2.5 sec
T = 3.0 sec
T = 4.0 sec
0.6
γ = Fy / W

0.4

0.2

0
0 1 4 8 12 16 20
Ductility Ratio, µ

(a) γ versus µ

µ Relationship
SDOF R-µ
NR94rrs, Bilinear, α = 3%, ξ = 2%
16
T=0.5 sec
T=1.0 sec
T=1.5 sec
T=2.0 sec
12 T=2.5 sec
T=3.0 sec
T=4.0 sec
R = Fy,e / Fy(µ)

0
0 4 8 12 16 20
Ductility Ratio, µ

(b) R versus µ=

Figure 4.8 Ductility Demands for SDOF Systems Subjected to Record NR94rrs,
Various Periods

Chapter 4 Response of Structures to Near-Fault …

52
 SDOF Ductility Demands
KB95kobj, Bilinear, α = 3%, ξ = 2%
1
T = 0.5 sec
T = 1.0 sec
T = 1.5 sec
0.8 T = 2.0 sec
T = 2.5 sec
T = 3.0 sec
T = 4.0 sec
0.6
γ = Fy / W

0.4

0.2

0
0 1 4 8 12 16 20
Ductility Ratio, µ

(a)=γ versus µ

µ Relationship
SDOF R-µ
KB95kobj, Bilinear, α = 3%, ξ = 2%
16

12
R = Fy,e / Fy(µ)

T=0.5 sec
T=1.0 sec
4 T=1.5 sec
T=2.0 sec
T=2.5 sec
T=3.0 sec
T=4.0 sec
0
0 4 8 12 16 20
Ductility Ratio, µ

(b) R versus µ=
=
Figure 4.9 Ductility Demands for SDOF Systems Subjected to Record KB95kobj,
Various Periods

Chapter 4 Response of Structures to Near-Fault …

53
 Story Ductility Demands
∆
NR94rrs, T = 0.5 sec, without P-∆
1
γ===2.00
γ===1.50
γ===1.00
0.8
γ===0.80
γ===0.60
γ===0.50
Relative Height

0.6 γ===0.40
γ===0.30
γ===0.25

0.4

0.2

0
0 4 8 12 16 20
Story Ductility Ratio, µi = δmax,i / δy,i

(a) T = 0.5 sec.

Story Ductility Demands

∆
NR94rrs, T = 2.0 sec, without P-∆
1
γ===0.80
γ===0.50
γ===0.40
0.8
γ===0.25
γ===0.20
γ===0.15
Relative Height

0.6 γ===0.10
γ===0.07
γ===0.05

0.4

0.2

0
0 4 8 12 16 20
Story Ductility Ratio, µi = δmax,i / δy,i

(b) T = 2.0 sec.

Figure 4.10 Dependence of Distributions of Story Ductility Demands on Base Shear

Strength for Record NR94rrs

Chapter 4 Response of Structures to Near-Fault …

54
 Story Ductility Demands
∆
KB95kobj, T = 0.5 sec, without P-∆
1
γ===2.00
γ===1.50
γ===1.00
0.8
γ===0.80
γ===0.60
γ===0.50
Relative Height

0.6 γ===0.40
γ===0.30
γ===0.25

0.4

0.2

0
0 4 8 12 16 20
Story Ductility Ratio, µi = δmax,i / δy,i

(a) T = 0.5 sec.

Story Ductility Demands

∆
KB95kobj, T = 2.0 sec, without P-∆
1
γ===0.80
γ===0.50
γ===0.40
0.8
γ===0.25
γ===0.20
γ===0.15
Relative Height

0.6 γ===0.10
γ===0.07
γ===0.05

0.4

0.2

0
0 4 8 12 16 20
Story Ductility Ratio, µi = δmax,i / δy,i

(b) T = 2.0 sec.

Figure 4.11 Dependence of Distributions of Story Ductility Demands on Base Shear

Strength for Record KB95kobj

Chapter 4 Response of Structures to Near-Fault …

55
 Story Ductility Demands
15-D* vs. Recorded Near-Fault, T = 2.0 sec, γ = 0.40
1
LP89lgpc
LP89lex
EZ92erzi
0.8 LN92lucr
NR94rrs
NR94sylm
Relative Height

0.6 KB95kobj
KB95tato
15-D* (mean)

0.4

0.2

0
0 1 2 3 4 5
Story Ductility Ratio, µi = δmax,i / δy,i

(a) Strong Structure, γ = 0.40

Story Ductility Demands

15-D* vs. Recorded Near-Fault, T = 2.0 sec, γ = 0.15
1
LP89lgpc
LP89lex
EZ92erzi
0.8 LN92lucr
NR94rrs
NR94sylm
Relative Height

0.6 KB95kobj
KB95tato
15-D* (mean)

0.4

0.2

0
0 2 4 6 8 10 12
Story Ductility Ratio, µi = δmax,i / δy,i

(b) Weak Structure, γ = 0.15

Figure 4.12 Story Ductility Demands for Near-Fault Ground Motions and Reference
Ground Motions, Structure Period T = 2.0 sec

Chapter 4 Response of Structures to Near-Fault …

56
 Maximum MDOF Ductility Demands
NR94rrs, α = 3%, without P-∆
∆
1
T = 0.5 sec
T = 1.0 sec
T = 1.5 sec
0.8
T = 2.0 sec
T = 2.5 sec
T = 3.0 sec
0.6 T = 4.0 sec
γ== Vy / W

0.4

0.2

0
0 1 4 8 12 16 20
Maximum Story Ductility, µmax

(a) without P-delta Effects

Maximum MDOF Ductility Demands

NR94rrs, α = 3%, with P-∆
∆
1
T = 0.5 sec
T = 1.0 sec
T = 1.5 sec
0.8
T = 2.0 sec
T = 2.5 sec
T = 3.0 sec
0.6 T = 4.0 sec
γ== Vy / W

0.4

0.2

0
0 1 4 8 12 16 20
Maximum Story Ductility, µmax

(b) with P-delta Effects

Figure 4.13 Base Shear Strength vs. Maximum Story Ductility Demands for Record
NR94rrs

Chapter 4 Response of Structures to Near-Fault …

57
 Maximum MDOF Ductility Demands
KB95kobj, α = 3%, without P-∆
∆
1
T = 0.5 sec
T = 1.0 sec
T = 1.5 sec
0.8
T = 2.0 sec
T = 2.5 sec
T = 3.0 sec
0.6 T = 4.0 sec
γ== Vy / W

0.4

0.2

0
0 1 4 8 12 16 20
Maximum Story Ductility, µmax

(a) without P-delta Effects

Maximum MDOF Ductility Demands

KB95kobj, α = 3%, with P-∆
∆
1
T = 0.5 sec
T = 1.0 sec
T = 1.5 sec
0.8
T = 2.0 sec
T = 2.5 sec
T = 3.0 sec
0.6 T = 4.0 sec
γ== Vy / W

0.4

0.2

0
0 1 4 8 12 16 20
Maximum Story Ductility, µmax

(b) with P-delta Effects

Figure 4.14 Base Shear Strength vs. Maximum Story Ductility Demands for Record
KB95kobj

Chapter 4 Response of Structures to Near-Fault …

58
 Maximum MDOF Ductility Demands
NR94rrs, α = 3%, without P-∆
∆
30
T = 0.5 sec
T = 1.0 sec
25 T = 1.5 sec
T = 2.0 sec
T = 2.5 sec
20 T = 3.0 sec
1 / γ == W / V y

T = 4.0 sec

15

10

0
0 4 8 12 16 20
Maximum Story Ductility, µmax

(a) without P-delta Effects

Maximum MDOF Ductility Demands

NR94rrs, α = 3%, with P-∆
∆
30
T = 0.5 sec
T = 1.0 sec
25 T = 1.5 sec
T = 2.0 sec
T = 2.5 sec
20 T = 3.0 sec
1 / γ == W / V y

T = 4.0 sec

15

10

0
0 4 8 12 16 20
Maximum Story Ductility, µmax

(b) with P-delta Effects

Figure 4.15 Inverse of Base Shear Strength vs. Maximum Story Ductility Demands for
Record NR94rrs

Chapter 4 Response of Structures to Near-Fault …

59
 Maximum MDOF Ductility Demands
KB95kobj, α = 3%, without P-∆
∆
30
T = 0.5 sec
T = 1.0 sec
25 T = 1.5 sec
T = 2.0 sec
T = 2.5 sec
20 T = 3.0 sec
T = 4.0 sec
1 / γ == W / V y

15

10

0
0 4 8 12 16 20
Maximum Story Ductility, µmax

(a) without P-delta Effects

Maximum MDOF Ductility Demands

KB95kobj, α = 3%, with P-∆
∆
30
T = 0.5 sec
T = 1.0 sec
25 T = 1.5 sec
T = 2.0 sec
T = 2.5 sec
20 T = 3.0 sec
T = 4.0 sec
1 / γ == W / V y

15

10

0
0 4 8 12 16 20
Maximum Story Ductility, µmax

(b) with P-delta Effects

Figure 4.16 Inverse of Base Shear Strength vs. Maximum Story Ductility Demands for
Record KB95kobj

Chapter 4 Response of Structures to Near-Fault …

60
 Maximum MDOF Ductility Demands
NR94rrs, α = 3%
1
T=1.0 sec, without P-delta
T=1.0 sec, with P-delta
0.8 T=2.0 sec, without P-delta
T=2.0 sec, with P-delta
T=3.0 sec, without P-delta
T=3.0 sec, with P-delta
0.6
γ= = Vy / W

0.4

0.2

0
0 4 8 12 16 20
Maximum Story Ductility, µmax

(a) γ versus µmax

Maximum MDOF Ductility Demands

NR94rrs, α = 3%
30
T=1.0 sec, without P-delta
T=1.0 sec, with P-delta
25 T=2.0 sec, without P-delta
T=2.0 sec, with P-delta
T=3.0 sec, without P-delta
20
T=3.0 sec, with P-delta
1 / γ= = W / V y

15

10

0
0 4 8 12 16 20
Maximum Story Ductility, µmax

(b) 1/γ versus µmax

Figure 4.17 Effect of P-Delta on Maximum Story Ductility Demands for Record
NR94rrs

Chapter 4 Response of Structures to Near-Fault …

61
 Maximum MDOF Ductility Demands
∆
15-D* vs. Recorded Near-Fault, T = 1.0 sec, without P-∆
1
LP89lgpc
LP89lex
0.8 EZ92erzi
LN92lucr
NR94rrs
0.6 NR94sylm
γ = Vy / W

KB95kobj
KB95tato
0.4 15-D* (mean)

0.2

0
0 4 8 12 16 20
Maximum Story Ductility, µmax

(a) T = 1.0 sec.

Maximum MDOF Ductility Demands

∆
15-D* vs. Recorded Near-Fault, T = 2.0 sec, without P-∆
1
LP89lgpc
LP89lex
0.8 EZ92erzi
LN92lucr
NR94rrs
0.6 NR94sylm
γ = Vy / W

KB95kobj
KB95tato
0.4 15-D* (mean)

0.2

0
0 4 8 12 16 20
Maximum Story Ductility, µmax

(b) T = 2.0 sec.

Figure 4.18 Base Shear Strength vs. Maximum Ductility Demand for Near-Fault Ground
Motions and Reference Ground Motions

Chapter 4 Response of Structures to Near-Fault …

62
 MDOF Strength Demands for Constant Ductility
NR94rrs, Fault-Normal, ξ = 2%, without P-∆
∆
3
µ===1
µ===2
2.5 µ===3
µ===4
µ===6
2 µ===8
γ = Vy / W

1.5

0.5

0
0 0.5 1 1.5 2 2.5 3
T (sec)

(a) Record NR94rrs

MDOF Strength Demands for Constant Ductility

KB95kobj, Fault-Normal, ξ = 2%, without P-∆
∆
5
µ===1
µ===2
µ===3
4
µ===4
µ===6
µ===8
3
γ = Vy / W

0
0 0.5 1 1.5 2 2.5 3
T (sec)

(b) Record KB95kobj

Figure 4.19 MDOF Base Shear Strength Demand Spectra for Target Maximum Story
Ductility, Near-Fault Ground Motions

Chapter 4 Response of Structures to Near-Fault …

63
 MDOF Strength Demand Ratios for Constant Ductility
NR94rrs, 0.707(FN+FP) vs. Fault-Normal, ξ = 2%, no P-∆
∆
1.2

0.8
γ 45°=/=γ FN

0.6

0.4 µ===1
µ===2
µ===3
0.2 µ===4
µ===6
µ===8
0
0 0.5 1 1.5 2 2.5 3
T (sec)

(a) 0.707(FN+FP) Component

MDOF Strength Demand Ratios for Constant Ductility

NR94rrs, 0.707(FN-FP) vs. Fault-Normal, ξ = 2%, no P-∆
∆
1.2

0.8
γ 45°=/=γ FN

0.6

0.4 µ===1
µ===2
µ===3
0.2 µ===4
µ===6
µ===8
0
0 0.5 1 1.5 2 2.5 3
T (sec)

(b) 0.707(FN-FP) Component

Figure 4.20 Comparison of Base Shear Strength Demands for Rotated and Fault-Normal
Components of Near-Fault Record NR94rrs

Chapter 4 Response of Structures to Near-Fault …

64
 MDOF Strength Demand Ratios for Constant Ductility
KB95kobj, 0.707(FN+FP) vs. Fault-Normal, ξ = 2%, no P-∆
∆
1.6

1.2
γ 45°=/=ηFN

0.8

µ===1
µ===2
0.4 µ===3
µ===4
µ===6
µ===8
0
0 0.5 1 1.5 2 2.5 3
T (sec)

(a) 0.707(FN+FP) Component

MDOF Strength Demand Ratios for Constant Ductility

KB95kobj, 0.707(FN-FP) vs. Fault-Normal, ξ = 2%, no P-∆
∆
1.6

1.2
γ 45°=/=γ FN

0.8

µ===1
µ===2
0.4 µ===3
µ===4
µ===6
µ===8
0
0 0.5 1 1.5 2 2.5 3
T (sec)

(b) 0.707(FN-FP) Component

Figure 4.21 Comparison of Base Shear Strength Demands for Rotated and Fault-Normal
Components of Near-Fault Record KB95kobj

Chapter 4 Response of Structures to Near-Fault …

65
 Normalized Inelastic SDOF Displacement Demands
NR94rrs, Bilinear, α = 3%, ξ = 2%
5
µ===2
µ===3
4 µ===4
µ===6
µ===8
δinelastic=/=δ elastic

0
0 0.5 1 1.5 2 2.5 3 3.5 4
T (sec)

(a) Record NR94rrs

Normalized Inelastic SDOF Displacement Demands

KB95kobj, Bilinear, α = 3%, ξ = 2%
5
µ===2
µ===3
4 µ===4
µ===6
µ===8
δinelastic=/=δ elastic

0
0 0.5 1 1.5 2 2.5 3 3.5 4
T (sec)

(b) Record KB95kobj

Figure 4.22 Ratios of Inelastic to Elastic SDOF Displacement Demands for Near-Fault
Ground Motions

Chapter 4 Response of Structures to Near-Fault …

66
 Normalized Inelastic Roof Displacement Demands
∆
NR94rrs, without P-∆
5
γ===0.05
γ===0.10
γ===0.20
4 γ===0.30
γ===0.40
γ===0.50
γ===0.60
δinelastic=/=δ elastic

3 γ===0.80
γ===1.00

0
0 0.5 1 1.5 2 2.5 3 3.5 4
T (sec)

(a) Record NR94rrs

Normalized Inelastic Roof Displacement Demands

∆
KB95kobj, without P-∆
5
γ===0.05
γ===0.10
γ===0.20
4 γ===0.30
γ===0.40
γ===0.50
γ===0.60
δinelastic=/=δ elastic

3 γ===0.80
γ===1.00

0
0 0.5 1 1.5 2 2.5 3 3.5 4
T (sec)

(b) Record KB95kobj

Figure 4.23 Ratios of Inelastic to Elastic MDOF Roof Displacement Demands for Near-
Fault Ground Motions

Chapter 4 Response of Structures to Near-Fault …

67
 CHAPTER 5

PULSE-TYPE SEISMIC INPUT

As discussed in Chapters 2 and 4, pulse-type characteristics are discernible in both the

ground motion time history traces and the response of structures to near-fault ground
motions with forward directivity. It was also shown that near-fault ground motions come
in large variations, which make a consistent evaluation of near-fault effects difficult and
cumbersome. If simple pulse models can be found that represent near-fault ground
motions with reasonable accuracy, the process of design or response evaluation will be
significantly facilitated. Furthermore, the study of simple pulses along with real ground
motions provides a more transparent picture of near-fault response properties, and leads
to a better understanding of the near-fault phenomena.

The objective of this part of the study is to introduce such simplified representations of
near-fault ground motions. Various pulse shapes and their spectral values are evaluated.
Three basic pulse shapes are utilized for this purpose. Other variations are also
investigated in addition to the three basic pulses. A comprehensive evaluation of the
response of structures to several of these pulse-type input motions is presented in the next
chapter.

5.1. Historical Perspective

The study of the dynamic response of structures to impulsive loading dates back to the
mid 20th century, when input pulses were first utilized to model the impact forces
imposed by bomb blasts. Biggs (1964) evaluated the response of elastic and inelastic
SDOF systems to one-sided force pulses of various shapes, i.e., rectangular, triangular
and ramp-like pulses. He observed that for each pulse shape, the maximum elastic
response is a function of the pulse intensity and the td/T ratio, where td is the duration of
the pulse and T is the natural period of the system. For undamped elastic systems

Chapter 5 Pulse-Type Seismic Input

68
subjected to the one-sided force pulses, the displacement response factor (maximum
dynamic displacement normalized by static displacement) does not exceed 2.0. For
elasto-plastic systems, he illustrated the dependence of the ductility demand on the td/T
ratio and the strength of the system relative to the intensity of the input pulse.

The response of elastic SDOF systems to a half-cycle sinusoidal force pulse is discussed
in Chopra (1995). Chopra demonstrated that the effect of damping on the maximum
response to pulse-type forces is not significant unless the system is highly damped. The
reason is that the energy dissipated by damping is small when the system is subjected to
pulse-type excitations with short duration. For the sinusoidal pulse the displacement
response factor is more sensitive to the damping ratio when the pulse duration is shorter
than the natural period of the system.

Simple pulse shapes have been used in the literature also to represent earthquake ground
motions (Veletsos et al., 1965). In the past, when ground motion records were scarce and
an evaluation of dynamic response was computationally expensive and time consuming,
simple pulses were very useful. Veletsos et al. (1965) presented elastic response spectra
for half- and full-cycle sinusoidal pulses of ground velocity. Their plots illustrated the
spectra for various damping ratios, indicating that the largest effect of damping is
obtained in the medium period range. They also evaluated inelastic strength demand
spectra for various target ductility ratios and an undamped elasto-plastic system subjected
to a half-cycle ground velocity pulse. The full-cycle velocity pulse was used in their
study to represent the Eureka ground motion, recorded in the California earthquake of
December 21, 1954. The results of their pulse study are in general agreement with the
response of SDOF systems to the pulses introduced in this chapter.

As discussed in the next section, triangular velocity pulses, whose acceleration time
histories are represented by square waves, are utilized primarily in this study. The effect
of rise time will be investigated later by studying triangular acceleration pulses.

5.2. Basic Pulse Shapes Used in this Study

The following three pulses (Figs. 5.1 to 5.3) are used as the basis for a representation of
the impulsive characteristics of near-fault ground motions. These pulses are fully defined
by a pulse shape and two parameters, i.e., the pulse period Tp and a pulse intensity

Chapter 5 Pulse-Type Seismic Input

69
parameter, which can be either the maximum pulse acceleration ag,max or the maximum
pulse velocity, vg,max.

Half-pulse, P1. In this pulse the ground experiences a non-reversing displacement history
that is generated through a single cycle of acceleration input, as shown in Fig. 5.1. In the
basic pulse P1 the acceleration input is represented by a single square wave, which results
in a triangular velocity half-cycle and a second-order one-directional displacement
history. The peak ground acceleration, PGA or ag,max, the peak ground velocity, PGV or
vg,max, and the peak ground displacement, PGD or ug,max, are related as follows:

a g ,max Tp
v g ,max = (5.1)
4

a g ,max Tp 2
u g ,max = (5.2)
16

Full pulse, P2. In this pulse the ground experiences a reversing displacement history that
is generated through a double cycle of acceleration input. In the basic pulse P2 the
acceleration input is represented by the square wave shown in Fig. 5.2, which results in a
triangular velocity cycle and a second-order reversing displacement history. The peak
ground velocity, vg,max, and the peak ground displacement, ug,max, are also given by Eqs.
5.1 and 5.2.

Multiple pulses, P3. This pulse sequence is generated by the acceleration history shown
in Fig. 5.3. It is utilized to investigate the effect of repeated pulses on response
parameters. The peak ground velocity, vg,max, is given by Eq. 5.1, but the peak ground
displacement, ug,max, is only half of that of the previous pulses, i.e.,

a g ,max Tp 2
u g ,max = (5.3)
32

In all three cases the pulse period Tp is defined as the duration of a full velocity cycle.
Thus, in P1 the duration of motion is only Tp/2. The effect of P1 and P2 on the response
of a continuous and elastic shear building was studied by Hall et al. (1995). In this study
the pulses will be used for a comprehensive evaluation of the elastic and inelastic
response of SDOF and MDOF systems.

Chapter 5 Pulse-Type Seismic Input

70
5.2.1. Elastic Response Spectra

Elastic strength (acceleration), velocity, and displacement demand spectra for pulses P1,
P2, and P3 are presented in Fig. 5.4. Each graph shows the spectra for all three pulses.
The structure period T is normalized by the pulse period Tp, and the spectral ordinates are
normalized by the corresponding ground motion peak value. All spectra are computed
for 2% damping. The following observations are made from the spectra:

•= The elastic strength demand (acceleration response) spectrum for P1 exhibits

closely spaced peaks and valleys for small T/Tp ratios and attains a maximum
dynamic amplification factor (spectral acceleration normalized by maximum
ground acceleration) of 3.75 at T/Tp = 0.5, i.e., at a period equal to the duration of
the one-directional half-pulse.

•= The spectral acceleration values for P2 are equal to those of P1 at periods for
which the maximum response occurs at times ≤ Tp/2, whereas they exceed the P1
values at all periods for which the maximum response occurs at times > Tp/2. The
maximum dynamic amplification factor is attained at T/Tp = 0.75 and is equal to
4.7.

•= The largest dynamic amplification factors are obtained for P3, and they occur at
periods for which maximum response is attained during the second velocity cycle.
The harmonic nature of P3 leads to very large dynamic amplification factors
around T/Tp = 1.0.

•= The damaging nature of P3 is evident also from the velocity and displacement
spectra. Assuming that for a given period the expected damage is proportional to
displacement, the presented displacement spectra provide a means to rank the
damage potential of the three pulses. When doing this, it must be considered that
the normalizing peak ground displacement ug,max is equal for P1 and P2, but is
only half as large for P3. Thus, if equal ground acceleration or velocity is used as
a basis for comparison of pulse effects, the normalized displacement spectral
values for P3 should be divided by a factor of 2.

Chapter 5 Pulse-Type Seismic Input

71
 •= The displacement spectra for P1 and P2 do not show a clear spectral peak. Only
for P3, which consists of two displacement cycles, the duration of motion is long
enough to generate high dynamic amplification around T/Tp = 1.0.

The general conclusion to be drawn from the pulse response spectra is that the structural
response is sensitive to the pulse shape and the relative value of structure to pulse period,
T/Tp. It should be noted, however, that the elastic spectra are inadequate to assess
damage potential and need to be supplemented by inelastic spectra and by the response
evaluation of MDOF systems in which multi-mode effects are present. This is the subject
of Chapter 6.

5.3. Other Pulse Shapes

A number of assumptions were made to define the basic pulses. In the three basic pulses
discussed previously, the acceleration history is described by square waves. Moreover,
each pulse consists of a certain number of velocity cycles, i.e., P1, P2, and P3 contain
one, two, and five velocity half-cycles, respectively. An important issue that needs to be
addressed is whether these three basic pulses can represent a reasonably large variety of
pulse-type ground motions, or whether more pulse shapes need to be considered.

The objective of this part of the study is to assess the response characteristics of SDOF
systems subjected to pulse inputs of different shapes and properties than the three basic
ones, and to evaluate the necessity of considering additional pulse shapes in the study.
For this purpose, elastic spectral quantities are evaluated for pulses with triangular
acceleration histories and pulses with different numbers of cycles from those of the three
basic pulses. The inelastic response of MDOF structure to these alternative pulse shapes
will be investigated in the next chapter.

5.3.1. Triangular Pulses

The use of acceleration square waves in the three basic pulses results in a very low
PGA/PGV ratio of 4/Tp, and assumes that the rise time of the acceleration pulse
approaches zero. Even for distinct near-fault pulses these are extreme and likely
unrepresentative conditions. Modified versions of basic pulses P1 and P2 are utilized to
study the effect of rise time on response parameters. These modified pulses, denoted as
P1.1 (a variation of P1), and P2.1 (a variation of P2) are illustrated in Figs. 5.5 and 5.6.

Chapter 5 Pulse-Type Seismic Input

72
If the peak parameters (PGA, PGV, and PGD) of the pulse variations are normalized by
the corresponding parameter of the basic pulses and the PGV ratio is set equal to 1.0, it is
observed that no difference exists in the peak ground displacements, but that in both
cases the PGA of the modified pulse is twice that of the basic pulse. Thus, the PGA/PGV
ratio is 8/Tp, which is in the range of values observed in “ordinary” records from past
earthquakes (Lawson, 1996) – if Tp is on the order of 1 sec. or shorter.

Elastic Response Spectra:

The effect of these pulse modifications on spectral response is documented in Figs. 5.7
and 5.8. The spectra of the modified pulses are shown in heavy lines and the spectra of
the corresponding basic pulses are shown in light lines. Each pulse response spectrum is
normalized by the peak value of the corresponding pulse input. These peak values are
shown in Figs. 5.1, 5.2, 5.5, and 5.6. Since the PGV and PGD values of both modified
pulses are identical to those of the basic pulses, and the normalized velocity and
displacement spectra do not differ by much, it is concluded that the pulse modifications
have a relatively small effect on spectral velocity and displacement responses. This
appears to be not so for the spectral acceleration responses. The normalized spectral
accelerations (dynamic amplification factors) for the modified pulses are much smaller
than those for the basic pulses. However, the PGA values of the modified pulses are
twice as large as those of the basic pulses. Once this is considered, it is observed that the
difference in spectral accelerations between the modified and basic pulses is not drastic at
most periods.

It is concluded that the spectral response of the modified pulses with triangular
acceleration histories can be adequately represented by their corresponding basic pulses.
The effect of these triangular pulses on the response of MDOF structures is studied in the
next chapter.

5.3.2. Pulse Histories with Different Duration

The period of the basic pulses (Tp) is defined as the time needed to complete a full
velocity cycle. Based on this definition, the duration of basic pulse histories P1, P2, and
P3 is 0.5, 1.0, and 2.5 times the pulse period, respectively. There are pulse histories with
different duration whose response properties may need to be evaluated. Two additional

Chapter 5 Pulse-Type Seismic Input

73
pulses, P4 and P5, are introduced and studied here for this purpose. The acceleration
time histories for P4 and P5 are described by the square waves shown in Figs. 5.9 and
5.10. The duration of these two pulse histories is 1.5Tp for P4 and 2.0Tp for P5. It should
be noted that the duration of these new pulses is between those of P2 and P3. Thus, it is
useful to compare the response characteristics of P4 and P5 with those of basic pulses P2
and P3.

Elastic Response Spectra:

The elastic strength (acceleration), velocity, and displacement demand spectra of all five
pulses P1 to P5 are compared in Fig. 5.11. The spectral ordinates are normalized by the
corresponding time history peak values. The following observations can be made from
the spectra:

•= The spectra of P4 and P5 exhibit patterns similar to those of P3 but with smaller
peak values.

•= The normalized spectral displacements for P4 and P5 are bounded by the

corresponding values for P2 and P3. However, this does not hold true at all
periods for the normalized spectral acceleration and velocity values.

Since the spectral responses for P4 and P5 are not always close to those for basic pulses
P1, P2, and P3 in the full range of T/Tp, more investigation is necessary to quantify the
MDOF response characteristics of pulses P4 and P5. The inelastic demands of MDOF
structures subjected to pulse shapes P4 and P5 are evaluated in the next chapter.

Chapter 5 Pulse-Type Seismic Input

74
 Ground Acceleration Time History
Pulse P1
1.5

0.5

ag / ag,max
0

-0.5

-1

-1.5
0 0.5 1 1.5 2 2.5 3 3.5
t / Tp

Ground Velocity Time History

Pulse P1
1.5
vg / vg,max = vg / (ag,[Link] / 4)

0.5

-0.5

-1

-1.5
0 0.5 1 1.5 2 2.5 3 3.5
t / Tp

Ground Displacement Time History

Pulse P1
1.5
ug / ug,max = ug / (ag,[Link] / 16)

1
2

0.5

-0.5

-1

-1.5
0 0.5 1 1.5 2 2.5 3 3.5
t / Tp

Figure 5.1 Pulse P1 Ground Acceleration, Velocity, and Displacement Time Histories

Chapter 5 Pulse-Type Seismic Input

75
 Ground Acceleration Time History
Pulse P2
1.5

0.5
ag / ag,max
0

-0.5

-1

-1.5
0 0.5 1 1.5 2 2.5 3 3.5
t / Tp

Ground Velocity Time History

Pulse P2
1.5
vg / vg,max = vg / (ag,[Link] / 4)

0.5

-0.5

-1

-1.5
0 0.5 1 1.5 2 2.5 3 3.5
t / Tp

Ground Displacement Time History

Pulse P2
1.5
ug / ug,max = ug / (ag,[Link] / 16)

1
2

0.5

-0.5

-1

-1.5
0 0.5 1 1.5 2 2.5 3 3.5
t / Tp

Figure 5.2 Pulse P2 Ground Acceleration, Velocity, and Displacement Time Histories

Chapter 5 Pulse-Type Seismic Input

76
 Ground Acceleration Time History
Pulse P3
1.5

0.5
ag / ag,max
0

-0.5

-1

-1.5
0 0.5 1 1.5 2 2.5 3 3.5
t / Tp

Ground Velocity Time History

Pulse P3
1.5
vg / vg,max = vg / (ag,[Link] / 4)

0.5

-0.5

-1

-1.5
0 0.5 1 1.5 2 2.5 3 3.5
t / Tp

Ground Displacement Time History

Pulse P3
1.5
ug / ug,max = ug / (ag,[Link] / 32)

1
2

0.5

-0.5

-1

-1.5
0 0.5 1 1.5 2 2.5 3 3.5
t / Tp

Figure 5.3 Pulse P3 Ground Acceleration, Velocity, and Displacement Time Histories

Chapter 5 Pulse-Type Seismic Input

77
 Elastic SDOF Strength Demands
Pulses P1, P2, P3, ξ = 2%
8
P1
P2
P3
6

Sa / ag,max 4

0
0 0.5 1 1.5 2 2.5 3
T / Tp

Elastic SDOF Velocity Demands

Pulses P1, P2, P3, ξ = 2%
5
P1
P2
4 P3
Sv / vg,max

0
0 0.5 1 1.5 2 2.5 3
T / Tp

Elastic SDOF Displacement Demands

Pulses P1, P2, P3, ξ = 2%
6
P1
P2
5 P3

4
Sd / ug,max

0
0 0.5 1 1.5 2 2.5 3
T / Tp

Figure 5.4 Elastic Strength (Acceleration), Velocity, and Displacement Demand Spectra
for P1, P2, and P3

Chapter 5 Pulse-Type Seismic Input

78
 Ground Acceleration Time History
Pulses P1 and P1.1
3
P1.1
2 P1

ag / ag,max [P1]
1

-1

-2

-3
0 0.25 0.5 0.75 1 1.25 1.5 1.75 2
t / Tp

Ground Velocity Time History

Pulses P1 and P1.1
1.5
P1.1
1 P1
vg / vg,max [P1]

0.5

-0.5

-1

-1.5
0 0.25 0.5 0.75 1 1.25 1.5 1.75 2
t / Tp

Ground Displacement Time History

Pulses P1 and P1.1
1.5

1
ug / ug,max [P1]

0.5

-0.5

-1 P1.1
P1
-1.5
0 0.25 0.5 0.75 1 1.25 1.5 1.75 2
t / Tp

Figure 5.5 Pulse P1.1 (Modification to Pulse P1)

Chapter 5 Pulse-Type Seismic Input

79
 Ground Acceleration Time History
Pulses P2 and P2.1
3
P2.1
2 P2

ag / ag,max [P2]
1

-1

-2

-3
0 0.25 0.5 0.75 1 1.25 1.5 1.75 2
t / Tp

Ground Velocity Time History

Pulses P2 and P2.1
1.5
P2.1
1 P2
vg / vg,max [P2]

0.5

-0.5

-1

-1.5
0 0.25 0.5 0.75 1 1.25 1.5 1.75 2
t / Tp

Ground Displacement Time History

Pulses P2 and P2.1
1.5

1
ug / ug,max [P2]

0.5

-0.5

-1 P2.1
P2
-1.5
0 0.25 0.5 0.75 1 1.25 1.5 1.75 2
t / Tp

Figure 5.6 Pulse P2.1 (Modification to Pulse P2)

Chapter 5 Pulse-Type Seismic Input

80
 Elastic SDOF Strength Demands
Pulses P1 and P1.1, ξ = 2%
4
P1.1
P1

Sa / ag,max 2

0
0 0.25 0.5 0.75 1 1.25 1.5 1.75 2
T / Tp

Elastic SDOF Velocity Demands

Pulses P1 and P1.1, ξ = 2%
1.6
P1.1
P1

1.2
Sv / vg,max

0.8

0.4

0
0 0.25 0.5 0.75 1 1.25 1.5 1.75 2
T / Tp

Elastic SDOF Displacement Demands

Pulses P1 and P1.1, ξ = 2%
1

0.8
Sd / ug,max

0.6

0.4

0.2
P1.1
P1
0
0 0.25 0.5 0.75 1 1.25 1.5 1.75 2
T / Tp

Figure 5.7 Comparison of Elastic Strength (Acceleration), Velocity, and Displacement

Demands for P1 and P1.1

Chapter 5 Pulse-Type Seismic Input

81
 Elastic SDOF Strength Demands
Pulses P2 and P2.1, ξ = 2%
5
P2.1
P2
4

Sa / ag,max
3

0
0 0.25 0.5 0.75 1 1.25 1.5 1.75 2
T / Tp

Elastic SDOF Velocity Demands

Pulses P2 and P2.1, ξ = 2%
3
P2.1

2.5 P2

2
Sv / vg,max

1.5

0.5

0
0 0.25 0.5 0.75 1 1.25 1.5 1.75 2
T / Tp

Elastic SDOF Displacement Demands

Pulses P2 and P2.1, ξ = 2%
1.6

1.2
Sd / ug,max

0.8

0.4
P2.1
P2
0
0 0.25 0.5 0.75 1 1.25 1.5 1.75 2
T / Tp

Figure 5.8 Comparison of Elastic Strength (Acceleration), Velocity, and Displacement

Demands for P2 and P2.1

Chapter 5 Pulse-Type Seismic Input

82
 Ground Acceleration Time History
Pulse P4
1.5

0.5
ag / ag,max
0

-0.5

-1

-1.5
0 0.5 1 1.5 2 2.5 3 3.5
t / Tp

Ground Velocity Time History

Pulse P4
1.5
vg / vg,max = vg / (ag,[Link] / 4)

0.5

-0.5

-1

-1.5
0 0.5 1 1.5 2 2.5 3 3.5
t / Tp

Ground Displacement Time History

Pulse P4
1.5
ug / ug,max = ug / (ag,[Link] / 32)

1
2

0.5

-0.5

-1

-1.5
0 0.5 1 1.5 2 2.5 3 3.5
t / Tp

Figure 5.9 Pulse P4 Ground Acceleration, Velocity, and Displacement Time Histories

Chapter 5 Pulse-Type Seismic Input

83
 Ground Acceleration Time History
Pulse P5
1.5

0.5
ag / ag,max
0

-0.5

-1

-1.5
0 0.5 1 1.5 2 2.5 3 3.5
t / Tp

Ground Velocity Time History

Pulse P5
1.5
vg / vg,max = vg / (ag,[Link] / 4)

0.5

-0.5

-1

-1.5
0 0.5 1 1.5 2 2.5 3 3.5
t / Tp

Ground Displacement Time History

Pulse P5
1.5
ug / ug,max = ug / (ag,[Link] / 32)

1
2

0.5

-0.5

-1

-1.5
0 0.5 1 1.5 2 2.5 3 3.5
t / Tp

Figure 5.10 Pulse P5 Ground Acceleration, Velocity, and Displacement Time Histories

Chapter 5 Pulse-Type Seismic Input

84
 Elastic SDOF Strength Demands
Different Pulses, ξ = 2%
8
P1
P2
P3
6
P4
P5

Sa / ag,max
4

0
0 0.25 0.5 0.75 1 1.25 1.5 1.75 2
T / Tp

Elastic SDOF Velocity Demands

Different Pulses, ξ = 2%
5
P1
P2
4 P3
P4
P5
3
Sv / vg,max

0
0 0.25 0.5 0.75 1 1.25 1.5 1.75 2
T / Tp

Elastic SDOF Displacement Demands

Different Pulses, ξ = 2%
6
P1
5 P2
P3
P4
4 P5
Sd / ug,max

0
0 0.25 0.5 0.75 1 1.25 1.5 1.75 2
T / Tp

Figure 5.11 Elastic Strength, Velocity, and Displacement Demand Spectra for P1 to P5

Chapter 5 Pulse-Type Seismic Input

85
 CHAPTER 6

RESPONSE OF STRUCTURES TO PULSE-TYPE

SEISMIC INPUT

This Chapter is devoted to evaluating the response of elastic and inelastic structures to
pulse-type seismic input, and achieving a fundamental understanding of response
characteristics that can be utilized later (in Chapter 7) in the representation of near-fault
ground motions by equivalent pulses. The SDOF and MDOF structures introduced in
Chapter 3 and the pulse-type ground motions defined in Chapter 5 are employed in the
pulse response evaluations.

First, the elastic response of MDOF frame structures to the pulse-type input motions is
investigated. Elastic base shear demands and the distribution of story shear demands
over the height of the structure are evaluated. The goal is to identify the response
characteristics and patterns that near-fault ground motions and simple pulses have in
common. Then, comprehensive response studies are carried out on inelastic SDOF
systems and MDOF frame structures subjected to the basic pulses. Story ductility
demands are utilized to evaluate the performance of inelastic MDOF systems. Inelastic
roof displacement and story drift demands are also quantified. Generic 3- and 9-story
frame structures are introduced and utilized to assess the sensitivity of the response of
MDOF structures to the number of stories. Finally, a pilot study is carried out whose
objective is to quantify the response to pulse-type ground motions of frame structures in
which plastic hinges can develop only at the end of columns.

6.1. Elastic Response of MDOF Structures to Pulse-Type Input

The basic assumptions and procedures employed in the design of the generic 20-story
structure were introduced in Section 3.2. In the pulse study, MDOF dynamic analyses

Chapter 6 Response of Structures to Pulse-Type …

86
are performed primarily for T/Tp = 0.375, 0.50, 0.75, 1.0, 1.5, 2.0, and 3.0, where T is the
structure fundamental period and Tp represents the pulse period defined in Section 5.2.

6.1.1. Deflected Shapes of Structure

Snapshots of the deflected shapes of the generic MDOF structures subjected to the basic
pulses at intervals of Tp/4 are shown in Figs. 6.1 to 6.4. The deflected shapes serve to
illustrate the effect of pulse loading on the structural response at specific time steps.
Figure 6.1 corresponds to a structure with T = 0.5Tp subjected to P1. It can be seen that
for this short period structure the deflected shape is similar to a first mode shape,
signifying that higher-mode effects are insignificant and the dynamic behavior is
controlled mostly by the first mode. The maximum roof displacement value of 0.55ug,max
occurs at t = Tp/2, i.e., the beginning of the free vibration phase.

Figures 6.2 to 6.4 illustrate deflected shapes for a structure with T = 2Tp subjected to all
three basic pulses. The following observations are made:

•= For all pulses the deflected shape, which is far from a straight line, clearly shows
higher mode contributions and the effect of a transient wave traveling up the
structure.

•= The deflected shapes for P1 and P2 are identical until Tp/2, at which time the
structure subjected to P1 enters the free vibration phase. The roof displacement at
this instance is very close to the ground displacement, which is at its maximum.
Under P1 the structure reaches its maximum roof displacement of 1.3ug,max in the
free vibration phase at t = Tp because enough energy is stored in the structure
during the half-pulse input to increase the roof displacement by 30% during the
first free vibration reversal.

•= The structure subjected to P2 also reaches its maximum roof displacement at t =

Tp, but because of the reversal of ground displacement, the maximum roof
displacement of 2.3ug,max is significantly larger than the corresponding value for
P1.

•= The largest roof displacement of 3.9ug,max is observed for the structure subjected
to P3. It occurs at t = 1.5Tp, the time of the peak of the second ground

Chapter 6 Response of Structures to Pulse-Type …

87
 displacement cycle. Thus, the harmonic nature of P3 has a larger effect on the
roof displacement compared to P2, which has only one displacement pulse.

6.1.2. Maximum Elastic Base Shear Force

Today's design procedures are based on specification of a base shear and a distribution of
story shear forces over the height of the structure. The base shear is usually evaluated
from the first mode spectral strength demand, with modifications applied to account for
higher-mode effects. Thus, an evaluation of base shear demands and their relation to the
first mode spectral values contributes to the understanding of pulse response.

Elastic base shear strength demands for the generic structures with various periods
subjected to all three basic pulses are presented in Fig. 6.5. The base shear demands are
normalized by [Link],max, so that the results are comparable with the SDOF strength
demand spectra presented in Fig. 5.4. This comparison is made in Fig. 6.6, which
illustrates ratios of MDOF base shear to SDOF strength demands at the first mode period.
The following observations are made:

•= The ratios are sensitive to the pulse type and T/Tp, and may easily exceed 1.0 for
larger T/Tp ratios, signifying large higher-mode effects.

•= The variation of the elastic base shear demands for each pulse typically follows
the same pattern as its corresponding SDOF strength demand spectrum. The base
shear demand for P1 and P2 is highest at relatively small T/Tp ratios.

•= The maximum base shear demand for P3 occurs at T/Tp = 1.0, which can be
attributed to the harmonic nature of this pulse. A similar observation has been
made by Rahnama and Krawinkler (1993) in their study on response to soft soil
motions, which have characteristics similar to those of P3.

Study on Continuous Shear Building with Uniform Properties:

Hall et al. (1995) studied the effects of pulses P1 and P2 on an undamped continuous
shear building with uniform shear stiffness over the height, based on elastic wave
propagation theory. They presented their results for specific T/Tp values in terms of
shear strain at the base of the building. In this study their approach is generalized to

Chapter 6 Response of Structures to Pulse-Type …

88
include also pulse P3, and the maximum base shear is calculated for the range of T/Tp
from 0 to 3.0. The results for base shear demands obtained from a closed form solution
are presented in Fig. 6.7, along with the time history of the base shear force for P2 and
T/Tp = 0.75 in Fig. 6.8.

A relatively consistent correlation is observed between the demands for the continuous
shear building and those for the generic structure. The elastic base shear demands for the
continuous shear building are typically larger than the corresponding demands for the
generic 20-story structure. In part, this can be attributed to the fact that no damping is
considered in the shear building model. However, it is believed that the effect of
damping is not large (see Section 6.2.2), and that the difference in the elastic base shear
demands has to do with the fact that unlike the generic structure, the continuous shear
building has constant stiffness over the height.

6.1.3. Distribution of Elastic Story Shear Over Height

Figure 6.9 compares the SRSS story shear distribution used in the design of the structures
(denoted as “Design”), with elastic story shear distributions obtained from (a) time
history analyses and (b) SRSS modal combinations using the corresponding pulse
spectrum for pulses P1 to P3 and various T/Tp ratios. The following observations are
made:

•= For structures with T/Tp ≤ 1.0, the dynamic shear force distributions are relatively
smooth, but for structures with T/Tp > 1.0 the distributions show a clear effect of a
wave traveling up the structure. This effect is evident for P1, more evident for P2,
and most evident for P3. In the last case the maximum elastic story shear demand
for T/Tp = 2.0 occurs about 2/3rd up the height of the structure rather than at the
base. Thus, for long period structures (T/Tp > 1.0) designed according to a
standard SRSS story shear strength distribution, early yielding has to be expected
in the upper stories.

•= For MDOF structures with T/Tp > 1.0, the SRSS modal combination is not a good
substitute for the dynamic time history analysis. This observation provides an
indication that spectral analysis may not capture all important response
characteristics of pulse-type ground motions, once higher-mode effects become
important.

Chapter 6 Response of Structures to Pulse-Type …

89
These results confirm observations made in Section 4.1.2 for near-fault ground motions,
and indicate that for pulse-type ground motions and structures whose fundamental period
is larger than the effective pulse period, the design story shear strength distribution over
the height may need to be modified compared to presently employed design patterns
(triangular, parabolic, or SRSS). However, it should be considered that this conclusion
applies only to elastic or nearly elastic structures, and that the distributions of demands
over the height of the structure may change significantly at lower performance levels
(highly inelastic systems). The issue of design story shear strength distributions is
pursued in Chapter 9.

6.1.4. Maximum Elastic Roof Displacement

Seneviratna and Krawinkler (1997) have performed statistical correlation studies between
the roof displacement of elastic MDOF structures and the spectral displacement of the
first mode SDOF systems. Their conclusion was that – for the ordinary ground motions
used in their study – there is a strong correlation with small scatter between these two
quantities. For all periods, and for frames as well as wall systems, the mean roof
displacement is very close to and usually slightly larger than the spectral displacement
multiplied by the first mode participation factor, PF1. This means that roof displacement
is dominated by first mode vibrations.

It turns out that the same conclusion cannot be drawn for structures subjected to pulse-
type excitations. This is illustrated in Fig. 6.10, which shows plots of the ratio of elastic
MDOF roof displacement to the first mode spectral displacement, δroof,max/Sd, for various
period ratios T/Tp and pulses. Each graph shows two curves, one for MDOF systems in
which P-delta effects are neglected, and one in which they are accounted for. The ratio is
close to the value of PF1 (1.37) for T/Tp ≤ 1.0, and consistently exceeds this value by a
significant amount for large T/Tp ratios. These observations are in agreement with the
results of the study on near-fault records summarized in Section 4.1.3. This demonstrates
that for pulse-type and near-fault input motions and large T/Tp ratios, higher-mode effects
play a larger role than for ordinary ground motions.

6.2. Ductility Demands for Inelastic Structures

The properties of the SDOF and MDOF structures used in the inelastic response
evaluations were summarized in Chapter 3. In the pulse study a strength parameter η is

Chapter 6 Response of Structures to Pulse-Type …

90
utilized to define yield strength values. For MDOF systems the base shear strength
coefficient η is defined as:

Vy
η= (6.1)
m.a g , max

where Vy = base shear yield strength of the MDOF system

m = total mass of the system
ag,max = maximum ground acceleration of the input pulse

The yield strength of SDOF systems is defined using the same coefficient, η, but
substituting the SDOF yield strength, Fy, for the base shear strength, Vy, in Eq. 6.1.

Results obtained from nonlinear time history analyses are presented mostly as η−µ
(strength-ductility) diagrams, and for MDOF systems with a given strength parameter η,
by means of plots that show the distribution of story ductility demands over the height of
the structure.

6.2.1. SDOF Systems

Displacement Time History:

Normalized displacement response time histories for SDOF systems subjected to typical
near-fault ground motions were presented in Fig. 4.5. It was shown that the near-fault
response has clear pulse-type characteristics. It is expected that pulse-type ground
motions will cause a response that is characterized by either a single large excursion in
one direction or by a single large cycle with comparable positive and negative excursions.
The question is how the period ratio T/Tp and pulse type affect this response behavior.
Typical normalized displacement response time histories for inelastic SDOF systems (for
a pre-defined maximum ductility of 8) subjected to the three basic pulses are illustrated in
Fig. 6.11 for periods of T/Tp = 0.5 and 1.0. The responses differ significantly between
the two selected periods and among the three pulse types. The following observations
can be made from the time histories:

Chapter 6 Response of Structures to Pulse-Type …

91
 •= For systems with T/Tp = 0.5, pulse P1 causes a one-sided response with a full
displacement reversal, which results in a very small residual displacement. Much
of the inelastic displacement reversal occurs during the free vibration phase after
the time Tp/2. The response to pulse P2 reaches a maximum during the second
half of the pulse, in the direction opposite to the maximum response to P1, and a
significant residual displacement is observed. The response to pulse P3 exhibits
two cycles with large inelastic displacements, but very little residual
displacement.

•= For systems with T/Tp = 1.0, the response to all three pulses results in significant
residual displacements. Pulses P1 and P2 cause only one cycle of large inelastic
response, whereas P3 generates two large inelastic cycles.

Constant Ductility Strength Demand Spectra:

Inelastic strength demand spectra for the three basic pulses are shown in Fig. 6.12. The
spectra are presented for targeted ductility ratios of µ = 1 (elastic), 2, 3, 4, 6, and 8. The
spectral ordinates are defined in terms of the strength coefficient η defined previously.
As discussed in Section 3.1, the results are computed for non-degrading bilinear
hysteretic systems with a strain-hardening ratio of 3%.

The strength demand spectra follow expected patterns insofar that the inelastic spectra for
larger µ values become much smoother than the elastic ones. This implies that the effects
of the large peaks and valleys in the short period range are smoothed because of the
period elongation of the inelastic systems. These patterns are observed in the inelastic
strength demand spectra of near-fault as well as ordinary ground motions.

A sensitivity analysis, similar to the one discussed in Section 4.2.1 for typical near-fault
ground motions, is performed to evaluate the effects of the hysteresis model on inelastic
strength demands for pulse-type ground motions. The inelastic SDOF strength demand
spectra are computed using the modified Clough model (see Rahnama and Krawinkler,
1993, for hysteresis rules), and are compared with the demands obtained from the bilinear
model shown in Fig. 6.12. Figure 6.13 illustrates the ratio of the strength demands
obtained from the modified Clough model to the corresponding demands obtained from
the bilinear model for various target ductility and period ratios, and pulses P2 and P3.
The following observations can be made:

Chapter 6 Response of Structures to Pulse-Type …

92
 •= The ratio for pulse P2 is ≥ 1.0 at all T/Tp values, but it does not exceed 1.3. This
indicates that for pulse P2 the strength demands are not very sensitive to the
hysteresis model. The sensitivity is larger at short periods, and diminishes to zero
for T/Tp ratios larger than a target ductility dependent value. For large T/Tp
values the maximum ductility is attained in the first inelastic excursion, in which
case the response is identical for the bilinear and Clough hysteresis models.

•= The sensitivity to the hysteresis model is larger for pulse P3 and the period range
of T/Tp < 1.0, in which the strength demand ratio reaches 1.5.

These observations typically agree with those made for near-fault ground motions, and
emphasize that for input motions with multi-pulse characteristics, the sensitivity of
inelastic SDOF strength demands to the hysteresis model may not be negligible, while for
motions of the type represented by a single pulse, the sensitivity is limited.

Ductility Demands for Specific Periods:

Figures 6.14 to 6.16 present η−µ=(strength vs. ductility demand) diagrams for selected
values of T/Tp = 0.375, 0.50, 0.75, 1.0, 1.5, 2.0, and 3.0 for pulses P1 to P3. These
graphs illustrate the ductility demand for an SDOF system as a function of its yield
strength, for given T/Tp values. The results exhibit a similar pattern for all three pulses,
and indicate a relationship of the type µ ∝ 1/η, which corresponds to a linear increase in
ductility with the inverse of strength.

A clearer picture can be obtained from R-µ diagrams, R being the strength reduction
factor defined as R = Fy,e/Fy(µ), where Fy,e is the elastic strength demand and Fy(µ) is the
inelastic strength demand corresponding to the ductility µ. Figure 6.17 presents the R-µ
diagrams for pulse P2 and various T/Tp ratios. The following observations are made:

•= The relationship between R and µ is linear in many cases, particularly in the large
ductility range.

•= For short period structures (T/Tp < 0.75) µ is larger than R, which indicates that
the inelastic displacement is larger than the elastic displacement (δin/δel = µ/R).
At T/Tp = 0.75, R is almost equal to µ for all ductility values.

Chapter 6 Response of Structures to Pulse-Type …

93
 •= For structures with 0.75 < T/Tp < 3.0, R becomes larger than µ, indicating that the
elastic displacement is larger than the inelastic displacement. For very long
period structures (T/Tp = 3.0), R approaches µ for all ductility values. The reason
is that at very long periods both elastic and inelastic displacements approach the
maximum ground displacements regardless of the structure strength.

The patterns observed previously for R-µ-T relationships can be seen also from δin/δel
ratios (see Section 6.3.1). A one-to-one comparison of the ductility demands for the three
basic pulses is illustrated in Fig. 6.18, which shows η−µ and R−µ graphs for T/Tp = 0.5
and 2.0. The following observations are made:

•= For given strength (η value) the results show significant pulse-type dependence at
T/Tp = 0.5, whereas the demands at T/Tp = 2.0 are less dependent on pulse type.
A general comparison of the results for different pulses indicates that for a given
strength level and T/Tp ratio, P2 always causes a larger or at least equal ductility
demand compared to P1. The reason is that pulses P1 and P2 have identical
ground time histories up to t = Tp/2, and P1 comes to rest thereafter.

•= At T/Tp = 0.5 and for a given ductility ratio, R is largest for P1 and smallest for
P3. At T/Tp = 2.0 pulse P2 corresponds to the largest R value regardless of µ, but
the pulse type corresponding to the smallest R value depends on the ductility
ratio. The pulse-type dependence of the R-µ relationship is significant for short
period structures, and becomes less so for long period structures.

6.2.2. MDOF Systems

Story Ductility Demands Over Height:

As shown in Section 4.2.2, the inelastic response of MDOF systems to near-fault ground
motions with forward directivity has special characteristics, which set these motions apart
from ordinary ground motions. The results presented in Section 6.1 for elastic MDOF
structures also provide evidence that the response of structures with T > Tp to pulse-type
ground motions is characterized by a traveling wave effect. The objective of this section
is to evaluate the inelastic response of MDOF structures to the basic pulse inputs, and to
identify similarities between the MDOF response to pulse-type and near-fault ground

Chapter 6 Response of Structures to Pulse-Type …

94
motions. Inelastic response is described here by the story ductility ratio defined as µi =
δmax,i/δy,i, where δmax,i is the story drift demand, and δy,i denotes the story yield drift.

Distributions of story ductility demands over the height of the structure, neglecting P-
delta effects, are illustrated in Fig. 6.19 for pulses P1 to P3 and T/Tp values of 0.5 and
2.0. Each graph clearly shows the variation of the ductility distribution with the structure
strength coefficient, η. Ductility ratios less than unity imply that the story shear force
demand is smaller than the provided shear strength in the story. The following
observations can be made for long period structures (Fig. 6.19(b)):

•= Since the strength of the MDOF systems is tuned to a story shear strength
distribution corresponding to the SRSS modal combination for a 1/T-type
acceleration spectrum (“design” in Fig. 6.9), the dynamic story shear force
distributions shown in Fig. 6.9(a) indicate that for T/Tp > 1.0 yielding will start in
upper stories. Figure 6.19(b) confirms that the ductility demands are highest in
upper stories for relatively strong structures (large η values).

•= As the structure strength is reduced, the maximum ductility demands in the top
portion of the structure seem to stabilize, whereas in the lower stories the ductility
demands grow rapidly.

•= For weak structures (small η values) a clear migration occurs of maximum

ductility demands to the bottom of the structure.

The consistency of this phenomenon is one of the fundamental MDOF response

characteristics of pulse-type and near-fault ground motions with forward directivity. It
has to do with the traveling wave effect that occurs primarily for T/Tp > 1.0. At T/Tp =
0.5 this phenomenon does not occur (see Fig. 6.19(a)). For the latter case the maximum
ductility demands are at or near the base also for strong structures.

Figure 6.20 compares the variation of story ductility demands with the inverse of η for
the first and 15th stories, using P2 and T/Tp = 0.5 and 2.0. The inverse of η has a linear
relationship with ag,max, which is a measure of the intensity of the ground motion, i.e., 1/η
= (W/Fy)ag,max/g. This equation shows that weakening the structure is equivalent to
intensifying the ground motion. The following observations are made:

Chapter 6 Response of Structures to Pulse-Type …

95
 •= For a structure with a fundamental period twice the pulse period, the maximum
ductility demands occur in the top portion of the structure when the input motion
is not severe (or the structure is strong). However, as the intensity of the input
pulse increases (or the structure becomes weaker) the ductility demand in this
portion stabilizes around 3.0 to 4.0 and even becomes smaller, whereas the first
story ductility demand continues to grow rapidly, so that beyond 1/η = 4.0 the
ductility demands at the bottom are larger (Fig. 6.20(b)).

•= The distribution of story ductility demands for strong structures with T/Tp = 0.5 is
more uniform than that for T/Tp = 2.0, and the early yielding of the upper stories
does not occur in this case (Fig. 6.20(a)).

•= Figure 6.20(b) demonstrates that the stabilization of the ductility demands in the
upper stories is not necessarily permanent; at very low strength levels the ductility
in the 15th story starts to grow again.

•= The slope of the 1/η−µ line for the first story and µ > 3 is constant and smaller
than 1.0, which indicates that the rate of increase in ductility is larger than the rate
of decrease in strength.

The statement made about stabilization of story ductility demands has to be put in
perspective. The results presented here are obtained for structures whose relative story
shear strengths follow the SRSS shear force distribution for a 1/T-type design spectrum.
However, a strength design according to this shear force distribution may not be the best
choice. Desirable story shear strength distributions are discussed in Chapter 9. Also, the
results presented here for pulse-type inputs are relevant only if actual near-fault ground
motions can be represented by equivalent pulses. This issue is the main focus of Chapter
7.

Maximum Story Ductility Demands:

A comprehensive picture of the maximum story ductility demand for all stories can be
obtained from the η−µmax diagrams illustrated for various T/Tp ratios in Fig. 6.21. These
plots are presented in the same manner as the SDOF demands shown in Figs. 6.14 to
6.16. The results are presented for MDOF structures without P-delta and with P-delta
effects.

Chapter 6 Response of Structures to Pulse-Type …

96
Many curves, particularly those for T/Tp > 1.0, have a close to vertical range around a
ductility of 3 to 4, which is the range of migration of the maximum ductility demand
from upper stories to the bottom story. For instance, for pulse P2 and T/Tp = 1.5 to 3.0,
there is a range in which η (base shear strength) can be reduced to half without leading to
an increase in the maximum ductility demands. A clearer picture of this phenomenon
may be obtained from Fig. 6.22, which shows 1/η−µmax plots for P2, using systems
without and with P-delta effects. The type of analysis this diagram represents is called
Incremental Dynamic Analysis (IDA). The range of stabilization of ductility demands is
clearly evident, as is the linear relationship between 1/η and µmax once the maximum
ductility demand has migrated to the first story. Again, the slope of these lines is less
than unity for T/Tp between 1.5 and 3.0.

P-delta effects can be assessed by comparing part (a) and part (b) of Fig. 6.21. The
results indicate a consistent pattern for T/Tp > 1.0, as follows:

•= As long as the MDOF system is strong enough to prevent migration of maximum

ductility demands to the first story, the inclusion of P-delta effects does not make
a significant difference.

•= For weaker structures, in which the maximum ductility demand occurs in the first
story, the effect of P-delta suddenly gains much on importance and may lead to
significant amplification of maximum ductility demands.

This provides a strong argument for making structures sufficiently strong to prevent
migration of maximum ductility demands to the first story. Figure 6.23 permits a direct
assessment of P-delta effects for P2 and various T/Tp values. It is noted that the P-delta
effect is by far largest for T/Tp = 0.50, for which maximum ductility demands are always
largest in the lower stories.

The effect of pulse type on the maximum ductility demand is illustrated in Fig. 6.24, in
which the demands for the three basic pulses are compared for T/Tp = 0.5 and 2.0. For
T/Tp = 2.0 there are considerable differences in the demands for the range of strength in
which the maximum story ductilities are in the upper stories, but the differences become
small once the maximum demands have migrated to the bottom of the structure.

Chapter 6 Response of Structures to Pulse-Type …

97
Base Shear Strength Demands for Target Ductility:

The η−µmax diagrams presented earlier provide comprehensive information on the

maximum ductility demand for MDOF structures subjected to pulse-type ground
motions. However, in design it is more useful to rearrange this information in order to
evaluate the base shear strength, η, required to limit the maximum story ductility
demand, µmax, to specific target values. Vertical cuts through the η−µmax diagrams shown
in Fig. 6.21(a), and using a linear interpolation scheme, provide values for the MDOF
strength demands. This representation is analogous to the SDOF strength demand spectra
for constant ductility, presented in Fig. 6.12. Figure 6.25 presents the MDOF base shear
strength demand spectra of the basic pulses for target maximum story ductility ratios of
µmax = 1, 2, 3, 4, 6, and 8. Inherent in these spectra are the assumptions that the structure
has 20 stories and the story shear strength distribution over the height follows the SRSS
story shear force distribution. The MDOF spectra for 3- and 9-story structures will be
investigated in Section 6.5. These spectra are very useful for near-fault design because
they provide the required strength for various targeted ductility levels, provided that the
basic pulses introduced in this study can represent near-fault ground motions. The
MDOF strength demand spectra are utilized here to study the following effects on the
response of MDOF structures:

Higher Mode Effects. In design it is often attempted to use SDOF strength and
displacement demands to deduce corresponding MDOF demands. If the ordinates of the
MDOF strength demand spectra, presented in Fig. 6.25, are divided by the corresponding
values of the SDOF strength demand spectra, presented in Fig. 6.12, for the same
ductility and period values, the ratio will quantify the MDOF/SDOF strength demand
relationships for the pulse-type ground motions. These strength demand ratios are
illustrated in Fig. 6.26 for the three basic pulses. This figure, which provides a
comprehensive comparison of MDOF and SDOF strength demands, verifies many past
observations such as:

•= In the short period range (T/Tp ≤ 1.0), the strength ratio does not differ much from
1.0 for all target ductility ratios and input pulses, which indicates that in short
period structures the effect of higher modes is insignificant. Furthermore, the
migration of maximum ductilities from the upper stories to the bottom does not
take place, and therefore the responses of MDOF and SDOF structures to the
pulse-type motions are comparable.

Chapter 6 Response of Structures to Pulse-Type …

98
 •= In the long period range (T/Tp > 1.0), the strength ratio can grow rapidly and
reach large values, implying that SDOF systems may greatly underestimate the
base shear strength demand for MDOF structures. For a structures with a
fundamental period of T = 3Tp subjected to P2 or P3, the base shear strength
required to limit the maximum ductility to a specific low value can be as high as
five times the required strength for the corresponding first mode SDOF system.
However, the MDOF/SDOF strength ratio for the same structure subjected to P1
does not exceed 2.8. At large T/Tp values the ratio is smaller for highly inelastic
structures (large µ) than for close to elastic structures (small µ).

P-Delta Effects. The MDOF base shear strength demand spectra can also be utilized to
quantify P-delta effects. The spectra illustrated in Fig. 6.25 are obtained from the η−µmax
diagrams that neglect P-delta effects. Similar spectra can be produced from the η−µmax
diagrams in which P-delta effects are included (Fig. 6.21(b)). Then, the ratios of the
MDOF spectral values with P-delta effects to the corresponding spectral values without
P-delta effects can be used to evaluate the effects of P-delta on base shear strength
demands. These ratios versus the T/Tp ratio are illustrated in Fig. 6.27 for various
targeted maximum story ductilities. The following observations can be made:

•= The P-delta amplifications are larger for short period structures. As shown in Fig.
3.4, for the P-delta case studied (see Section 3.2.3), gravity loads are large enough
to cause a negative post-yield stiffness for the generic 20-story structure. In each
inelastic cycle of response, when the displacement demands enter this negative-
stiffness region, the displacement response is amplified by a certain irreversible
amount. Short period structures typically experience a larger number of inelastic
cycles compared to long period structures for a given input motion, resulting in a
larger cumulative amplification caused by P-delta effects. However, this
conclusion is drawn under the assumption that the stability coefficient of 0.1,
which is used here to evaluate P-delta effects, applies to both long and short
period structures. This may lead to overestimating P-delta effects for short frame
structures.

•= The larger number of inelastic cycles is also the reason that pulse P3 produces the
largest P-delta amplifications of the three basic pulses. Pulse P3 consists of more
cycles than the other two pulses, and for the same base shear strength, the

Chapter 6 Response of Structures to Pulse-Type …

99
 structure that is subjected to P3 experiences more drifting (ratcheting) of
displacement response.

•= It is also important to note that P-delta amplifications are more significant when
maximum ductility demands occur in the bottom story because the vertical gravity
loads are largest at the base.

Damping Effects. As discussed in Section 3.2.1, Rayleigh damping is used to obtain a

damping ratio of 2% at the first mode period T and at 0.1T of the generic structure.
Advantage can be taken of the MDOF base shear strength demand spectra to assess the
sensitivity of the demands to the damping ratio of the system. If the ordinates of the
MDOF strength demand spectra for 2% damping are divided by the corresponding values
for 5% damping, for the same ductility and period values, the ratio can be used to
evaluate the significance of damping effects for structures subjected to pulse-type
motions. These strength demand ratios are illustrated in Fig. 6.28 for the three basic
pulses. As can be seen, with few exceptions, the ratio oscillates between 1.0 and 1.2,
with an average of 1.1, for all periods, target ductilities, and pulses. The relatively
insignificant effect of damping can be attributed to the short duration of pulse-type
ground motions. The small effect of damping for impulsive excitations has been reported
by other researchers, e.g., Veletsos et al. (1965) and Chopra (1995).

6.3. Displacement Demands for Inelastic Structures

The SDOF and MDOF ductility demands presented in the previous section were used to
identify the salient response attributes of pulse-type ground motions. However, ductility
demands alone will not provide a comprehensive understanding of near-fault and pulse-
type response characteristics. In this section, displacement demands for inelastic systems
are evaluated. In addition to SDOF demands, roof displacement and story drift demands
are investigated for the generic MDOF frame structure introduced in Section 3.2
subjected to the three basic pulses defined in Section 5.2.

Chapter 6 Response of Structures to Pulse-Type …

100
6.3.1. SDOF Systems

Inelastic Displacement Demands for Constant Ductility:

Figure 6.29 illustrates the ratio of inelastic to elastic spectral displacement demands for
the basic pulses. These ratios follow patterns observed for ordinary ground motions. In
the short period range (small T/Tp), the inelastic spectral displacements are usually larger
than the elastic ones, whereas the reverse is observed in the intermediate period range,
and at very long periods the inelastic and elastic displacements both approach the
maximum ground displacement. This pattern is particularly clear for pulse P3, for which
the ratio of δinelastic/δelastic is significantly smaller than 1.0 around T/Tp = 1.0, passes
through 1.0 at T/Tp = 0.75 (for all ductility ratios), and is larger than 1.0 for smaller T/Tp
ratios. A similar pattern was reported by Rahnama and Krawinkler (1993) for spectra of
soft soil ground motions. This is no surprise since the harmonic motion of pulse P3
closely replicates a soft soil motion whose frequency content is dominated by a soil
period Ts.

Inelastic Displacement Demands for Specific Periods:

A comprehensive picture of the displacement demands for SDOF systems subjected to

pulse P2 is presented in Fig. 6.31, which illustrates, for specific periods, the variation of
the displacement demand with the strength of the system. The displacement demand is
normalized by the pulse peak ground displacement ug,max, and the strength coefficient η=
= Fy/([Link],max) is used on the vertical axis. The results again indicate that for the period
range of 0.75 < T/Tp < 3.0, the inelastic displacement demands are typically smaller than
the elastic ones, and that for T/Tp < 0.75 this pattern is reversed. It is also observed that
at a strength level corresponding to η ≅ 0.25, the displacement demand is about 0.9ug,max
regardless of the period. The displacement demands approach the peak ground
displacement as the structure becomes very weak or the ground motion intensity becomes
very large.

Chapter 6 Response of Structures to Pulse-Type …

101
6.3.2. MDOF Systems

Inelastic Roof Displacement Demands for Specific Strength:

The evaluation of inelastic roof displacement demands is carried out through a two-step
procedure. In the first step, elastic roof displacement demands are related to spectral
displacement values. Then, the effort is devoted to the assessment of the relationship
between elastic and inelastic roof displacement demands. Information regarding the first
step of this procedure was provided in Section 6.1.4.

The relative values of inelastic to elastic roof displacements generally follow patterns
observed for inelastic to elastic spectral displacements (Fig. 6.29). This is illustrated in
Fig. 6.30, which shows, for the basic pulses, the inelastic roof displacement of MDOF
systems with various base shear strength coefficients η, normalized by the corresponding
elastic roof displacement. The inelastic roof displacements are larger than the elastic
ones in the short period range (T/Tp < 0.75), and become much larger with decreasing
T/Tp and η. The reverse is observed in the long period range (T/Tp > 0.75), where the
inelastic roof displacement demand is smaller than the elastic one and decreases when the
structure becomes weaker. This pattern is consistently observed for all three pulses.

Inelastic Roof Displacement Demands for Specific Periods:

Analogous to Fig. 6.31 for SDOF displacement demands, Fig. 6.32 provides
comprehensive information on elastic and inelastic roof displacement demands for
MDOF structures subjected to P2. The graph illustrates the variation of the roof
displacement demand as a function of the base shear strength coefficient, η. The roof
displacement demands are normalized by the peak ground displacement of the pulse. It
can be seen that in the short period range (T/Tp ≤ 1.0) the corresponding MDOF and
SDOF displacement demands are close, whereas in the long period range (T/Tp > 1.0) the
roof demands are significantly larger than the SDOF demands. This again highlights the
significance of higher-mode effects for long period structures subjected to pulse-type
ground motions.

Chapter 6 Response of Structures to Pulse-Type …

102
Inelastic Story Drift Demands Over Height:

Figure 6.33 illustrates the strength dependence of story drift demand distributions over
the height of the generic 20-story frame structure subjected to pulse P2. These story drift
demands, which are normalized by ug,max, are obtained for a 20-story structure, and,
unlike roof displacement demands, cannot be directly compared to the corresponding
demands for structures with a different number of stories (see Section 6.5.3). A
comparison between the story drift demands (Fig. 6.33) with the corresponding story
ductility demands (Fig. 6.19) reveals that drift and ductility demands do not always
follow the same patterns. The following observations serve to shed light on the
differences:

•= For short period structures (T/Tp = 0.5), the maximum story drift demand occurs
in the bottom story regardless of base shear strength, and increases when the
strength (η) is reduced (or pulse intensity increases). Even though the story
ductility demand exhibits a similar pattern, its rate of increase is much higher than
that of the story drift demands.

•= For long period structures (T/Tp = 2.0), as with ductility, the maximum story drift
demand occurs in the upper portion of strong structures, and migrates to the base
for weak structures. However, there is a radical difference in the relative
magnitude of drift and ductility demands: the maximum story drift demand is
much larger for a strong structure (e.g. η = 0.75) than for a weak structure (e.g. η
= 0.07), whereas the opposite is observed for the maximum story ductility
demand. This pattern for the drift demands is related to the observation made
from Fig. 6.32 for roof displacements, i.e., for long period structures the inelastic
roof displacement demands are larger for stronger systems.

•= For a given base shear strength value (η), the maximum story drift demand is
always larger for T/Tp = 2.0 compared to the corresponding demand for T/Tp =
0.5, whereas this pattern is reversed for the maximum ductility demand.

These observations indicate that both story ductility and story drift demands are
necessary for a complete evaluation of MDOF response characteristics, and that ductility
demands alone can provide only a partial assessment.

Chapter 6 Response of Structures to Pulse-Type …

103
6.4. Investigation of Other Pulse Shapes

Pulses of different shape and duration than the three basic ones were introduced in
Section 5.3, and their spectral values were evaluated. This section provides a
complementary investigation, which assesses the inelastic response of MDOF structures
to those alternative pulse-type motions. The objective is to determine whether the three
basic pulses can also represent other pulse shapes with sufficient accuracy. If so, it will
not be necessary to further investigate the response characteristics of other pulse shapes.

6.4.1. Triangular Pulses

Variations of pulses P1 and P2 were introduced in Section 5.3.1 in order to investigate

the effect of rise time on response parameters. These modified pulses, which have a
triangular rather than square acceleration history, were denoted as P1.1 (variation of P1)
and P2.1 (variation of P2) (Figs. 5.5 and 5.6). It was shown that the pulse modifications
have a relatively small effect on spectral response. The effect of these pulse
modifications on the inelastic response of MDOF structures is discussed here.

Distributions of story ductility demands over the height of the structure for the modified
pulses are compared with those for the corresponding basic pulses in Figs. 6.34 and 6.35.
This comparison is made for the period ratio T/Tp = 2.0. In order to study the effect of
structure strength, the top graph in each figure shows the story ductility demands for a
strong structure (large η), while the bottom graph illustrates the story ductility demands
for a weak system (small η). For all pulses the peak acceleration of the basic pulses,
ag,max, is used to determine η, even though the peak acceleration of the modified pulses is
2ag,max. This means that the presented results in each graph are story ductility demands
for an identical structure (with the same Vy) subjected to a basic pulse and its modified
version.

The results indicate that although there are some differences, the basic pulse represents
the major response characteristics of the modified version with sufficient accuracy. In
other words, the responses to square acceleration pulses with the peak of ag,max are not
significantly different from the responses to triangular acceleration pulses with the peak
of 2ag,max, provided that they have the same pulse period.

Chapter 6 Response of Structures to Pulse-Type …

104
A similar comparison is presented in Figs. 6.36 and 6.37 for the maximum story ductility
demand. The results exhibit a good agreement between the maximum story ductility
demands for structures subjected to a basic pulse and its modified version.

On the basis of the results presented here for inelastic MDOF demands and those
provided in Section 5.3.1 for spectral responses, the conclusion is that the three basic
pulses characterized by square-shaped acceleration histories are able to represent their
modified versions. Therefore, there appears to be no need to further study the effect of
pulse shapes with different acceleration rise times.

6.4.2. Pulse Input Motions with Different Duration

Pulse input motions of different duration from the basic pulses, i.e., P4 and P5, were
introduced in Section 5.3.2, and their spectral quantities were evaluated. The elastic
response does not provide sufficient evidence that the inelastic response properties of P4
and P5 can be represented by the three basic pulses. Thus, the inelastic response of
MDOF structure to these pulse-type motions is investigated here.

Figure 6.38 illustrates distributions of story ductility demands over the height of the
structure for pulses P4 and P5 and T/Tp values of 0.5, 1.0, and 2.0. The distributions are
presented for structures with various base shear strength values to show the effect of
structure strength on the story ductility distributions. A comparison between the ductility
distributions shown in this figure and the distributions obtained for pulses P2 and P3 (Fig.
6.19) reveals that P2 and P3 can also represent story ductility demands for structures
subjected to pulses P4 and P5 with reasonable accuracy.

This conclusion is verified by a comparison of the maximum ductility demands for pulses
P2, P3, P4, and P5 presented in Fig. 6.39, which shows the η−µmax diagrams for T/Tp =
0.5 and 2.0. The results indicate that at most strength values the maximum story ductility
demands for P4 and P5 are adequately represented by pulse P3. This exempts P4 and P5
from further scrutiny and implies that the three basic pulses P1, P2, and P3 are reasonable
representations of a variety of pulse-type ground motions of different shapes and
duration. Therefore, later in Chapter 7, only these three pulse shapes are used as
equivalent pulses to represent near-fault ground motions.

Chapter 6 Response of Structures to Pulse-Type …

105
6.5. Sensitivity of Inelastic Demands to Number of Stories

The generic 20-story frame model introduced in Section 3.2 has been used extensively in
the MDOF response evaluations, even for short-period structures (structures with small
T/Tp ratios). However, using a 20-story model to quantify the response of structures with
a short fundamental period (e.g., T < Tp) may be questionable. Stiff frame structures
normally have a small number of stories and therefore few degrees of freedom, which
may lead to a different contribution of higher modes to the response compared to the 20-
story frame. Furthermore, for a structure with a small number of stories, distributions of
story stiffness and shear strength over the height are not as smooth as those for the
generic 20-story structure. This could result in different response characteristics in short-
period structures that may not be represented by a model whose properties vary almost
continuously over the height. The results of a sensitivity analysis are summarized here
that is intended to assess the effect of the number of stories on seismic demands.

As will be shown in Chapter 7, the period of the pulse contained in near-fault ground
motions, Tp, is typically longer than about 1.0 second. This makes the period range of
T/Tp ≤ 1.0 the emphasis of this part of the study because frame structures with a
fundamental period shorter than one second are likely to be misrepresented by a 20-story
frame model. Generic 3- and 9-story structures are developed for this purpose, and their
response to pulse-type ground motions is compared to that of the 20-story structure.

6.5.1. Generic 3-Story and 9-Story Structures

The generic 3- and 9-story structures utilized in this part of the study are designed
according to the same rules and assumptions as used for the 20-story structures discussed
in Section 3.2. As with the 20-story model, the story stiffnesses are tuned such that a
straight-line deflected shape is achieved under a load pattern that is based on the story
shear forces obtained from an SRSS modal combination. Figure 6.40 compares the SRSS
story shear force distributions for the generic 3-story, 9-story, and 20-story structures.
The figure indicates a relatively large lateral force at the roof level for the 3-story frame.
The SRSS distributions for the 9-story and 20-story structures are very close.

Various demands for the 3- and 9-story structures subjected to pulse P2 are evaluated and
compared with the corresponding demands for the 20-story frame in order to assess the

Chapter 6 Response of Structures to Pulse-Type …

106
usefulness of the results presented previously in the short period range (primarily T/Tp ≤
1.0), and to investigate the effect of the number of stories on the demands.

6.5.2. Ductility Demands

Story Ductility Demands Over Height:

Figure 6.41 illustrates the effect of the number of stories on the distributions of story
ductility demands over the height for T/Tp = 0.375, 0.75, 1.0 and 2.0, and pulse P2. Each
graph presents the distributions for all three generic structures with various base shear
strength coefficients η, ranging from almost elastic to highly inelastic behavior. The
following observations can be made:

•= For the short period range (T/Tp ≤ 1.0), no migration of maximum ductilities from
the upper stories to the bottom is observed for all three structures. In strong
systems the story ductility demands for all three structures vary almost linearly
over the height, and the demands are close for systems with the same strength
coefficients.

•= When the strength of the system is reduced, the ductility distributions in the 20-
story system become highly non-uniform with rapidly growing demands in the
bottom story, whereas the distributions in the 3-story structure remain more
uniform. For a given period and base shear strength coefficient η, the distribution
in the 9-story structure is typically more uniform than that in the 20-story
structure, and less uniform than that in the 3-story structure. The general
observation is that for frames with fewer stories, the distribution of story ductility
demands over the height is more uniform than for the 20-story structure.

•= For weak systems (small η) the maximum story ductility in the 20-story structure
always occurs in the bottom story. The same holds true for the 9-story structure
except for very weak systems, but for weak 3-story structures the maximum story
ductility demands occur mostly in the top story. But the maximum demands for
the three structures are comparable for a given η, even if they occur at different
locations.

Chapter 6 Response of Structures to Pulse-Type …

107
 •= The migration of maximum ductilities from the upper stories to the bottom of
long period structures (T/Tp = 2.0), which is very evident in the 20-story structure,
is also observed for the 9-story structure, although to a lesser degree. For the 3-
story structure the migration does not occur. This observation, as well as the
previous one, is rationalized by differences in higher mode effects, recognizing
that a bulge in the upper portion of the deflected shape of a 3-story structure
cannot develop.

Maximum Story Ductility Demands:

Figure 6.42 compares η−µmax and 1/η−µmax diagrams for the 3-story, 9-story, and 20-
story systems with T/Tp = 0.375 and 0.75 subjected to pulse P2. For a given strength
coefficient η, each curve provides the maximum story ductility demand (maximum over
all stories). For strong systems (large η) close proximity is observed among the
maximum story ductility demands for the three systems. When the structures become
weaker, the differences between the demands for the systems increase, but the differences
are not very large in all cases. It appears that unlike the distributions of story ductility
over the height, the maximum story ductility demands do not depend strongly on the
number of stories. However, the location of the maximum demand is not identical for
different structures; e.g., for weak systems the maximum ductility occurs in the bottom
story of the 20-story structure, whereas the location of the maximum demand in the 3-
story structure is the upper story.

Base Shear Strength Demands for Target Ductility:

Similar to the procedure followed for the 20-story structure, the base shear strength of the
3- and 9-story structures required to limit the maximum story ductility demands to
specific target values can be obtained from the η−µmax diagrams of the type shown in Fig.
6.42(a) by using a linear interpolation algorithm. Such base shear strength demand
spectra are compared in Fig. 6.43 for the 3-story, 9-story, and 20-story structures, for
small and large target story ductility values. This comparison evaluates the extent to
which the generic 20-story structure can be utilized to estimate strength demands for stiff
structures with fewer stories. Although the graphs illustrate the spectra for fundamental
periods up to T = 2Tp, the primary range of interest for stiff structures is T/Tp ≤ 1.0.

Chapter 6 Response of Structures to Pulse-Type …

108
The results indicate small differences among the base shear strength demands for the
three structures at small target ductilities. At large ductilities the differences increase
somewhat, with the maximum difference between the demands for the 3- and 20-story
structures being at T/Tp = 0.75. The differences between the demands for the 9- and 20-
story structures are typically very small. For given period and target ductility values, the
strength demands for the 3-story system are always smaller than those for the 20-story
frame, with the exception of T/Tp > 1.25 and µmax = 4, which is associated with the
migration phenomenon in the 20-story frame. Overall, the results imply that the generic
20-story structure can reasonably represent base shear strength demands for stiff
structures with fewer stories, and that the base shear strength demands for target story
ductilities are not very sensitive to the number of stories.

6.5.3. Inelastic Displacement Demands

Roof Displacements:

Roof displacement demands can be taken advantage of to investigate the extent to which
the generic 20-story structure represents the global response of stiff structures that have
fewer stories. Figure 6.44 compares the variations of inelastic roof displacement
demands with base shear strength for the three structures under consideration, with T/Tp
ratios ranging from 0.375 to 2.0, subjected to pulse P2. The following observations can
be made:

•= In the long period range (T/Tp > 1.0), for a given period and large η values
(elastic behavior), the 20-story structure has the largest, and the 3-story structure
has the smallest roof displacement demands. This can be attributed to larger
contributions of higher-mode effects in structures with a larger number of stories.
For structures with inelastic behavior (small η), the differences in roof
displacement demands are very small.

•= In the short period range (T/Tp ≤ 1.0), the results show small to moderate
differences among the demands at all strength levels, indicating that the generic
20-story frame can be utilized to estimate roof displacement demands for stiff
structures with fewer stories, and that global displacement demands are not very
sensitive to the number of stories.

Chapter 6 Response of Structures to Pulse-Type …

109
Story Drift Angle:

The story ductility ratio is defined as the story drift normalized by the story yield drift.
The focus of this part of the study is on the story drift angle, which is defined as the story
drift normalized by the height of the story. The story drift angle demands are not directly
comparable between structures that have the same fundamental period but different
heights. For example, in a simple case of two elastic SDOF structures with the same
period and different heights, subjected to the same ground motion, the displacement
demands are expected to be identical rather than the drift angles. If the story drift angles
are normalized by the roof drift angle (roof displacement divided by the structure height),
a picture is obtained of the relative drift distribution over the height, which can be
compared between structures with different numbers of stories provided the difference in
roof drift is accounted for (see Fig. 6.44).

Figures 6.45 compares distributions of the normalized story drift angle demands (θi/θroof)
over the height of the 3-story, 9-story, and 20-story structures subjected to P2 in the short
period range, i.e., T/Tp = 0.375 and 0.75. The distributions are presented for various η
values to illustrate the effect of base shear strength. The figure shows that the
distributions of story drift angle demands over the height of the structure are much more
uniform for the 3-story frame than for the 20-story frame. In other words, in the 3-story
frame all stories contribute to the roof displacement almost equally, whereas in the 20-
story frame the contribution of the bottom stories is much larger than that of the top
stories.

Summary:

The results of this limited sensitivity study on the effects of the number of stories on
seismic demands indicates that different parameters show various degrees of sensitivity
to the number of stories. The maximum story ductility demand over the structure height
appears to be not very sensitive to the number of stories, which is reflected in the similar
shape of the MDOF strength demand spectra for constant ductility shown in Fig. 6.43.
The roof drift also appears to be comparable, regardless of the number of stories.
However, the distributions of ductility demands and story drifts over the height appear to
be rather sensitive to the number of stories, and may differ significantly from that of the
20-story generic structure if the number of stories becomes very small (e.g., the 3-story

Chapter 6 Response of Structures to Pulse-Type …

110
structure). The results presented in Figs. 6.41 and 6.45 provide insight into the sensitivity
of these parameters to the number of stories.

6.6. Sensitivity of Inelastic Demands to Plastic Hinge Locations

As discussed in Section 3.2.1, inelastic deformations in the structures utilized in this

study are permitted only at the ends of the beam in each story and at the base of the
columns. Thus, no individual story mechanism is allowed to form (see Fig. 3.1). This
condition is sometimes difficult to meet in practice. The objective of this section is to
investigate seismic demands for structures in which individual story mechanisms can
form under lateral loads. A generic 20-story structure with plastic hinges in columns is
developed for this purpose, and its response to pulse P2 is compared with the response of
the structure defined in Section 3.2, which is used as a reference.

6.6.1. MDOF System Investigated in this Study

The generic 20-story frame structure utilized in this part of the study is designed
according to the rules and assumptions discussed in Section 3.2, except that plastic hinges
are allowed to develop only at the ends of the columns in each story, leading to story
mechanisms under lateral loads. In the remainder of the section, this structure is referred
to as “CH” (Column Hinge), and the reference structure is referred to as “BH” (Beam
Hinge).

As with the reference structure, the story stiffnesses are tuned such that a straight-line
deflected shape is achieved under a load pattern that is based on the story shear forces
obtained from an SRSS modal combination. The column bending strength in each story
is tuned such that under the SRSS lateral load pattern simultaneous yielding occurs in all
stories.

6.6.2. Story Ductility Demands Over Height

Story ductility demands are utilized here to describe the inelastic response of the CH
structure to pulse P2. Figure 6.46 compares distributions of story ductility demands over
the height for the CH and BH structures. The distributions are presented for structures
with T/Tp = 0.5, 1.0, and 2.0, and various base shear strength coefficients η. The
following observations can be made:

Chapter 6 Response of Structures to Pulse-Type …

111
 •= In the short period range (T/Tp ≤ 1.0), the maximum ductility demands for both
structures are typically located in the bottom story, regardless of strength, and
grow with a reduction in the strength coefficient η. The difference is that the
growth rate is higher for the CH structure, whereas in the top portion of the
structure the ductility demand hardly exceeds a value of about 3.0, even when the
demands at the bottom are excessively large. The reason is that developing story
mechanisms in the bottom stories of the CH structure causes large ductility
demands, with a large concentration of energy dissipation at the bottom.

•= In the long period range (T/Tp > 1.0), the maximum story ductility demand occurs
in the upper portion of strong structures. In the BH structure the migration of
large demands to the bottom, as a result of decreasing η, does not take place
before the demands stabilize in the upper stories around a story ductility of 3.3,
whereas in the CH structure the ductility demands increase simultaneously in the
upper and bottom stories. The stabilization of the demands in the upper stories of
the CH structure happens at a larger ductility (about 5.0) compared to the BH
structure. These observations can be rationalized by the formation of story
mechanisms in the CH structure. Again, for the same strength (η value) the
maximum ductility demands for the CH structure are larger.

•= For relatively strong structures (large η) the distributions of story ductility

demands for the CH and BH structures are very close.

6.6.3. Maximum Story Ductility Demands

A comprehensive evaluation of maximum story ductility demands is presented in Fig.

6.47, which illustrates η−µmax diagram for the CH structure subjected to pulse P2. This
graph can be compared directly with the corresponding graph for the BH structure
subjected to P2 shown in Fig. 6.21(a). A comparison between the η−µmax curves for
given T/Tp values reveals that for long period structures (T/Tp > 1.0) the close to vertical
ranges in the BH curves around a ductility of 3 are not observed in the corresponding CH
curves. As explained in the previous section, the reason is that the migration of the
maximum ductility demand from the upper stories to the bottom occurs within a very
narrow range of η for the CH structure.

Chapter 6 Response of Structures to Pulse-Type …

112
6.6.4. Base Shear Strength Demands for Target Ductility

The base shear strength of the CH structure required to limit the maximum story ductility
demand to specific target values can be obtained from the η−µmax diagram shown in Fig.
6.47 using a linear interpolation scheme. The ratios of the CH to BH base shear demands
for given T/Tp and targeted maximum story ductility values are illustrated in Fig. 6.48.
The figure provides a comprehensive comparison between base shear strength demands
for the CH and BH structures. The following observations are made:

•= The ratios are typically larger than 1.0, but for small target ductilities (close to
elastic structures), they are close to 1.0 in the full range of period. This is not
surprising because the locations of plastic hinges do not have a significant effect
on the response of strong structures with close to elastic behavior.

•= Very large ratios are observed for a target ductility of µmax = 4 and T/Tp > 1.5,
which indicate that in order to limit story ductility demands to a target value of 4,
the CH structure should be several times as strong as the BH structure with the
same period. The reason is that µmax = 4 is associated with the migration of
ductility demands in the BH structure, during which the maximum ductility
demand remains almost unchanged with a large reduction in the base shear
strength. This phenomenon appears as close to vertical ranges in the 1/η−µmax
curves illustrated in Fig. 6.22(a) around a ductility of 3.5.

•= Overall, the ratio decrease when T/Tp becomes smaller. For very stiff structures
the ratio is close to 1.0 regardless of the target story ductility.

Summary:

On the basis of the results presented here for pulse-type ground motions, and those
presented in past studies for ordinary ground motions (Nassar and Krawinkler, 1991), it is
concluded that formation of plastic hinges in columns has undesirable consequences, and
should be avoided by making columns stronger than beams. If inelastic deformations are
allowed in the columns of a moment resisting frame, additional base shear strength needs
to be provided for ductility control.

Chapter 6 Response of Structures to Pulse-Type …

113
 Deflection with Respect to Ground Deflection with Respect to Ground
t = Tp / 4 t = Tp / 2
1 1

0.8 0.8
Relative Height

Relative Height
0.6 0.6

0.4 0.4

0.2 0.2

0 0
-0.6 -0.4 -0.2 0 0.2 0.4 0.6 -0.6 -0.4 -0.2 0 0.2 0.4 0.6
ui / ug,max ui / ug,max

Deflection with Respect to Ground Deflection with Respect to Ground

t = 3 Tp / 4 t = Tp
1 1

0.8 0.8
Relative Height

Relative Height
0.6 0.6

0.4 0.4

0.2 0.2

0 0
-0.6 -0.4 -0.2 0 0.2 0.4 0.6 -0.6 -0.4 -0.2 0 0.2 0.4 0.6
ui / ug,max ui / ug,max

Deflection with Respect to Ground Deflection with Respect to Ground

t = 5 Tp / 4 t = 3 Tp / 2
1 1

0.8 0.8
Relative Height
Relative Height

0.6 0.6

0.4 0.4

0.2 0.2

0 0
-0.6 -0.4 -0.2 0 0.2 0.4 0.6 -0.6 -0.4 -0.2 0 0.2 0.4 0.6
ui / ug,max ui / ug,max

Figure 6.1 Elastic Deflected Shapes of MDOF Structure with T/Tp = 0.5, Pulse P1

Chapter 6 Response of Structures to Pulse-Type …

114
 Deflection with Respect to Ground Deflection with Respect to Ground
t = Tp / 4 t = Tp / 2
1 1

0.8 0.8
Relative Height

Relative Height
0.6 0.6

0.4 0.4

0.2 0.2

0 0
-1.4 -1 -0.6 -0.2 0.2 0.6 1 1.4 -1.4 -1 -0.6 -0.2 0.2 0.6 1 1.4
ui / ug,max ui / ug,max

Deflection with Respect to Ground Deflection with Respect to Ground

t = 3 Tp / 4 t = Tp
1 1

0.8 0.8
Relative Height

Relative Height
0.6 0.6

0.4 0.4

0.2 0.2

0 0
-1.4 -1 -0.6 -0.2 0.2 0.6 1 1.4 -1.4 -1 -0.6 -0.2 0.2 0.6 1 1.4
ui / ug,max ui / ug,max

Deflection with Respect to Ground Deflection with Respect to Ground

t = 5 Tp / 4 t = 3 Tp / 2
1 1

0.8 0.8
Relative Height

Relative Height

0.6 0.6

0.4 0.4

0.2 0.2

0 0
-1.4 -1 -0.6 -0.2 0.2 0.6 1 1.4 -1.4 -1 -0.6 -0.2 0.2 0.6 1 1.4
ui / ug,max ui / ug,max

Figure 6.2 Elastic Deflected Shapes of MDOF Structure with T/Tp = 2.0, Pulse P1

Chapter 6 Response of Structures to Pulse-Type …

115
 Deflection with Respect to Ground Deflection with Respect to Ground
t = Tp / 4 t = Tp / 2
1 1

0.8 0.8
Relative Height

Relative Height
0.6 0.6

0.4 0.4

0.2 0.2

0 0
-2.4 -1.6 -0.8 0 0.8 1.6 2.4 -2.4 -1.6 -0.8 0 0.8 1.6 2.4
ui / ug,max ui / ug,max

Deflection with Respect to Ground Deflection with Respect to Ground

t = 3 Tp / 4 t = Tp
1 1

0.8 0.8
Relative Height

Relative Height
0.6 0.6

0.4 0.4

0.2 0.2

0 0
-2.4 -1.6 -0.8 0 0.8 1.6 2.4 -2.4 -1.6 -0.8 0 0.8 1.6 2.4
ui / ug,max ui / ug,max

Deflection with Respect to Ground Deflection with Respect to Ground

t = 5 Tp / 4 t = 3 Tp / 2
1 1

0.8 0.8
Relative Height

Relative Height

0.6 0.6

0.4 0.4

0.2 0.2

0 0
-2.4 -1.6 -0.8 0 0.8 1.6 2.4 -2.4 -1.6 -0.8 0 0.8 1.6 2.4
ui / ug,max ui / ug,max

Figure 6.3 Elastic Deflected Shapes of MDOF Structure with T/Tp = 2.0, Pulse P2

Chapter 6 Response of Structures to Pulse-Type …

116
 Deflection with Respect to Ground Deflection with Respect to Ground
t = Tp / 4 t = Tp / 2
1 1

0.8 0.8
Relative Height

Relative Height
0.6 0.6

0.4 0.4

0.2 0.2

0 0
-4 -3 -2 -1 0 1 2 3 4 -4 -3 -2 -1 0 1 2 3 4
ui / ug,max ui / ug,max

Deflection with Respect to Ground Deflection with Respect to Ground

t = 3 Tp / 4 t = Tp
1 1

0.8 0.8
Relative Height

0.6 Relative Height 0.6

0.4 0.4

0.2 0.2

0 0
-4 -3 -2 -1 0 1 2 3 4 -4 -3 -2 -1 0 1 2 3 4
ui / ug,max ui / ug,max

Deflection with Respect to Ground Deflection with Respect to Ground

t = 5 Tp / 4 t = 3 Tp / 2
1 1

0.8 0.8
Relative Height

Relative Height

0.6 0.6

0.4 0.4

0.2 0.2

0 0
-4 -3 -2 -1 0 1 2 3 4 -4 -3 -2 -1 0 1 2 3 4
ui / ug,max ui / ug,max

Figure 6.4 Elastic Deflected Shapes of MDOF Structure with T/Tp = 2.0, Pulse P3

Chapter 6 Response of Structures to Pulse-Type …

117
 Elastic MDOF Base Shear Demands
Generic 20-Story, ξ = 2%, without P-∆
∆
8
P1
P2
P3
6
Vbase,max / (m . ag,max)

0
0 0.5 1 1.5 2 2.5 3
T / Tp

Figure 6.5 Maximum Base Shear for Elastic Structures Subjected to Basic Pulses

Normalized Elastic Base Shear Demands

Generic 20-Story, ξ = 2%, without P-∆
∆
3

P1
P2
2.5
P3
VMDOF / VSDOF

1.5

0.5
0 0.5 1 1.5 2 2.5 3
T / Tp

Figure 6.6 Ratios of Elastic MDOF Base Shear to SDOF Strength Demands

Chapter 6 Response of Structures to Pulse-Type …

118
 Elastic MDOF Base Shear Demands
Continuous Shear Building, ξ = 0%
8
P1
P2
P3
6
Vbase,max / (m . ag,max)

0
0 0.5 1 1.5 2 2.5 3
T / Tp

Figure 6.7 Maximum Base Shear for Elastic Continuous Shear Building Subjected to
Basic Pulses

Elastic Base Shear Force Time History

Pulse P2, Continuous Shear Building, T / Tp = 0.75, ξ = 0%
6

4
Vbase / (m . ag,max)

-2

-4

-6
0 0.5 1 1.5 2 2.5 3
t / Tp

Figure 6.8 Base Shear Time History for Shear Building with T/Tp = 0.75, Pulse P2

Chapter 6 Response of Structures to Pulse-Type …

119
 Normalized Elastic Story Shear Demands Normalized Elastic Story Shear Demands
∆, Time History
Pulse P1, without P-∆ ∆, SRSS
Pulse P1, without P-∆
1

Relative Height 0.8

0.6
Design Design
T/Tp = 0.375 T/Tp = 0.375
T/Tp = 0.50 T/Tp = 0.50
0.4
T/Tp = 0.75 T/Tp = 0.75
T/Tp = 1.00 T/Tp = 1.00

0.2 T/Tp = 1.50 T/Tp = 1.50

T/Tp = 2.00 T/Tp = 2.00
T/Tp = 3.00 T/Tp = 3.00

0
0 0.2 0.4 0.6 0.8 1 1.2 0 0.2 0.4 0.6 0.8 1 1.2
Vi,max / Vbase,max Vi,max / Vbase,max

Pulse P1

Normalized Elastic Story Shear Demands Normalized Elastic Story Shear Demands
∆, Time History
Pulse P2, without P-∆ ∆, SRSS
Pulse P2, without P-∆
1

0.8
Relative Height

0.6
Design Design
T/Tp = 0.375 T/Tp = 0.375
T/Tp = 0.50 T/Tp = 0.50
0.4
T/Tp = 0.75 T/Tp = 0.75
T/Tp = 1.00 T/Tp = 1.00

0.2 T/Tp = 1.50 T/Tp = 1.50

T/Tp = 2.00 T/Tp = 2.00
T/Tp = 3.00 T/Tp = 3.00

0
0 0.2 0.4 0.6 0.8 1 1.2 0 0.2 0.4 0.6 0.8 1 1.2
Vi,max / Vbase,max Vi,max / Vbase,max

Pulse P2

Normalized Elastic Story Shear Demands Normalized Elastic Story Shear Demands
∆, Time History
Pulse P3, without P-∆ ∆, SRSS
Pulse P3, without P-∆
1

0.8
Relative Height

0.6
Design Design
T/Tp = 0.375 T/Tp = 0.375
T/Tp = 0.50 T/Tp = 0.50
0.4
T/Tp = 0.75 T/Tp = 0.75
T/Tp = 1.00 T/Tp = 1.00

0.2 T/Tp = 1.50 T/Tp = 1.50

T/Tp = 2.00 T/Tp = 2.00
T/Tp = 3.00 T/Tp = 3.00

0
0 0.2 0.4 0.6 0.8 1 1.2 1.4 0 0.2 0.4 0.6 0.8 1 1.2 1.4
Vi,max / Vbase,max Vi,max / Vbase,max

Pulse P3
(a) Time History (b) SRSS

Figure 6.9 Normalized Elastic Story Shear Demands for Basic Pulses,
without P-Delta Effects

Chapter 6 Response of Structures to Pulse-Type …

120
 Elastic Roof Displacement Demands
Pulse P1, ξ = 2%
1.7
without P-delta

with P-delta
1.6
First Mode

δ roof,max / Sd
1.5

1.4

1.3

1.2
0 0.5 1 1.5 2 2.5 3 3.5
T / Tp

Elastic Roof Displacement Demands

Pulse P2, ξ = 2%
2.8

without P-delta

with P-delta
2.4 First Mode
δ roof,max / Sd

1.6

1.2
0 0.5 1 1.5 2 2.5 3 3.5
T / Tp

Elastic Roof Displacement Demands

Pulse P3, ξ = 2%
2.8
without P-delta

with P-delta
2.4 First Mode
δ roof,max / Sd

1.6

1.2
0 0.5 1 1.5 2 2.5 3 3.5
T / Tp

Figure 6.10 Ratios of Elastic MDOF Displacement to First Mode Spectral Displacement
for Basic Pulses

Chapter 6 Response of Structures to Pulse-Type …

121
 Normalized Inelastic SDOF Displacement Time History
Bilinear, T / Tp = 0.5, µ = 8, α = 0%, ξ = 2%
8
P1

P2

4 P3
u / uy

-4

-8
0 0.5 1 1.5 2 2.5 3 3.5 4
t / Tp

(a) T/Tp = 0.5

Normalized Inelastic SDOF Displacement Time History

Bilinear, T / Tp = 1.0, µ = 8, α = 0%, ξ = 2%
8

4
u / uy

-4 P1

P2

P3
-8
0 0.5 1 1.5 2 2.5 3 3.5 4
t / Tp

(b) T/Tp = 1.0

Figure 6.11 Displacement Response of Inelastic SDOF Systems to Basic Pulses

Chapter 6 Response of Structures to Pulse-Type …

122
 SDOF Strength Demands for Constant Ductility
Pulse P1, Bilinear, α = 3%, ξ = 2%
5
µ===1
µ===2
4 µ===3
µ===4

η = Fy / ([Link],max)
µ===6
3 µ===8

0
0 0.5 1 1.5 2 2.5 3
T / Tp

SDOF Strength Demands for Constant Ductility

Pulse P2, Bilinear, α = 3%, ξ = 2%
5
µ===1
µ===2
4 µ===3
µ===4
µ===6
η = Fy / ([Link],max)

3 µ===8

0
0 0.5 1 1.5 2 2.5 3
T / Tp

SDOF Strength Demands for Constant Ductility

Pulse P3, Bilinear, α = 3%, ξ = 2%
8
µ===1
µ===2
µ===3
6 µ===4
η = Fy / ([Link],max)

µ===6
µ===8

0
0 0.5 1 1.5 2 2.5 3
T / Tp

Figure 6.12 Elastic and Inelastic Strength Demand Spectra for Basic Pulses

Chapter 6 Response of Structures to Pulse-Type …

123
 SDOF Strength Demands for Constant Ductility
Pulse P2, Clough / Bilinear, α = 3%, ξ = 2%
2
µ===2
µ===3
µ===4
1.5 µ===6
µ===8
η Clough=/=ηBi

0.5

0
0 0.5 1 1.5 2 2.5 3
T / Tp

(a) Pulse P2

SDOF Strength Demands for Constant Ductility

Pulse P3, Clough / Bilinear, α = 3%, ξ = 2%
2
µ===2
µ===3
µ===4
1.5 µ===6
µ===8
η Clough=/=ηBi

0.5

0
0 0.5 1 1.5 2 2.5 3
T / Tp

(b) Pulse P3

Figure 6.13 Comparison of Inelastic SDOF Strength Demand Spectra for Bilinear and
Modified Clough Hysteresis Models, Basic Pulses

Chapter 6 Response of Structures to Pulse-Type …

124
 SDOF Ductility Demands
Pulse P1, Bilinear, α = 3%, ξ = 2%
2
T/Tp = 0.375
T/Tp = 0.50
1.6 T/Tp = 0.75
T/Tp = 1.00
T/Tp = 1.50
η = Fy / ([Link],max)

T/Tp = 2.00
1.2 T/Tp = 3.00

0.8

0.4

0
0 4 8 12 16 20
Ductility,,=µ

Figure 6.14 Ductility Demands for SDOF Systems Subjected to Pulse P1

SDOF Ductility Demands

Pulse P2, Bilinear, α = 3%, ξ = 2%
2
T/Tp = 0.375
T/Tp = 0.50
1.6 T/Tp = 0.75
T/Tp = 1.00
T/Tp = 1.50
η = Fy / ([Link],max)

T/Tp = 2.00
1.2
T/Tp = 3.00

0.8

0.4

0
0 4 8 12 16 20
Ductility,,=µ

Figure 6.15 Ductility Demands for SDOF Systems Subjected to Pulse P2

Chapter 6 Response of Structures to Pulse-Type …

125
 SDOF Ductility Demands
Pulse P3, Bilinear, α = 3%, ξ = 2%
2
T/Tp = 0.375
T/Tp = 0.50
1.6 T/Tp = 0.75
T/Tp = 1.00
T/Tp = 1.50
η = Fy / ([Link],max)

T/Tp = 2.00
1.2 T/Tp = 3.00

0.8

0.4

0
0 4 8 12 16 20
Ductility,,=µ

Figure 6.16 Ductility Demands for SDOF Systems Subjected to Pulse P3

µ Relationship
SDOF R-µ
Pulse P2, Bilinear, α = 3%, ξ = 2%
16
T/Tp = 0.375
T/Tp = 0.50
T/Tp = 0.75
T/Tp = 1.00
12
T/Tp = 1.50
T/Tp = 2.00
R = Fy,e / Fy(µ)

T/Tp = 3.00

0
0 4 8 12 16 20
Ductility Ratio, µ

Figure 6.17 R−µ Diagram for SDOF Systems Subjected to Pulse P2

Chapter 6 Response of Structures to Pulse-Type …

126
 SDOF Ductility Demands
Different Pulses, Bilinear, α = 3%, ξ = 2%
2
P1, T/Tp = 0.5
P2, T/Tp = 0.5
1.6
P3, T/Tp = 0.5
P1, T/Tp = 2.0
η = Fy / ([Link],max)

P2, T/Tp = 2.0

1.2
P3, T/Tp = 2.0

0.8

0.4

0
0 4 8 12 16 20
Ductility Ratio,,=µ

(a) η−µ Domain

µ Relationship
SDOF R-µ
Different Pulses, Bilinear, α = 3%, ξ = 2%
16
P1, T/Tp = 0.50
P2, T/Tp = 0.50
P3, T/Tp = 0.50
12 P1, T/Tp = 2.00
P2, T/Tp = 2.00
R = Fy,e / Fy(µ )

P3, T/Tp = 2.00

0
0 4 8 12 16 20
Ductility Ratio, µ

(b) R–µ=Domain

Figure 6.18 Effect of Pulse Type on Ductility Demands for SDOF Systems

Chapter 6 Response of Structures to Pulse-Type …

127
 Story Ductility Demands Story Ductility Demands
∆
Pulse P1, SRSS Pattern, T / Tp = 0.50, without P-∆ ∆
Pulse P1, SRSS Pattern, T / Tp = 2.0, without P-∆
1
η===2.50 η==1.25
η===2.00 η==1.00
η===1.50 η===0.75
0.8 η===1.25 η===0.50
η===1.00 η===0.25
η===0.75 η===0.15
Relative Height

η===0.60 η===0.10
0.6 η===0.50 η===0.07
η===0.40 η===0.05

0.4

0.2

0
0 4 8 12 16 20 0 4 8 12 16 20
Story Ductility Ratio, µi = δmax,i / δy,i Story Ductility Ratio, µi = δmax,i / δy,i

Pulse P1

Story Ductility Demands Story Ductility Demands

∆
Pulse P2, SRSS Pattern, T / Tp = 0.50, without P-∆ ∆
Pulse P2, SRSS Pattern, T / Tp = 2.0, without P-∆
1
η===2.50 η===1.25
η===2.00 η===1.00
η===1.50 η===0.75
0.8 η===1.25 η===0.50
η===1.00 η===0.25
η===0.75 η===0.15
Relative Height

η===0.60 η===0.10
0.6 η===0.50 η===0.07
η===0.40 η===0.05

0.4

0.2

0
0 4 8 12 16 20 0 4 8 12 16 20
Story Ductility Ratio, µi = δmax,i / δy,i Story Ductility Ratio, µi = δmax,i / δy,i

Pulse P2

Story Ductility Demands Story Ductility Demands

∆
Pulse P3, SRSS Pattern, T / Tp = 0.50, without P-∆ ∆
Pulse P3, SRSS Pattern, T / Tp = 2.0, without P-∆
1
η===2.50 η==1.25
η===2.00 η==1.00
η===1.50 η===0.75
0.8 η===1.25 η==0.50
η===1.00 η===0.25
η===0.75 η===0.15
Relative Height

η===0.60 η===0.10
0.6 η===0.50 η===0.07
η===0.40 η===0.05

0.4

0.2

0
0 4 8 12 16 20 0 4 8 12 16 20
Story Ductility Ratio, µi = δmax,i / δy,i Story Ductility Ratio, µi = δmax,i / δy,i

Pulse P3
(a) T/Tp = 0.5 (b) T/Tp = 2.0

Figure 6.19 Story Ductility Demands Over Height for Various Values of η, Basic Pulses,
without P-Delta Effects

Chapter 6 Response of Structures to Pulse-Type …

128
 Story Ductility Demands
∆
Pulse P2, SRSS Pattern, T / Tp = 0.5, without P-∆
4
Story 1

Story 15

3
1 / η = (W / Vy) ag,max / g

0
0 4 8 12 16 20
Story Ductility Ratio,,=µi = δmax,i / δy,i

(a) T/Tp = 0.5

Story Ductility Demands

∆
Pulse P2, SRSS Pattern, T / Tp = 2.0, without P-∆
20
Story 1

Story 15
16
1 / η = (W / Vy) ag,max / g

12

0
0 4 8 12 16 20
Story Ductility Ratio,,=µi = δmax,i / δy,i

(b) T/Tp = 2.0

Figure 6.20 Comparison of Story Ductility Demands in Different Stories for Pulse P2

Chapter 6 Response of Structures to Pulse-Type …

129
 Maximum MDOF Story Ductility Demands Maximum MDOF Story Ductility Demands
∆
Pulse P1, SRSS Pattern, without P-∆ ∆
Pulse P1, SRSS Pattern, with P-∆
4
T/Tp=0.375 T/Tp = 0.375
T/Tp=0.50 T/Tp = 0.50
T/Tp=0.75 T/Tp = 0.75
3 T/Tp=1.00 T/Tp = 1.00
T/Tp=1.50 T/Tp = 1.50
η = Vy / (m.a g,max)

T/Tp=2.00 T/Tp = 2.00

T/Tp=3.00 T/Tp = 3.00

0
0 4 8 12 16 20 0 4 8 12 16 20
Maximum Story Ductility,,=µmax Maximum Story Ductility,,=µmax

Pulse P1

Maximum MDOF Story Ductility Demands Maximum MDOF Story Ductility Demands
∆
Pulse P2, SRSS Pattern, without P-∆ ∆
Pulse P2, SRSS Pattern, with P-∆
4
T/Tp = 0.375 T/Tp = 0.375
T/Tp = 0.50 T/Tp = 0.50
T/Tp = 0.75 T/Tp = 0.75
3 T/Tp = 1.00 T/Tp = 1.00
T/Tp = 1.50 T/Tp = 1.50
η = Vy / (m.a g,max)

T/Tp = 2.00 T/Tp = 2.00

T/Tp = 3.00 T/Tp = 3.00

0
0 4 8 12 16 20 0 4 8 12 16 20
Maximum Story Ductility,,=µmax Maximum Story Ductility,,=µmax

Pulse P2

Maximum MDOF Story Ductility Demands Maximum MDOF Story Ductility Demands
∆
Pulse P3, SRSS Pattern, without P-∆ ∆
Pulse P3, SRSS Pattern, with P-∆
4
T/Tp=0.375 T/Tp = 0.375
T/Tp=0.50 T/Tp = 0.50
T/Tp=0.75 T/Tp = 0.75
3 T/Tp=1.00 T/Tp = 1.00
T/Tp=1.50 T/Tp = 1.50
η = Vy / (m.a g,max)

T/Tp=2.00 T/Tp = 2.00

T/Tp=3.00 T/Tp = 3.00

0
0 4 8 12 16 20 0 4 8 12 16 20
Maximum Story Ductility,,=µmax Maximum Story Ductility,,=µmax

Pulse P3
(a) without P-delta (b) with P-delta

Figure 6.21 Base Shear Strength versus Maximum Story Ductility Demands for Basic
Pulses

Chapter 6 Response of Structures to Pulse-Type …

130
 Maximum MDOF Story Ductility Demands
∆
Pulse P2, SRSS Pattern, without P-∆
20
T/Tp = 0.375
T/Tp = 0.50
16 T/Tp = 0.75
T/Tp = 1.00
1 / η = (W / Vy) ag,max / g

T/Tp = 1.50
T/Tp = 2.00
12 T/Tp = 3.00

0
0 4 8 12 16 20
Maximum Story Ductility,,=µmax

(a) without P-delta

Maximum MDOF Story Ductility Demands

∆
Pulse P2, SRSS Pattern, with P-∆
20
T/Tp = 0.375
T/Tp = 0.50
16 T/Tp = 0.75
T/Tp = 1.00
1 / η = (W / V y) ag,max / g

T/Tp = 1.50
T/Tp = 2.00
12 T/Tp = 3.00

0
0 4 8 12 16 20
Maximum Story Ductility,,=µmax

(b) with P-delta

Figure 6.22 1/η−µmax Diagrams for MDOF Structures Subjected to Pulse P2

Chapter 6 Response of Structures to Pulse-Type …

131
 Maximum MDOF Story Ductility Demands
Pulse P2, SRSS Pattern
4
T/Tp = 0.5, no P-delta
T/Tp = 0.5, with P-delta
T/Tp = 1.0, no P-delta
T/Tp = 1.0, with P-delta
3
T/Tp = 2.0, no P-delta
η = Vy / ([Link],max)

T/Tp = 2.0, with P-delta

0
0 4 8 12 16 20
Maximum Story Ductility,,=µmax

Figure 6.23 Effect of P-Delta on Ductility Demands for MDOF Structures, Pulse P2

Maximum MDOF Story Ductility Demands

∆
Basic Pulses, SRSS Pattern, without P-∆
4
P1, T/Tp = 0.5
P2, T/Tp = 0.5
P3, T/Tp = 0.5
3 P1, T/Tp = 2.0
P2, T/Tp = 2.0
η = Vy / ([Link],max)

P3, T/Tp = 2.0

0
0 4 8 12 16 20
Maximum Story Ductility,,=µmax

Figure 6.24 Effect of Pulse Type on Maximum Ductility Demands for MDOF Structures

Chapter 6 Response of Structures to Pulse-Type …

132
 MDOF Strength Demands for Constant Ductility
∆
Pulse P1, SRSS Pattern, without P-∆
4
µ===1
µ===2
µ===3
3 µ===4

η = Vy / ([Link],max)
µ===6
µ===8

0
0 0.5 1 1.5 2 2.5 3
T / Tp

MDOF Strength Demands for Constant Ductility

∆
Pulse P2, SRSS Pattern, without P-∆
5
µ===1
µ===2
4 µ===3
µ===4
η = Vy / ([Link],max)

µ===6
3 µ===8

0
0 0.5 1 1.5 2 2.5 3
T / Tp

MDOF Strength Demands for Constant Ductility

∆
Pulse P3, SRSS Pattern, without P-∆
6
µ===1
µ===2
5
µ===3
µ===4
η = Vy / ([Link],max)

4 µ===6
µ===8

0
0 0.5 1 1.5 2 2.5 3
T / Tp

Figure 6.25 MDOF Base Shear Strength Demand Spectra of Basic Pulses for Target
Story Ductility Ratios from 1 to 8

Chapter 6 Response of Structures to Pulse-Type …

133
 Strength Demand Ratios for Constant Ductility
Pulse P1, SRSS Pattern
5
µ===1
µ===2
4 µ===3
µ===4
µ===6

η MDOF=/=ηSDOF
3 µ===8

0
0 0.5 1 1.5 2 2.5 3
T / Tp

Strength Demand Ratios for Constant Ductility

Pulse P2, SRSS Pattern
5
µ===1
µ===2
4 µ===3
µ===4
µ===6
η MDOF=/=ηSDOF

3 µ===8

0
0 0.5 1 1.5 2 2.5 3
T / Tp

Strength Demand Ratios for Constant Ductility

Pulse P3, SRSS Pattern
5
µ===1
µ===2
4 µ===3
µ===4
µ===6
η MDOF=/=ηSDOF

3 µ===8

0
0 0.5 1 1.5 2 2.5 3
T / Tp

Figure 6.26 Ratios of MDOF to SDOF Strength Demands for Basic Pulses

Chapter 6 Response of Structures to Pulse-Type …

134
 MDOF Strength Demand Ratios for Constant Ductility
Pulse P1, SRSS Pattern
2.2

µ===1
µ===2
µ===3
1.8 µ===4
µ===6

η(P−∆)=/=η(no P−∆)
µ===8

1.4

0.6
0 0.5 1 1.5 2 2.5 3
T / Tp

MDOF Strength Demand Ratios for Constant Ductility

Pulse P2, SRSS Pattern
2.2

µ===1
µ===2
µ===3
1.8 µ===4
µ===6
η(P−∆)=/=η(no P−∆)

µ===8

1.4

0.6
0 0.5 1 1.5 2 2.5 3
T / Tp

MDOF Strength Demand Ratios for Constant Ductility

Pulse P3, SRSS Pattern
2.2

µ===1
µ===2
µ===3
1.8 µ===4
µ===6
η(P−∆)=/=η(no P−∆)

µ===8

1.4

0.6
0 0.5 1 1.5 2 2.5 3
T / Tp

Figure 6.27 Effect of P-Delta on MDOF Base Shear Strength Demands for Basic Pulses

Chapter 6 Response of Structures to Pulse-Type …

135
 MDOF Strength Demand Ratios for Constant Ductility
∆
Pulse P1, SRSS Pattern, no P-∆
1.6
µ===1
µ===2
µ===3
1.4 µ===4
µ===6

η (ξ = 2%) / η (ξ = 5%)
µ===8

1.2

0.8
0 0.5 1 1.5 2 2.5 3
T / Tp

MDOF Strength Demand Ratios for Constant Ductility

∆
Pulse P2, SRSS Pattern, no P-∆
1.6
µ===1
µ===2
µ===3
1.4 µ===4
µ===6
η(ξ = 2%) / η (ξ = 5%)

µ===8

1.2

0.8
0 0.5 1 1.5 2 2.5 3
T / Tp

MDOF Strength Demand Ratios for Constant Ductility

∆
Pulse P3, SRSS Pattern, no P-∆
1.6
µ===1
µ===2
µ===3
1.4 µ===4
µ===6
η(ξ = 2%) / η(ξ = 5%)

µ===8

1.2

0.8
0 0.5 1 1.5 2 2.5 3
T / Tp

Figure 6.28 Effect of Damping on MDOF Strength Demands for Basic Pulses

Chapter 6 Response of Structures to Pulse-Type …

136
 Normalized Inelastic SDOF Displacement Demands
Pulse P1, Bilinear, α = 3%, ξ = 2%
5
µ===2
µ===3
4
µ===4
µ===6

δinelastic=/=δelastic
µ===8
3

0
0 0.5 1 1.5 2 2.5 3
T / Tp

Normalized Inelastic SDOF Displacement Demands

Pulse P2, Bilinear, α = 3%, ξ = 2%
5
µ===2
µ===3
4 µ===4
µ===6
δ inelastic=/=δelastic

µ===8
3

0
0 0.5 1 1.5 2 2.5 3
T / Tp

Normalized Inelastic SDOF Displacement Demands

Pulse P3, Bilinear, α = 3%, ξ = 2%
5
µ===2
µ===3
4 µ===4
µ===6
δ inelastic=/=δelastic

µ===8
3

0
0 0.5 1 1.5 2 2.5 3
T / Tp

Figure 6.29 Ratios of Inelastic to Elastic SDOF Displacement Demands for Basic Pulses

Chapter 6 Response of Structures to Pulse-Type …

137
 Normalized Inelastic Roof Displacement Demands
Pulse P1, without P-∆
6
η===0.10
η===0.15
5 η===0.25
η===0.50
η===0.75
4 η===1.00

δ inelastic=/=δelastic
η===1.25
η===1.50
3 η===2.00

0
0 0.5 1 1.5 2 2.5 3
T / Tp

Normalized Inelastic Roof Displacement Demands

Pulse P2, without P-∆
6
η===0.10
η===0.15
5 η===0.25
η===0.50
η===0.75
4 η===1.00
δ inelastic=/=δelastic

η===1.25
η===1.50
3 η===2.00

0
0 0.5 1 1.5 2 2.5 3
T / Tp

Normalized Inelastic Roof Displacement Demands

Pulse P3, without P-∆
6
η===0.10
η===0.15
5 η===0.25
η===0.50
η===0.75
4 η===1.00
δ inelastic=/=δelastic

η===1.25
η===1.50
3 η===2.00

0
0 0.5 1 1.5 2 2.5 3
T / Tp

Figure 6.30 Ratios of Inelastic to Elastic Roof Displacement Demands for Basic Pulses

Chapter 6 Response of Structures to Pulse-Type …

138
 SDOF Displacement Demands
Pulse P2, Bilinear, α = 3%, ξ== 2%
4
T/Tp=0.375
T/Tp=0.50
T/Tp=0.75
3 T/Tp=1.00
T/Tp=1.50
η = Fy / ([Link],max)

T/Tp=2.00
T/Tp=3.00
2

0
0 0.5 1 1.5 2 2.5
δmax / ug,max

Figure 6.31 Inelastic SDOF Displacement Demands for Pulse P2

Roof Displacement Demands

∆
Pulse P2, 20-Story, SRSS Pattern, without P-∆
4
T/Tp=0.375
T/Tp=0.50
T/Tp=0.75
3 T/Tp=1.00
T/Tp=1.50
η = Vy / ([Link],max)

T/Tp=2.00
T/Tp=3.00
2

0
0 0.5 1 1.5 2 2.5
δroof,max / ug,max

Figure 6.32 Inelastic Roof Displacement Demands for Pulse P2

Chapter 6 Response of Structures to Pulse-Type …

139
 Story Drift Demands
∆
Pulse P2, 20-Story, SRSS Pattern, T / Tp = 0.5, without P-∆
1
η===2.50
η===2.00
η===1.50
0.8 η===1.25
η===1.00
η===0.75
η===0.60
Relative Height

0.6 η===0.50
η===0.40

0.4

0.2

0
0 0.02 0.04 0.06 0.08 0.1
δmax,i / ug,max

(a) T/Tp = 0.5

Story Drift Demands

∆
Pulse P2, 20-Story, SRSS Pattern, T / Tp = 2.0, without P-∆
1
η===1.25
η===1.00
η===0.75
0.8 η===0.50
η===0.25
η===0.15
η===0.10
Relative Height

0.6 η===0.07
η===0.05

0.4

0.2

0
0 0.1 0.2 0.3 0.4 0.5
δmax,i / ug,max

(b) T/Tp = 2.0

Figure 6.33 Story Drift Demands for Generic 20-Story Structure with Various Values of
η Subjected to Pulse P2, without P-Delta Effects

Chapter 6 Response of Structures to Pulse-Type …

140
 Story Ductility Demands
Pulse P1 and P1.1, T / Tp = 2.0, η = 0.5, without P-∆
∆
1

P1

0.8 P1.1
Relative Height

0.6

0.4

0.2

0
0 0.4 0.8 1.2 1.6 2
Story Ductility Ratio, µi = δmax,i / δy,i

(a) Strong Structure, η = 0.5

Story Ductility Demands

Pulse P1 and P1.1, T / Tp = 2.0, η = 0.15, without P-∆
∆
1

P1

P1.1
0.8
Relative Height

0.6

0.4

0.2

0
0 1 2 3 4 5 6 7 8
Story Ductility Ratio, µi = δmax,i / δy,i

(b) Weak Structure, η = 0.15

Figure 6.34 Comparison of Story Ductility Demands for Pulses P1 and P1.1, T/Tp = 2.0

Chapter 6 Response of Structures to Pulse-Type …

141
 Story Ductility Demands
Pulse P2 and P2.1, T / Tp = 2.0, η = 1.0, without P-∆
∆
1

P2
P2.1
0.8
Relative Height

0.6

0.4

0.2

0
0 0.5 1 1.5 2 2.5 3
Story Ductility Ratio, µi = δmax,i / δy,i

(a) Strong Structure, η = 1.0

Story Ductility Demands

Pulse P2 and P2.1, T / Tp = 2.0, η = 0.1, without P-∆
∆
1
P2

P2.1
0.8
Relative Height

0.6

0.4

0.2

0
0 2 4 6 8 10 12 14
Story Ductility Ratio, µi = δmax,i / δy,i

(b) Weak Structure, η = 0.1

Figure 6.35 Comparison of Story Ductility Demands for Pulses P2 and P2.1, T/Tp = 2.0

Chapter 6 Response of Structures to Pulse-Type …

142
 Maximum MDOF Ductility Demands
∆
Pulse P1, P1.1, without P-∆
2
T/Tp = 0.5, P1
T/Tp = 0.5, P1.1
1.6 T/Tp = 1.0, P1
T/Tp = 1.0, P1.1
T/Tp = 2.0, P1
η = Vy / [Link],max

1.2 T/Tp = 2.0, P1.1

0.8

0.4

0
0 4 8 12 16 20
Maximum Story Ductility, µmax

Figure 6.36 Comparison of Maximum Story Ductility Demands for Pulses P1 and P1.1

Maximum MDOF Ductility Demands

∆
Pulse P2, P2.1, without P-∆
2
T/Tp = 0.5, P2
T/Tp = 0.5, P2.1
1.6 T/Tp = 1.0, P2
T/Tp = 1.0, P2.1
T/Tp = 2.0, P2
η = Vy / [Link],max

1.2 T/Tp = 2.0, P2.1

0.8

0.4

0
0 4 8 12 16 20
Maximum Story Ductility, µmax

Figure 6.37 Comparison of Maximum Story Ductility Demands for Pulses P2 and P2.1

Chapter 6 Response of Structures to Pulse-Type …

143
 Story Ductility Demands Story Ductility Demands
∆
Pulse P4, SRSS Pattern, T / Tp = 0.5, without P-∆ ∆
Pulse P5, SRSS Pattern, T / Tp = 0.50, without P-∆
1
η===2.50 η===2.50
η===2.00 η===2.00
η===1.50 η===1.50
0.8 η===1.25 η===1.25
η===1.00 η===1.00
η===0.75 η===0.75
Relative Height

η===0.60 η===0.60
0.6 η===0.50 η===0.50
η===0.40 η===0.40

0.4

0.2

0
0 4 8 12 16 20 0 4 8 12 16 20
Story Ductility Ratio, µi = δmax,i / δy,i Story Ductility Ratio, µi = δmax,i / δy,i

T/Tp = 0.5

Story Ductility Demands Story Ductility Demands

∆
Pulse P4, SRSS Pattern, T / Tp = 1.0, without P-∆ ∆
Pulse P5, SRSS Pattern, T / Tp = 1.0, without P-∆
1
η===2.00 η===2.00
η===1.50 η===1.50
η===1.25 η===1.25
0.8 η===1.00 η===1.00
η===0.75 η===0.75
η===0.50 η===0.50
Relative Height

η===0.25 η===0.25
0.6 η===0.15 η===0.15
η===0.10 η===0.10

0.4

0.2

0
0 4 8 12 16 20 0 4 8 12 16 20
Story Ductility Ratio, µi = δmax,i / δy,i Story Ductility Ratio, µi = δmax,i / δy,i

T/Tp = 1.0

Story Ductility Demands Story Ductility Demands

∆
Pulse P4, SRSS Pattern, T / Tp = 2.0, without P-∆ ∆
Pulse P5, SRSS Pattern, T / Tp = 2.0, without P-∆
1
η===1.25 η===1.25
η===1.00 η===1.00
η===0.75 η===0.75
0.8 η===0.50 η===0.50
η===0.25 η===0.25
η===0.15 η===0.15
Relative Height

η===0.10 η===0.10
0.6 η===0.07 η===0.07
η===0.05 η===0.05

0.4

0.2

0
0 4 8 12 16 20 0 4 8 12 16 20
Story Ductility Ratio, µi = δmax,i / δy,i Story Ductility Ratio, µi = δmax,i / δy,i

T/Tp = 2.0
(a) Pulse P4 (b) Pulse P5

Figure 6.38 Story Ductility Demands for Pulses P4 and P5, Various Values of η

Chapter 6 Response of Structures to Pulse-Type …

144
 Maximum MDOF Story Ductility Demands
∆
Different Pulses, SRSS Pattern, T / Tp = 0.5, without P-∆
4
P2
P3
P4
3
P5
η = Vy / ([Link],max)

0
0 4 8 12 16 20
Maximum Story Ductility,,=µmax
=
(a) T/Tp = 0.5
=
Maximum MDOF Story Ductility Demands
∆
Different Pulses, SRSS Pattern, T / Tp = 2.0, without P-∆
4
P2
P3
P4
3
P5
η = Vy / ([Link],max)

0
0 4 8 12 16 20
Maximum Story Ductility,,=µmax
=
(b) T/Tp = 2.0

Figure 6.39 Comparison of Maximum Ductility Demands for MDOF Structures

Subjected to Pulses P2, P3, P4, and P5

Chapter 6 Response of Structures to Pulse-Type …

145
 SRSS Story Shear Force Pattern
Equal Interstory Drift
1
20-Story
9-Story
3-Story
0.8
Relative Height

0.6

0.4

0.2

0
0 0.2 0.4 0.6 0.8 1 1.2
Vi / Vbase

Figure 6.40 SRSS Story Shear Force Distributions for Generic 3-Story, 9-Story, and
20-Story Structures

Chapter 6 Response of Structures to Pulse-Type …

146
 Story Ductility Demands Story Ductility Demands
∆
Pulse P2, 3-Story, SRSS Pattern, T / Tp = 0.375, without P-∆ ∆
Pulse P2, 3-Story, SRSS Pattern, T / Tp = 0.75, without P-∆
1

Relative Height 0.8

0.6

0.4 η===3.00 η===2.00

η===2.00 η===1.25
η===1.50 η===1.00
η===1.20 η===0.50
0.2 η===0.30
η===1.00
η===0.75 η===0.25
η===0.60 η===0.20
0
0 4 8 12 16 20 0 4 8 12 16 20
Story Ductility Ratio, µi = δmax,i / δy,i Story Ductility Ratio, µi = δmax,i / δy,i

3-Story Structure

Story Ductility Demands Story Ductility Demands

∆
Pulse P2, 9-Story, SRSS Pattern, T / Tp = 0.375, without P-∆ ∆
Pulse P2, 9-Story, SRSS Pattern, T / Tp = 0.75, without P-∆
1
η===3.00 η===2.00
η===2.00 η===1.25
0.8 η===1.50 η===1.00
η===1.20 η===0.50
η===1.00 η===0.30
Relative Height

η===0.75 η===0.25
0.6
η===0.60 η===0.20

0.4

0.2

0
0 4 8 12 16 20 0 4 8 12 16 20
Story Ductility Ratio, µi = δmax,i / δy,i Story Ductility Ratio, µi = δmax,i / δy,i

9-Story Structure

Story Ductility Demands Story Ductility Demands

∆
Pulse P2, 20-Story, SRSS Pattern, T / Tp = 0.375, without P-∆ ∆
Pulse P2, 20-Story, SRSS Pattern, T / Tp = 0.75, without P-∆
1
η===3.00 η===2.00
η===2.00 η===1.25
0.8 η===1.50 η===1.00
η===1.20 η===0.50
η===1.00 η===0.30
Relative Height

η===0.75 η===0.25
0.6
η===0.60 η===0.20

0.4

0.2

0
0 4 8 12 16 20 0 4 8 12 16 20
Story Ductility Ratio, µi = δmax,i / δy,i Story Ductility Ratio, µi = δmax,i / δy,i

20-Story Structure
(a) T/Tp = 0.375 (b) T/Tp = 0.75

Figure 6.41 Comparison of Story Ductility Demands for 3-Story, 9-Story, and 20-Story
Structures, without P-Delta Effects

Chapter 6 Response of Structures to Pulse-Type …

147
 Story Ductility Demands Story Ductility Demands
∆
Pulse P2, 3-Story, SRSS Pattern, T / Tp = 1.0, without P-∆ ∆
Pulse P2, 3-Story, SRSS Pattern, T / Tp = 2.0, without P-∆
1
η===2.00
η===1.50
0.8 η===1.00
η===0.50
η===0.25
Relative Height

η===0.15
0.6
η===0.10

0.4 η===1.00
η===0.75
η===0.50
η===0.25
0.2
η===0.10
η===0.07
η===0.05
0
0 4 8 12 16 20 0 4 8 12 16 20
Story Ductility Ratio, µi = δmax,i / δy,i Story Ductility Ratio, µi = δmax,i / δy,i

3-Story Structure

Story Ductility Demands Story Ductility Demands

∆
Pulse P2, 9-Story, SRSS Pattern, T / Tp = 1.0, without P-∆ ∆
Pulse P2, 9-Story, SRSS Pattern, T / Tp = 2.0, without P-∆
1
η===2.00 η===1.00
η===1.50 η===0.75
0.8 η===1.00 η===0.50
η===0.50 η===0.25
η===0.25 η===0.10
Relative Height

η===0.15 η===0.07
0.6
η===0.10 η===0.05

0.4

0.2

0
0 4 8 12 16 20 0 4 8 12 16 20
Story Ductility Ratio, µi = δmax,i / δy,i Story Ductility Ratio, µi = δmax,i / δy,i

9-Story Structure

Story Ductility Demands Story Ductility Demands

∆
Pulse P2, 20-Story, SRSS Pattern, T / Tp = 1.0, without P-∆ ∆
Pulse P2, 20-Story, SRSS Pattern, T / Tp = 2.0, without P-∆
1
η===2.00 η===1.00
η===1.50 η===0.75
0.8 η===1.00 η===0.50
η===0.50 η===0.25
η===0.25 η===0.10
Relative Height

η===0.15 η===0.07
0.6
η===0.10 η===0.05

0.4

0.2

0
0 4 8 12 16 20 0 4 8 12 16 20
Story Ductility Ratio, µi = δmax,i / δy,i Story Ductility Ratio, µi = δmax,i / δy,i

20-Story Structure
(c) T/Tp = 1.0 (d) T/Tp = 2.0

Figure 6.41 (Cont’d) Comparison of Story Ductility Demands for 3-Story, 9-Story, and
20-Story Structures, without P-Delta Effects

Chapter 6 Response of Structures to Pulse-Type …

148
 Maximum Story Ductility Demands
∆
Pulse P2, T / Tp = 0.375, 0.75, without P-∆
5
T/Tp=0.375, 20-story
T/Tp=0.375, 9-story
4 T/Tp=0.375, 3-story
T/Tp=0.75, 20-story
η = Vy / ([Link],max)

T/Tp=0.75, 9-story
3
T/Tp=0.75, 3-story

0
0 4 8 12 16 20
Maximum Story Ductility,,=µmax

(a) η-µmax Domain

Maximum Story Ductility Demands

∆
Pulse P2, T / Tp = 0.375, 0.75, without P-∆
5
T/Tp=0.375, 20-story
T/Tp=0.375, 9-story
4 T/Tp=0.375, 3-story
1 / η = (W / Vy) ag,max / g

T/Tp=0.75, 20-story
T/Tp=0.75, 9-story
3
T/Tp=0.75, 3-story

0
0 4 8 12 16 20
Maximum Story Ductility,,=µmax

(b) 1/η-µmax Domain

Figure 6.42 Comparison of Maximum Story Ductility Demands for 3-Story, 9-Story, and
20-Story Structures, T/Tp = 0.375 and 0.75

Chapter 6 Response of Structures to Pulse-Type …

149
 MDOF Strength Demands for Constant Ductility
Pulse P2, SRSS Pattern, µmax = 1 and 2, ξ = 2%, without P-∆
∆
5
mu=1, 20-story
mu=1, 9-story
4 mu=1, 3-story
mu=2, 20-story
mu=2, 9-story
η = Vy / ([Link],max)

3 mu=2, 3-story

0
0 0.25 0.5 0.75 1 1.25 1.5 1.75 2
T / Tp

(a) µmax = 1 and 2

MDOF Strength Demands for Constant Ductility

Pulse P2, SRSS Pattern, µmax = 4 and 8, ξ = 2%, without P-∆
∆
1.6
mu=4, 20-story
mu=4, 9-story
mu=4, 3-story
1.2 mu=8, 20-story
mu=8, 9-story
η = Vy / ([Link],max)

mu=8, 3-story

0.8

0.4

0
0 0.25 0.5 0.75 1 1.25 1.5 1.75 2
T / Tp

(b) µmax = 4 and 8

Figure 6.43 Comparison of MDOF Base Shear Strength Demand Spectra for 3-Story,
9-Story, and 20-Story Structures

Chapter 6 Response of Structures to Pulse-Type …

150
 Roof Displacement Demands
∆
Pulse P2, T / Tp = 0.375 and 0.75, without P-∆
4
T/Tp=0.375, 20-story
T/Tp=0.375, 9-story
T/Tp=0.375, 3-story
3
T/Tp=0.75, 20-story
η = Vy / ([Link],max)

T/Tp=0.75, 9-story
T/Tp=0.75, 3-story
2

0
0 0.4 0.8 1.2 1.6
δroof,max / ug,max

(a) T/Tp = 0.375 and 0.75

Roof Displacement Demands

∆
Pulse P2, T / Tp = 1.0 and 2.0, without P-∆
4
T/Tp=1.0, 20-story
T/Tp=1.0, 9-story
T/Tp=1.0, 3-story
3
T/Tp=2.0, 20-story
η = Vy / ([Link],max)

T/Tp=2.0, 9-story
T/Tp=2.0, 3-story
2

0
0 0.5 1 1.5 2 2.5
δroof,max / ug,max

(a) T/Tp = 1.0 and 2.0

Figure 6.44 Comparison of Inelastic Roof Displacement Demands for 3-Story, 9-Story,
and 20-Story Structures

Chapter 6 Response of Structures to Pulse-Type …

151
 Normalized Story Drift Angle Demands Normalized Story Drift Angle Demands
∆
Pulse P2, 3-Story, SRSS Pattern, T / Tp = 0.375, without P-∆ ∆
Pulse P2, 3-Story, SRSS Pattern, T / Tp = 0.75, without P-∆
1
η===3.00 η===2.00
η===2.00 η===1.25
0.8 η===1.50 η===1.00
η===1.20 η===0.50
η===1.00 η===0.30
Relative Height

η===0.75 η===0.25
0.6 η===0.20
η===0.60

0.4

0.2

0
0 0.5 1 1.5 2 2.5 0 0.5 1 1.5 2 2.5
θi / θroof θi / θroof

3-Story Structure

Normalized Story Drift Angle Demands Normalized Story Drift Angle Demands
∆
Pulse P2, 9-Story, SRSS Pattern, T / Tp = 0.375, without P-∆ ∆
Pulse P2, 9-Story, SRSS Pattern, T / Tp = 0.75, without P-∆
1
η===3.00 η===2.00
η===2.00 η===1.25
0.8 η===1.50 η===1.00
η===1.20 η===0.50
η===1.00 η===0.30
Relative Height

η===0.75 η===0.25
0.6 η===0.20
η===0.60

0.4

0.2

0
0 0.5 1 1.5 2 2.5 0 0.5 1 1.5 2 2.5
θi / θroof θi / θroof

9-Story Structure

Normalized Story Drift Angle Demands Normalized Story Drift Angle Demands
∆
Pulse P2, 20-Story, SRSS Pattern, T / Tp = 0.375, without P-∆ ∆
Pulse P2, 20-Story, SRSS Pattern, T / Tp = 0.75, without P-∆
1
η===3.00 η===2.00
η===2.00 η===1.25
0.8 η===1.50 η===1.00
η===1.20 η===0.50
η===1.00 η===0.30
Relative Height

η===0.75 η===0.25
0.6
η===0.60 η===0.20

0.4

0.2

0
0 0.5 1 1.5 2 2.5 0 0.5 1 1.5 2 2.5
θi / θroof θi / θroof

20-Story Structure
(a) T/Tp = 0.375 (b) T/Tp = 0.75

Figure 6.45 Comparison of Normalized Story Drift Angle Demands for 3-Story, 9-Story,
and 20-Story Structures for T/Tp = 0.375 and 0.75

Chapter 6 Response of Structures to Pulse-Type …

152
 Story Ductility Demands Story Ductility Demands
∆
Pulse P2, CH Frame, SRSS Pattern, T / Tp = 0.5, without P-∆ ∆
Pulse P2, BH Frame, SRSS Pattern, T / Tp = 0.5, without P-∆
1
η===2.50 η===2.50
η===2.00 η===2.00
η===1.50 η===1.50
0.8 η===1.25 η===1.25
η===1.00 η===1.00
η===0.75 η===0.75
Relative Height

η===0.60 η===0.60
0.6 η===0.50 η===0.50
η===0.40 η===0.40

0.4

0.2

0
0 4 8 12 16 20 0 4 8 12 16 20
Story Ductility Ratio, µi = δmax,i / δy,i Story Ductility Ratio, µi = δmax,i / δy,i

T/Tp = 0.5

Story Ductility Demands Story Ductility Demands

∆
Pulse P2, CH Frame, SRSS Pattern, T / Tp = 1.0, without P-∆ ∆
Pulse P2, BH Frame, SRSS Pattern, T / Tp = 1.0, without P-∆
1
η===2.00 η===2.00
η===1.50 η===1.50
η===1.25 η===1.25
0.8 η===1.00 η===1.00
η===0.75 η===0.75
η===0.50 η===0.50
Relative Height

η===0.25 η===0.25
0.6 η===0.15 η===0.15
η===0.10 η===0.10

0.4

0.2

0
0 4 8 12 16 20 0 4 8 12 16 20
Story Ductility Ratio, µi = δmax,i / δy,i Story Ductility Ratio, µi = δmax,i / δy,i

T/Tp = 1.0

Story Ductility Demands Story Ductility Demands

∆
Pulse P2, CH Frame, SRSS Pattern, T / Tp = 2.0, without P-∆ ∆
Pulse P2, BH Frame, SRSS Pattern, T / Tp = 2.0, without P-∆
1
η===1.25 η===1.25
η===1.00 η===1.00
η===0.75 η===0.75
0.8 η===0.50 η===0.50
η===0.25 η===0.25
η===0.15 η===0.15
Relative Height

η===0.10 η===0.10
0.6 η===0.07 η===0.07
η===0.05 η===0.05

0.4

0.2

0
0 4 8 12 16 20 0 4 8 12 16 20
Story Ductility Ratio, µi = δmax,i / δy,i Story Ductility Ratio, µi = δmax,i / δy,i

T/Tp = 2.0
(a) CH Structure (b) BH Structure

Figure 6.46 Comparison of Story Ductility Demands for CH and BH Structures, without
P-Delta Effects

Chapter 6 Response of Structures to Pulse-Type …

153
 Maximum MDOF Story Ductility Demands
∆
Pulse P2, CH Frame, SRSS Pattern, without P-∆
4
T/Tp = 0.375
T/Tp = 0.50
T/Tp = 0.75
3 T/Tp = 1.00
T/Tp = 1.50
η = Vy / ([Link],max)

T/Tp = 2.00
T/Tp = 3.00

0
0 4 8 12 16 20
Maximum Story Ductility,,=µmax

Figure 6.47 Base Shear Strength vs. Max. Story Ductility Demands for CH Structure

MDOF Strength Demand Ratios for Constant Ductility

Pulse P2, SRSS Pattern
3
µ===1
µ===2
2.5 µ===3
µ===4
µ===6
2 µ===8
ηCH=/=ηBH

1.5

0.5

0
0 0.5 1 1.5 2 2.5 3
T / Tp

Figure 6.48 Comparison of Base Shear Strength Demands for BH and CH Structures

Chapter 6 Response of Structures to Pulse-Type …

154
 CHAPTER 7

REPRESENTATION OF NEAR-FAULT GROUND

MOTIONS BY EQUIVALENT PULSES

In Chapters 2 and 4 it was shown that near-fault ground motions come in large variations.
This variety very much complicates the evaluation or prediction of structural response
unless near-fault ground motions can be represented by a small number of simplified
motions that can reasonably replicate important near-fault response characteristics. An
inspection of the time history records (especially velocity and displacement) of near-fault
ground motions reveals their impulsive characteristics (see Appendix A). The study of
similarities between the response of structures subjected to near-fault records and simple
pulses also provides much evidence that, within limitations, near-fault records can be
represented by equivalent pulses of the type introduced in Chapter 5.

However, it is not reasonable to expect that perfect equivalence can be established

between near-fault ground motions and simple pulses for the full range of interest. Near-
fault records usually contain high frequency components that have little to do with the
characteristics of the long-period high-energy pulse generated by the propagation of fault
rupture. As will be shown in this chapter, in the very short period range these high
frequency contents make the pulse-record equivalence questionable. Furthermore, in the
very long period range it is also likely that other phenomena (e.g., basin effects or
instrument limitations) contaminate the record.

This chapter presents procedures that can be used to identify the parameters of the
predominant pulse contained in near-fault ground motions. Using these procedures,
equivalent pulses are established for the fault-normal component of the records with
forward directivity introduced in Chapter 2. Then, the capability of the equivalent pulses
to replicate the salient SDOF and MDOF response attributes of near-fault ground motions

Chapter 7 Representation of Near-Fault Ground …

155
is evaluated. Finally, equivalent pulses are established and evaluated for the rotated
components of the near-fault records.

7.1. Matching of Near-Fault Ground Motions to Equivalent Pulses

It is necessary to estimate the period range in which simple pulses can represent near-
fault records with reasonable confidence. As shown in Fig. 5.4, the elastic strength
demand (acceleration) spectra of the basic pulses demonstrate a relatively deep valley at
T/Tp = 0.25, where T is the fundamental period of the structure and Tp is the period of the
pulse. The valley, however, disappears in the inelastic strength spectra on account of
post-yield period elongation, and even turns into a wide peak at very large ductilities.
Furthermore, there are a number of close peaks and valleys in the elastic spectrum of the
simple pulses in the period range of T/Tp < 0.25. These observations are incompatible
with the spectra of recorded near-fault ground motions (see Appendix A), indicating
questionable equivalence between near-fault records and simple pulses in the very short
period range, i.e., T/Tp ≤ 0.25. Hence, it is postulated that the equivalence between a
near-fault record and a basic pulse can be reasonably established within the range of T/Tp
from 0.375 to 3.0.

Even though much effort has been devoted to establishing such equivalence through a
systematic procedure, it cannot be claimed that the outcome of this procedure is an
equivalent pulse that can perfectly represent the near-fault ground motion. For many of
the records used in this study an equivalent pulse has been successfully established, but
not all near-fault records have become part of this success. It is important to note that the
use of equivalent pulses is an approximation to a very complex problem. If this
approximation proves to be reasonably effective, it can significantly simplify the process
of predicting near-fault demands. The extent of this approximation is evaluated in
Section 7.2.

7.1.1. Parameters of Equivalent Pulses

The basic pulses introduced in Chapter 5, i.e., P1, P2, and P3 are utilized here as
equivalent pulses for near-fault ground motions. It was shown that the other pulse
motions (triangular pulses and pulses with different duration) are adequately represented
by the three basic pulses for practical purposes. In order to establish an equivalent pulse
for a record, three parameters need to be evaluated; pulse type (P1, P2, or P3), pulse

Chapter 7 Representation of Near-Fault Ground …

156
period Tp, and pulse intensity. The peak ground acceleration of the square wave history,
ag,max, is used for the latter purpose (see Figs. 5.1 to 5.3). These three parameters
completely characterize the equivalent pulse, and therefore its time history and response
properties can be derived accordingly. In the following discussion the ag,max and vg,max of
the equivalent pulse are referred to as the effective acceleration, aeff, and the effective
velocity, veff, respectively.

The equivalent pulse intensity can also be quantified by the (effective) peak velocity
because the pulse peak acceleration and velocity are related, i.e., veff = aeffTp/4. As shown
in the following section, the effective velocity is a more convenient pulse intensity
measure since it is usually close to the recorded peak ground velocity of the ground
motion. Nevertheless, the effective acceleration is more useful for design purposes due
to its direct relationship with the structure strength coefficient η = Vy/([Link]).

7.1.2. Procedure for Matching

In this study the equivalent pulse parameters are estimated using engineering rather than
seismological considerations. As pointed out before, the equivalent pulse is by no means
a precise representation for near-fault ground motions. In many cases, inspection,
common sense, and judgmental decisions need to be employed in order to arrive at a final
value for the equivalent pulse parameters. Some of these decisions are validated using a
sensitivity analysis.

Pulse Type and Pulse Period:

Mostly judgment is employed to decide on the pulse type, based on an inspection of the
time history trace, and on a comparison between ground motion and pulse spectral shapes
(primarily velocity and displacement spectra). An inspection of the ground displacement
and velocity time histories of the records (fault-normal component) investigated in this
research reveals that none of the near-fault ground motions exhibits motions of the type
represented by pulse P1, which has a half velocity cycle with a permanent ground
displacement. Therefore, only equivalent pulses of type P2 (full cycle) and P3 (multi-
cycle) are employed to represent the records.

Since the velocity response spectra of basic pulses P2 and P3 show a clear hump at T/Tp
= 1 (see Fig. 5.4), the pulse period Tp for a near-fault record is identified from the

Chapter 7 Representation of Near-Fault Ground …

157
location of a global and clear peak in the velocity response spectrum. Typical examples
are illustrated in Figs. 7.1 to 7.3, which show the velocity and displacement spectra of the
three basic pulses superimposed on the velocity and displacement spectra of near-fault
records NR94rrs, KB95kobj, and KB95tato. In these figures the pulse and ground motion
spectra are compared in a normalized domain. The period of the structure is normalized
by the pulse period and the spectral ordinates are normalized by corresponding time
history peak values. In most cases a narrow range for Tp could be identified rigorously.
But in some other cases, in which the peak is not clear or there are two or more peaks
(e.g., Fig. 7.3), judgment has to be employed to decide on a final value. A sensitivity
analysis is performed in Section 7.2.3 to evaluate these decisions.

Pulse Intensity:

Various procedures were investigated to determine the effective pulse acceleration aeff;
the simplest one being the estimation of aeff from the elastic displacement spectra
(equating pulse and ground motion spectral displacement at T = Tp). However,
inconsistent results were obtained when the so estimated values were used to compute
ductility demands for MDOF systems. The reason is that when aeff is determined based
on elastic spectra only, no consideration is given to inelastic MDOF response
characteristics. Ultimately, a rigorous process is employed whose objective is to
minimize the differences between the maximum story ductility demands obtained from
the near-fault ground motion and the equivalent pulse. In summary the procedure
includes the following steps:

1. Compute the η−µmax curves for the appropriate pulse type for T/Tp = 0.375, 0.5,
0.75, 1.0, 1.5, 2.0, and 3.0.
2. Compute the γ−µmax curves for the near-fault record for T/Tp = 0.375, 0.5, 0.75,
1.0, 1.5, 2.0, and 3.0.
3. For each T/Tp value, convert the η-µmax curve of the pulse into a γ−µmax curve [γ =
Vy/(m.g) = (aeff/g)η] and find best-fit values for aeff by minimizing the relative
differences between the two γ−µmax curves. The minimization technique is
explained in more detail next.
4. Obtain final values for aeff by averaging the best-fit values for the seven period
ratios.

Chapter 7 Representation of Near-Fault Ground …

158
The relative difference between the γ−µmax curves of the pulse and near-fault record at a
given ductility value of µi is defined as:

γ i − γ iP
E
γ iP ηiP
ei = = 1− = 1 − (a eff / g ) (7.1)
γi γi γi
E E E

where γiE and γiP represent the strength coefficients corresponding to the ground motion
and the pulse, respectively, at µi. The difference (error) values ei are calculated for
discrete ductility ratios in the range of interest. Then, using the least-squares method, the
best-fit value for aeff is evaluated such that the differences between the two curves are
minimized, i.e., Σ(ei)2 is minimum. But, minimizing the sum of the squared errors is the
same as minimizing the mean of the squared errors, which is equal to:

Mean[(ei)2] = Var[ei] + (Mean[ei])2 = (Mean[ei])2(1+Ve2) (7.2)

where Var[ei] is the variance of the error and Ve represents the coefficient of variation of
the error in the range in which the difference between the two curves is minimized. This
is again equivalent to minimizing the square root of the quantity shown in Eq. 7.2, which
can be written as:

Mean[(e i ) 2 ] = (Mean[e i ]) 1 + Ve2 (7.3)

This shows that by using the least-squares method, the quantity given by Eq. 7.3, which
is a combination of the mean error and the error dispersion, is minimized. This quantity
can be used as a good measure to assess the final proximity of the two γ-µmax curves
resulting from the error minimization procedure. Smaller values for this quantity imply a
closer match between the two curves.

Through the procedure described above, values for aeff can be obtained that minimize the
differences between the γ-µmax curves corresponding to the pulse and record in a given
range of ductility, µmax. In this study the following ranges of µmax are investigated:

•= µmax = 1 to 10, which covers the full range of interest.

•= µmax = 4 to 10, which represents behavior at a low performance level.
•= µmax = 1.0, which represents behavior at a high performance level.

Chapter 7 Representation of Near-Fault Ground …

159
The results of this procedure for the fault-normal component of the recorded near-fault
ground motions introduced in Chapter 2 are presented in Table 7.1. For each ductility
range and T/Tp ratio, the best-fit value for aeff is shown as well as the quantity given by
Eq. 7.3, the mean error, and the error coefficient of variation. The results are tabulated
only for the near-fault records with forward directivity. The Landers (Lucerne) record
(LN92lucr) is omitted from Table 7.1 because it has a pulse period longer than 4 seconds
(see the velocity spectrum in Fig. A.2), which may be contaminated by instrument errors.
Besides, the matching procedure for this record would require computing the demands
for structures with unreasonably long periods (e.g., T = 3Tp > 12.0 sec.).

As explained in the last step of the procedure, the final value for aeff is arrived at by
averaging the best-fit values for the seven period ratios. Figure 7.4 illustrates the
variations of the aeff values obtained for the recorded ground motions using different T/Tp
values and the full period range of interest (1 to 10). For each record the aeff values are
normalized by their average. To make the picture more readable, the results are
presented in two separate graphs, each corresponding to 7 near-fault records. As can be
seen, the aeff values obtained from the matching procedure are not very sensitive to T/Tp.
This justifies averaging the values in order to arrive at a single pulse intensity measure
for a given near-fault ground motion.

The results presented in Table 7.1 are summarized in Table 7.2 by averaging the aeff
values obtained for various T/Tp ratios for the recorded ground motions. The table lists
pulse type and period, as well as effective acceleration and peak velocity of the
equivalent pulse for the three ductility ranges. As expected, the values of aeff differ
somewhat but not by a large amount among the three ductility ranges. Having estimates
of Tp and aeff, the peak velocity of the pulse can be computed (veff = aeffTp/4). Comparing
the so computed veff with the peak ground velocity PGV listed in the last column of Table
7.2 indicates that in most of the cases the peak velocity of the equivalent pulse is within
20% of the PGV of the near-fault record. Thus, it appears to be feasible to use the peak
ground velocity of the near-fault record, PGV, to estimate the pulse intensity parameter
(i.e., aeff = 4PGV/Tp) rather than employing the complex procedure outlined here.

A summary of the results for the fault-normal component of the simulated ground
motions with forward directivity (see Chapter 2), derived in the same manner as for the
recorded ground motions, is listed in Table 7.3. Again, the results indicate that in most of

Chapter 7 Representation of Near-Fault Ground …

160
the cases the peak velocity of the equivalent pulse is very close to the peak ground
velocity (PGV) of the ground motion record.

7.2. Evaluation of Equivalent Pulses

In the previous section the parameters of equivalent pulses were identified for the near-
fault records investigated in this study. An important issue that needs to be addressed is
to what extent this representation is reasonable, practical, and reliable. It should be
considered that simple pulse representation of complex near-fault records is not expected
to be precise because near-fault ground motions are affected by many complex
seismological phenomena. The objective is not to develop pulses that can accurately
replicate recorded ground motions, but to develop pulses that can reasonably simulate
predominant response characteristics of structures located in the near-fault region of a
seismic source.

In this section demands for structures subjected to near-fault records are compared to the
corresponding demands obtained from the equivalent pulses of those records. The
equivalent pulse parameters identified for the ductility range from 1 to 10 are utilized for
this purpose. The main goal is to evaluate the quality of the pulse-record equivalence
established previously.

7.2.1. Comparison of SDOF Response Time Histories

In this section elastic and inelastic responses of SDOF systems subjected to the near-fault
record NR94rrs are compared qualitatively with the corresponding responses for Pulse
P2. As listed in Table 7.2, the near-fault record NR94rrs has an equivalent pulse of type
P2 with a period of Tp = 1.0 sec. Figure 7.5 illustrates the inelastic displacement
response time history normalized by the maximum ground displacement for Pulse P2 and
T/Tp ratios of 1.0 and 2.0. Each graph represents the ground displacement time history
(denoted as “Ugr”) together with the displacement response time history (denoted as
“Ust”) of a system with a ductility of µ = 6. Figure 7.6 shows the ground and response
time histories for the near-fault record NR94rrs in a similar manner. The following
observations can be made:

•= The ground displacement time history of NR94rrs looks different from the time
history of P2, with the greatest difference being its two-sided displacement.

Chapter 7 Representation of Near-Fault Ground …

161
 However, it should be noted that the ground displacement value at the first
negative peak (at t = 2.0 sec.) is almost half the value at the following positive
peak. Furthermore, the slight slope of the ground displacement time history in the
time period t ≤ 2.0 sec. implies very low ground velocity, whereas the following
impulsive motion (2.0 sec. < t < 3.0 sec.) is associated with a much higher ground
velocity. Therefore, the positive ground displacement pulse, which has a peak at
about t = 2.7 sec., will most likely dominate over the initial negative motion, and
control the response of the structure. The response time histories, superimposed
on the ground time histories, support this hypothesis. It is observed that for both
T/Tp values, before the positive pulse strikes the structure (t < 2.0 sec.), the
response displacement is negligible, and then rapidly increases.

•= An inspection of the displacement response time histories for pulse P2 (Fig. 7.5)
and the record NR94rrs (Fig. 7.6) reveals that there is a clear correlation between
the response to the near-fault ground motion and its equivalent pulse. Since Tp is
1.0 sec. for NR94rrs, the time scales of the two displacement time histories are
directly comparable. There are, however, some differences, which should be
expected when such simple pulse shapes are utilized. For example, for T/Tp = 2.0
the residual displacement for the pulse is positive, whereas the residual
displacement for the record is negative. The reason is that unlike the pulse, the
ground in the record does not return rapidly to its original position (ug = 0) after
the positive peak.

Figure 7.7 provides a comprehensive comparison of the elastic displacement response

time histories for pulse P2 and near-fault record NR94rrs. The elastic displacement time
history of each structure is normalized by its maximum value. The time scales of the two
graphs are comparable since Tp = 1.0 sec. for NR94rrs. To facilitate the comparison of
the two sets of time histories, the time scale for the record starts at t = 2.1 sec., the
moment at which the predominant displacement pulse of the near-fault ground motion
comes into play. Despite existing differences, a relatively close correlation is evident
between the corresponding time histories at all T/Tp values.

A similar picture is presented in Fig. 7.8 for inelastic SDOF systems with a ductility of 6.
The inelastic displacement time history for each structure is normalized by its yield
displacement. The similarities between the displacement time histories for the near-fault
record and pulse are apparent. A closer match between the responses to the near-fault

Chapter 7 Representation of Near-Fault Ground …

162
record and pulse is observed for more flexible structures (particularly elastic), which is
verified by an inspection of the response spectra.

7.2.2. Comparison of Story Ductility Demands

Figure 7.9 compares the converted γ-µmax curve for the equivalent pulse with the γ-µmax
curve for records NR94rrs and KB95tato. The averaged aeff value obtained in the
ductility range of 1 to 10 is used to convert the pulse η-µmax curve into a γ-µmax curve.
This comparison is made for three period ratios T/Tp = 0.5, 1.0, and 2.0. In general the
results show a reasonable agreement between the two γ-µmax curves indicating that the
equivalent pulse appears to be capable of replicating maximum story ductility demands
for MDOF systems subjected to near-fault records with reasonable accuracy. This
compatibility is assessed more comprehensively in the following section using MDOF
strength demand spectra for given story ductility ratios.

Examples of the distribution of story ductility demands over the height of the structure
obtained from a near-fault record and the equivalent pulse are presented in Figs. 7.10 and
7.11 for cases of small and large ductility demands. The results are presented for
structures with different T/Tp values, subjected to near-fault records NR94rrs and
KB95kobj. Although some differences exist, it is believed that the equivalent pulse
captures the important response characteristics of the near-fault records, particularly the
migration of ductility demands from the top to the bottom portion of flexible structures
(T/Tp = 2.0).

7.2.3. Sensitivity to Pulse Type and Period

In the process of identifying equivalent pulses, judgment had to be employed in many

cases to decide on the pulse type and a final value for the pulse period Tp. Those
decisions are evaluated here using a sensitivity analysis, which assesses the quality of the
pulse-record equivalence when different pulse types or different values for Tp are
adopted. In each case the pulse intensity is computed again using the selected pulse type
and Tp, and the matching procedure described in Section 7.1.2. The final objective is to
determine which alternative for the pulse type or Tp value leads to a closer match
between the record and equivalent pulse.

Chapter 7 Representation of Near-Fault Ground …

163
In order to assess the quality of the match, base shear strength demands for target story
ductility ratios are utilized. As shown in Sections 4.2.2 and 6.2.2, the base shear strength
required to limit story ductility demands to a given value can be obtained from γ-µmax (for
records) or η-µmax (for pulses) curves employing a linear interpolation scheme. As a
result, base shear strength demand spectra (γ-T curves), such as those presented in Fig.
4.19, are obtained. If the parameters of the equivalent pulse (Tp and aeff) are known, the
γ-T curves can be converted into η-T/Tp curves using the relation between γ and η, i.e., η
= γ/(aeff/g). Examples of so converted η-T/Tp curves are presented in Fig. 7.12 for near-
fault records NR94rrs and KB95kobj.

These graphs are directly comparable with base shear strength demand spectra (η-T/Tp
curves) for the basic pulses, i.e., Fig. 6.25. Thus, if the ordinates of the record η-T/Tp
curves are divided by the corresponding values of the pulse η-T/Tp curves, the resulting
ratios can be used to assess the capability of the equivalent pulse to estimate the base
shear strength demands for a structure subjected to the near-fault record for various T/Tp
and story ductility values. Examples of these ratios are shown in Fig. 7.13, which
compares the record KB95kobj with equivalent pulses with a Tp value of 0.9 sec. and of
type P2 and P3. A ratio larger than 1.0 for given T/Tp and µ values indicates that in order
to limit story ductility demands to µ, more base shear strength is needed for the record
than for the equivalent pulse. In other words, the equivalent pulse underestimates the
required base shear strength in this particular case. A perfect match between the record
and equivalent pulse is implied when the record-to-pulse strength ratio is equal to one.
The following observations can be made from Fig. 7.13:

•= Figure 7.13 compares P2 and P3 as the equivalent pulse for the near-fault record
KB95kobj to determine which one better represents this ground motion. As can
be seen, there are differences between the near-fault record and equivalent pulses
of both types. However, overall P3 is a better representative for the near-fault
record KB95kobj.

•= Similar to this record, many other ground motions investigated in this study
exhibit large record-to-pulse strength ratios for a story ductility of 4 and T/Tp >
1.0, indicating that the equivalent pulse underestimates the base shear strength
demands. The reason lies in the migration phenomenon that occurs in long period
structures. As the strength of the structure is reduced, the migration of high
ductility demands to the bottom of the structure, which starts at a story ductility

Chapter 7 Representation of Near-Fault Ground …

164
 ratio between 3 to 4, occurs at a somewhat faster rate for the pulse than for the
ground motion. Therefore, the story ductility of 4 at the bottom of the structure is
reached at a smaller base shear strength value for the pulse. This is only a local
discrepancy and does not have an impact on the strength demands for other
ductility values.

Another issue that needs to be addressed is the sensitivity of the equivalent pulse
representation to the value chosen for Tp. As discussed in Section 7.1.2, the equivalent
pulse period, Tp, is identified based on the location of a clear and global peak in the
velocity response spectrum. But not all near-fault ground motions exhibit a single clear
peak in their velocity spectra. For example, the velocity spectrum of the fault-normal
component of the record NR94rrs contains two humps at periods around 1.0 and 1.35 sec.
(see Fig. 2.5). In order to choose one of these two candidates for the period, Fig. 7.14
illustrates the record-to-pulse strength demand ratios for this record assuming that the
value of Tp is (a) 1.0, and (b) 1.35 sec. It appears that Tp = 1.0 sec. leads to a closer
match between the equivalent pulse P2 and the near-fault record NR94rrs.

7.3. Equivalent Pulse for Rotated Components

In Chapter 2 it was shown that rotated components of near-fault ground motions (with
respect to the fault direction) have spectral values nearly as large as those associated with
the fault-normal component. Time history traces indicate that the rotated components
have pulse-type characteristics similar to those of the fault-normal component. These
pulse-type characteristics and their effects on MDOF response can be investigated more
rigorously using the procedures introduced in this chapter for detection and identification
of equivalent pulses. In this section attempts are made to identify equivalent pulses for
the 45° components of the near-fault record NR94rrs.

Figure 2.5 displays a global peak in the velocity spectra of both 45° components of
NR94rrs, indicating that as with the fault-normal component, the rotated components also
contain pulses of a period around Tp = 1.0 sec. Similar to the fault-normal component,
the rotated components are represented by equivalent pulses of type P2. Following the
procedure presented in Section 7.1.2 for estimating the equivalent pulse intensity,
effective pulse acceleration values of 0.58g and 0.47g are obtained for the 45°
components (aeff = 0.72g for the fault-normal component). In other words, the intensity
of the pulse contained in one of the rotated components is almost 80% of the intensity of

Chapter 7 Representation of Near-Fault Ground …

165
the pulse contained in the fault-normal component. This signifies that 45° components of
near-fault records may still be quite severe.

Figure 7.15 presents the record-to-pulse strength demand ratios for the rotated
components of NR94rrs to illustrate the quality of the match between these records and
the equivalent pulses identified previously. A comparison of these results with those
corresponding to the fault-normal component (Fig. 7.14(a)) reveals that the response
characteristics of the rotated record with aeff = 0.58 are very similar to those of the fault-
normal component. Furthermore, this rotated component is relatively well represented by
the equivalent pulse.

7.4. Estimation of Structure Response to Near-Fault Ground Motions

An accurate evaluation of structure response to near-fault ground motions requires

dynamic analyses that are computationally expensive. Since every near-fault record has
unique properties, the response results for one ground motion are not useful in evaluating
the response to a different ground motion. The equivalence established in this chapter
between near-fault ground motions and simple pulses can be utilized in order to estimate
seismic demands for MDOF structures with little effort. A multi-step procedure is
presented in this section that can be used to estimate maximum story ductility, roof drift,
and story drift angle demands for a given frame structure subjected to a specific near-
fault record. The response of structures to simple pulses presented in Chapter 6 provides
basic information for this procedure, which consists of the following steps:

1. Obtain the following structural properties:

•= Fundamental period, T
•= Base shear strength coefficient, γ = Vy/W

2. Obtain the following ground motion properties:

•= Peak ground velocity, PGV
•= Pulse type (from the time history and spectral shapes, see Section 7.1.2)
•= Pulse period, Tp (from the velocity spectrum, see Section 7.1.2)
•= Pulse intensity, aeff = 4veff/Tp (use veff = PGV)

Chapter 7 Representation of Near-Fault Ground …

166
 3. Calculate the following quantities:
•= Period ratio, T/Tp
•= Pulse strength parameter, η = (g/aeff)γ

Presuming that the given near-fault ground motion is best represented by pulse P2, the
demands for structures subjected to P2, as presented in Chapter 6, are used to estimate
seismic demands as follows:

4. Using the T/Tp and η values obtained in step 3, estimate the maximum story
ductility demands from Fig. 7.16 (repetition of part of Fig. 6.21). For T/Tp values
between those given in the graph, the maximum story ductility demand can be
obtained using linear interpolation.

5. Using the T/Tp and η values, and computing ug,max = [Link]/4, estimate the roof
displacement demand, δroof,max, from Fig. 7.17 (repetition of Fig. 6.32). For T/Tp
values between those given in the graph, the roof displacement demand can be
obtained using linear interpolation.

6. Using the T/Tp and η values, estimate story drift demands, ui, as follows:
•= Calculate roof drift angle, θroof = δroof,max/H, where H is the structure
height.
•= Obtain the drift angle ratio for the story under consideration, λ = θi/θroof,
from Fig. 7.18. Interpolate for T/Tp values between those given in the
graphs.
•= Calculate ui = λhiθroof, where hi is the height of the story under
consideration.

It is important to note that P-delta effects are not accounted for in the figures referred to
in this section. While maximum story ductility demands and especially roof
displacement demands are not very sensitive to the number of stories, the distribution of
story drift demands over the height can be significantly affected by the number of stories
(see Section 6.5). Therefore, for structures with a small number of stories, Fig. 7.18,
which represents demands for a 20-story structure, may not provide a good estimate of
the story drift demand. Figure 6.45 helps to provide improved estimates for structures
with fewer than 20 stories.

Chapter 7 Representation of Near-Fault Ground …

167
Table 7.1 Properties of Equivalent Pulses for Recorded Near-Fault Ground Motions

µ=1 µ = (4 - 10) µ = (1 - 10)

Rec/Pls T / Tp aeff (g) µe^2) µe, %
aeff (g) sqr(µ Ve, % µe^2) µe, %
aeff (g) sqr(µ Ve, %
0.375 0.35 0.24 0.109 10.0 41.5 0.25 0.130 9.9 85.5
LP89lgpc 0.50 0.73 0.20 0.094 6.5 103.5 0.22 0.301 23.2 82.8
0.75 0.24 0.25 0.042 2.9 102.8 0.23 0.121 10.7 52.4
1.00 0.18 0.27 0.044 3.9 47.7 0.22 0.236 20.6 56.2
P3 1.50 0.33 0.21 0.089 6.7 87.8 0.23 0.214 18.1 63.3
Tp = 3.0 sec 2.00 0.25 0.20 0.116 9.0 79.7 0.21 0.195 15.9 71.3
3.00 0.21 0.28 0.138 8.6 125.5 0.28 0.210 14.7 102.3
0.375 0.20 0.50 0.047 3.4 94.7 0.38 0.297 24.6 68.2
LPP89lex 0.50 0.28 0.52 0.044 3.4 79.9 0.46 0.159 12.9 72.0
0.75 0.48 0.55 0.049 4.3 54.1 0.55 0.088 7.1 73.2
1.00 0.85 0.84 0.039 3.0 80.1 0.76 0.130 11.4 54.9
P2 1.50 0.75 1.26 0.023 2.0 63.4 0.99 0.253 24.3 28.7
Tp = 1.0 sec 2.00 0.69 1.37 0.010 0.8 74.0 0.98 0.321 30.8 29.8
3.00 0.78 1.18 0.019 1.6 65.8 0.79 0.369 34.4 39.4
0.375 0.22 0.24 0.224 20.0 49.9 0.24 0.218 19.6 48.0
EZ92erzi 0.50 0.26 0.16 0.111 8.4 86.1 0.18 0.211 18.9 50.6
0.75 0.18 0.12 0.123 10.3 66.0 0.13 0.186 16.5 52.4
1.00 0.18 0.19 0.166 14.8 50.9 0.17 0.205 17.2 65.3
P2 1.50 0.19 0.24 0.016 1.3 79.5 0.18 0.296 26.3 51.6
Tp = 2.3 sec 2.00 0.18 0.23 0.021 1.8 57.1 0.17 0.317 28.3 50.4
3.00 0.17 0.23 0.037 3.2 58.0 0.15 0.383 35.4 41.7
0.375 0.45 0.74 0.030 2.5 66.0 0.67 0.164 12.4 86.9
NR94rrs 0.50 0.52 0.70 0.047 4.2 54.1 0.66 0.090 7.6 63.0
0.75 0.56 0.58 0.033 2.5 86.1 0.57 0.051 4.0 81.6
1.00 0.60 0.63 0.050 4.3 61.9 0.64 0.045 3.7 70.3
P2 1.50 0.65 0.91 0.023 2.0 55.6 0.83 0.165 13.9 63.0
Tp = 1.0 sec 2.00 0.65 0.95 0.026 2.4 43.9 0.84 0.191 17.0 50.9
3.00 0.65 0.93 0.038 2.3 136.9 0.81 0.206 18.5 48.8
0.375 0.26 0.19 0.025 2.0 79.5 0.21 0.186 15.9 60.0
NR94sylm 0.50 0.45 0.16 0.191 14.3 87.8 0.18 0.359 31.3 56.2
0.75 0.17 0.13 0.048 3.0 123.9 0.13 0.151 10.1 111.6
1.00 0.13 0.16 0.057 4.7 70.6 0.15 0.119 9.9 67.5
P3 1.50 0.22 0.21 0.052 4.4 61.6 0.21 0.103 7.0 108.2
Tp = 2.4 sec 2.00 0.19 0.20 0.077 7.0 47.6 0.19 0.110 8.4 85.0
3.00 0.19 0.24 0.079 6.6 64.4 0.19 0.258 23.5 44.6
0.375 0.66 0.76 0.073 5.7 78.0 0.69 0.124 11.2 46.5
KB95kobj 0.50 1.32 0.83 0.031 2.6 65.2 0.89 0.143 11.3 78.0
0.75 0.79 0.70 0.109 9.9 45.2 0.70 0.109 9.4 57.7
1.00 0.67 0.79 0.048 4.3 49.1 0.81 0.086 6.7 80.2
P3 1.50 0.86 1.03 0.016 1.4 66.9 0.98 0.106 7.5 100.3
Tp = 0.9 sec 2.00 1.05 1.07 0.124 9.4 87.2 1.08 0.134 10.4 80.4
3.00 0.82 0.86 0.094 5.1 155.5 0.90 0.145 11.6 74.7
0.375 0.45 0.62 0.038 3.3 53.2 0.58 0.086 6.7 82.3
KB95tato 0.50 0.76 0.52 0.080 7.4 43.8 0.55 0.125 10.6 62.6
0.75 0.36 0.26 0.023 1.3 155.3 0.28 0.119 10.3 58.9
1.00 0.30 0.29 0.148 11.9 74.7 0.31 0.210 17.5 65.8
P3 1.50 0.44 0.33 0.029 2.1 96.1 0.35 0.123 10.1 70.5
Tp = 2.0 sec 2.00 0.37 0.31 0.050 4.3 56.8 0.33 0.128 9.5 90.8
3.00 0.44 0.36 0.097 7.9 70.4 0.39 0.225 18.2 72.9

Chapter 7 Representation of Near-Fault Ground …

168
Table 7.1 (Cont’d) Properties of Equiv. Pulses for Recorded Near-Fault Ground Motions

µ=1 µ = (4 - 10) µ = (1 - 10)

Rec/Pls T / Tp aeff (g) µe^2) µe, %
aeff (g) sqr(µ Ve, % µe^2) µe, %
aeff (g) sqr(µ Ve, %
0.375 0.10 0.15 0.021 1.8 57.9 0.13 0.156 13.3 61.6
IV79ar06 0.50 0.12 0.14 0.063 5.4 57.8 0.13 0.104 9.4 46.3
0.75 0.12 0.11 0.036 3.3 48.1 0.11 0.049 3.6 90.3
1.00 0.15 0.14 0.140 12.6 48.0 0.14 0.130 11.8 45.6
P2 1.50 0.13 0.18 0.028 2.5 54.1 0.13 0.325 28.1 58.5
Tp = 3.4 sec 2.00 0.12 0.18 0.090 7.4 68.9 0.13 0.333 29.3 53.3
3.00 0.13 0.18 0.133 11.8 52.0 0.13 0.368 30.6 67.1
0.375 0.11 0.17 0.087 7.8 48.5 0.15 0.173 12.8 90.6
IV79melo 0.50 0.16 0.15 0.060 4.7 78.2 0.15 0.081 6.8 66.8
0.75 0.17 0.12 0.141 12.1 61.0 0.12 0.152 13.5 52.0
1.00 0.16 0.18 0.173 15.5 48.9 0.15 0.246 21.9 50.8
P2 1.50 0.15 0.21 0.051 4.0 81.9 0.15 0.307 28.6 38.8
Tp = 2.8 sec 2.00 0.14 0.20 0.061 5.3 58.4 0.16 0.259 24.0 41.0
3.00 0.14 0.22 0.128 10.0 80.6 0.18 0.257 23.7 42.5
0.375 0.17 0.30 0.033 2.5 83.9 0.25 0.232 18.4 77.2
KB95kpi1 0.50 0.31 0.27 0.028 2.2 77.0 0.27 0.059 4.4 91.3
0.75 0.32 0.18 0.097 8.7 49.2 0.18 0.162 12.2 87.9
1.00 0.27 0.23 0.146 12.0 70.2 0.20 0.189 17.1 46.0
P2 1.50 0.26 0.33 0.096 8.9 40.8 0.28 0.230 19.7 60.5
Tp = 1.8 sec 2.00 0.30 0.30 0.055 4.0 93.9 0.29 0.172 11.6 109.3
3.00 0.27 0.32 0.079 6.9 55.8 0.26 0.243 21.1 57.6
0.375 0.18 0.18 0.049 2.9 139.0 0.19 0.099 7.6 83.4
MH84andd 0.50 0.27 0.17 0.059 5.4 46.2 0.17 0.133 9.4 100.1
0.75 0.14 0.10 0.153 11.6 84.5 0.11 0.179 16.8 38.3
1.00 0.21 0.11 0.105 9.1 57.4 0.12 0.211 14.4 107.0
P2 1.50 0.21 0.16 0.051 3.6 101.7 0.15 0.196 11.6 136.2
Tp = 0.8 sec 2.00 0.13 0.21 0.070 6.0 60.5 0.17 0.256 23.6 42.8
3.00 0.19 0.27 0.066 5.1 83.9 0.24 0.159 14.0 54.8
0.375 0.42 0.43 0.082 7.5 43.2 0.44 0.074 6.4 58.4
MH84cyld 0.50 0.72 0.38 0.158 15.0 34.1 0.42 0.233 19.3 67.6
0.75 0.43 0.47 0.062 4.4 99.9 0.42 0.153 13.0 62.3
1.00 0.37 0.57 0.033 2.5 84.8 0.48 0.217 17.6 72.2
P3 1.50 0.36 0.59 0.021 1.7 76.4 0.49 0.218 18.9 58.0
Tp = 0.8 sec 2.00 0.51 0.62 0.029 2.5 59.9 0.55 0.145 13.0 48.8
3.00 0.37 0.52 0.179 11.9 113.2 0.52 0.219 15.9 94.9
0.375 0.44 0.49 0.114 10.0 55.1 0.49 0.121 10.9 46.7
NR94newh 0.50 0.65 0.38 0.031 2.8 51.6 0.41 0.166 11.4 106.1
0.75 0.36 0.20 0.078 6.3 73.6 0.21 0.199 13.3 110.9
1.00 0.42 0.27 0.201 18.1 48.6 0.25 0.279 25.1 48.6
P2 1.50 0.43 0.43 0.076 6.9 48.2 0.35 0.271 23.6 55.6
Tp = 1.3 sec 2.00 0.39 0.52 0.048 4.0 68.6 0.39 0.308 26.2 61.4
3.00 0.36 0.58 0.052 4.7 49.9 0.48 0.260 23.4 49.0
0.375 0.19 0.10 0.289 26.9 38.6 0.11 0.337 32.7 24.7
NR94spva 0.50 0.30 0.07 0.131 8.9 108.8 0.08 0.352 29.6 64.3
0.75 0.10 0.06 0.043 3.4 76.7 0.06 0.222 14.4 116.8
1.00 0.07 0.08 0.063 5.6 52.0 0.07 0.134 10.8 74.6
P3 1.50 0.11 0.10 0.068 5.2 83.0 0.10 0.080 6.2 82.5
Tp = 2.7 sec 2.00 0.11 0.09 0.226 21.0 40.4 0.10 0.261 23.2 50.6
3.00 0.10 0.07 0.065 6.2 29.6 0.08 0.097 8.5 52.9

Chapter 7 Representation of Near-Fault Ground …

169
 Table 7.2 Equivalent Pulses for Recorded Near-Fault Ground Motions (Summary)

µ=1 µ = (4-10) µ = (1-10)

Designation Mw R (km) Pulse Type Tp (sec) aeff (g) veff (cm/sec) aeff (g) veff (cm/sec) aeff (g) veff (cm/sec) PGV
LP89lgpc 7.0 3.5 P3 3.0 0.33 241 0.24 173 0.23 169 173
LP89lex 7.0 6.3 P2 1.0 0.58 141 0.89 218 0.70 172 179
EZ92erzi 6.7 2.0 P2 2.3 0.20 111 0.20 114 0.17 96 119
NR94rrs 6.7 7.5 P2 1.0 0.58 143 0.78 191 0.72 177 174
NR94sylm 6.7 6.4 P3 2.4 0.23 135 0.18 108 0.18 106 122
KB95kobj 6.9 0.6 P3 0.9 0.88 194 0.86 190 0.86 190 160
KB95tato 6.9 1.5 P3 2.0 0.45 219 0.38 188 0.40 196 174
IV79ar06 6.5 1.2 P2 3.4 0.12 104 0.15 129 0.13 108 110
IV79melo 6.5 0.0 P2 2.8 0.15 101 0.18 123 0.15 103 117
KB95kpi1 6.9 3.7 P2 1.8 0.27 120 0.28 122 0.25 110 100
MH84andd 6.2 4.5 P2 0.8 0.19 37 0.17 34 0.16 31 27
MH84cyld 6.2 0.1 P3 0.8 0.45 89 0.51 100 0.47 92 65
NR94newh 6.7 7.1 P2 1.3 0.44 139 0.41 131 0.37 118 119
NR94spva 6.7 8.9 P3 2.7 0.14 93 0.08 54 0.09 60 63

Table 7.3 Equivalent Pulses for Simulated Near-Fault Ground Motions (Summary)

µ=1 µ = (4-10) µ = (1-10)

Station Mw R (km) Pulse Type Tp (sec) aeff (g) veff (cm/sec) aeff (g) veff (cm/sec) aeff (g) veff (cm/sec) PGV
f6 6.5 3 P2 1.7 0.18 76 0.19 79 0.17 71 68
f8 6.5 3 P2 1.2 0.31 90 0.47 138 0.41 121 116
f6 6.5 5 P2 2.0 0.11 52 0.11 53 0.10 49 46
f8 6.5 5 P2 2.1 0.18 93 0.20 104 0.19 98 70
f6 6.5 10 P2 2.6 0.04 28 0.04 24 0.03 19 25
f8 6.5 10 P2 3.0 0.05 39 0.05 36 0.05 37 32
f6 7.0 3 P2 3.2 0.14 107 0.15 118 0.13 102 98
f8 7.0 3 P2 3.4 0.20 169 0.22 181 0.21 175 146
f6 7.0 5 P2 3.5 0.13 113 0.12 107 0.11 94 89
f8 7.0 5 P2 3.6 0.16 143 0.17 150 0.16 141 124
f6 7.0 10 P2 5.0 0.05 58 0.04 46 0.04 49 51
f8 7.0 10 P2 3.3 0.11 88 0.11 89 0.10 81 76
f6 7.5 3 P3 3.2 0.21 161 0.22 169 0.21 165 148
f8 7.5 3 P3 3.2 0.31 241 0.28 218 0.27 212 210
f6 7.5 5 P3 3.2 0.18 144 0.20 157 0.19 149 138
f8 7.5 5 P3 3.2 0.26 207 0.25 194 0.24 188 201
f6 7.5 10 P3 3.2 0.13 101 0.14 107 0.14 110 84
f8 7.5 10 P3 3.2 0.18 145 0.19 149 0.18 141 131

Chapter 7 Representation of Near-Fault Ground …

170
 Elastic SDOF Velocity Demands
Pulses and NR94rrs, Fault-Normal, Tp = 1.0 sec
5
NR94rrs
P1
4 P2
P3

3
Sv / vg,max

0
0 0.5 1 1.5 2 2.5 3
T / Tp

(a) Velocity

Elastic SDOF Displacement Demands

Pulses and NR94rrs, Fault-Normal, Tp = 1.0 sec
6
NR94rrs
P1
5
P2
P3
4
Sd / ug,max

0
0 0.5 1 1.5 2 2.5 3
T / Tp

(b) Displacement

Figure 7.1 Determination of Pulse Period (and Pulse Type) from Elastic Spectra,
Record NR94rrs

Chapter 7 Representation of Near-Fault Ground …

171
 Elastic SDOF Velocity Demands
Pulses and KB95kobj, Fault-Normal, Tp = 0.9 sec
5
KB95kobj
P1
4 P2
P3

3
Sv / vg,max

0
0 0.5 1 1.5 2 2.5 3
T / Tp

(a) Velocity

Elastic SDOF Displacement Demands

Pulses and KB95kobj, Fault-Normal, Tp = 0.9 sec
6
KB95kobj
P1
5
P2
P3
4
Sd / ug,max

0
0 0.5 1 1.5 2 2.5 3
T / Tp

(b) Displacement

Figure 7.2 Determination of Pulse Period (and Pulse Type) from Elastic Spectra,
Record KB95kobj

Chapter 7 Representation of Near-Fault Ground …

172
 Elastic SDOF Velocity Demands
Pulses and KB95tato, Fault-Normal, Tp = 2.0 sec
5
KB95tato
P1
4 P2
P3

3
Sv / vg,max

0
0 0.5 1 1.5 2 2.5 3
T / Tp

(a) Velocity

Elastic SDOF Displacement Demands

Pulses and KB95tato, Fault-Normal, Tp = 2.0 sec
6
KB95tato
P1
5
P2
P3
4
Sd / ug,max

0
0 0.5 1 1.5 2 2.5 3
T / Tp

(b) Displacement

Figure 7.3 Determination of Pulse Period (and Pulse Type) from Elastic Spectra,
Record KB95tato

Chapter 7 Representation of Near-Fault Ground …

173
 Effective Acceleration of Equivalent Pulse
Recorded Ground Motions, µ = (1-10)
2.5
LP89lgpc
Lp89lex
EZ92erzi
2
NR94rrs
NR94sylm
KB95kobj
aeff / average(aeff)

KB95tato
1.5

0.5

0
0 0.5 1 1.5 2 2.5 3
T / Tp

(a)

Effective Acceleration of Equivalent Pulse

Recorded Ground Motions, µ = (1-10)
2.5
IV79ar06
IV79melo
KB95kpi1
2 MH84andd
MH84cyld
NR94newh
aeff / average(aeff)

NR94spva
1.5

0.5

0
0 0.5 1 1.5 2 2.5 3
T / Tp

(b)

Figure 7.4 Variations of Equivalent Pulse Effective Acceleration with T/Tp for Recorded
Ground Motions, Ductility Range of 1 to 10

Chapter 7 Representation of Near-Fault Ground …

174
 Normalized Inelastic SDOF Displacement Time History
Pulse P2, T / Tp = 1.0, µ = 6, α = 3%, ξ = 2%
1.2

0.8

0.4
u / ug,max

-0.4

-0.8 Ugr

Ust
-1.2
0 1 2 3 4 5 6 7 8
t / Tp

(a) T/Tp = 1.0

Normalized Inelastic SDOF Displacement Time History

Pulse P2, T / Tp = 2.0, µ = 6, α = 3%, ξ = 2%
1.2

Ugr
0.8
Ust

0.4
u / ug,max

-0.4

-0.8

-1.2
0 1 2 3 4 5 6 7 8
t / Tp

(b) T/Tp = 2.0

Figure 7.5 Ground and Inelastic SDOF Displacement Time Histories for
Pulse P2 and µ = 6

Chapter 7 Representation of Near-Fault Ground …

175
 Inelastic SDOF Displacement Time History
NR94rrs, T = 1.0 sec, µ = 6, α = 3%, ξ = 2%
60

40

20
u (cm)

-20

-40 Ugr

Ust

-60
0 1 2 3 4 5 6 7 8
t (sec)

(a) T = 1.0 sec.

Inelastic SDOF Displacement Time History

NR94rrs, T = 2.0 sec, µ = 6, α = 3%, ξ = 2%
60

40

20
u (cm)

-20

-40 Ugr

Ust

-60
0 1 2 3 4 5 6 7 8
t (sec)

(b) T = 2.0 sec.

Figure 7.6 Ground and Inelastic SDOF Displacement Time Histories for
NR94rrs and µ = 6

Chapter 7 Representation of Near-Fault Ground …

176
 Normalized Elastic SDOF Displacement Time History
Pulse P2, α = 3%, ξ = 2%
1
T/Tp = 0.375
T/Tp = 0.50
T/Tp = 1.00
0.5
T/Tp = 2.00
u / umax

-0.5

-1
0 0.5 1 1.5 2 2.5
t / Tp

(a) Pulse P2

Normalized Elastic SDOF Displacement Time History

NR94rrs, α = 3%, ξ = 2%
1
T = 0.375 sec
T = 0.50 sec
T = 1.00 sec
0.5 T = 2.00 sec
u / umax

-0.5

-1
2.1 2.6 3.1 3.6 4.1 4.6
t (sec)

(b) Record NR94rrs

Figure 7.7 Normalized Elastic SDOF Displacement Time Histories for Various Periods

Chapter 7 Representation of Near-Fault Ground …

177
 Normalized Inelastic SDOF Displacement Time History
Pulse P2, µ = 6, α = 3%, ξ = 2%
6

2
u / uy

-2
T/Tp = 0.375
T/Tp = 0.50
-4
T/Tp = 1.00
T/Tp = 2.00
-6
0 0.5 1 1.5 2 2.5
t / Tp

(a) Pulse P2

Normalized Inelastic SDOF Displacement Time History

NR94rrs, µ = 6, α = 3%, ξ = 2%
6

2
u / uy

-2
T = 0.375 sec
T = 0.50 sec
-4
T = 1.00 sec
T = 2.00 sec
-6
2.1 2.6 3.1 3.6 4.1 4.6
t (sec)

(b) Record NR94rrs

Figure 7.8 Normalized Inelastic SDOF Displacement Time Histories for Various Periods
and µ = 6

Chapter 7 Representation of Near-Fault Ground …

178
 Maximum Story Ductility Demands Maximum Story Ductility Demands
T / Tp = 0.50, aeff = 0.72 g, µ = (1,10), without P-∆
∆ T / Tp = 0.50, aeff = 0.86 g, µ = (1,10), without P-∆
∆
3 4

NR94rrs KB95kobj

2.5 Pulse P3
Pulse P2
3

2
γ = Vy / W

y
1.5 2

1
1

0.5

0 0
1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10
Maximum Story Ductility, µmax Maximum Story Ductility, µmax

T/Tp = 0.5

Maximum Story Ductility Demands Maximum Story Ductility Demands

T / Tp = 1.00, aeff = 0.72 g, µ = (1,10), without P-∆
∆ T / Tp = 1.00, aeff = 0.86 g, µ = (1,10), without P-∆
∆
3 4
NR94rrs KB95kobj
2.5
Pulse P2 Pulse P3
3
2
γ = Vy / W

1.5 2

1
0.5

0 0
1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10
Maximum Story Ductility, µmax Maximum Story Ductility, µmax

T/Tp = 1.0

Maximum Story Ductility Demands Maximum Story Ductility Demands

T / Tp = 2.00, aeff = 0.72 g, µ = (1,10), without P-∆
∆ T / Tp = 2.00, aeff = 0.86 g, µ = (1,10), without P-∆
∆
1 2

NR94rrs KB95kobj

0.8 Pulse P2 Pulse P3

1.5
γ = Vy / W

0.6
y

1
γ

0.4

0.5
0.2

0 0
1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10
Maximum Story Ductility, µmax Maximum Story Ductility, µmax

T/Tp = 2.0
(a) NR94rrs and Pulse P2 (b) KB95kobj and Pulse P3

Figure 7.9 Matching of γ-µmax Curves for Identification of Best Fit aeff for
NR94rrs and KB95kobj

Chapter 7 Representation of Near-Fault Ground …

179
 Story Ductility Demands Story Ductility Demands
Low Ductility, T / Tp = 0.5, γ = 1.2, without P-∆
∆ High Ductility, T / Tp = 0.5, γ = 0.6, without P-∆
∆
1
NR94rrs NR94rrs

Pulse P2 Pulse P2
Relative Height 0.8

0.6

0.4

0.2

0
0 1 2 3 4 0 2 4 6 8 10 12
Story Ductility Ratio, µi = δdyn,i / δy,i Story Ductility Ratio, µi = δdyn,i / δy,i

T/Tp = 0.5

Story Ductility Demands Story Ductility Demands

Low Ductility, T / Tp = 1.0, γ = 0.8, without P-∆
∆ High Ductility, T / Tp = 1.0, γ = 0.2, without P-∆
∆
1
NR94rrs NR94rrs

Pulse P2 Pulse P2
0.8
Relative Height

0.6

0.4

0.2

0
0 1 2 3 4 0 2 4 6 8 10 12
Story Ductility Ratio, µi = δdyn,i / δy,i Story Ductility Ratio, µi = δdyn,i / δy,i

T/Tp = 1.0

Story Ductility Demands Story Ductility Demands

Low Ductility, T / Tp = 2.0, γ = 0.4, without P-∆
∆ High Ductility, T / Tp = 2.0, γ = 0.1, without P-∆
∆
1
NR94rrs NR94rrs

Pulse P2
0.8 Pulse P2
Relative Height

0.6

0.4

0.2

0
0 1 2 3 4 0 2 4 6 8 10 12
Story Ductility Ratio, µi = δdyn,i / δy,i Story Ductility Ratio, µi = δdyn,i / δy,i

T/Tp = 2.0
(a) Low Ductility, High Strength (b) High Ductility, Low Strength

Figure 7.10 Comparison of Story Ductility Demands Obtained from Record NR94rrs and
its Equivalent Pulse

Chapter 7 Representation of Near-Fault Ground …

180
 Story Ductility Demands Story Ductility Demands
Low Ductility, T / Tp = 0.5, γ = 1.2, without P-∆
∆ High Ductility, T / Tp = 0.5, γ = 0.75, without P-∆
∆
1
KB95kobj KB95kobj

Pulse P3 Pulse P3
Relative Height 0.8

0.6

0.4

0.2

0
0 1 2 3 4 0 2 4 6 8 10 12
Story Ductility Ratio, µi = δdyn,i / δy,i Story Ductility Ratio, µi = δdyn,i / δy,i

T/Tp = 0.5

Story Ductility Demands Story Ductility Demands

Low Ductility, T / Tp = 1.0, γ = 0.8, without P-∆
∆ High Ductility, T / Tp = 1.0, γ = 0.3, without P-∆
∆
1
KB95kobj KB95kobj

Pulse P3 Pulse P3
0.8
Relative Height

0.6

0.4

0.2

0
0 1 2 3 4 0 2 4 6 8 10
Story Ductility Ratio, µi = δdyn,i / δy,i Story Ductility Ratio, µi = δdyn,i / δy,i

T/Tp = 1.0

Story Ductility Demands Story Ductility Demands

Low Ductility, T / Tp = 2.0, γ = 0.4, without P-∆
∆ High Ductility, T / Tp = 2.0, γ = 0.1, without P-∆
∆
1
KB95kobj KB95kobj

Pulse P3
0.8 Pulse P3
Relative Height

0.6

0.4

0.2

0
0 1 2 3 4 0 2 4 6 8 10 12
Story Ductility Ratio, µi = δdyn,i / δy,i Story Ductility Ratio, µi = δdyn,i / δy,i

T/Tp = 2.0
(a) Low Ductility, High Strength (b) High Ductility, Low Strength

Figure 7.11 Comparison of Story Ductility Demands Obtained from Record KB95kobj
and its Equivalent Pulse

Chapter 7 Representation of Near-Fault Ground …

181
 MDOF Strength Demands for Constant Ductility
NR94rrs, Fault-Normal, Tp = 1.0 sec, aeff = 0.72 g, ξ = 2%, without P-∆
∆
5
µ===1
µ===2

4 µ===3
µ===4
µ===6
µ===8
η = Vy / ([Link])

0
0 0.5 1 1.5 2 2.5 3
T / Tp

(a) Record NR94rrs

MDOF Strength Demands for Constant Ductility

KB95kobj, Fault-Normal, Tp = 0.9 sec, aeff = 0.86 g, ξ = 2%, no P-∆
∆
5
µ===1
µ===2
µ===3
4
µ===4
µ===6
µ===8
η = Vy / ([Link])

0
0 0.5 1 1.5 2 2.5 3
T / Tp

(b) Record KB95kobj

Figure 7.12 Base Shear Strength Demand Spectra for Records NR94rrs and KB95kobj

Chapter 7 Representation of Near-Fault Ground …

182
 MDOF Strength Demand Ratios for Constant Ductility
KB95kobj, Fault-Normal, Tp = 0.9 sec, aeff = 0.85 g, ξ = 2%, no P-∆
∆
2
µ===1
µ===2
µ===3
1.5 µ===4
µ===6
µ===8
η NF=/=ηP2

0.5

0
0 0.5 1 1.5 2 2.5 3
T / Tp

(a) Equivalent Pulse P2

MDOF Strength Demand Ratios for Constant Ductility

KB95kobj, Fault-Normal, Tp = 0.9 sec, aeff = 0.86 g, ξ = 2%, no P-∆
∆
2
µ===1
µ===2
µ===3
1.5 µ===4
µ===6
µ===8
η NF=/=ηP3

0.5

0
0 0.5 1 1.5 2 2.5 3
T / Tp

(b) Equivalent Pulse P3

Figure 7.13 Record-to-Pulse Strength Demand Ratios for Specific Target Ductilities,
Different Equivalent Pulse Types, Record KB95kobj

Chapter 7 Representation of Near-Fault Ground …

183
 MDOF Strength Demand Ratios for Constant Ductility
NR94rrs, Fault-Normal, Tp = 1.0 sec, aeff = 0.72 g, ξ = 2%, without P-∆
∆
2
µ===1
µ===2
µ===3
1.5 µ===4
µ===6
µ===8
η NF=/=ηP2

0.5

0
0 0.5 1 1.5 2 2.5 3
T / Tp

(a) T/Tp = 1.0 sec.

MDOF Strength Demand Ratios for Constant Ductility

NR94rrs, SRSS Pattern, Tp = 1.35 sec, aeff = 0.54 g, ξ = 2%, without P-∆
∆
2

1.5
η NF=/=ηP2

µ===1
µ===2
0.5 µ===3
µ===4
µ===6
µ===8
0
0 0.5 1 1.5 2 2.5 3
T / Tp

(b) T/Tp = 1.35 sec.

Figure 7.14 Record-to-Pulse Strength Demand Ratios for Specific Target Ductilities,
Different Equivalent Pulse Periods, Record NR94rrs

Chapter 7 Representation of Near-Fault Ground …

184
 MDOF Strength Demand Ratios for Constant Ductility
NR94rrs, 0.707(FN+FP), Tp = 1.0 sec, aeff = 0.58 g, ξ = 2%, no P-∆
∆
2

1.5
η45°=/=ηP2

µ===1
µ===2
0.5
µ===3
µ===4
µ===6
µ===8
0
0 0.5 1 1.5 2 2.5 3
T / Tp

(a) 0.707(FN+FP) Component

MDOF Strength Demand Ratios for Constant Ductility

NR94rrs, 0.707(FN-FP), Tp = 1.0 sec, aeff = 0.47 g, ξ = 2%, no P-∆
∆
2

1.5
η45°=/=ηP2

µ===1
µ===2
0.5
µ===3
µ===4
µ===6
µ===8

0
0 0.5 1 1.5 2 2.5 3
T / Tp

(b) 0.707(FN-FP) Component

Figure 7.15 Evaluation of Equivalent Pulse for Rotated Components of Near-Fault

Record NR94rrs

Chapter 7 Representation of Near-Fault Ground …

185
 Maximum MDOF Story Ductility Demands
∆
Pulse P2, SRSS Pattern, without P-∆
4
T/Tp = 0.375
T/Tp = 0.50
T/Tp = 0.75
3 T/Tp = 1.00
T/Tp = 1.50
η = Vy / ([Link],max)

T/Tp = 2.00
T/Tp = 3.00

0
0 4 8 12 16 20
Maximum Story Ductility,,=µmax

Figure 7.16 Estimation of Maximum Story Ductility Demands Based on Pulse P2

Roof Displacement Demands

∆
Pulse P2, SRSS Pattern, without P-∆
4
T/Tp=0.375
T/Tp=0.50
T/Tp=0.75
3 T/Tp=1.00
T/Tp=1.50
η = Vy / ([Link],max)

T/Tp=2.00
T/Tp=3.00
2

0
0 0.5 1 1.5 2 2.5
δroof,max / ug,max

Figure 7.17 Estimation of Roof Displacement Demands Based on Pulse P2

Chapter 7 Representation of Near-Fault Ground …

186
 Normalized Story Drift Angle Demands Normalized Story Drift Angle Demands
∆
Pulse P2, 20-Story, SRSS Pattern, T / Tp = 0.5, without P-∆ ∆
Pulse P2, 20-Story, SRSS Pattern, T / Tp = 0.75, without P-∆
1
η===3.00 η===2.00
η===2.00 η===1.25
0.8 η===1.50 η===1.00
η===1.25 η===0.50
η===1.00 η===0.30
Relative Height

η===0.75 η===0.25
0.6
η===0.50 η===0.20

0.4

0.2

0
0 0.5 1 1.5 2 2.5 0 0.5 1 1.5 2 2.5
θi / θroof θi / θroof

(a) T/Tp = 0.5 (b) T/Tp = 0.75

Normalized Story Drift Angle Demands Normalized Story Drift Angle Demands
∆
Pulse P2, 20-Story, SRSS Pattern, T / Tp = 1.0, without P-∆ ∆
Pulse P2, 20-Story, SRSS Pattern, T / Tp = 1.5, without P-∆
1
η===2.00 η===1.50
η===1.50 η===1.00
0.8 η===1.00 η===0.50
η===0.50 η===0.25
η===0.25 η===0.15
Relative Height

η===0.15 η===0.10
0.6
η===0.10 η===0.07

0.4

0.2

0
0 0.5 1 1.5 2 2.5 0 0.5 1 1.5 2 2.5 3 3.5
θi / θroof θi / θroof

(c) T/Tp = 1.0 (d) T/Tp = 1.5

Normalized Story Drift Angle Demands Normalized Story Drift Angle Demands
∆
Pulse P2, 20-Story, SRSS Pattern, T / Tp = 2.0, without P-∆ ∆
Pulse P2, 20-Story, SRSS Pattern, T / Tp = 3.0, without P-∆
1
η===1.00 η===0.75
η===0.75 η===0.50
0.8 η===0.50 η===0.25
η===0.25 η===0.10
η===0.10 η===0.07
Relative Height

η===0.07 η===0.05
0.6
η===0.05 η===0.03

0.4

0.2

0
0 0.5 1 1.5 2 2.5 3 3.5 4 0 1 2 3 4 5
θi / θroof θi / θroof

(e) T/Tp = 2.0 (f) T/Tp = 3.0

Figure 7.18 Estimation of Story Drift Demands Based on Pulse P2, Various T/Tp values

Chapter 7 Representation of Near-Fault Ground …

187
 CHAPTER 8

STUDY OF MODELS OF STEEL STRUCTURES

The results presented in the previous chapters will be of value only if the generic
structures introduced in Section 3.2 can represent a variety of real-world MDOF frame
structures. In addition to frame geometrical configuration parameters such as the number
of bays and stories, span length, or story height, there are a large number of factors that
differentiate the simple generic frames from real structures. The design of the generic
frames is based on many simplifying assumptions that do not necessarily represent real
conditions in all cases. For instance, contributions of panel zones, floor slabs, and non-
structural elements to the stiffness and strength of the system are ignored, and the generic
design is based on particular distributions of story stiffness and strength over the height
of the structure.

On the other hand, employing sophisticated structural models that can fully account for
all existing elements and effects of the system is not practical in many cases. If
advantage can be taken of generic structures to predict the seismic demands of real
structures with reasonable accuracy, a great deal of computational time and effort can be
saved. Therefore, it is important to assess the usefulness of predictions obtained from the
generic structures for various seismic demand quantities. In this chapter the correlation
between demands is investigated for the generic frames and more realistic structural
models subjected to near-fault ground motions, in order to justify and validate the use of
generic models in this study.

8.1. Models of Steel Structures Used in this Study

Two steel moment resisting frame (SMRF) models (3- and 9-story) are utilized, which
were thoroughly analyzed in a past study carried out as part of the SAC Steel Program
(Gupta and Krawinkler, 1999). The 2-dimensional frame models, which are referred to

Chapter 8 Study of Models of Steel Structures

188
as “LA 3-story” and “LA 9-story”, correspond to the perimeter moment resisting frames
of two office buildings located in Los Angeles. The buildings have been designed to
meet all prevailing gravity, wind, and seismic requirements of UBC’94 guidelines. The
geometrical configurations of these two models are illustrated in Fig. 8.1. The moment
resisting frames are shown in solid bold lines in the plan views. The LA 3-story frame
has four bays, while the LA 9-story frame contains five bays. Moreover, the LA 9-story
structure has a basement, and is laterally restrained at the ground level.

Most of the member sizes are controlled by drift rather than strength considerations. The
sections used in the LA 3-story and LA 9-story frames are summarized in Table 8.1. All
columns in the perimeter frames bend about the strong axis. In the analyses, expected
yield stress values of 339 MPa (49.2 ksi) and 397 MPa (57.6 ksi) are used for beams and
columns, respectively, rather than the nominal values of 248 MPa (36 ksi) and 345 Mpa
(50 ksi).

The structure models used in this study are bare frames of type M2 in the SAC project, in
which the dimensions, stiffness, strength, and inelastic shear distortion of panel zones are
considered. The panel zones are modeled using a combination of standard beam-column
elements and trilinear rotational springs at each joint. All other contributions to strength
and stiffness, such as those due to the floor slabs, gravity columns, simple (shear)
connections, and nonstructural elements, are neglected. Even though the interior
(gravity) frames, which include shear (simple) connections, are not modeled, second-
order (P-delta) effects due to gravity loads on these frames are taken into account by
linking a virtual column to the main frame, which carries the vertical loads tributary to
the interior gravity frames. For more modeling details the reader is referred to Gupta and
Krawinkler (1999).

8.2. Inelastic Static Analysis and Calibration of Generic Structures

Inelastic static (pushover) analysis is utilized here to calibrate the properties of the
generic frames such that their demands can be directly compared with those of the LA 3-
story and LA 9-story structures. The 3-story and 9-story frames introduced in Section
6.5.1 are selected as generic counterparts of the LA 3-story and LA 9-story structures,
respectively. In addition to the SRSS lateral load pattern discussed previously, the 1994
NEHRP load pattern with k = 2 is used here in the pushover analyses. In this load
pattern, the lateral load applied to the i-th story of the frame is given as:

Chapter 8 Study of Models of Steel Structures

189
 w i h 2i
Fi = n V (8.1)
2
w jh j
j=1

where hi and wi are the height (from the base) and seismically effective weight of the i-th
floor, respectively, and V represents the base shear. The power k = 2 in the NEHRP load
pattern will be too large for the design of a 3-story structure, but in this study it is used
only to compare pushover curves for the generic and LA 3-story structures.

The global pushover curves, i.e., normalized base shear force (base shear normalized by
structure seismic weight, V/W) versus roof displacement are shown in Figs. 8.2 and 8.3
for the LA 3-story and LA 9-story structures. Each graph illustrates the global nonlinear
behavior of the structure subjected to the SRSS and NEHRP load patterns with and
without consideration of P-delta effects. The pushover curve for the LA 3-story structure
exhibits a bilinear shape, which indicates that all plastic hinges develop within a
relatively narrow range of displacement. On the other hand, the plastic hinge formation
in the LA 9-story frame occurs over a much wider range of displacement, resulting in a
smoother transition from elastic behavior to a mechanism. Furthermore, the SRSS load
pattern leads to higher yield strengths compared to the NEHRP pattern. This illustrates
the sensitivity of the inelastic static behavior of multi-story frame structures to the shape
of load patterns employed in the pushover analysis.

Even though a post-yield strain-hardening ratio of 3.0% is assigned at the element level,
the global strain-hardening ratio amounts to about 3.7% for the LA 3-story frame, and
4.1% for the LA 9-story frame in the absence of P-delta effects. However, once second-
order effects are taken into account, the global strain-hardening ratio decreases to 0.3%
for the LA 3-story frame, and to negative values for LA 9-story. The reason is that the
gravity load in the 3-story structure is not large enough to cause a negative post-yield
stiffness, whereas for the 9-story frame, beyond a certain displacement value, negative
post-yield slopes are observed. An indication of the significance of P-delta effects on the
global behavior of frame structures can be obtained from the first story “stability
coefficient” defined as:

P∆ 1
θ1 = (8.2)
Vh 1

Chapter 8 Study of Models of Steel Structures

190
where P is the total vertical gravity load, ∆1 is the elastic first-order drift in the first story
caused by the base shear V, and h1 denotes the height of the first story. The value of θ1
for LA 3-story and LA 9-story structures is 3.4% and 7.1%, respectively. These values
help to explain the difference in the post-yield stiffness of the two frames.

For comparison purposes, the generic 3-story structure is tuned to have a fundamental
period equal to the fundamental period of the LA 3-story structure, i.e., 0.99 sec. (without
P-delta effects). Likewise, the generic 9-story structure with a fundamental period of
2.17 sec. is compared with the LA 9-story structure. Since the story strengths of the
generic structures follow a distribution that is obtained from the SRSS load pattern, only
the pushover curves that are based on the SRSS load pattern are used for strength
calibration. Even though the pushover curves of the LA structures, unlike those of the
generic frames, do not exhibit a distinct yield point, the base shear strength can be
estimated from the intersection of the extended elastic and inelastic branches. This
results in strength coefficient values γ = Vy/W = 0.35 and 0.21 for the generic 3- and 9-
story structures, respectively.

In order for the response of the generic frames to be comparable with the response of the
SAC structures, P-delta effects should also be compatible between the two systems. For
this purpose, the amount of the vertical load acting on the generic frame is determined
such that the first story stability coefficient is equal to the corresponding value for the
SAC structure, i.e., θ1 = 3.4% and 7.1% for the 3- and 9-story frames. However, this
merely means that static P-delta effects in the first story of the SAC and generic frames
are at the same level as long as the systems behave elastically. The stability coefficients
in higher stories of the two structures are not necessarily identical. Besides, once plastic
hinges form, the deflected shape of the structure will deviate from the elastic deflected
shape, and therefore the distribution of secondary effects over the height of the structure
may change significantly (Gupta and Krawinkler, 1999).

Figures 8.4 and 8.5 compare the global pushover curves (base shear vs. roof
displacement) for the calibrated generic structures with the corresponding pushover
curves for the SAC structures under NEHRP and SRSS load patterns (a) ignoring P-delta
effects, and (b) considering P-delta effects. A close correlation between the results of the
SAC structures and their corresponding generic frames are observed, which demonstrates
the similarities of their global static behavior in both elastic and inelastic ranges. The

Chapter 8 Study of Models of Steel Structures

191
reflection of these similarities in dynamic response at global and story levels is
investigated in the following section.

8.3. Inelastic Dynamic Analysis

In this section the dynamic response of the SAC structures to near-fault ground motions
is compared with the response obtained from the generic models. The objective is to
assess the degree to which the response of generic structures can be used to estimate the
response of real frame structures to near-fault ground motions.

8.3.1. Roof Displacement

Figure 8.6 compares roof displacement time histories for the generic and LA 3-story
structures subjected to near-fault records LP89lex, and KB95tato, considering P-delta
effects. As can be seen, the correlation is very good, which is not surprising when
similarity of the pushover curves for these two structures is considered (see Fig. 8.4).
Figure 8.7 compares the roof displacement time histories, in the same manner as Fig. 8.6,
for the generic and LA 9-story structures. Again, a very close agreement between the
responses of the SAC and generic structures is observed, indicating that the generic
model can capture global dynamic response with good accuracy.

In order to investigate the response differences between the SAC and generic structures at
different performance (inelasticity) levels, the input ground motion (i.e., LP89lex) is
scaled using an intensity multiplier. A sequence of dynamic analyses that utilizes the
same record with gradually increasing intensity is called “incremental dynamic analysis”
(IDA). Figure 8.8 illustrates the roof displacement demands for the generic and LA 3-
story frames versus the intensity multiplier (which equals unity for the original record)
(a) ignoring P-delta effects, and (b) accounting for P-delta effects. Figure 8.9 compares
the generic and LA 9-story structures in the same manner. The following observations
can be made:

•= Despite some differences between the roof displacement demands for the generic
and LA 9-story structures in the presence of P-delta effects, the demands obtained
from the generic models are close to those obtained from the SAC models. This
demonstrates that generic structures can estimate the global demands (roof

Chapter 8 Study of Models of Steel Structures

192
 displacement) with sufficient accuracy and much less computational effort
compared to more realistic, but complicated, models.

•= When P-delta effects are taken into account, the roof displacement demand grows
more rapidly with the ground motion intensity. This decrease in the slope of the
IDA curve is especially evident for the 9-story structures (see Fig. 8.9) whose
post-yield stiffness is negative giving rise to large secondary effects.

•= The reason why the generic 9-story structure exhibits larger P-delta effects than
the LA 9-story structure is that it enters a negative stiffness range at a smaller
global drift, see Fig. 8.5(b).

8.3.2. Story Drift Angles

In order to avoid ambiguity associated with the definition of story yield drift, story drift
angle demands (story drift demand divided by story height) rather than story ductility
demands are used here to represent story-level displacement demands. To allow a direct
comparison of structures with different heights, as discussed in Section 6.5.3, the drift
angle demands are normalized by the roof drift angle (roof displacement divided by
structure height).

Figure 8.10 compares the normalized drift angle demands for the calibrated generic and
LA 3-story structures subjected to records LP89lex and KB95tato in the presence of P-
delta effects. The distribution of the demands over the height is almost uniform, which
implies that the deflected shapes are close to a straight line. This is not surprising when it
is considered that the response of the 3-story structures is mainly controlled by the first
mode, and that the first mode shape is close to a straight line. The results show that the
normalized drift angle demands for the generic and SAC structures are relatively close.
This is an indication that the generic 3-story structure can be utilized to estimate story
drift demands with good accuracy.

Figure 8.11 compares normalized drift angle demands for the generic and LA 9-story
structures in the same manner. Although some differences are observed, particularly in
the bottom portion of the structure, the general patterns of the two distributions are
similar. The results indicate that overall, the generic model can represent story drift angle

Chapter 8 Study of Models of Steel Structures

193
demands. However, it should be noted that the distribution of story drift angle demands
over the height of the structure is dependent on the number of stories (see Fig. 6.45).

If the input ground motion is scaled using an intensity multiplier, as in the previous
section, the distributions of normalized story drift angle demands over the height of the
structure can be investigated at different performance levels. Figures 8.12 and 8.13
compare these distributions for the SAC and generic structures subjected to the record
LP89lex with P-delta effects. The following observations are made:

•= Figure 8.12 shows that for the generic and LA 3-story structures the distributions
are not far from uniform, indicating that the deflected shapes are close to a
straight line at all performance levels. For the generic structure the maximum
drift demand occurs in the bottom story at high performance levels (low severity
ground motion), whereas at low performance levels (severe ground motion) the
maximum demand shifts to the top story. This upward shift of the maximum drift
demand was also observed in the response of the generic 3-story structure to the
basic pulse P2 (see Fig. 6.45). This effect, however, is not as clear in the response
of the LA 3-story structure to the same near-fault ground motion.

•= Figure 8.13 illustrates the migration phenomenon for the generic and LA 9-story
structures, which was previously identified as a near-fault response characteristic
of MDOF structures with T/Tp > 1.0 (T/Tp ≅ 2.2 in this case). That is, for low
intensity ground motions (or strong structures) the maximum story drift demand
occurs in the upper portion of the structure, and as the ground motion intensity
increases (or the structure strength decreases), the maximum demand migrates to
the bottom. Although the generic 9-story structure can overall capture the
distribution of drift demands for the LA 9-story structure, at low performance
levels the distributions are much less uniform for the SAC structure than for the
generic one. The reason is that the story shear strengths of the SAC structures are
not fine-tuned to the SRSS shear force distribution used to design the generic
structure.

Summary:

The results of the study presented in this chapter indicate that generic structural models
can be utilized to estimate the global demands for MDOF structures with good accuracy.

Chapter 8 Study of Models of Steel Structures

194
However, due to the dependence of the distribution of story-level demands on the number
of stories, as well as on the distribution of stiffness and strength over the height of the
structure, estimating these demands from a single generic model, such as the generic 20-
story frame used extensively in this study, should be done with caution and with due
consideration to the peculiarities caused by subjective design decisions.

Chapter 8 Study of Models of Steel Structures

195
Table 8.1 Beam and Column Sections for LA 3- and 9-Story Frame Structures
(Krawinkler and Gupta, 1998)

3-STORY
Story/Floor Ext. Columns Int. Columns Girder
1/2 W14X257 W14X311 W33X118
2/3 W14X257 W14X311 W30X116
3/4 W14X257 W14X311 W24X68
9-STORY
Story/Floor Ext. Columns Int. Columns Girder
-1/1 W14X370 W14X500 W36X160
1/2 W14X370 W14X500 W36X160
2/3 W14X370 W14X455 W36X160
3/4 W14X370 W14X455 W36X135
4/5 W14X283 W14X370 W36X135
5/6 W14X283 W14X370 W36X135
6/7 W14X257 W14X283 W36X135
7/8 W14X257 W14X283 W30X99
8/9 W14X233 W14X257 W27X84
9/10 W14X233 W14X257 W24X68
6 bays @ 9.15m

5 bays @ 9.15m

4 bays @ 9.15m 5 bays @ 9.15m

MRF
Gravity Frame
8 @ 3.96m

3 @ 3.96m 5.49m
3.66m

Figure 8.1 Floor Plans and Elevations of LA 3- and 9-Story Model Buildings
(Gupta and Krawinkler, 1999)

Chapter 8 Study of Models of Steel Structures

196
 Roof Displacement vs. Normalized Base Shear
SAC LA 3-Story, Model M2, SRSS & NEHRP 94 k=2, αelement = 3%
0.5

α = 3.6%
0.4 α = 3.7%

α = 0.3%
α = 0.3%
0.3
Vbase / W

0.2

NEHRP with P-delta

0.1 NEHRP without P-delta

SRSS with P-delta
SRSS without P-delta
0
0 20 40 60 80 100
δroof (cm)

Figure 8.2 Global Pushover Curves for LA 3-Story Structure, Different Load Patterns

Roof Displacement vs. Normalized Base Shear

SAC LA 9-Story, Model M2, SRSS & NEHRP 94 k=2, αelement = 3%
0.3

α = 4.1%
0.25
α = 4.1%

0.2
Vbase / W

0.15

0.1
NEHRP with P-delta
NEHRP without P-delta
0.05 SRSS with P-delta
SRSS without P-delta
0
0 50 100 150 200 250 300
δroof (cm)

Figure 8.3 Global Pushover Curves for LA 9-Story Structure, Different Load Patterns

Chapter 8 Study of Models of Steel Structures

197
 Roof Displacement vs. Normalized Base Shear
Generic and LA 3-Story, γ = 0.35, without P-∆
∆=
0.5

0.4

0.3
Vbase / W

0.2

NEHRP, SAC

0.1 NEHRP, Generic

SRSS, SAC
SRSS, Generic
0
0 20 40 60 80 100
δroof (cm)

(a) without P-delta

Roof Displacement vs. Normalized Base Shear

Generic and LA 3-Story, γ = 0.35, with P-∆
∆=
0.5

0.4

0.3
Vbase / W

0.2

NEHRP, SAC

0.1 NEHRP, Generic

SRSS, SAC
SRSS, Generic
0
0 20 40 60 80 100
δroof (cm)

(b) with P-delta

Figure 8.4 Comparison of Pushover Curves for Generic and LA 3-Story Structures

Chapter 8 Study of Models of Steel Structures

198
 Roof Displacement vs. Normalized Base Shear
Generic and LA 9-Story, γ = 0.21, without P-∆
∆=
0.3

0.25

0.2
Vbase / W

0.15

0.1
NEHRP, SAC
NEHRP, Generic
0.05 SRSS, SAC
SRSS, Generic
0
0 50 100 150 200 250 300
δroof (cm)

(a) without P-delta

Roof Displacement vs. Normalized Base Shear

Generic and LA 9-Story, γ = 0.21, with P-∆
∆=
0.3

0.25

0.2
Vbase / W

0.15

0.1
NEHRP, SAC
NEHRP, Generic
0.05 SRSS, SAC
SRSS, Generic
0
0 50 100 150 200 250 300
δroof (cm)

(b) with P-delta

Figure 8.5 Comparison of Pushover Curves for Generic and LA 9-Story Structures

Chapter 8 Study of Models of Steel Structures

199
 Roof Displacement Time History
LP89lex, Generic and LA 3-Story, γ = 0.35, with P-∆
∆=
100
SAC
Generic
75

50
δroof (cm)

25

-25

-50
0 2 4 6 8 10 12 14 16 18 20
Time (sec)

(a) Record LP89lex

Roof Displacement Time History

KB95tato, Generic and LA 3-Story, γ = 0.35, with P-∆
∆=
50
SAC
Generic
25

0
δroof (cm)

-25

-50

-75

-100
0 2 4 6 8 10 12 14 16 18 20
Time (sec)

(b) Record KB95tato

Figure 8.6 Comparison of Roof Displacement Time Histories for Generic and
LA 3-Story Structures

Chapter 8 Study of Models of Steel Structures

200
 Roof Displacement Time History
LP89lex, Generic and LA 9-Story, γ = 0.21, with P-∆
∆=
25
SAC

0 Generic

-25
δroof (cm)

-50

-75

-100

-125
0 2 4 6 8 10 12 14 16 18 20
Time (sec)

(a) Record LP89lex

Roof Displacement Time History

KB95tato, Generic and LA 9-Story, γ = 0.21, with P-∆
∆=
100
SAC
Generic
50

0
δroof (cm)

-50

-100

-150

-200
0 2 4 6 8 10 12 14 16 18 20
Time (sec)

(b) Record KB95tato

Figure 8.7 Comparison of Roof Displacement Time Histories for Generic and
LA 9-Story Structures

Chapter 8 Study of Models of Steel Structures

201
 Roof Displacement Demands
∆
LP89lex, Generic and LA 3-Story, without P-∆
3
SAC

Record Multiplier 2.5 Generic

1.5

0.5

0
0 50 100 150 200 250 300
δroof (cm)

(a) without P-delta

Roof Displacement Demands

∆
LP89lex, Generic and LA 3-Story, with P-∆
3
SAC

2.5 Generic
Record Multiplier

1.5

0.5

0
0 50 100 150 200 250 300
δroof (cm)

(b) with P-delta

Figure 8.8 Comparison of Roof Displacement Demands for Generic and LA 3-Story
Structures, Record LP89lex

Chapter 8 Study of Models of Steel Structures

202
 Roof Displacement Demands
∆
LP89lex, Generic and LA 9-Story, without P-∆
3
SAC
Generic
Record Multiplier 2.5

1.5

0.5

0
0 100 200 300 400 500 600
δroof (cm)

(a) without P-delta

Roof Displacement Demands

∆
LP89lex, Generic and LA 9-Story, with P-∆
3
SAC
Generic
2.5
Record Multiplier

1.5

0.5

0
0 100 200 300 400 500 600
δroof (cm)

(b) with P-delta

Figure 8.9 Comparison of Roof Displacement Demands for Generic and LA 9-Story
Structures, Record LP89lex

Chapter 8 Study of Models of Steel Structures

203
 Normalized Story Drift Angle Demands
LP89lex, Generic and LA 3-Story, γ = 0.35, with P-∆
∆
1
SAC

Generic
0.8
Relative Height

0.6

0.4

0.2

0
0 0.25 0.5 0.75 1 1.25 1.5 1.75 2
θi / θroof

(a) Record LP89lex

Normalized Story Drift Angle Demands

KB95tato, Generic and LA 3-Story, γ = 0.35, with P-∆
∆
1
SAC

Generic
0.8
Relative Height

0.6

0.4

0.2

0
0 0.25 0.5 0.75 1 1.25 1.5 1.75 2
θi / θroof

(b) Record KB95tato

Figure 8.10 Comparison of Normalized Story Drift Angle Demands for Generic and
LA 3-Story Structures

Chapter 8 Study of Models of Steel Structures

204
 Normalized Story Drift Angle Demands
LP89lex, Generic and LA 9-Story, γ = 0.21, with P-∆
∆
1
SAC

Generic
0.8
Relative Height

0.6

0.4

0.2

0
0 0.5 1 1.5 2 2.5 3
θi / θroof

(a) Record LP89lex

Normalized Story Drift Angle Demands

KB95tato, Generic and LA 9-Story, γ = 0.21, with P-∆
∆
1

SAC

0.8 Generic
Relative Height

0.6

0.4

0.2

0
0 0.25 0.5 0.75 1 1.25 1.5
θi / θroof

(b) Record KB95tato

Figure 8.11 Comparison of Normalized Story Drift Angle Demands for

Generic and LA 9-Story Structures

Chapter 8 Study of Models of Steel Structures

205
 Normalized Story Drift Angle Demands
∆
LP89lex with Various Severity, LA 3-Story, with P-∆
1
LP89lex * 0.3

LP89lex * 0.5
0.8
LP89lex * 1.0

LP89lex * 1.5
Relative Height

0.6 LP89lex * 2.0

LP89lex * 2.5

0.4

0.2

0
0 0.25 0.5 0.75 1 1.25 1.5 1.75 2
θi / θroof

(a) LA 3-Story Structure

Normalized Story Drift Angle Demands

∆
LP89lex with Various Severity, Generic 3-Story, with P-∆
1
LP89lex * 0.3

LP89lex * 0.5
0.8
LP89lex * 1.0

LP89lex * 1.5
Relative Height

0.6 LP89lex * 2.0

LP89lex * 2.5

0.4

0.2

0
0 0.25 0.5 0.75 1 1.25 1.5 1.75 2
θi / θroof

(b) Generic 3-Story Structure

Figure 8.12 Comparison of Normalized Story Drift Angle Demands for Generic and LA
3-Story Structures, Record LP89lex

Chapter 8 Study of Models of Steel Structures

206
 Normalized Story Drift Angle Demands
∆
LP89lex with Various Severity, LA 9-Story, with P-∆
1
LP89lex * 0.3

LP89lex * 0.5
0.8
LP89lex * 1.0

LP89lex * 1.5
Relative Height

0.6 LP89lex * 2.0

LP89lex * 3.0

0.4

0.2

0
0 0.5 1 1.5 2 2.5 3
θi / θroof

(a) LA 9-Story Structure

Normalized Story Drift Angle Demands

∆
LP89lex with Various Severity, Generic 9-Story, with P-∆
1
LP89lex * 0.3

LP89lex * 0.5
0.8
LP89lex * 1.0

LP89lex * 1.5
Relative Height

0.6 LP89lex * 2.0

LP89lex * 3.0

0.4

0.2

0
0 0.5 1 1.5 2 2.5 3
θi / θroof

(b) Generic 9-Story Structure

Figure 8.13 Comparison of Normalized Story Drift Angle Demands for Generic and LA
9-Story Structures, Record LP89lex

Chapter 8 Study of Models of Steel Structures

207
 CHAPTER 9

DESIGN CONSIDERATIONS FOR NEAR-FAULT

GROUND MOTIONS

In the previous chapters it was shown that structures designed according to current
seismic guidelines might experience excessively large demands or undesirable
distributions of demands over their height when subjected to near-fault ground motions.
Various techniques are investigated in this chapter that can provide improved protection
for frame structures located in the near-fault region of a seismic source. A summary is
provided first of the results of a study that relates the parameters of the equivalent pulse
discussed in Chapter 7 to earthquake moment magnitude, Mw, and closest distance to the
fault rupture plane, R. Once these relationships are established, advantage is taken of the
response of structures to simple pulse-type motions in order to evaluate the Mw and R
dependence of seismic demands imposed by near-fault ground motions and to develop
design recommendations. Strengthening techniques are investigated that serve to reduce
the seismic demands either by modifying the distribution of story shear strength over the
height of the structure or by adding walls of different properties to frame structures. The
beneficial effects of strengthening with walls and the effects of different wall properties
are discussed in detail.

9.1. Relationships Between Equivalent Pulse Parameters and Earthquake

Magnitude and Distance

In Chapter 7 it was demonstrated that within the period range of T/Tp from 0.375 to 3.0 it
appears to be reasonable to represent near-fault ground motions by an equivalent pulse.
Although not precise, such a representation is sufficiently accurate for practical purposes.
The values of the equivalent pulse parameters were summarized in Tables 7.2 and 7.3 for
the fault-normal component of the recorded and simulated near-fault ground motions
with forward directivity used in this study. Regression analysis is performed on these

Chapter 9 Design Considerations for Near-Fault …

208
data to evaluate the magnitude (Mw) and distance (R) dependence of these parameters.
Since relatively small sets of near-fault ground motions are available for this purpose, the
results of such regression analysis should be interpreted with caution.

9.1.1. Pulse Period

Somerville showed that the period of the pulse (Tp) contained in a near-fault ground
motion is mostly affected by slip rise time, which is defined as the duration of slip at a
given point on the fault (Somerville, 1998). He established a relationship between rise
time and the moment magnitude (Mw) of the event. He also proposed a preliminary
model that expresses log10Tp as a linear function of Mw, independent of distance.
Employing the formulation proposed by Somerville for Tp and performing a linear
regression analysis on equivalent pulse periods obtained for the combination of the
recorded and simulated ground motions, the following regression equation is obtained:

log10Tp = -1.76 + 0.31 Mw (9.1)

Figure 9.1 illustrates this equation together with the data points (solid circles) to which
the line is fitted for (a) log10Tp and (b) Tp. It should be noted that in this figure some of
the circles represent more than one data point with identical Mw and Tp values. Two
points corresponding to the recent Turkey and Taiwan earthquakes, which were not
included in the regression analysis, are also shown. Superimposed on each graph are
90%, 80%, and 70% prediction bands for Tp, which illustrate the confidence levels for the
regressed model. For example, for a given value of Mw, there is a 90% likelihood that the
Tp value falls within the 90% prediction band that surrounds the regressed line. These
prediction bands are obtained based on the assumption that the errors and the dependent
parameter, i.e. log10Tp, are normally distributed (Ramsey and Schafer, 1996), which is to
say that Tp follows a lognormal distribution with its median being the Tp value obtained
from Eq. 9.1.

If only pulse periods for the recorded ground motions are used in the regression analysis,
the following equation is obtained:

log10Tp = -2.06 + 0.34 Mw (9.2)

Chapter 9 Design Considerations for Near-Fault …

209
Figure 9.2 illustrates the same information as presented in Fig. 9.1, but for only the
recorded ground motions. The following observations can be made from these figures:

•= The slopes of the regressed lines are very close for the two record sets, but the
intercept is somewhat smaller for the recorded set, resulting in consistently
smaller predictions based on the recorded set compared to the combined set of
ground motions.

•= Large scatter in data points is observed for both record sets. In part this can be
attributed to the fact that the ground motions used in the regression analysis come
from different events with different faulting mechanisms and geological
conditions. The large scatter translates into wide prediction bands, which indicate
large uncertainties in predicting Tp from the regression equations.

The large scatter may be interpreted as a lack of confidence in predicting Tp values from
the regression equations. Later in Section 9.2 it will be investigated how a limited
variation of Tp affects the base shear strength demands obtained from the equivalent
pulse approach.

As shown in figure 9.1, the right tail of the data used in the derivation of Eq. 9.1 consists
of data points corresponding to one record from the Landers (1992) earthquake with
magnitude 7.3 and six records from a simulated earthquake with magnitude 7.5 (with
identical Tp). Any extrapolation beyond the magnitude range of the data points should be
carried out with caution. The following example serves to evaluate to what extent the
regression equation can be employed for large earthquakes.

Somerville studied near-fault ground motions recorded in the magnitude 7.4 Izmit
(Turkey, 1999) and magnitude 7.6 Chi-Chi (Taiwan, 1999) earthquakes (Somerville,
2000). Two of the pulse-type records used in his study are Yarimca (Turkey) and
Tsaotun (Taiwan). The period of the pulse contained in the fault-normal component of
these ground motions (from the location of a global peak in the velocity spectrum; see
Somerville, 2000, for the spectra) is estimated to be 3.7 sec. and 4.2 sec., respectively.
Figure 9.1 compares these values with those predicted by Eq. 9.1, indicating a good
agreement.

Chapter 9 Design Considerations for Near-Fault …

210
9.1.2. Pulse Intensity

Pulse effective velocity (veff = aeffTp/4) is utilized here to measure the intensity of the
equivalent pulse contained in near-fault ground motions. Somerville developed a model
that expresses the peak ground velocity (PGV) of the record as a function of moment
magnitude and distance (Somerville, 1998). In Section 7.1.2 it was shown that pulse
effective velocity can be approximated by peak ground velocity with good accuracy.
Therefore, the formulation proposed by Somerville for PGV is used in a linear two-
variable regression analysis that relates veff to magnitude (Mw) and shortest distance from
the site to the fault rupture (R). Records with R values smaller than 3 km are not used in
the derivation of the regression equation on account of the logarithmic form of this
equation, which results in unreasonably large values for veff at small R values. A
combination of records, also used by Somerville, is utilized that includes the fault-normal
component of the recorded ground motions with forward directivity as well as a subset of
the simulated ground motions. This subset, which consists of simulations for Mw = 6.5
and 7.0 with strong forward directivity effects (station f8), is most compatible with the
recorded time histories (Somerville, 1998). The following equation results from the
regression analysis:

log10veff = -2.03 + 0.65 Mw – 0.47 log10R (9.3)

If only veff values for the recorded ground motions are used in the regression analysis, the
following equation is obtained:

log10veff = -4.04 + 0.88 Mw + 0.20 log10R (9.4)

Figure 9.3 illustrates these equations together with the data points to which the surfaces
are fitted for (a) combined and (b) recorded sets. The solid circles identify the data points
used in the regression analysis, and the empty circles represent points on the regression
surface with the same R and Mw values as the solid circles. A short distance between a
solid circle and an empty circle on the same vertical line indicates a close match between
the regressed surface and the data point. The following observations can be made:

•= Equation 9.3 represents a relationship that shows attenuation of pulse intensity

with distance and an increase in intensity with magnitude. However, Eq. 9.4
indicates an increase in pulse intensity (veff) with distance (R), which is not

Chapter 9 Design Considerations for Near-Fault …

211
 expected. As Fig. 9.3(b) demonstrates, after omitting the recorded ground
motions with R < 3 km, there remain only eight data points in the regression
analysis leading to Eq. 9.4. The insufficient number of data points may be the
main reason for this questionable observation.

•= From Fig. 9.3(a) it appears that the quality of the fit for veff from Eq. 9.3 is
relatively good. However, as pointed out previously, the data points used in the
derivation of this equation contain a relatively large number of simulations, which
may bias the outcome of the regression analysis. Thus, it is essential that Eq. 9.3
be improved as more recorded near-fault ground motions with forward directivity
become available.

•= As shown in Fig. 9.3(a), only data points with Mw ≤ 7.0 are used in the regression
analysis leading to Eq. 9.3. The regressed surface exhibits a relatively steep slope
at Mw = 7.0, indicating that the veff values obtained from Eq. 9.3 are very large for
events with Mw > 7.0. Table 9.1 compares recorded PGV values for three near-
fault ground motions with forward directivity from large earthquakes (recent
Turkey and Taiwan events) with the corresponding veff values obtained from Eq.
9.3. Presuming that the PGV of the recorded ground motion is close to the veff of
the equivalent pulse, it is evident that Eq. 9.3 overestimates the equivalent pulse
velocity by large amounts.

The graphs in Figs. 9.3(a) and (b), together with the data from recent large earthquakes
(Turkey and Taiwan), indicate that there are good reasons to question both Eqs. 9.3 and
9.4, particularly for magnitudes exceeding 7.0. Equation 9.4 is based on an insufficient
number of data points, and Eq. 9.3 contains simulated records whose relation to “reality”
has not been established. Recordings from recent large earthquakes give a strong
indication that the near-fault ground motion intensity, as measured by an equivalent
velocity or the PGV, saturates at a yet to be established magnitude. It is still believed that
the effective velocity (or even the PGV), together with the pulse period, is a good
measure of the pulse intensity, but more data points are needed to develop a more reliable
regression relationship between this intensity parameter and magnitude and distance.
Nevertheless, Eqs. 9.1 and 9.3 are used in the following sections to estimate the period
and intensity of the equivalent pulse for specific Mw and R values. Then, the response
properties of simple pulses are utilized to predict demands for structures subjected to
near-fault ground motions. These demands are very large for magnitudes approaching

Chapter 9 Design Considerations for Near-Fault …

212
7.0 and larger, and have to be viewed with caution, as modifications to Eqs. 9.1 and
particularly 9.3 may be needed as the magnitude approaches and exceeds 7.0.

9.2. Base Shear Strength Demands for Targeted Maximum Ductilities

Results from Eqs. 9.1 and 9.3 are tabulated in Table 9.2 for several Mw and R values.
The table shows that veff increases strongly with magnitude, whereas Tp and aeff are
moderately dependent on magnitude. The attenuation of aeff and veff with distance is the
same because for a given magnitude the pulse period, which relates aeff and veff, remains
constant. Values for Mw = 7.5 are also listed, but they are not used in the subsequent
evaluation because they are based on extrapolation beyond the range of available data
points. In fact, the values of veff and aeff at this large magnitude are believed to be much
too large.

Given the site/source parameters R and Mw, the pulse parameters Tp and ag,max (= aeff =
4veff/Tp) can be estimated from Eqs. 9.1 and 9.3 (see Table 9.2), and the η–T/Tp curves
presented in Fig. 6.25 can be converted into γ-T curves [γ== Vy/W = (ag,max/g)η]. The γ-T
curves represent MDOF base shear strength demand spectra for specified target
ductilities. Examples of such base shear strength demand spectra are presented in Figs.
9.4 and 9.5 for various combinations of magnitude and distance, using the η–T/Tp curves
for pulses P2 and P3, respectively, obtained for the generic 20-story structure without P-
delta effects. Superimposed on each graph are two UBC‘97 soil type SD design spectra
for Seismic Zone 4, one with and one without the code specified near-fault factor. In
calculating the near-fault factor according to the UBC guidelines the following source
types are used: Type A for Mw = 7.0 and 7.5, Type B for Mw = 6.5, and Type C for Mw =
6.0. The code spectrum is scaled down by a factor of 4, which is arrived at by choosing a
strength reduction factor of 8 (special moment frame) and assuming an overstrength
factor of 2 for each story. The graphs illustrate the magnitude and distance dependence
of the base shear strength demands obtained from the equivalent pulse approach, and put
these demands in perspective with the values used in present design practice. It must be
emphasized that the MDOF base shear strength demand spectra have been obtained for
generic 20-story frame structures whose relative story shear strengths are tuned to an
SRSS lateral load pattern.

The following observations can be made from the MDOF constant ductility strength
demand spectra presented in Figs. 9.4 and 9.5:

Chapter 9 Design Considerations for Near-Fault …

213
 •= The base shear strength required to limit story ductility demands to a target value
strongly depends on magnitude, distance and the fundamental period of the
structure.

•= A comparison of the strength demand spectra with the UBC/4 curves reveals that
for given magnitude and distance values, a structure designed according to present
code provisions will experience very different levels of inelasticity depending on
the fundamental period. For instance, Fig. 9.4(b) indicates that under a ground
motion of the type represented by pulse P2 for a magnitude 7.0 event at R = 3 km,
structures designed according to the UBC’97 (including the near-fault factor) are
expected to experience story ductility demands less than 4 when the period is
longer than 3.4 sec., but experience ductility demands larger than 8 if the period is
between 1.0 and 2.2 sec.

•= The shapes of the MDOF strength demand spectra differ significantly from the
UBC/4 curves. Assuming that the µ = 4 spectrum for Mw = 7.0 and R = 3 km is
representative of design conditions, it is observed that this spectrum has a much
wider constant strength plateau than the UBC/4 curve (up to about 2.0 sec.
compared to 0.8 sec. for UBC/4). This plateau is followed by a decreasing
strength demand region that falls below the UBC/4 curve around 3.4 sec. This
indicates that long period structures are well protected by the UBC’97 provisions,
short period structures are adequately protected, but structures of intermediate
period (from about 0.8 to 3.4 sec. with the postulated scenario) are inadequately
protected, i.e., the ductility demands exceed 4.0. In the period range from about
1.5 to 2.5 sec. the strength should be increased by a factor larger than 2.0 in order
to keep the maximum ductility demands below 4.0. The incorporation of P-delta
effects, which are neglected here, will make matters even worse.

•= As far as maximum ductility is concerned, the differences in strength demands for

ground motions represented by either pulse P2 or pulse P3 are not very large.
However, pulse P3 will cause more cumulative damage than pulse P2.

The introduction of near-fault factors in the UBC’97 provisions is an improvement that

gives recognition to the existence of the problem. But, the results presented here indicate
that the code criteria do not provide a consistent level of protection against near-fault
ground motions, and in certain period ranges provide an inadequate protection. The

Chapter 9 Design Considerations for Near-Fault …

214
problem cannot be solved by introducing additional factors to conventional design
spectra, and a more rigorous approach appears to be necessary. The spectra presented in
Figs. 9.4 and 9.5 provide the basis for such an approach.

However, there are a number of issues that need to be addressed before definite
conclusions can be drawn and an approach based on magnitude and distance dependent
pulse response spectra can be implemented. They have to do with the evaluation of the
magnitude dependent hazard at the site, and with the confidence that can be placed in the
prediction of veff and Tp. In Section 9.1.2 it was shown that Eq. 9.3 overestimates the
values of veff for large earthquakes (Mw > 7.0). Therefore, caution should be exercised in
estimating the pulse intensity parameter for large events.

As mentioned earlier, the large scatter of data shown in Figs. 9.1 and 9.2 provides little
confidence in estimating Tp as a function of magnitude. The question that needs to be
addressed is that to what extent the uncertainty in Tp can affect the base shear strength
demands obtained from the equivalent pulse approach (Figs. 9.4 and 9.5). This is
investigated in Fig. 9.6, which illustrates the sensitivity of the base shear strength
demands to a variation in Tp. The figure shows the strength demands for pulse P2, Mw =
7.0, R = 3 km, and target story ductility ratios of µmax = 2 and 8. Tp is varied around the
value obtained from Eq. 9.1 for Mw = 7.0 (i.e., Tp = 2.6 sec), while veff is evaluated from
Eq. 9.3. As can be seen, the equivalent pulse strength demands for target ductilities of 2
and 8 are affected to various degrees, depending on T and µmax, and it becomes a matter
of judgment in what range the scatter in strength demands is acceptable.

In the following discussion it is assumed that the demands predicted from pulse
parameters based on the regression equations 9.1 and 9.3 are realistic representations.
Since these demands are very high in specific period regions, the development of
strengthening procedures is called for. Several alternatives are discussed in the next two
sections.

9.3. Effect of Story Shear Strength Distribution on Ductility Demands

As discussed in Chapter 3, an SRSS-based story shear strength distribution over the

height has been assigned to the generic structure. All the results presented so far are
based on this strength distribution. Figures 4.12 and 6.19, among others, indicate that
this strength distribution leads to large variations of ductility demands over the height of

Chapter 9 Design Considerations for Near-Fault …

215
structures subjected to near-fault and pulse-type ground motions. Therefore, the standard
SRSS distribution may not be the most suitable one for consistent protection against near-
fault effects. In this section, other story shear strength distributions are investigated, and
their advantages and disadvantages compared to the SRSS distribution are demonstrated.

9.3.1. Story Shear Strength Distribution for Uniform Story Ductility

It is postulated that an ideal story shear strength distribution would result in a uniform
distribution of story ductility over the height of the structure, in order to efficiently utilize
the energy dissipation capacity available in all elements. The story shear strength
distribution that induces uniform ductility over the height can be found through an
iterative process in which story shear strengths are varied until a targeted uniform story
ductility is achieved. Such shear strength distributions are investigated for various T/Tp
ratios and target ductilities ranging from 1 (elastic behavior) to 8. Representative results
are shown in Fig. 9.7 for pulse P2 and T/Tp = 1.0 and 2.0. The story shear strength
values are normalized by the base shear strength required to achieve the desired uniform
ductility. The SRSS story shear strength distribution is also superimposed. The trends
are distinctly different for T/Tp = 1.0 and 2.0. The following observations can be made:

•= For T/Tp = 2.0 and close to elastic behavior (µ = 1 and 2), relatively high strength
is required around 2/3rd up the structure to control ductility demands in the top
portion, whereas for a uniform ductility of 3 or larger the strength demands are
high at the base and decrease rapidly with height. Thus, the story shear strength
distribution for uniform ductility changes radically with the target ductility ratio,
and the change takes place within the narrow ductility range from 2 to 3. The
results indicate that there is no ideal story shear strength distribution for long
period structures (T/Tp > 1.0) at all performance levels. Unless the story shear
strength in the upper portion of the structure is relatively large (compared to the
base shear), early yielding at about 2/3rd of the height has to be expected.

•= The situation is considerably different for T/Tp = 1.0, where the story shear
strength distribution that causes uniform ductility does not strongly depend on the
target ductility ratio. The distribution is close to the SRSS distribution for small
ductility ratios and gradually approaches a linear shape for large ductility ratios.

Chapter 9 Design Considerations for Near-Fault …

216
The base shear strength required to limit the maximum story ductility to specific target
values for pulse P2 and T/Tp = 2.0 and 1.0, respectively, is shown in Figs. 9.8(a) and (b).
The normalized base shear strength is represented by the parameter η = Vy/([Link]).
Results for two story shear strength distributions are shown: the solid line illustrates the
base shear demands for the shear strength distribution that causes uniform story ductility,
and the dotted line illustrates the base shear demands for the SRSS distribution. The
following observations can be made:

•= For flexible structures (T/Tp = 2.0), the base shear strength demands strongly
depend on the story shear strength distribution for a maximum story ductility of
up to about 3.5, whereas for larger ductilities the demands are insensitive to the
distribution of strength over the height. For these structures ductility control to a
large extent is a function of both base shear strength and story shear strength
distribution, except for relatively weak structures (µmax > 3.5) in which the first
story is the critical one and the story shear strength distribution no longer matters.

•= The observations made for flexible structures do not apply to relatively stiff
structures (T/Tp = 1.0). The base shear strength required to limit the maximum
ductility to a specific target value is not very sensitive to the selected story shear
strength distribution, for all levels of ductility. The more important concern is the
large η value required to limit the maximum story ductility to a reasonable value.

The conclusion is that no single story shear strength distribution will provide consistent
protection at all performance levels and for structures with different periods. For long
period structures either early damage in upper stories has to be tolerated (caused by
ductility demands on the order of 3 to 4) or very high strength has to be provided over a
large portion of the height. The strengthening of the lower portion of the structure will be
very effective in reducing excessive ductility demands at the bottom.

9.3.2. Strengthening Schemes Based on Base Shear Strength and Story Shear
Strength Distribution

The results presented in the previous section demonstrate that the ductility demands in
the upper portion of frame structures stabilize around 3 to 4, regardless of the story shear
strength distribution employed in design. For weak structures (or very severe near-fault
ground motions) the large demands usually occur in the bottom stories. Thus, in order to

Chapter 9 Design Considerations for Near-Fault …

217
prevent excessive ductility demands in severe events, it appears to be appropriate to
strengthen the bottom portion of the structure compared to the standard SRSS story
strength distribution. However, since such strengthening is associated with extra cost,
cost effectiveness becomes an overriding issue, with the objective being to minimize the
additional cost while providing adequate ductility control. This section summarizes
several attempts to modify story shear strength distributions in a manner that will reduce
the maximum story ductility demands while keeping down the cost of strengthening.
Again, pulse P2 is assumed to represent near-fault ground motions. The following four
story shear strength distributions, which are illustrated in Fig. 9.9, are investigated:

a) SRSS: the SRSS story shear strength distribution forming the baseline for this
study, which is summarized in Section 3.2.2
b) SRSS + Strengthening: the SRSS distribution with the lower 30% of the structure
strengthened with a linear strength increase leading to 40% extra strength at the
base
c) Linear: a linear story shear strength distribution with a shear strength in the top
story equal to 20% of the base shear
d) Linear + Strengthening: the linear distribution with a strengthened bottom portion
defined in the same manner as in (b)

In all four cases the story stiffness is assumed to remain unchanged and is equal to that of
the SRSS base case. The “SRSS + Strengthening” distribution requires extra steel weight
compared to the base case, whereas the “Linear” distribution results in weight savings. If
for simplicity it is assumed that weight is proportional to strength demand, the weight of
the “Linear + Strengthening” distribution will be almost equal to that of the base case.
The base shear strength for all four distributions is defined in terms of the parameter η =
Vy/([Link]) of the SRSS base case without strengthening, i.e., the actual base shear of the
cases with strengthening is defined by 1.4η. Since this η=is only a reference value for the
strengthened frames, it is denoted as η* in Figs. 9.10 and 9.11.

The story ductility demands for structures with T/Tp = 0.5, 1.0, and 2.0, and the four story
strength distributions previously defined, are shown in Fig. 9.10. The graphs can be
compared with the base case (Fig. 9.10(a)) to evaluate the benefits and shortcomings of
each distribution. The following observations can be made:

Chapter 9 Design Considerations for Near-Fault …

218
 •= For long period structures (T/Tp = 2.0) using the linear rather than SRSS strength
distribution makes essentially no difference in the bottom stories, but causes
somewhat larger ductility demands in the upper stories. Adding strength to the
bottom portion of the structure does not much affect the ductility demands in the
upper stories but has a significant benefit in the lower stories; the demands
become more uniform over the bottom portion and the maximum ductility
demand decreases by a factor that is in most cases larger than the strength
increase factor of 1.4. (Since the story yield drift increases by the same factor as
the story shear strength, the reduction in the story drift demands is smaller than
that in the story ductility demands.) Thus, if the main objective is to protect the
bottom stories from excessive ductility demands, then the use of a bilinear story
shear strength distribution (Linear + Strengthening) appears to be very cost
effective. The negative aspect is that the reduced strength in the upper stories will
lead to earlier yielding in these stories compared to the base case. Better control
of upper story ductility demands is achieved by using the SRSS + Strengthening
distribution.

•= For short period structures (T/Tp = 0.5 and 1.0) the relative benefits are not well
established. Using the linear rather than SRSS strength distribution affects the
ductility distribution over the height but has relatively little effect on the
maximum story ductility demand. Providing additional strength in the bottom
portion greatly reduces the ductility demand in the first story and leads to an
upward shift of the critical story, but does not decrease the maximum demand
significantly; for relatively strong structures, it even increases the demands
compared to the unstrengthened structure. This applies particularly for structures
with the linear strength distribution. Again, the use of the SRSS + Strengthening
distribution appears to be a desirable compromise.

The conclusion is that strengthening of the bottom portion is an effective technique to

control excessive ductility demands in the bottom stories, especially for flexible (T/Tp >
1.0) and relatively weak structures (or severe pulses), in which the migration of the
critical story from the upper portion to the bottom story has occurred. The beneficial
effects of this technique are limited to the strengthened stories and do not extend beyond.

Chapter 9 Design Considerations for Near-Fault …

219
P-Delta Effects:

P-delta effects are usually largest in the first story, where the cumulative gravity load is
maximum. Thus, there could be a considerable benefit in moving the maximum ductility
demand away from the first story. The strengthening of the bottom portion, which moves
the critical story upwards, is likely to decrease the P-delta effects considerably. This
hypothesis is investigated in Fig. 9.11, which illustrates the distributions of story ductility
demands in the presence of P-delta effects for structures with T/Tp = 0.5, 1.0, and 2.0, and
story strengths distributed according to the (a) SRSS and (b) SRSS + Strengthening
distributions. The following observations can be made:

•= For T/Tp = 2.0 and maximum ductility demands occurring in the bottom story, the
strengthening of the bottom portion is even more beneficial than for the case of
strengthening in the absence of P-delta effects. For relatively strong structures, in
which the critical story is in the top portion, no beneficial effect is observed by
adding strength to the bottom stories.

•= For T/Tp = 1.0 the strengthening scheme is again very effective in controlling high
ductility demands in the bottom story, but somewhat increases the maximum
demand for strong structures.

•= As discussed in Section 6.2.2, P-delta effects are largest for short period structures
[provided that the stability coefficient of 0.1, which is used here to evaluate P-
delta effects, applies to both long and short period structures]. This can also be
observed from a comparison of Figs. 9.10(a) and 9.11(a) for T/Tp = 0.5. The
results show that strengthening of the bottom stories is very effective in reducing
the maximum ductility demand in short period structures at almost all
performance levels – if the P-delta effect is significant.

The general conclusion is that in cases in which P-delta effects are significant, providing
additional strength to the bottom portion of the structure is an efficient technique to
control ductility demands for structures in all period ranges.

Chapter 9 Design Considerations for Near-Fault …

220
9.4. Strengthening of Frames with Walls

The previous section provided information on frame strengthening schemes consisting of

an increase in the base shear strength and modifications to the distribution of story shear
strength over the height of the structure. As the results indicate, despite many successes,
no single strengthening scheme can provide consistent improved protection of frames at
all performance levels and for the full range of periods. As an alternative to those
strengthening schemes, shear walls can be added to the frames to form a dual system.
This section summarizes a study that investigates the demands for such dual systems
subjected to pulse P2. The objective of this study is to evaluate the extent to which
strengthening of frames with walls can improve the performance of structures in the near-
fault region of an earthquake. In addition to conventional fixed-base walls, walls that can
freely rotate about their bases are also studied.

9.4.1. Dual Systems Investigated in this Study

The generic 20-story frame introduced in Section 3.2 is strengthened by adding a

prismatic wall (constant cross-section over height of structure) that is either fixed or
hinged at the base. Typical deflected shapes of these two dual systems are shown in Fig.
9.12. As illustrated, the wall is horizontally linked to the frame at story levels.

The wall to frame stiffness ratio, Kw/Kf, is varied to investigate the effects of wall
stiffness on the seismic demands of the system. The wall stiffness Kw is defined as the
point load applied at the top of the wall corresponding to a unit displacement at the top,
provided that the base is fixed and the wall behavior is elastic. If only prismatic walls are
considered and shear deformations are neglected, Kw = 3EwIw/H3, where Ew, Iw, and H are
the wall elastic modulus, moment of inertia, and height, respectively. Likewise, the
frame stiffness Kf is defined as the point load at the roof level that causes a unit roof
displacement. Elastic walls as well as inelastic walls with constant shear and bending
strength over the height are considered. The base shear strength coefficient introduced in
Chapter 6, i.e., ηf = Vframe,y/([Link],max), is used to define the frame strength. A similar
strength coefficient, ηw = Vwall,y/([Link],max), is utilized to define the shear strength of the
wall.

Figure 9.13 shows base shear versus roof displacement diagrams obtained from pushover
analyses of systems with (a) fixed and (b) hinged walls, in which the wall is assumed to

Chapter 9 Design Considerations for Near-Fault …

221
behave elastically. The SRSS lateral load pattern, to which the story stiffness and
strength of the frame are tuned, is used in the pushover analyses. Adding a fixed wall to
the frame increases the stiffness of the system significantly, and hence shortens the
fundamental period of the system. The fundamental period of the so strengthened
system, normalized by the period of the unstrengthened frame, is listed in Table 9.3 for
various values of the Kw/Kf ratio. On the other hand, as the results of the pushover
analysis indicate, if the wall is hinged at the base, it affects the stiffness of the system
neither before nor after the formation of plastic hinges in the frame. The addition of a
hinged wall also does not affect the pushover strength of the system. The reason is that
the deflected shape of the frame subjected to the SRSS load pattern is a straight line (see
Section 3.2), which corresponds to the rigid body motion of the hinged wall. Thus,
regardless of wall stiffness, the stiffness of the system with the hinged wall does not
change if the SRSS lateral load pattern is applied. The strength of the system also is
unaffected because all plastic hinges in the frame form simultaneously under the SRSS
lateral load pattern. It is also important to note that, since the first mode shape of the
frame is close to a straight line, adding the hinged wall to the frame does not affect the
first mode period of the system.

9.4.2. Demands for Dual Systems with Elastic Walls

In this section, drift and force demands for systems with elastic fixed and hinged walls
subjected to pulse P2 are evaluated. Both strong frames (low ductility demands) and
weak frames (high ductility demands) with fundamental periods of T = 0.5 Tp, 1.0 Tp, and
2.0 Tp are considered. In each case the Kw/Kf ratio is varied within the range from 0.0
(no wall) to 2.0 (stiff wall) to study the effect of wall stiffness on the demands.

Story Drift Demands:

Figure 9.14 illustrates the effect of adding elastic fixed walls with various stiffnesses on
the distribution of story drift demands for systems subjected to pulse P2. The story drift
demands are normalized by the maximum ground displacement ug,max. Figure 9.15 shows
corresponding results for systems strengthened with a wall that is hinged at the base. The
drift demands for systems with strong (large η) and weak (small η) frames are presented
side by side. Results for the unstrengthened system (Kw = 0) are shown in heavy solid
lines, on which the maximum story ductility demand is indicated. For the systems with
fixed walls, the period T corresponds to the frame alone. As explained earlier, the

Chapter 9 Design Considerations for Near-Fault …

222
fundamental period of the strengthened system is considerably shortened by adding a
fixed wall. The following observations can be made from Figs. 9.14 and 9.15:

•= Fixed walls are very effective in reducing the drift demands for structures with
T/Tp < 1.0, but become much less effective in reducing the demands for T/Tp ≥
1.0. In some cases the addition of a fixed wall even increases the maximum drift
demand. The reason for the latter is higher mode effects as well as the fact that
for T/Tp > 0.75 the elastic strength demand spectrum of pulse P2 is descending,
and thus the period shortening of the system due to strengthening with a fixed
wall leads to a rapid increase in base shear demands.

•= Hinged walls are effective in reducing the drift demands for frames with T/Tp
both smaller and larger than 1.0. For systems with T/Tp = 0.5 the reduction is
significant for weak systems and can be accomplished with relatively flexible
walls. Similar observations also hold true for systems with T/Tp = 1.0. The
effectiveness of adding hinged walls is largest for T/Tp = 2.0, where walls
significantly reduce the large drift demands in the upper portion of strong
structures and in the bottom stories of weak structures. However, relatively stiff
walls are needed to accomplish significant drift reduction in the case of weak
frames with high ductility demands at the base.

•= In a system strengthened with a hinged wall, three components contribute to the

horizontal displacement of the wall: 1) bending deformations, 2) shear
deformations, and 3) rotation about the base. If the wall is stiff compared to the
frame, bending and shear deformations will become small and a rigid body
rotation will dominate the wall displacement, resulting in a close to uniform
distribution of drift demands over the height. Shear deformations are ignored
here, but their effects will be studied later in this section.

•= The wall stiffness (relative to frame stiffness) needed for effective drift control
increases with the T/Tp ratio. For strengthening with a hinged wall, minimum
required Kw/Kf values for effective drift control are deduced to be 0.25, 0.5, and
2.0 (a value as low as 1.0 could have been chosen) for T/Tp ratios of 0.5, 1.0, and
2.0, respectively. Hence, considering that the frame stiffness is proportional to
1/T2, for a given Tp value, the absolute wall stiffness (Kw) values for T/Tp = 0.5,
1.0, and 2.0 will be proportional to 1.0:0.5:0.5.

Chapter 9 Design Considerations for Near-Fault …

223
 •= For all three weak frames (with different T/Tp) strengthened with stiff hinged
walls, the drift demand is almost the same (about 0.06 ug,max) regardless of the
frame period. The reason is that when the hinged wall is stiff compared to the
frame, the dual system acts as an SDOF oscillator, whose only mode of vibration
is a straight line. As shown in Fig. 6.31, the inelastic displacement demand of
weak SDOF systems is almost independent of the T/Tp ratio because a weak
system yields rapidly resulting in a large reduction of the stiffness. The
displacement demand for systems with small stiffness is close to the peak ground
displacement and has little to do with the initial elastic period.

Wall Shear and Moment Demands:

From the study of drift demands it appears that for a given wall stiffness, elastic fixed
walls are more effective in reducing drift demands for short period frames (T/Tp < 1.0),
while elastic hinged walls are more effective for long period frames (T/Tp > 1.0).
However, these results do not address force (shear and moment) demands imposed on the
wall. The force demands will determine the feasibility of making the walls sufficiently
strong to justify the assumption of elastic wall behavior. Figures 9.16 and 9.17 illustrate
elastic shear strength demands for fixed and hinged walls of different stiffness used to
strengthen frames of various periods and strength subjected to pulse P2. The results are
organized in the same manner as in Figs. 9.14 and 9.15. The wall shear demands are
normalized by [Link],max, where m is the total mass of the structure, and ag,max is the
maximum ground acceleration of the pulse. Similarly, elastic moment demands for the
fixed and hinged walls are illustrated in Figs. 9.18 and 9.19. The wall moment demands
are normalized by [Link],max, where H is the total height of the structure. The following
observations can be made from these figures:

•= For the stiff fixed walls the shear strength demands become very large in the
bottom stories regardless of frame strength and T/Tp. These demands are much
larger than those for the elastic hinged walls. This is in part because of the period
shift that occurs by adding a fixed shear wall to the system.

•= When a fixed wall is used for strengthening, the shear demands increase rather
consistently with the stiffness of the wall for T/Tp ≥ 1.0, whereas this pattern is
not observed consistently for T/Tp = 0.5. This aberration can be rationalized by
large and frequent peaks and valleys in the elastic strength demand spectrum of

Chapter 9 Design Considerations for Near-Fault …

224
 pulse P2 for T/Tp < 0.5, which is the period range that contains the shifted
fundamental period of the strengthened system.

•= Even though there is only a negligible shift of the fundamental period caused by
adding a hinged wall, a similar aberration is also observed in the hinged wall case,
i.e., the shear demand does not necessarily increase with wall stiffness. The
reason is that while the first mode period of the system remains unchanged,
adding a hinged wall to the system affects its higher mode periods, shortening
them significantly. If the shifted higher mode periods – especially the second
mode – coincide with one of the large peaks of the elastic strength spectrum,
higher mode effects become more significant, resulting in a larger shear demand
for the strengthened system.

•= The distributions and values of the shear demands over the height for hinged
walls are similar, regardless of T/Tp and frame strength. This provides a great
advantage in design, considering the large uncertainties in predicting the period of
the equivalent pulse (see Section 9.1.1). The maximum base shear strength
demand in almost all cases studied here is about [Link],max. The demands show
little sensitivity to wall stiffness for T/Tp = 0.5, but this sensitivity increases for
longer periods.

•= The wall moment demands for the hinged walls are merely a small fraction of the
demands for the fixed wall. While the moment demands for the fixed walls
consistently increase from top to bottom, the maximum demands for the hinged
walls occur at around 40% of the height and do not exceed [Link],max by
much.

In the results presented so far, both weak and strong frames were considered in order to
explore the effects of adding walls to frames of different strength. However, for the
purpose of practicality, frame strength should be looked at in the context of realistic
design requirements. Assuming that the frame is designed for an event with Mw = 7.0
and R = 3 km, the period and intensity of the equivalent pulse are estimated from Eqs. 9.1
and 9.3, i.e., Tp = 2.6 sec. and ag,max = 4veff/Tp = 0.31g. Then, considering a strength
reduction factor of 8 and an overstrength factor of 2, the base shear strength (γ = Vy/W)
of the frame is obtained from the UBC design spectrum with the near-fault factor for Mw
= 7.0 and R = 3 km (see Fig. 9.4). For T/Tp = 0.5, 1.0 and 2.0 [or T = (0.5)(2.6) = 1.3,

Chapter 9 Design Considerations for Near-Fault …

225
2.6, and 5.2 sec.] the so obtained strength coefficient γ is 0.23, 0.15, and 0.15,
respectively. Finally, the base shear strength coefficient, η, is computed from the
equation η = γ/(ag,max/g), i.e., η = 0.74, 0.48, and 0.48. In the remainder of this chapter,
for each of the three T/Tp values, results are presented only for the frame whose base
shear strength coefficient is closer to these η values (i.e., η = 0.75, 0.2, and 0.5, for T/Tp
= 0.5, 1.0, and 2.0, respectively).

Effect of Shear Deformations:

The effects of wall shear deformations on drift demands have been ignored so far.
Typically, shear deformations contribute significantly to the total deformation if the wall
M/V ratio is small and the wall is stocky. For the hinged walls the M/V ratio is
approximately ([Link],max)/([Link],max) ≅ 0.25H, which indeed is small. Figure
9.20 illustrates distributions of story drift demands, incorporating the effects of shear
deformations, for frames strengthened with (a) fixed and (b) hinged walls with aspect
ratios (b/H) ranging from 0.0 (no shear deformations) to 0.5 (stocky wall). In each case a
(Kw/Kf) value is chosen that corresponds to the minimum hinged wall stiffness required
to achieve a relatively uniform distribution of drift over the height in the absence of shear
deformations. A rectangular wall cross-section and a Poisson’s ratio of ν = 0.17 are
considered in computing the shear deformations. The following observations can be
made:

•= For the fixed walls the effects of shear deformations are small in the top portion
and large in the bottom portion of the wall. Nevertheless, the drift demands at the
bottom of the structure are still much smaller than the demands at the top, and are
not a matter of concern.

•= For the hinged walls the effects of shear deformations are relatively small
everywhere, even for stocky walls. The small shear deformation effects may
come as a surprise especially in the lower portion of the wall, where shear
demands are large and moment demands are very small. However, it should be
noted that these results are obtained for walls in which rotation about the base
dominates over flexural and shear modes of deformation.

The general conclusion is that it is acceptable to neglect the effects of wall shear
deformations on drift demands for frames strengthened with hinged walls.

Chapter 9 Design Considerations for Near-Fault …

226
P-Delta Effects:

To illustrate the beneficial effects of strengthening in the presence of P-delta effects, Fig.
9.21 shows distributions of story drift demands over the height for systems strengthened
with (a) fixed and (b) hinged walls when P-delta effects are taken into account. The
heavy solid lines represent the distribution for unstrengthened frames with the gravity
loads introduced in Section 3.2.3. The same gravity loads are used for strengthened
frames to trigger P-delta effects. The following observations can be made:

•= As discussed in Section 6.2.2, P-delta effects typically increase maximum story

ductility (or drift) demands for unstrengthened frames. However, a comparison of
the drift distribution for T/Tp = 0.5 in Fig. 9.21 with the corresponding
distribution in Fig. 9.14 indicates that for a frame with T/Tp = 0.5 and η = 0.75 the
maximum drift is smaller in the presence of P-delta effects. This aberration,
which appears as a sharp kink in the η-µmax curve shown in Fig. 6.21(b) for P2, is
limited to short period frames in a narrow range of strength, and can be explained
by period shifts due to P-delta effects in a very sensitive region of the elastic
spectrum of pulse P2. Nonetheless, adding a fixed wall reduces the maximum
demand effectively. Strengthening with a hinged wall is also beneficial, although
to a lesser degree.

•= For T/Tp = 1.0 and η = 0.2, a comparison shows that P-delta effects lead to a
considerable increase in maximum story drift demands for unstrengthened frames.
While adding a fixed wall may not be effective in reducing the drifts in the top
stories, strengthening with a hinged wall proves to be very effective.

•= For T/Tp = 2.0 and η = 0.5, P-delta effects for the unstrengthened frame are not
large because the maximum story drift does not occur in the bottom story, where
gravity loads are largest. Nevertheless, similar to the case without P-delta effects,
adding hinged walls effectively reduces the maximum drift demands.

The conclusion from these results is that strengthening frames with walls reduces story
drift demands of frame structures, especially excessive demands in bottom stories, and
therefore can provide improved protection against P-delta effects.

Chapter 9 Design Considerations for Near-Fault …

227
9.4.3. Demands for Inelastic Walls

From the results presented so far, it appears that strengthening with a wall hinged at the
base is an attractive technique for reducing the drift demands of frame structures
subjected to pulse-type ground motions. However, the results have been obtained under
the assumption that the wall behaves elastically, which may be difficult to achieve when
the shear or moment demands of the wall become large. The formation of plastic regions
in fixed walls is almost inevitable during a major event. If a fixed wall is designed
carefully, a brittle shear failure, to which a reinforced concrete wall is most susceptible,
can be prevented. However, shear yielding in the bottom portion of a hinged wall, where
shear demands are large and moment demands are small, may not be avoidable.

In this section demands for systems with inelastic fixed and hinged walls subjected to
pulse P2 are investigated. Flexural yielding in fixed walls and shear yielding in hinged
walls are studied to assess the effects of wall inelastic behavior on the benefits of
strengthening, highlighted previously for elastic walls. Three frames with fundamental
periods of T = 0.5 Tp, 1.0 Tp, and 2.0 Tp and base shear strength coefficients of ηf = 0.75,
0.2, and 0.5 are considered. For each T/Tp the smallest Kw/Kf ratio is chosen that can
provide a relatively uniform distribution of drift over the height based on the results
presented in the previous section for elastic hinged walls. For comparison purposes, the
same Kw/Kf values also are used for the fixed walls.

Hinged Walls:

Figure 9.22 illustrates distributions of various demands over the height for systems
strengthened with hinged walls. The plots serve to compare demands for systems with
elastic and inelastic walls. The heavy solid line in Fig. 9.22(a) represents story drift
demands for unstrengthened frames. The inelastic walls can yield in shear in any story in
which the shear demand reaches the shear capacity, which is assumed to be constant over
the height. The wall shear capacity is varied using the shear strength parameter ηw =
Vwall,y/([Link],max), which is denoted as “eta(w)” in the figure. A bilinear hysteretic story
shear-drift relationship is used with no strain hardening. Wall elastic shear stiffness is
determined assuming a rectangular cross-section, an aspect ratio (b/H) of 0.25, and a
Poisson’s ratio of 0.17. The following observations can be made:

Chapter 9 Design Considerations for Near-Fault …

228
 •= As Fig. 9.22(a) indicates, shear yielding of the wall does not cause a significant
increase in the total story drifts of the system (for T/Tp = 1.0 it even decreases the
demands). A comparison with the demands for the unstrengthened frame shows
that adding a hinged wall with limited shear strength can still effectively reduce
the drift demands of the frame and provide improved protection.

•= Figure 9.22(b) shows that the wall shear drift demands (contribution of wall shear
deformations to drift) for elastic walls are negligible (compared to the total drifts),
but the demands increase rapidly in the bottom portion of the wall when shear
yielding occurs. Of all the cases studied here, the wall shear drift demand is
largest for T/Tp = 2.0, ηf = 0.5, and ηw = 0.2, i.e., δshear = 0.047ug,max. Using Eqs.
9.1 and 9.3 for a scenario event with Mw = 7 and R = 3 km, and considering the
relationship ug,max = vg,maxTp/4 for pulse P2, the peak ground displacement can be
estimated as ug,max = 51 in. Thus, the drift angle due to wall shear deformations is
computed to be δshear/h = (0.047)(51)/(144) = 0.017, where h is the story height.
This drift angle is large, but later in this section it will be shown that ηw = 0.2
corresponds to a low shear strength capacity of the wall, and hence the shear drift
demands will be smaller than the value calculated here.

•= Figure 9.22(c) demonstrates that for very weak walls, the top portion of the wall
may also yield in shear. However, unlike the bottom, the shear yielding at the top
is not associated with large inelastic shear deformations in the wall.

•= Figure 9.22(d) shows that the shear yielding of the wall consistently reduces the
wall moment demands.

This pilot study indicates that some shear yielding in hinged walls used to strengthen
frames against near-fault ground motions has to be expected. Some shear yielding should
be acceptable for reinforced concrete walls, particularly if the axial force in the wall is
kept low. Steel shear walls may provide an effective alternative that can accommodate
large inelastic shear deformations.

Fixed Walls:

It is assumed that the fixed walls can only yield in bending with no shear plastification.
Thus, wall shear deformations are neglected as with the elastic walls in the preceding

Chapter 9 Design Considerations for Near-Fault …

229
section. Flexural yielding is allowed to occur at any point over the height at which the
moment demand reaches the wall moment capacity, which is considered to be constant
over the height. The wall moment capacity is defined using the parameter κ =
My/([Link],max), where My is the moment capacity and H is the total height. The moment
strength parameter κ is varied in order to demonstrate the effects of wall strength on
demands.

To model distributed plasticity, which can take place over a portion of the height of the
wall, the wall model is divided into small segments, each capable of forming plastic
hinges at its ends with a bilinear hysteretic moment-rotation relationship with no strain
hardening. Curvature demands are employed to represent the flexural deformations of
the wall. The wall curvature in each segment is computed from the difference between
the end rotations of the segment, i.e., φ = dθ/dx ≅ ∆θ/∆x, where ∆x is the length of the
segment.

Figure 9.23 illustrates distributions of various demands over the height for systems
strengthened with fixed walls. Each graph compares the demands for elastic and inelastic
walls. The heavy solid line in Fig. 9.23(a) represents story drift demands for
unstrengthened frames. The following observations can be made:

•= As Fig. 9.23(a) demonstrates, for short period frames (T/Tp = 0.5) the elastic wall
reduces drifts very effectively (better than a hinged wall). This effectiveness
slightly declines for very weak walls (κ = 0.1), but the benefits are still
considerable. The opposite is true for long period frames, i.e., the elastic wall is
less effective in reducing the drift demands (for T/Tp = 1.0 it even increases the
maximum drift) than the inelastic walls. This is not surprising because when the
fixed wall plastifies at the bottom, its behavior approaches that of a hinged wall
discussed previously.

•= Figure 9.23(b) shows that the curvature demands for elastic walls consistently
increase with period, owing to larger displacements in more flexible systems.
Even though a large portion of the weak walls plastifies (see Fig. 9.23(c)),
inelastic curvature demands are smaller than their elastic counterparts everywhere
over the height except the very bottom story, in which the inelastic curvature
demands exceed the elastic demands by far. This pattern is very important
because it shows that an inelastic fixed wall with sufficient initial stiffness

Chapter 9 Design Considerations for Near-Fault …

230
 deflects roughly as a straight line over most of its height and only bends at the
very bottom considerably. This concentration of rotation at the base indicates that
weak fixed walls behave in a similar manner as hinged walls, and therefore can
reduce drift demands effectively, provided that shear failure at the base is
prevented.

•= Figure 9.23(c) indicates that in flexible frames (T/Tp = 1.0 and 2.0), for which the
elastic wall moment demands are large, a reduction of the wall bending strength
will lead to plastification that is not limited to a small zone near the base but
spreads over a large portion of the height. For the weakest wall studied here, i.e.
κ = 0.1, the plastified zone comprises the entire bottom half of the structure. On
the other hand, in stronger walls (κ = 0.4), only a small portion of the wall
adjacent to the base plastifies.

•= Figure 9.23(d) shows that plastification of the wall due to flexural yielding
consistently reduces the wall shear demands. This reduction is larger for more
flexible systems, in which flexural yielding extends over a larger portion of the
wall. However, the bad news is that the reduced shear demands are still as high
as, or higher than, the shear demands for the elastic hinged wall. Thus, it will be
very difficult to provide sufficient shear resistance to avoid undesirable shear
yielding in conjunction with flexural yielding.

The conclusion is that strengthening with hinged walls that yield in shear is still
beneficial in reducing story drift demands, and that strengthening with fixed walls that
yield in flexure will lead to a concentration of large plastic rotations in the bottom story.
As the wall moment strength capacity decreases, the behavior of the fixed wall
approaches that of a hinged wall. It will be difficult in such cases to provide the shear
resistance necessary to avoid shear yielding. Thus, strengthening with hinged walls
appears to be a more effective technique.

Use of Reinforced Concrete Walls for Strengthening:

The values used in the previous discussion for the wall shear and flexural strength
parameters (ηw and κ) need to be put in perspective. The following example serves to
provide reasonable estimates for the shear and moment strength capacity of a reinforced
concrete shear wall that can effectively strengthen frames subjected to pulse-type ground

Chapter 9 Design Considerations for Near-Fault …

231
motions. The wall cross-section is assumed to be rectangular with a width of lw = 360 in.
and an unknown thickness of t. The results of a modal analysis for the generic 20-story
frame used in this study indicate that the stiffness (Kf [kip/in]), as defined in Section
9.4.1, total seismic mass (m [kip.sec2/in]), and period (T [sec]) of the structure are related
as follows:

m
K f = 12.50 (9.5)
T2

As discussed in Section 9.4.2, a desirable wall stiffness for T/Tp = 0.5, 1.0, and 2.0 is Kw
= 0.25 Kf, 0.5 Kf, and 2.0 Kf, respectively. Substituting for Kf from Eq. 9.5, and using
the equation for wall flexural stiffness, i.e. Kw = 3EwIw/Hw3 (Ew is modulus of elasticity,
Iw is moment of inertia, and Hw is total height), values for a desirable Iw can be
determined for the three T/Tp cases. For T/Tp = 0.5 the result is:

mH 3w
I w = 1.04 (9.6)
EwT2

Continuing the illustrative example for T/Tp = 0.5, for a scenario event with Mw = 7 and
R = 3 km, the pulse parameters are Tp = 2.6 sec. and ag,max = 0.31g, and hence T = 1.3
sec. The total frame height for a 20-story building is Hw = 2880 in. The modulus of
elasticity for reinforced concrete can be estimated from the ACI equation (f’c = 4000 psi)
as follows:

E w = 57 f c' = 3605 ksi (9.7)

Substituting for T, Hw, and Ew in Eq. 9.6, the desirable wall moment of inertia, in terms
of the mass m, is found to be Iw = 4.08×106m. For a rectangular section with lw = 360 in.,
the required thickness can be computed as t = 12Iw/lw3 = 1.05m. Similar calculations for
T/Tp = 1.0 and 2.0 results in t = 0.53m for both cases. These wall thickness-mass
relationships can be used to estimate the wall shear strength capacity as follows:

According to ACI 318-99, the wall shear strength provided by the concrete for aspect
ratios Hw/lw ≥ 2.0 can be estimated as:

Chapter 9 Design Considerations for Near-Fault …

232
 2 4,000 (360)t
Vc = 2 f c' A cv = = 46t (9.8)
1000

If minimum shear reinforcement (ρn = 0.0025) with fy = 60,000 psi is used, the shear
strength provided by the reinforcement is:

(360)t(0.0025)(60,000)
Vs = A cv ρ n f y = = 54t (9.9)
1000

Therefore, the total minimum shear strength of the wall is Vn = Vc + Vs = 100t. Using the
t-m relationships found above and normalizing the shear strength by [Link],max, where ag,max
= 0.31g for the scenario event under consideration, ηw = Vn/([Link],max) = 0.88, 0.44, and
0.44 for T/Tp = 0.5, 1.0, and 2.0, respectively. As shown in Fig. 9.17, the maximum base
shear strength demand for elastic hinged walls in almost all cases is about [Link],max,
which is only slightly larger than the values estimated for the minimum shear capacity of
the reinforced concrete wall used in the example. Considering that adding more
reinforcement will further increase the shear strength capacity, it appears that shear
yielding, which may result in a brittle mode of failure in reinforced concrete walls, can be
prevented. This conclusion supports the use of hinged reinforced concrete walls as a
beneficial strengthening technique.

To estimate the bending strength capacity of the example wall, boundary zones are
considered to be fully imbedded in the wall, each 40 in. wide. As an example, the
reinforcement ratio in each boundary zone is assumed to be 3%, amounting to As =
0.03(40)t = 1.2t. For this case the nominal bending strength is conservatively estimated
as Mn = 24,000t. After substituting for t from the t-m relationships and normalizing by
[Link],max, the normalized moment capacity for T/Tp = 0.5, 1.0, and 2.0 is computed as κ
= Mn/([Link],max) = 0.07, 0.04, and 0.04. These estimates are much smaller than the
elastic moment demands for fixed walls. They are even smaller than the moment
demands for elastic hinged walls. Thus, under the assumptions made in this example, it
appears that avoiding flexural yielding in the wall may be difficult (but not impossible for
hinged walls) and requires heavy flexural reinforcement.

If the hinged wall yields in bending, as the previous example suggested, the effectiveness
of the wall in reducing the maximum story drift demand needs to be reevaluated. This
reevaluation is illustrated in Fig. 9.24, which shows distributions of story drift and wall

Chapter 9 Design Considerations for Near-Fault …

233
moment demands for frames strengthened by inelastic hinged walls. Flexural yielding is
allowed to occur at any point over the height of the wall at which the moment demand
reaches the wall moment capacity, which is assumed to be constant over the height. The
wall moment capacity is defined by the parameter κ = My/([Link],max), where My is the
moment capacity and H is the total height. The moment strength parameter κ is varied in
order to investigate the effect of wall flexural strength on story drift demands. The
technique and assumptions discussed in Section 9.4.3 for fixed walls are employed here
to model distributed plasticity due to bending in the hinged wall. The following
observations are made:

•= Figure 9.24(a) shows that the effectiveness of the hinged wall only slightly
declines when the wall yields in bending. Weak hinged walls that yield in
bending still can reduce the maximum drift demand significantly for both short
and long period frames, and therefore can be utilized effectively to strengthen
frames.

•= Figure 9.24(b) indicates that flexural yielding in the wall initiates at the location
of the maximum elastic moment demand (around 40% of the height), and that for
very weak walls the plastified region spreads over a large portion of the height.

The overall conclusion is that reinforced concrete shear walls that are hinged at the base
can be employed efficiently to strengthen frame structures subjected to pulse-type ground
motions.

Chapter 9 Design Considerations for Near-Fault …

234
 Table 9.1 Evaluation of Predicted Pulse Intensity for Large Earthquakes

Earthquake Station Mw R (km) veff* (cm/sec) PGV (cm/sec)

Izmit, 1999 Yarimca 7.4 5.0 283 96
Izmit, 1999 Gebze 7.4 14.5 171 41
Chi-Chi, 1999 Tsaotun 7.6 5.9 353 116
* From Eq. 9.3

Table 9.2 Predictions of Pulse Properties from Regression Models

R = 3 km R = 5 km R = 10 km
Mw Tp (sec) veff aeff veff aeff veff aeff
(cm/sec) (g) (cm/sec) (g) (cm/sec) (g)
6.0 1.3 44 0.14 35 0.11 25 0.08
6.5 1.8 93 0.21 74 0.17 53 0.12
7.0 2.6 198 0.31 155 0.25 112 0.18
7.5 3.7 418 0.46 328 0.36 237 0.26

Table 9.3 Fundamental Period of Dual Systems with Fixed Wall

Kw / Kf 0.00 0.10 0.25 0.50 1.00 1.50 2.00

T / T0 1.000 0.881 0.813 0.738 0.639 0.573 0.524

Chapter 9 Design Considerations for Near-Fault …

235
 Pulse Period - Magnitude Relationship
Combined Ground Motions with Forward Directivity
0.8

0.6

log10Tp = - 1.76 + 0.31 Mw

0.4
log10Tp

0.2

data points
regressed, mean
0 90% prediction band
80% prediction band
70% prediction band
Taiwan, Tsaotun
Turkey, Yarimca
-0.2
6 6.2 6.4 6.6 6.8 7 7.2 7.4 7.6 7.8
Magnitude, Mw

(a) log10Tp vs. Mw

Pulse Period - Magnitude Relationship

Combined Ground Motions with Forward Directivity
5

3
Tp (sec)

data points
regressed, median
1 90% prediction band
80% prediction band
70% prediction band
Taiwan, Tsaotun
Turkey, Yarimca
0
6 6.2 6.4 6.6 6.8 7 7.2 7.4 7.6 7.8
Magnitude, Mw

(b) Tp vs. Mw

Figure 9.1 Dependence of Equivalent Pulse Period on Magnitude, Combined Set of

Ground Motions

Chapter 9 Design Considerations for Near-Fault …

236
 Pulse Period - Magnitude Relationship
Recorded Ground Motions with Forward Directivity
0.8
data points
regressed, mean
90% prediction band
80% prediction band
0.6 70% prediction band

log10Tp = - 2.06 + 0.34 Mw

0.4
log10Tp

0.2

-0.2
6 6.2 6.4 6.6 6.8 7 7.2 7.4 7.6 7.8
Magnitude, Mw

(a) log10Tp vs. Mw

Pulse Period - Magnitude Relationship

Recorded Ground Motions with Forward Directivity
5
data points
regressed, median
90% prediction band
80% prediction band
4 70% prediction band

3
Tp (sec)

0
6 6.2 6.4 6.6 6.8 7 7.2 7.4 7.6 7.8
Magnitude, Mw

(b) Tp vs. Mw

Figure 9.2 Dependence of Equivalent Pulse Period on Magnitude, Recorded Set of

Ground Motions

Chapter 9 Design Considerations for Near-Fault …

237
 Effective Pulse Velocity − Combined
Data Points
Surface Points

200

150
Veff (cm / sec)

100

50

3
0 4
6
5
6.2 6
6.4 7
6.6 8
6.8 9
7 10
R (km)
Mw

(a) Combined Set

Effective Pulse Velocity − Recorded

Data Points
Surface Points

250

200
Veff (cm / sec)

150

100

50
3
4
0
6 5
6.2 6
6.4 7
6.6 8
6.8 9
7 10 R (km)
Mw

(b) Recorded Set

Figure 9.3 Dependence of Equivalent Pulse Velocity on Magnitude and Distance

Chapter 9 Design Considerations for Near-Fault …

238
 Inelastic Base Shear Strength Demands Inelastic Base Shear Strength Demands
UBC-97 Type SD vs. Pulse P2, Mw = 6.0, R = 10, Tp = 1.3, ξ = 5%, no P-∆
∆ UBC-97 Type SD vs. Pulse P2, Mw= 6.0, R = 3, Tp = 1.3, ξ = 5%, no P-∆
∆
1
UBC/4, no NFF UBC/4, no NFF
UBC/4, with NFF UBC/4, with NFF
µ===2 µ===2
0.8 µ===3 µ===3
µ===4 µ===4
µ===6 µ===6
µ===8
γ = Vy / W
µ===8
0.6

0.4

0.2

0
0 1 2 3 4 5 0 1 2 3 4 5
T (sec) T (sec)

Mw = 6.0

Inelastic Base Shear Strength Demands Inelastic Base Shear Strength Demands
UBC-97 Type SD vs. Pulse P2, Mw = 6.5, R = 10, Tp = 1.8, ξ = 5%, no P-∆
∆ UBC-97 Type SD vs. Pulse P2, Mw = 6.5, R = 3, Tp = 1.8, ξ = 5%, no P-∆
∆
1
UBC/4, no NFF UBC/4, no NFF
UBC/4, with NFF UBC/4, with NFF
µ===2 µ===2
0.8 µ===3 µ===3
µ===4 µ===4
µ===6 µ===6
µ===8 µ===8
γ = Vy / W

0.6

0.4

0.2

0
0 1 2 3 4 5 0 1 2 3 4 5
T (sec) T (sec)

Mw = 6.5

Inelastic Base Shear Strength Demands Inelastic Base Shear Strength Demands
UBC-97 Type SD vs. Pulse P2, Mw = 7.0, R = 10, Tp = 2.6, ξ = 5%, no P-∆
∆ UBC-97 Type SD vs. Pulse P2, Mw = 7.0, R = 3, Tp = 2.6, ξ = 5%, no P-∆
∆
1
UBC/4, no NFF UBC/4, no NFF
UBC/4, with NFF UBC/4, with NFF
µ===2 µ===2
0.8 µ===3 µ===3
µ===4 µ===4
µ===6 µ===6
µ===8 µ===8
γ = Vy / W

0.6

0.4

0.2

0
0 1 2 3 4 5 0 1 2 3 4 5
T (sec) T (sec)

Mw = 7.0
(a) R = 10 km (b) R = 3 km

Figure 9.4 Magnitude and Distance Dependence of MDOF Base Shear Strength Demand
Spectra for Constant Ductilities, Equivalent Pulse P2

Chapter 9 Design Considerations for Near-Fault …

239
 Inelastic Base Shear Strength Demands Inelastic Base Shear Strength Demands
UBC-97 Type SD vs. Pulse P3, Mw = 6.0, R = 10, Tp = 1.3, ξ = 5%, no P-∆
∆ UBC-97 Type SD vs. Pulse P3, Mw= 6.0, R = 3, Tp = 1.3, ξ = 5%, no P-∆
∆
1
UBC/4, no NFF UBC/4, no NFF
UBC/4, with NFF UBC/4, with NFF
µ===2 µ===2
0.8 µ===3 µ===3
µ===4 µ===4
µ===6 µ===6
µ===8
γ = Vy / W
µ===8
0.6

0.4

0.2

0
0 1 2 3 4 5 0 1 2 3 4 5
T (sec) T (sec)

Mw = 6.0

Inelastic Base Shear Strength Demands Inelastic Base Shear Strength Demands
UBC-97 Type SD vs. Pulse P3, Mw = 6.5, R = 10, Tp = 1.8, ξ = 5%, no P-∆
∆ UBC-97 Type SD vs. Pulse P3, Mw = 6.5, R = 3, Tp = 1.8, ξ = 5%, no P-∆
∆
1
UBC/4, no NFF UBC/4, no NFF
UBC/4, with NFF UBC/4, with NFF
µ===2 µ===2
0.8 µ===3 µ===3
µ===4 µ===4
µ===6 µ===6
µ===8 µ===8
γ = Vy / W

0.6

0.4

0.2

0
0 1 2 3 4 5 0 1 2 3 4 5
T (sec) T (sec)

Mw = 6.5

Inelastic Base Shear Strength Demands Inelastic Base Shear Strength Demands
UBC-97 Type SD vs. Pulse P3, Mw = 7.0, R = 10, Tp = 2.6, ξ = 5%, no P-∆
∆ UBC-97 Type SD vs. Pulse P3, Mw = 7.0, R = 3, Tp = 2.6, ξ = 5%, no P-∆
∆
1
UBC/4, no NFF UBC/4, no NFF
UBC/4, with NFF UBC/4, with NFF
µ===2 µ===2
0.8 µ===3 µ===3
µ===4 µ===4
µ===6 µ===6
µ===8 µ===8
γ = Vy / W

0.6

0.4

0.2

0
0 1 2 3 4 5 0 1 2 3 4 5
T (sec) T (sec)

Mw = 7.0
(a) R = 10 km (b) R = 3 km

Figure 9.5 Magnitude and Distance Dependence of MDOF Base Shear Strength Demand
Spectra for Constant Ductilities, Equivalent Pulse P3

Chapter 9 Design Considerations for Near-Fault …

240
 Base Shear Strength Demands
Equivalent P2 for (Mw = 7.0, R = 3.0), µmax = 2, Diff. Tp, ξ = 5%, no P-∆
∆
1
Tp=2.0 sec

Tp=2.3 sec
0.8 Tp=2.6 sec

Tp=3.0 sec

0.6
γ = Vy / W

0.4

0.2

0
0 1 2 3 4 5
T (sec)

(a) µmax = 2

Base Shear Strength Demands

Equivalent P2 for (Mw = 7.0, R = 3.0), µmax = 8, Diff. Tp, ξ = 5%, no P-∆
∆
0.4
Tp=2.0 sec

Tp=2.3 sec

Tp=2.6 sec
0.3
Tp=3.0 sec
γ = Vy / W

0.2

0.1

0
0 1 2 3 4 5
T (sec)

(b) µmax = 8

Figure 9.6 Sensitivity of Base Shear Strength Demands to Tp, Equivalent Pulse P2 for
Mw = 7.0 and R = 3 km

Chapter 9 Design Considerations for Near-Fault …

241
 Story Shear Strength Pattern for Uniform Ductility
∆
Pulse P2, 20-Story Frame, T / Tp = 2.0, without P-∆
1

0.8
Relative Height

0.6

0.4
SRSS Pattern
µ===1
µ===2
0.2 µ===3
µ===4
µ===6
µ===8
0
0 0.2 0.4 0.6 0.8 1
Vy,i / Vy,base

(a) T/Tp = 2.0

Story Shear Strength Pattern for Uniform Ductility

∆
Pulse P2, 20-Story Frame, T / Tp = 1.0, without P-∆
1

0.8
Relative Height

0.6

0.4
SRSS Pattern
µ===1=
µ===2
0.2 µ===3
µ===4=
µ===6
µ===8
0
0 0.2 0.4 0.6 0.8 1
Vy,i / Vy,base

(b) T/Tp = 1.0

Figure 9.7 Story Shear Strength Distributions for Uniform Ductility Over Height,
Pulse P2

Chapter 9 Design Considerations for Near-Fault …

242
 Base Shear Strength vs. Maximum Story Ductility
∆
Pulse P2, 20-Story Frame, T / Tp = 2.0, without P-∆
2
SRSS pattern
uniform ductility pattern

1.5
η = Vy / ([Link],max)

0.5

0
1 2 3 4 5 6 7 8
Maximum Story Ductility,,=µmax

(a) T/Tp = 2.0

Base Shear Strength vs. Maximum Story Ductility

∆
Pulse P2, 20-Story Frame, T / Tp = 1.0, without P-∆
4
SRSS pattern
uniform ductility pattern

3
η = Vy / ([Link],max)

0
1 2 3 4 5 6 7 8
Maximum Story Ductility,,=µmax

(b) T/Tp = 1.0

Figure 9.8 Base Shear Strength Demands for Specific Target Ductilities, SRSS and
Uniform Ductility Story Shear Strength Distributions, Pulse P2

Chapter 9 Design Considerations for Near-Fault …

243
 Story Shear Strength Pattern
Different Patterns
1
SRSS Pattern
Linear Pattern
SRSS+Strengthening
0.8
Linear+Strengthening
Relative Height

0.6

0.4

0.2

0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
Vy,i / Vy,base(SRSS)

Figure 9.9 SRSS and Modified Story Shear Strength Distributions

Chapter 9 Design Considerations for Near-Fault …

244
 Story Ductility Demands Story Ductility Demands
∆
Pulse P2, SRSS Pattern, T / Tp = 0.5, without P-∆ ∆
Pulse P2, SRSS+Strengthening, T / Tp = 0.50, without P-∆
1
η===2.50 η∗===2.50
η===2.00 η∗===2.00
η===1.50 η∗===1.50
0.8 η===1.25 η∗===1.25
η===1.00 η∗===1.00
η===0.75 η∗===0.75
Relative Height

η===0.60 η∗===0.60
0.6 η===0.50 η∗===0.50
η===0.40 η∗===0.40

0.4

0.2

0
0 4 8 12 16 20 0 4 8 12 16 20
Story Ductility Ratio, µi = δmax,i / δy,i Story Ductility Ratio, µi = δmax,i / δy,i

T/Tp = 0.5

Story Ductility Demands Story Ductility Demands

∆
Pulse P2, SRSS Pattern, T / Tp = 1.0, without P-∆ ∆
Pulse P2, SRSS+Strengthening, T / Tp = 1.0, without P-∆
1
η===2.00 η∗===2.00
η===1.50 η∗===1.50
η===1.25 η∗===1.25
0.8 η===1.00 η∗===1.00
η===0.75 η∗===0.75
η===0.50 η∗===0.50
Relative Height

η===0.25 η∗===0.25
0.6 η===0.15 η∗===0.15
η===0.10 η∗===0.10

0.4

0.2

0
0 4 8 12 16 20 0 4 8 12 16 20
Story Ductility Ratio, µi = δmax,i / δy,i Story Ductility Ratio, µi = δmax,i / δy,i

T/Tp = 1.0

Story Ductility Demands Story Ductility Demands

∆
Pulse P2, SRSS Pattern, T / Tp = 2.0, without P-∆ ∆
Pulse P2, SRSS+Strengthening, T / Tp = 2.0, without P-∆
1
η===1.25 η∗ ===1.25
η===1.00 η∗ ===1.00
η===0.75 η∗ ===0.75
0.8 η===0.50 η∗ ===0.50
η===0.25 η∗ ===0.25
η===0.15 η∗ ===0.15
Relative Height

η===0.10 η∗ ===0.10
0.6 η===0.07 η∗ ===0.07
η===0.05 η∗ ===0.05

0.4

0.2

0
0 4 8 12 16 20 0 4 8 12 16 20
Story Ductility Ratio, µi = δmax,i / δy,i Story Ductility Ratio, µi = δmax,i / δy,i

T/Tp = 2.0
(a) SRSS (b) SRSS + Strengthening

Figure 9.10 Story Ductility Demands for Unstrengthened and Strengthened Structures
Subjected to Pulse P2, without P-Delta Effects

Chapter 9 Design Considerations for Near-Fault …

245
 Story Ductility Demands Story Ductility Demands
∆
Pulse P2, Linear Pattern, T / Tp = 0.50, without P-∆ ∆
Pulse P2, Linear+Strengthening, T / Tp = 0.50, without P-∆
1
η===2.50 η∗===2.50
η===2.00 η∗===2.00
η===1.50 η∗===1.50
0.8 η===1.25 η∗===1.25
η===1.00 η∗===1.00
η===0.75 η∗===0.75
Relative Height

η===0.60 η∗===0.60
0.6 η===0.50 η∗===0.50
η===0.40 η∗===0.40

0.4

0.2

0
0 4 8 12 16 20 0 4 8 12 16 20
Story Ductility Ratio, µi = δmax,i / δy,i Story Ductility Ratio, µi = δmax,i / δy,i

T/Tp = 0.5

Story Ductility Demands Story Ductility Demands

∆
Pulse P2, Linear Pattern, T / Tp = 1.0, without P-∆ ∆
Pulse P2, Linear+Strengthening, T / Tp = 1.0, without P-∆
1
η===2.00 η∗===2.00
η===1.50 η∗===1.50
η===1.25 η∗===1.25
0.8 η===1.00 η∗===1.00
η===0.75 η∗===0.75
η===0.50 η∗===0.50
Relative Height

η===0.25 η∗===0.25
0.6 η===0.15 η∗===0.15
η===0.10 η∗===0.10

0.4

0.2

0
0 4 8 12 16 20 0 4 8 12 16 20
Story Ductility Ratio, µi = δmax,i / δy,i Story Ductility Ratio, µi = δmax,i / δy,i

T/Tp = 1.0

Story Ductility Demands Story Ductility Demands

∆
Pulse P2, Linear Pattern, T / Tp = 2.0, without P-∆ ∆
Pulse P2, Linear+Strengthening, T / Tp = 2.0, without P-∆
1
η===1.25 η∗===1.25
η===1.00 η∗===1.00
η===0.75 η∗===0.75
0.8 η===0.50 η∗===0.50
η===0.25 η∗===0.25
η===0.15 η∗===0.15
Relative Height

η===0.10 η∗===0.10
0.6 η===0.07 η∗===0.07
η===0.05 η∗===0.05

0.4

0.2

0
0 4 8 12 16 20 0 4 8 12 16 20
Story Ductility Ratio, µi = δmax,i / δy,i Story Ductility Ratio, µi = δmax,i / δy,i

T/Tp = 2.0
(c) Linear (d) Linear + Strengthening

Figure 9.10 (Cont’d) Story Ductility Demands for Unstrengthened and Strengthened
Structures Subjected to Pulse P2, without P-Delta Effects

Chapter 9 Design Considerations for Near-Fault …

246
 Story Ductility Demands Story Ductility Demands
∆
Pulse P2, SRSS Pattern, T / Tp = 0.50, with P-∆ ∆
Pulse P2, SRSS+Strengthening, T / Tp = 0.50, with P-∆
1
η===2.50 η∗===2.50
η===2.00 η∗===2.00
η===1.50 η∗===1.50
0.8 η===1.25 η∗===1.25
η===1.00 η∗===1.00
η===0.75 η∗===0.75
Relative Height

η===0.60 η∗===0.60
0.6 η===0.50 η∗===0.50
η===0.40 η∗===0.40

0.4

0.2

0
0 4 8 12 16 20 0 4 8 12 16 20
Story Ductility Ratio, µi = δmax,i / δy,i Story Ductility Ratio, µi = δmax,i / δy,i

T/Tp = 0.5

Story Ductility Demands Story Ductility Demands

∆
Pulse P2, SRSS Pattern, T / Tp = 1.0, with P-∆ ∆
Pulse P2, SRSS+Strengthening, T / Tp = 1.0, with P-∆
1
η===2.00 η∗===2.00
η===1.50 η∗===1.50
η===1.25 η∗===1.25
0.8 η===1.00 η∗===1.00
η===0.75 η∗===0.75
η===0.50 η∗===0.50
Relative Height

η===0.25 η∗===0.25
0.6 η===0.15 η∗===0.15
η===0.10 η∗===0.10

0.4

0.2

0
0 4 8 12 16 20 0 4 8 12 16 20
Story Ductility Ratio, µi = δmax,i / δy,i Story Ductility Ratio, µi = δmax,i / δy,i

T/Tp = 1.0

Story Ductility Demands Story Ductility Demands

∆
Pulse P2, SRSS Pattern, T / Tp = 2.0, with P-∆ ∆
Pulse P2, SRSS+Strengthening, T / Tp = 2.0, with P-∆
1
η===1.25 η∗===1.25
η===1.00 η∗===1.00
η===0.75 η∗===0.75
0.8 η===0.50 η∗===0.50
η===0.25 η∗===0.25
η===0.15 η∗===0.15
Relative Height

η===0.10 η∗===0.10
0.6 η===0.07 η∗===0.07
η===0.05 η∗===0.05

0.4

0.2

0
0 4 8 12 16 20 0 4 8 12 16 20
Story Ductility Ratio, µi = δmax,i / δy,i Story Ductility Ratio, µi = δmax,i / δy,i

T/Tp = 2.0
(a) SRSS (b) SRSS + Strengthening

Figure 9.11 Story Ductility Demands for Unstrengthened and Strengthened Structures
Subjected to Pulse P2, with P-Delta Effects

Chapter 9 Design Considerations for Near-Fault …

247
 (a) Fixed Wall (b) Hinged Wall

Figure 9.12 Typical Deflected Shapes of Dual Systems Used in this Study

Chapter 9 Design Considerations for Near-Fault …

248
 Roof Displacement vs. Base Shear
∆=
20-Story Frame + Fixed Wall, SRSS Load Pattern, no P-∆
2

1.5
Vbase / Vbase,y

Kw/Kf = 0.00
Kw/Kf = 0.10
Kw/Kf = 0.25
0.5 Kw/Kf = 0.50
Kw/Kf = 1.00
Kw/Kf = 1.50
Kw/Kf = 2.00
0
0 1 2 3 4 5
δroof=/=δroof,y

(a) Fixed Wall

Roof Displacement vs. Base Shear

20-Story Frame + Hinged Wall, SRSS Load Pattern, no P-∆=
2

1.5
Vbase / Vbase,y

Kw/Kf = 0.00
Kw/Kf = 0.10
Kw/Kf = 0.25
0.5 Kw/Kf = 0.50
Kw/Kf = 1.00
Kw/Kf = 1.50
Kw/Kf = 2.00
0
0 1 2 3 4 5
δroof=/=δroof,y

(b) Hinged Wall

Figure 9.13 Global Pushover Curves for Dual Systems without P-Delta Effects

Chapter 9 Design Considerations for Near-Fault …

249
 Story Drift Demands Story Drift Demands
Pulse P2, 20-Story Frame + Fixed Wall, T / Tp = 0.5, η = 1.7, no P-∆
∆ Pulse P2, 20-Story Frame + Fixed Wall, T / Tp = 0.5, η = 0.75, no P-∆
∆
1
Kw/Kf = 0.00 Kw/Kf = 0.00
Kw/Kf = 0.10 Kw/Kf = 0.10
Kw/Kf = 0.25 Kw/Kf = 0.25
0.8
Kw/Kf = 0.50 Kw/Kf = 0.50
Kw/Kf = 1.00 Kw/Kf = 1.00
Relative Height

Kw/Kf = 1.50 Kw/Kf = 1.50

0.6 Kw/Kf = 2.00 Kw/Kf = 2.00

0.4

0.2 µ = 11.7
µ = 3.4

0
0 0.02 0.04 0.06 0.08 0.1 0 0.02 0.04 0.06 0.08 0.1
δmax,i / ug,max δmax,i / ug,max

T/Tp = 0.5

Story Drift Demands Story Drift Demands

Pulse P2, 20-Story Frame + Fixed Wall, T / Tp = 1.0, η = 1.0, no P-∆
∆ Pulse P2, 20-Story Frame + Fixed Wall, T / Tp = 1.0, η = 0.2, no P-∆
∆
1
Kw/Kf = 0.00
Kw/Kf = 0.10
Kw/Kf = 0.25
0.8 Kw/Kf = 0.50
Kw/Kf = 1.00
Relative Height

Kw/Kf = 1.50
0.6 Kw/Kf = 2.00

µ = 3.8
0.4 Kw/Kf = 0.00
Kw/Kf = 0.10
Kw/Kf = 0.25
Kw/Kf = 0.50
0.2
Kw/Kf = 1.00
µ = 17.2 Kw/Kf = 1.50
Kw/Kf = 2.00
0
0 0.04 0.08 0.12 0.16 0 0.04 0.08 0.12 0.16
δmax,i / ug,max δmax,i / ug,max

T/Tp = 1.0

Story Drift Demands Story Drift Demands

Pulse P2, 20-Story Frame + Fixed Wall, T / Tp = 2.0, η = 0.5, no P-∆
∆ Pulse P2, 20-Story Frame + Fixed Wall, T / Tp = 2.0, η = 0.1, no P-∆
∆
1

0.8 µ = 3.3
Relative Height

0.6

0.4 Kw/Kf = 0.00

Kw/Kf = 0.00
Kw/Kf = 0.10
Kw/Kf = 0.10
Kw/Kf = 0.25
Kw/Kf = 0.25
Kw/Kf = 0.50 Kw/Kf = 0.50
0.2
Kw/Kf = 1.00 µ = 10.4 Kw/Kf = 1.00
Kw/Kf = 1.50 Kw/Kf = 1.50
Kw/Kf = 2.00 Kw/Kf = 2.00
0
0 0.05 0.1 0.15 0.2 0.25 0.3 0 0.05 0.1 0.15 0.2 0.25 0.3
δmax,i / ug,max δmax,i / ug,max

T/Tp = 2.0
(a) Strong Frames (b) Weak Frames

Figure 9.14 Distribution of Story Drift Demands Over Height for Dual Systems with
Elastic Fixed Wall, without P-Delta Effects

Chapter 9 Design Considerations for Near-Fault …

250
 Story Drift Demands Story Drift Demands
Pulse P2, 20-Story Frame + Hinged Wall, T / Tp = 0.5, η = 1.7, no P-∆
∆ Pulse P2, 20-Story Frame + Hinged Wall, T / Tp = 0.5, η = 0.75, no P-∆
∆
1
Kw/Kf = 0.00
Kw/Kf = 0.10
Kw/Kf = 0.25
0.8
Kw/Kf = 0.50
Kw/Kf = 1.00
Relative Height

Kw/Kf = 1.50
0.6 Kw/Kf = 2.00

0.4 Kw/Kf = 0.00

Kw/Kf = 0.10
Kw/Kf = 0.25
Kw/Kf = 0.50
0.2 µ = 11.7
µ = 3.4 Kw/Kf = 1.00
Kw/Kf = 1.50
Kw/Kf = 2.00
0
0 0.02 0.04 0.06 0.08 0.1 0 0.02 0.04 0.06 0.08 0.1
δmax,i / ug,max δmax,i / ug,max

T/Tp = 0.5

Story Drift Demands Story Drift Demands

Pulse P2, 20-Story Frame + Hinged Wall, T / Tp = 1.0, η = 1.0, no P-∆
∆ Pulse P2, 20-Story Frame + Hinged Wall, T / Tp = 1.0, η = 0.2, no P-∆
∆
1

0.8
Relative Height

0.6

µ = 3.8
0.4 Kw/Kf = 0.00 Kw/Kf = 0.00
Kw/Kf = 0.10 Kw/Kf = 0.10
Kw/Kf = 0.25 Kw/Kf = 0.25
Kw/Kf = 0.50 Kw/Kf = 0.50
0.2
Kw/Kf = 1.00 Kw/Kf = 1.00 µ = 17.2
Kw/Kf = 1.50 Kw/Kf = 1.50
Kw/Kf = 2.00 Kw/Kf = 2.00
0
0 0.04 0.08 0.12 0.16 0 0.04 0.08 0.12 0.16
δmax,i / ug,max δmax,i / ug,max

T/Tp = 1.0

Story Drift Demands Story Drift Demands

Pulse P2, 20-Story Frame + Hinged Wall, T / Tp = 2.0, η = 0.5, no P-∆
∆ Pulse P2, 20-Story Frame + Hinged Wall, T / Tp = 2.0, η = 0.1, no P-∆
∆
1
Kw/Kf = 0.00
Kw/Kf = 0.10
µ = 3.3 Kw/Kf = 0.25
0.8
Kw/Kf = 0.50
Kw/Kf = 1.00
Relative Height

Kw/Kf = 1.50
0.6 Kw/Kf = 2.00

0.4
Kw/Kf = 0.00
Kw/Kf = 0.10
Kw/Kf = 0.25
0.2 Kw/Kf = 0.50
µ = 10.4
Kw/Kf = 1.00
Kw/Kf = 1.50
Kw/Kf = 2.00
0
0 0.05 0.1 0.15 0.2 0.25 0.3 0 0.05 0.1 0.15 0.2 0.25 0.3
δmax,i / ug,max δmax,i / ug,max

T/Tp = 2.0
(a) Strong Frames (b) Weak Frames

Figure 9.15 Distribution of Story Drift Demands Over Height for Dual Systems with
Elastic Hinged Wall, without P-Delta Effects

Chapter 9 Design Considerations for Near-Fault …

251
 Wall Shear Force Demands Wall Shear Force Demands
Pulse P2, 20-Story Frame + Fixed Wall, T / Tp = 0.5, η = 1.7, no P-∆
∆ Pulse P2, 20-Story Frame + Fixed Wall, T / Tp = 0.5, η = 0.75, no P-∆
∆
1
Kw/Kf = 0.10 Kw/Kf = 0.10
Kw/Kf = 0.25 Kw/Kf = 0.25
0.8 Kw/Kf = 0.50 Kw/Kf = 0.50
Kw/Kf = 1.00 Kw/Kf = 1.00
Kw/Kf = 1.50 Kw/Kf = 1.50
Relative Height

Kw/Kf = 2.00 Kw/Kf = 2.00

0.6

0.4

0.2

0
0 0.5 1 1.5 2 2.5 3 0 0.5 1 1.5 2 2.5 3
Vi / ([Link],max) Vi / ([Link],max)

T/Tp = 0.5

Wall Shear Force Demands Wall Shear Force Demands

Pulse P2, 20-Story Frame + Fixed Wall, T / Tp = 1.0, η = 1.0, no P-∆
∆ Pulse P2, 20-Story Frame + Fixed Wall, T / Tp = 1.0, η = 0.2, no P-∆
∆
1
Kw/Kf = 0.10 Kw/Kf = 0.10
Kw/Kf = 0.25 Kw/Kf = 0.25
0.8 Kw/Kf = 0.50 Kw/Kf = 0.50
Kw/Kf = 1.00 Kw/Kf = 1.00
Kw/Kf = 1.50 Kw/Kf = 1.50
Relative Height

Kw/Kf = 2.00 Kw/Kf = 2.00

0.6

0.4

0.2

0
0 0.5 1 1.5 2 2.5 3 0 0.5 1 1.5 2 2.5 3
Vi / ([Link],max) Vi / ([Link],max)

T/Tp = 1.0

Wall Shear Force Demands Wall Shear Force Demands

Pulse P2, 20-Story Frame + Fixed Wall, T / Tp = 2.0, η = 0.5, no P-∆
∆ Pulse P2, 20-Story Frame + Fixed Wall, T / Tp = 2.0, η = 0.1, no P-∆
∆
1
Kw/Kf = 0.10 Kw/Kf = 0.10
Kw/Kf = 0.25 Kw/Kf = 0.25
0.8 Kw/Kf = 0.50 Kw/Kf = 0.50
Kw/Kf = 1.00 Kw/Kf = 1.00
Kw/Kf = 1.50
Relative Height

Kw/Kf = 1.50
Kw/Kf = 2.00 Kw/Kf = 2.00
0.6

0.4

0.2

0
0 0.5 1 1.5 2 2.5 3 0 0.5 1 1.5 2 2.5 3
Vi / ([Link],max) Vi / ([Link],max)

T/Tp = 2.0
(a) Strong Frames (b) Weak Frames

Figure 9.16 Distribution of Wall Shear Demands Over Height for Dual Systems with
Elastic Fixed Wall, without P-Delta Effects

Chapter 9 Design Considerations for Near-Fault …

252
 Wall Shear Force Demands Wall Shear Force Demands
Pulse P2, 20-Story Frame + Hinged Wall, T / Tp = 0.5, η = 1.7, no P-∆
∆ Pulse P2, 20-Story Frame + Hinged Wall, T / Tp = 0.5, η = 0.75, no P-∆
∆
1
Kw/Kf = 0.10 Kw/Kf = 0.10
Kw/Kf = 0.25 Kw/Kf = 0.25
Kw/Kf = 0.50 Kw/Kf = 0.50
0.8
Kw/Kf = 1.00 Kw/Kf = 1.00
Kw/Kf = 1.50 Kw/Kf = 1.50
Relative Height

Kw/Kf = 2.00
Kw/Kf = 2.00
0.6

0.4

0.2

0
0 0.2 0.4 0.6 0.8 0 0.2 0.4 0.6 0.8
Vi / ([Link],max) Vi / ([Link],max)

T/Tp = 0.5

Wall Shear Force Demands Wall Shear Force Demands

Pulse P2, 20-Story Frame + Hinged Wall, T / Tp = 1.0, η = 1.0, no P-∆
∆ Pulse P2, 20-Story Frame + Hinged Wall, T / Tp = 1.0, η = 0.2, no P-∆
∆
1
Kw/Kf = 0.10 Kw/Kf = 0.10
Kw/Kf = 0.25 Kw/Kf = 0.25
0.8 Kw/Kf = 0.50 Kw/Kf = 0.50
Kw/Kf = 1.00 Kw/Kf = 1.00
Kw/Kf = 1.50 Kw/Kf = 1.50
Relative Height

Kw/Kf = 2.00 Kw/Kf = 2.00

0.6

0.4

0.2

0
0 0.2 0.4 0.6 0.8 0 0.2 0.4 0.6 0.8
Vi / ([Link],max) Vi / ([Link],max)

T/Tp = 1.0

Wall Shear Force Demands Wall Shear Force Demands

Pulse P2, 20-Story Frame + Hinged Wall, T / Tp = 2.0, η = 0.5, no P-∆
∆ Pulse P2, 20-Story Frame + Hinged Wall, T / Tp = 2.0, η = 0.1, no P-∆
∆
1
Kw/Kf = 0.10 Kw/Kf = 0.10
Kw/Kf = 0.25 Kw/Kf = 0.25
0.8 Kw/Kf = 0.50 Kw/Kf = 0.50
Kw/Kf = 1.00 Kw/Kf = 1.00
Kw/Kf = 1.50 Kw/Kf = 1.50
Relative Height

Kw/Kf = 2.00 Kw/Kf = 2.00

0.6

0.4

0.2

0
0 0.2 0.4 0.6 0.8 0 0.2 0.4 0.6 0.8
Vi / ([Link],max) Vi / ([Link],max)

T/Tp = 2.0
(a) Strong Frames (b) Weak Frames

Figure 9.17 Distribution of Wall Shear Demands Over Height for Dual Systems with
Elastic Hinged Wall, without P-Delta Effects

Chapter 9 Design Considerations for Near-Fault …

253
 Shear Wall Moment Demands Shear Wall Moment Demands
Pulse P2, 20-Story Frame + Fixed Wall, T / Tp = 0.5, η = 1.7, no P-∆
∆ Pulse P2, 20-Story Frame + Fixed Wall, T / Tp = 0.5, η = 0.75, no P-∆
∆
1
Kw/Kf = 0.10 Kw/Kf = 0.10
Kw/Kf = 0.25 Kw/Kf = 0.25
0.8 Kw/Kf = 0.50 Kw/Kf = 0.50
Kw/Kf = 1.00 Kw/Kf = 1.00
Relative Height

Kw/Kf = 1.50 Kw/Kf = 1.50

0.6 Kw/Kf = 2.00 Kw/Kf = 2.00

0.4

0.2

0
0 0.2 0.4 0.6 0.8 1 1.2 0 0.2 0.4 0.6 0.8 1 1.2
M / ([Link],max) M / ([Link],max)

T/Tp = 0.5

Shear Wall Moment Demands Shear Wall Moment Demands

Pulse P2, 20-Story Frame + Fixed Wall, T / Tp = 1.0, η = 1.0, no P-∆
∆ Pulse P2, 20-Story Frame + Fixed Wall, T / Tp = 1.0, η = 0.2, no P-∆
∆
1
Kw/Kf = 0.10 Kw/Kf = 0.10
Kw/Kf = 0.25 Kw/Kf = 0.25
0.8 Kw/Kf = 0.50 Kw/Kf = 0.50
Kw/Kf = 1.00 Kw/Kf = 1.00
Kw/Kf = 1.50 Kw/Kf = 1.50
Relative Height

Kw/Kf = 2.00 Kw/Kf = 2.00

0.6

0.4

0.2

0
0 0.4 0.8 1.2 1.6 0 0.4 0.8 1.2 1.6
M / ([Link],max) M / ([Link],max)

T/Tp = 1.0

Shear Wall Moment Demands Shear Wall Moment Demands

Pulse P2, 20-Story Frame + Fixed Wall, T / Tp = 2.0, η = 0.5, no P-∆
∆ Pulse P2, 20-Story Frame + Fixed Wall, T / Tp = 2.0, η = 0.1, no P-∆
∆
1
Kw/Kf = 0.10 Kw/Kf = 0.10
Kw/Kf = 0.25 Kw/Kf = 0.25
0.8 Kw/Kf = 0.50 Kw/Kf = 0.50
Kw/Kf = 1.00 Kw/Kf = 1.00
Kw/Kf = 1.50 Kw/Kf = 1.50
Relative Height

Kw/Kf = 2.00 Kw/Kf = 2.00

0.6

0.4

0.2

0
0 0.2 0.4 0.6 0.8 1 1.2 0 0.2 0.4 0.6 0.8 1 1.2
M / ([Link],max) M / ([Link],max)

T/Tp = 2.0
(a) Strong Frames (b) Weak Frames

Figure 9.18 Distribution of Wall Moment Demands Over Height for Dual Systems with
Elastic Fixed Wall, without P-Delta Effects

Chapter 9 Design Considerations for Near-Fault …

254
 Shear Wall Moment Demands Shear Wall Moment Demands
Pulse P2, 20-Story Frame + Hinged Wall, T / Tp = 0.5, η = 1.7, no P-∆
∆ Pulse P2, 20-Story Frame + Hinged Wall, T / Tp = 0.5, η = 0.75, no P-∆
∆
1
Kw/Kf = 0.10 Kw/Kf = 0.10
Kw/Kf = 0.25 Kw/Kf = 0.25
0.8 Kw/Kf = 0.50 Kw/Kf = 0.50
Kw/Kf = 1.00 Kw/Kf = 1.00
Relative Height

Kw/Kf = 1.50 Kw/Kf = 1.50

0.6 Kw/Kf = 2.00 Kw/Kf = 2.00

0.4

0.2

0
0 0.04 0.08 0.12 0.16 0 0.04 0.08 0.12 0.16
M / ([Link],max) M / ([Link],max)

T/Tp = 0.5

Shear Wall Moment Demands Shear Wall Moment Demands

Pulse P2, 20-Story Frame + Hinged Wall, T / Tp = 1.0, η = 1.0, no P-∆
∆ Pulse P2, 20-Story Frame + Hinged Wall, T / Tp = 1.0, η = 0.2, no P-∆
∆
1
Kw/Kf = 0.10 Kw/Kf = 0.10
Kw/Kf = 0.25 Kw/Kf = 0.25
0.8 Kw/Kf = 0.50 Kw/Kf = 0.50
Kw/Kf = 1.00 Kw/Kf = 1.00
Kw/Kf = 1.50 Kw/Kf = 1.50
Relative Height

Kw/Kf = 2.00 Kw/Kf = 2.00

0.6

0.4

0.2

0
0 0.04 0.08 0.12 0.16 0 0.04 0.08 0.12 0.16
M / ([Link],max) M / ([Link],max)

T/Tp = 1.0

Shear Wall Moment Demands Shear Wall Moment Demands

Pulse P2, 20-Story Frame + Hinged Wall, T / Tp = 2.0, η = 0.5, no P-∆
∆ Pulse P2, 20-Story Frame + Hinged Wall, T / Tp = 2.0, η = 0.1, no P-∆
∆
1
Kw/Kf = 0.10 Kw/Kf = 0.10
Kw/Kf = 0.25 Kw/Kf = 0.25
0.8 Kw/Kf = 0.50 Kw/Kf = 0.50
Kw/Kf = 1.00 Kw/Kf = 1.00
Kw/Kf = 1.50 Kw/Kf = 1.50
Relative Height

Kw/Kf = 2.00 Kw/Kf = 2.00

0.6

0.4

0.2

0
0 0.04 0.08 0.12 0.16 0 0.04 0.08 0.12 0.16
M / ([Link],max) M / ([Link],max)

T/Tp = 2.0
(a) Strong Frames (b) Weak Frames

Figure 9.19 Distribution of Wall Moment Demands Over Height for Dual Systems with
Elastic Hinged Wall, without P-Delta Effects

Chapter 9 Design Considerations for Near-Fault …

255
 Story Drift Demands with Shear Deformations Story Drift Demands with Shear Deformations
P2, Frame + Fixed Wall, T / Tp = 0.5, ηf = 0.75, Kw / Kf = 0.25, no P-∆
∆ P2, Frame + Hinged Wall, T / Tp = 0.5, ηf = 0.75, Kw / Kf = 0.25, no P-∆
∆
1
b/H = 0.5
b/H = 0.4
0.8 b/H = 0.3
b/H = 0.2
Relative Height

b/H = 0.1
0.6
b/H = 0.0

0.4 b/H = 0.5

b/H = 0.4
b/H = 0.3
0.2 b/H = 0.2
b/H = 0.1
b/H = 0.0
0
0 0.01 0.02 0.03 0.04 0.05 0.06 0 0.01 0.02 0.03 0.04 0.05 0.06
δmax,i / ug,max δmax,i / ug,max

T/Tp = 0.5, ηf = 0.75, Kw/Kf = 0.25

Story Drift Demands with Shear Deformations Story Drift Demands with Shear Deformations
P2, Frame + Fixed Wall, T / Tp = 1.0, ηf = 0.2, Kw / Kf = 0.5, no P-∆
∆ P2, Frame + Hinged Wall, T / Tp = 1.0, ηf = 0.2, Kw / Kf = 0.5, no P-∆
∆
1
b/H = 0.5
b/H = 0.4
0.8 b/H = 0.3
b/H = 0.2
Relative Height

b/H = 0.1
0.6
b/H = 0.0

0.4 b/H = 0.5

b/H = 0.4
b/H = 0.3
0.2 b/H = 0.2
b/H = 0.1
b/H = 0.0
0
0 0.04 0.08 0.12 0.16 0 0.04 0.08 0.12 0.16
δmax,i / ug,max δmax,i / ug,max

T/Tp = 1.0, ηf = 0.2, Kw/Kf = 0.5

Story Drift Demands with Shear Deformations Story Drift Demands with Shear Deformations
P2, Frame + Fixed Wall, T / Tp = 2.0, ηf = 0.5, Kw / Kf = 2.0, no P-∆
∆ P2, Frame + Hinged Wall, T / Tp = 2.0, ηf = 0.5, Kw / Kf = 2.0, no P-∆
∆
1
b/H = 0.5
b/H = 0.4
0.8 b/H = 0.3
b/H = 0.2
Relative Height

b/H = 0.1
0.6
b/H = 0.0

0.4 b/H = 0.5

b/H = 0.4
b/H = 0.3
0.2 b/H = 0.2
b/H = 0.1
b/H = 0.0
0
0 0.04 0.08 0.12 0.16 0 0.04 0.08 0.12 0.16
δmax,i / ug,max δmax,i / ug,max

T/Tp = 2.0, ηf = 0. 5, Kw/Kf = 2.0

(a) Fixed Wall (b) Hinged Wall

Figure 9.20 Effect of Shear Deformations on Distribution of Story Drift Demands Over
Height for Dual Systems with Elastic Wall, without P-Delta Effects

Chapter 9 Design Considerations for Near-Fault …

256
 Story Drift Demands Story Drift Demands
Pulse P2, 20-Story Frame + Fixed Wall, T / Tp = 0.5, η = 0.75, with P-∆
∆ Pulse P2, 20-Story Frame + Hinged Wall, T / Tp = 0.5, η = 0.75, with P-∆
∆
1
Kw/Kf = 0.00
Kw/Kf = 0.10
Kw/Kf = 0.25
0.8
Kw/Kf = 0.50
Kw/Kf = 1.00
Relative Height

Kw/Kf = 1.50
0.6 Kw/Kf = 2.00

0.4 Kw/Kf = 0.00

Kw/Kf = 0.10
Kw/Kf = 0.25
Kw/Kf = 0.50
0.2
Kw/Kf = 1.00
Kw/Kf = 1.50
Kw/Kf = 2.00
0
0 0.02 0.04 0.06 0.08 0.1 0 0.02 0.04 0.06 0.08 0.1
δmax,i / ug,max δmax,i / ug,max

T/Tp = 0.5, ηf = 0.75

Story Drift Demands Story Drift Demands

Pulse P2, 20-Story Frame + Fixed Wall, T / Tp = 1.0, η = 0.2, with P-∆
∆ Pulse P2, 20-Story Frame + Hinged Wall, T / Tp = 1.0, η = 0.2, with P-∆
∆
1

0.8
Relative Height

0.6

Kw/Kf = 0.00
0.4 Kw/Kf = 0.00
Kw/Kf = 0.10
Kw/Kf = 0.10
Kw/Kf = 0.25
Kw/Kf = 0.25
Kw/Kf = 0.50
Kw/Kf = 0.50
0.2 Kw/Kf = 1.00
Kw/Kf = 1.00
Kw/Kf = 1.50
Kw/Kf = 1.50
Kw/Kf = 2.00
Kw/Kf = 2.00
0
0 0.04 0.08 0.12 0.16 0 0.04 0.08 0.12 0.16
δmax,i / ug,max δmax,i / ug,max

T/Tp = 1.0, ηf = 0.2

Story Drift Demands Story Drift Demands

Pulse P2, 20-Story Frame + Fixed Wall, T / Tp = 2.0, η = 0.5, with P-∆
∆ Pulse P2, 20-Story Frame + Hinged Wall, T / Tp = 2.0, η = 0.5, with P-∆
∆
1

0.8
Relative Height

0.6

0.4
Kw/Kf = 0.00 Kw/Kf = 0.00
Kw/Kf = 0.10 Kw/Kf = 0.10
Kw/Kf = 0.25 Kw/Kf = 0.25
0.2 Kw/Kf = 0.50 Kw/Kf = 0.50
Kw/Kf = 1.00 Kw/Kf = 1.00
Kw/Kf = 1.50 Kw/Kf = 1.50
Kw/Kf = 2.00 Kw/Kf = 2.00
0
0 0.05 0.1 0.15 0.2 0.25 0.3 0 0.05 0.1 0.15 0.2 0.25 0.3
δmax,i / ug,max δmax,i / ug,max

T/Tp = 2.0, ηf = 0. 5
(a) Fixed Wall (b) Hinged Wall

Figure 9.21 Distribution of Story Drift Demands Over Height for Dual Systems with
Elastic Wall, with P-Delta Effects

Chapter 9 Design Considerations for Near-Fault …

257
 Total Story Drift Demands Wall Shear Drift Demands
P2, Frame + Hinged Wall, T/Tp = 0.5, ηf = 0.75, Kw / Kf = 0.25, b/H = 0.25 P2, Frame + Hinged Wall, T/Tp = 0.5, ηf = 0.75, Kw / Kf = 0.25, b/H = 0.25
1
frame only elastic wall
elastic wall eta(w)=0.4
0.8 eta(w)=0.4
eta(w)=0.3 eta(w)=0.3
eta(w)=0.2 eta(w)=0.2
Relative Height

0.6

0.4

0.2

0
0 0.02 0.04 0.06 0.08 0.1 0 0.02 0.04 0.06 0.08 0.1
δi,total / ug,max δi,shear / ug,max

T/Tp = 0.5, ηf = 0.75, Kw/Kf = 0.25

Total Story Drift Demands Wall Shear Drift Demands

P2, Frame + Hinged Wall, T/Tp = 1.0, ηf = 0.2, Kw / Kf = 0.5, b/H = 0.25 P2, Frame + Hinged Wall, T / Tp = 1.0, ηf = 0.2, Kw / Kf = 0.5, b / H = 0.25
1
frame only elastic wall
elastic wall eta(w)=0.4
0.8 eta(w)=0.4
eta(w)=0.3 eta(w)=0.3
eta(w)=0.2 eta(w)=0.2
Relative Height

0.6

0.4

0.2

0
0 0.04 0.08 0.12 0.16 0 0.04 0.08 0.12 0.16
δi,total / ug,max δi,shear / ug,max

T/Tp = 1.0, ηf = 0.2, Kw/Kf = 0.5

Total Story Drift Demands Wall Shear Drift Demands

P2, Frame + Hinged Wall, T/Tp = 2.0, ηf = 0.5, Kw / Kf = 2.0, b/H = 0.25 P2, Frame + Hinged Wall, T / Tp = 2.0, ηf = 0.5, Kw / Kf = 2.0, b / H = 0.25
1
frame only elastic wall
elastic wall
eta(w)=0.4
eta(w)=0.4
0.8
eta(w)=0.3 eta(w)=0.3
eta(w)=0.2 eta(w)=0.2
Relative Height

0.6

0.4

0.2

0
0 0.05 0.1 0.15 0.2 0.25 0.3 0 0.05 0.1 0.15 0.2 0.25 0.3
δi,total / ug,max δi,shear / ug,max

T/Tp = 2.0, ηf = 0.5, Kw/Kf = 2.0

(a) Total Drift (b) Wall Shear Drift

Figure 9.22 Distribution of Various Demands Over Height for Dual Systems with
Inelastic Hinged Wall, without P-Delta Effects

Chapter 9 Design Considerations for Near-Fault …

258
 Wall Shear Force Demands Wall Moment Demands
P2, Frame + Hinged Wall, T/Tp = 0.5, ηf = 0.75, Kw / Kf = 0.25, b/H = 0.25 P2, Frame + Hinged Wall, T/Tp = 0.5, ηf = 0.75, Kw / Kf = 0.25, b/H = 0.25
1
elastic wall elastic wall
eta(w)=0.4 eta(w)=0.4
0.8 eta(w)=0.3 eta(w)=0.3
eta(w)=0.2 eta(w)=0.2
Relative Height

0.6

0.4

0.2

0
0 0.2 0.4 0.6 0.8 0 0.04 0.08 0.12 0.16
Vi / ([Link],max) M / ([Link],max)

T/Tp = 0.5, ηf = 0.75, Kw/Kf = 0.25

Wall Shear Force Demands Wall Moment Demands

P2, Frame + Hinged Wall, T / Tp = 1.0, ηf = 0.2, Kw / Kf = 0.5, b / H = 0.25 P2, Frame + Hinged Wall, T / Tp = 1.0, ηf = 0.2, Kw / Kf = 0.5, b / H = 0.25
1
elastic wall elastic wall
eta(w)=0.4 eta(w)=0.4
0.8 eta(w)=0.3 eta(w)=0.3
eta(w)=0.2 eta(w)=0.2
Relative Height

0.6

0.4

0.2

0
0 0.2 0.4 0.6 0.8 0 0.04 0.08 0.12 0.16
Vi / ([Link],max) M / ([Link],max)

T/Tp = 1.0, ηf = 0.2, Kw/Kf = 0.5

Wall Shear Force Demands Wall Moment Demands

P2, Frame + Hinged Wall, T / Tp = 2.0, ηf = 0.5, Kw / Kf = 2.0, b / H = 0.25 P2, Frame + Hinged Wall, T / Tp = 2.0, ηf = 0.5, Kw / Kf = 2.0, b / H = 0.25
1
elastic wall elastic wall
eta(w)=0.4 eta(w)=0.4
0.8 eta(w)=0.3 eta(w)=0.3
eta(w)=0.2 eta(w)=0.2
Relative Height

0.6

0.4

0.2

0
0 0.2 0.4 0.6 0.8 0 0.04 0.08 0.12 0.16
Vi / ([Link],max) M / ([Link],max)

T/Tp = 2.0, ηf = 0.5, Kw/Kf = 2.0

(c) Wall Shear (d) Wall Moment

Figure 9.22 (Cont’d) Distribution of Various Demands Over Height for Dual Systems
with Inelastic Hinged Wall, without P-Delta Effects

Chapter 9 Design Considerations for Near-Fault …

259
 Story Drift Demands Wall Curvature Demands
P2, Frame + Fixed Wall, T/Tp = 0.5, ηf = 0.75, Kw/Kf = 0.25, no P-∆
∆ P2, Frame + Fixed Wall, T / Tp = 0.5, ηf = 0.75, Kw / Kf = 0.25, no P-∆
∆
1
κ = My/([Link],max) κ = My/([Link],max)

frame only elastic wall

0.8
elastic wall κ===0.4
κ===0.4 κ===0.2
κ===0.1
Relative Height

κ===0.2
0.6 κ===0.1

0.4

0.2

0
0 0.02 0.04 0.06 0.08 0.1 0 20 40 60 80
δmax,i / ug,max (H / ug,max)==φmax
2

T/Tp = 0.5, ηf = 0.75, Kw/Kf = 0.25

Story Drift Demands Wall Curvature Demands

P2, Frame + Fixed Wall, T / Tp = 1.0, ηf = 0.2, Kw / Kf = 0.5, no P-∆
∆ P2, Frame + Fixed Wall, T / Tp = 1.0, ηf = 0.2, Kw / Kf = 0.5, no P-∆
∆
1
κ = My/([Link],max)

elastic wall
0.8
κ===0.4
κ===0.2
κ===0.1
Relative Height

0.6

0.4 κ = My/([Link],max)

frame only
elastic wall
0.2
κ===0.4
κ===0.2
κ===0.1
0
0 0.04 0.08 0.12 0.16 0 20 40 60 80
δmax,i / ug,max (H / ug,max)==φmax
2

T/Tp = 1.0, ηf = 0.2, Kw/Kf = 0.5

Story Drift Demands Wall Curvature Demands

P2, Frame + Fixed Wall, T / Tp = 2.0, ηf = 0.5, Kw / Kf = 2.0, no P-∆
∆ P2, Frame + Fixed Wall, T / Tp = 2.0, ηf = 0.5, Kw / Kf = 2.0, no P-∆
∆
1
κ = My/([Link],max)

elastic wall
0.8
κ===0.4
κ===0.2
κ===0.1
Relative Height

0.6

0.4 κ = My/([Link],max)

frame only
elastic wall
0.2 κ===0.4
κ===0.2
κ===0.1
0
0 0.05 0.1 0.15 0.2 0.25 0.3 0 20 40 60 80
δmax,i / ug,max (H / ug,max)==φmax
2

T/Tp = 2.0, ηf = 0.5, Kw/Kf = 2.0

(a) Story Drift (b) Wall Curvature

Figure 9.23 Distribution of Various Demands Over Height for Dual Systems with
Inelastic Fixed Wall, without P-Delta Effects

Chapter 9 Design Considerations for Near-Fault …

260
 Wall Moment Demands Wall Shear Demands
P2, Frame + Fixed Wall, T / Tp = 0.5, ηf = 0.75, Kw / Kf = 0.25, no P-∆
∆ P2, Frame + Fixed Wall, T / Tp = 0.5, ηf = 0.75, Kw / Kf = 0.25, no P-∆
∆
1
κ = My/([Link],max) κ = My/([Link],max)

elastic wall elastic wall

0.8
κ===0.4 κ===0.4
κ===0.2 κ===0.2
Relative Height

κ===0.1 κ===0.1
0.6

0.4

0.2

0
0 0.2 0.4 0.6 0.8 0 0.5 1 1.5 2 2.5 3
M / ([Link],max) Vi / ([Link],max)

T/Tp = 0.5, ηf = 0.75, Kw/Kf = 0.25

Wall Moment Demands Wall Shear Demands

P2, Frame + Fixed Wall, T / Tp = 1.0, ηf = 0.2, Kw / Kf = 0.5, no P-∆
∆ P2, Frame + Fixed Wall, T / Tp = 1.0, ηf = 0.2, Kw / Kf = 0.5, no P-∆
∆
1
κ = My/([Link],max) κ = My/([Link],max)

elastic wall elastic wall

0.8
κ===0.4 κ===0.4
κ===0.2 κ===0.2
Relative Height

κ===0.1 κ===0.1
0.6

0.4

0.2

0
0 0.2 0.4 0.6 0.8 1 0 0.5 1 1.5 2 2.5 3
M / ([Link],max) Vi / ([Link],max)

T/Tp = 1.0, ηf = 0.2, Kw/Kf = 0.5

Wall Moment Demands Wall Shear Demands

P2, Frame + Fixed Wall, T / Tp = 2.0, ηf = 0.5, Kw / Kf = 2.0, no P-∆
∆ P2, Frame + Fixed Wall, T / Tp = 2.0, ηf = 0.5, Kw / Kf = 2.0, no P-∆
∆
1
κ = My/([Link],max) κ = My/([Link],max)

elastic wall elastic wall

0.8
κ===0.4 κ===0.4
κ===0.2 κ===0.2
Relative Height

κ===0.1 κ===0.1
0.6

0.4

0.2

0
0 0.2 0.4 0.6 0.8 1 0 0.5 1 1.5 2 2.5 3
M / ([Link],max) Vi / ([Link],max)

T/Tp = 2.0, ηf = 0.5, Kw/Kf = 2.0

(c) Wall Moment (d) Wall Shear

Figure 9.23 (Cont’d) Distribution of Various Demands Over Height for Dual Systems
with Inelastic Fixed Wall, without P-Delta Effects

Chapter 9 Design Considerations for Near-Fault …

261
 Story Drift Demands Wall Moment Demands
P2, Frame + Hinged Wall, T/Tp = 0.5, ηf = 0.75, Kw/Kf = 0.25, no P-∆
∆ P2, Frame + Hinged Wall, T / Tp = 0.5, ηf = 0.75, Kw / Kf = 0.25, no P-∆
∆
1
κ = My/([Link],max) κ = My/([Link],max)

frame only elastic wall

0.8
elastic wall κ===0.10
κ===0.10 κ===0.08
Relative Height

κ===0.08
κ===0.06
0.6 κ===0.06

0.4

0.2

0
0 0.02 0.04 0.06 0.08 0.1 0 0.04 0.08 0.12 0.16
δmax,i / ug,max M / ([Link],max)

T/Tp = 0.5, ηf = 0.75, Kw/Kf = 0.25

Story Drift Demands Wall Moment Demands

P2, Frame + Hinged Wall, T / Tp = 1.0, ηf = 0.2, Kw / Kf = 0.5, no P-∆
∆ P2, Frame + Hinged Wall, T / Tp = 1.0, ηf = 0.2, Kw / Kf = 0.5, no P-∆
∆
1
κ = My/([Link],max)

elastic wall
0.8
κ===0.08
κ===0.06
Relative Height

κ===0.04
0.6

0.4 κ = My/([Link],max)

frame only
elastic wall
0.2
κ===0.08
κ===0.06
κ===0.04
0
0 0.04 0.08 0.12 0.16 0 0.04 0.08 0.12 0.16
δmax,i / ug,max M / ([Link],max)

T/Tp = 1.0, ηf = 0.2, Kw/Kf = 0.5

Story Drift Demands Wall Moment Demands

P2, Frame + Hinged Wall, T / Tp = 2.0, ηf = 0.5, Kw / Kf = 2.0, no P-∆
∆ P2, Frame + Hinged Wall, T / Tp = 2.0, ηf = 0.5, Kw / Kf = 2.0, no P-∆
∆
1
κ = My/([Link],max)

elastic wall
0.8
κ===0.06
κ===0.04
Relative Height

κ===0.03
0.6

0.4 κ = My/([Link],max)

frame only
elastic wall
0.2
κ===0.06
κ===0.04
κ===0.03
0
0 0.05 0.1 0.15 0.2 0.25 0.3 0 0.04 0.08 0.12 0.16
δmax,i / ug,max M / ([Link],max)

T/Tp = 2.0, ηf = 0.5, Kw/Kf = 2.0

(a) Story Drift (b) Wall Moment

Figure 9.24 Effects of Wall Flexural Yielding on Story Drift and Wall Moment Demands
for Dual Systems with Hinged Wall, without P-Delta Effects

Chapter 9 Design Considerations for Near-Fault …

262
 CHAPTER 10

NEAR-FAULT GROUND MOTIONS FROM

MODERATE EARTHQUAKES

The results of the studies presented in the previous chapters demonstrated the pulse-type
characteristics of near-fault ground motions with forward directivity, recorded during
relatively large earthquakes with magnitudes ranging from 6.2 to 7.4. The properties of
simple pulses were employed to represent such ground motions in Chapter 7, and the
pulse representation was taken advantage of in Chapter 9 to study the beneficial effects of
strengthening frame structures. An issue that remains to be addressed is whether near-
fault ground motions recorded in the forward directivity region of smaller events also
exhibit pulse-type characteristics.

The results presented in this chapter shed light on the response of SDOF and MDOF
structures located in the near-fault region of moderate earthquakes. These earthquakes
may not be severe enough to raise a major concern about collapse safety, but they occur
more frequently and can endanger the serviceability of the structure or contribute to
cumulative damage. Performance-based engineering guidelines require that the behavior
of structures subjected to moderate (but more frequent) ground motions also be
evaluated. Several near-fault ground motions from moderate events are introduced for
this purpose, and their elastic response spectra are investigated. The spectral shapes of
near-fault ground motions from moderate events are compared with those of “ordinary”
ground motions. Finally, MDOF demands for frame structures subjected to the ground
motion records are quantified.

10.1. Ground Motion Records Used in this Study

Five near-fault ground motions recorded during the Parkfield (1966) and Coyote Lake
(1979) earthquakes are used in the study of moderate events. The designation and basic

Chapter 10 Near-Fault Ground Motions from Moderate …

263
properties of these ground motions are listed in Table 10.1. All five recording stations
are located in the forward direction of the corresponding seismic sources. Two horizontal
components of each ground motion are obtained from the PEER Strong Motion Database
(Silva, 1999). The ground motions are recorded on USGS soil type C (roughly
equivalent to NEHRP soil type D), which is compatible with the soil conditions of the
records introduced in Chapter 2 for the study of near-fault effects in larger events.

Ground velocity time history traces are presented in Fig. 10.1 for the fault-normal and
fault-parallel components of the ground motions. The following observations are made:

•= The time history traces for the Parkfield records do not exhibit typical pulse-type
characteristics similar to those observed in Chapter 2 for near-fault records with
forward directivity. The maximum time history values for the two components
are quite comparable, and there is no dominant component.

•= The Coyote Lake records are more pulse-like in their time histories compared to
the Parkfield records. The fault-parallel component of these records contains a
more distinct pulse of motion than the fault-normal component. The PGV values
for the fault-parallel component are considerably larger than their fault-normal
counterparts.

10.2. Elastic Response Spectra

Figure 10.2 illustrates acceleration (elastic strength demand), velocity, and displacement
spectra of the near-fault ground motions introduced in the previous section. Each graph
shows the spectra for the fault-normal and fault-parallel components. The following
observations are made:

•= The spectra of the Parkfield records confirm the observations made from the time
history traces. The shape of the spectra is not like that of pulse-type ground
motions or near-fault ground motions recorded during large earthquakes. No
component of the record dominates in the spectra.

•= For the Coyote Lake ground motions the spectral values for the fault-parallel
component are larger than the values for the fault-normal component, which is
contrary to the observation made for near-fault ground motions from large

Chapter 10 Near-Fault Ground Motions from Moderate …

264
 earthquakes. Pulse-type characteristics observed in the time histories appear as a
global hump in the velocity spectra. However, the humps of the velocity spectra
are not as well-defined as those of the spectra of near-fault ground motions
recorded during large events (see Chapter 2).

The conclusion is that ground motions recorded in the forward direction of a moderate
earthquake may not be characterized by a distinct pulse. The reason may be that the
amount of energy released in moderate events is not sufficiently large to cause severe
directivity effects. Contrary to near-fault records from large earthquakes, ground motions
from moderate earthquakes have comparable fault-normal and fault-parallel components.

10.2.1 Comparison with Ordinary Ground Motions

It was demonstrated that near-fault records from moderate events might not contain
distinct pulse-type characteristics. The objective of this part of the study is to evaluate
the extent to which moderate-magnitude near-fault ground motions can be represented by
“ordinary” ground motions that are not affected by directivity in the near-fault region.

Individual Ground Motions:

The spectra of the ground motions introduced in the previous section are compared with
reference spectra derived from an attenuation relationship proposed by Abrahamson and
Silva (1997). The attenuation relationship, which expresses the spectral acceleration (Sa
[g]) at a specific period as a function of the moment magnitude (Mw) and the closest
distance from the site to the rupture plane (R [km]), is given by the following equation
for strike-slip faulting:

ln(Sa ) = f 1 (M w , R) + S.f 5 (PGA rock ) (10.1)

where Sa is spectral acceleration for the average horizontal component, S is a soil factor
with its value being zero for rock and shallow soil (Geomatrix A and B) and unity for
deep soil (Geomatrix C and D), PGArock is the peak ground acceleration (spectral
acceleration at a very short period) from Eq. (10.1) with S = 0, and the functions f1 (for
Mw ≤ 6.4) and f5 are given as follows:

Chapter 10 Near-Fault Ground Motions from Moderate …

265
f1 (M w , R) = a 1 + a 2 (M w − 6.4) + a 12 (8.5 − M w ) 2
(10.2)
+ [a 3 + a 13 (M w − 6.4)]ln R 2 + c 24

f 5 (PGA rock ) = a 10 + a 11ln(PGA rock + 0.03) (10.3)

The period dependent coefficients a1, a2, a3, a10, a11, a12, a13, and c4 are given in
Abrahamson and Silva (1997). Since this attenuation model has not been modified for
directivity effects at close distances, the spectrum obtained from Eq. 10.1 is considered to
represent ordinary ground motions.

Figure 10.3 illustrates a comparison between the reference spectrum (for the appropriate
magnitude and distance) and the spectra for the fault-normal and fault-parallel
components of the ground motion CL79ga4. A similar comparison is illustrated in Fig.
10.4 for the ground motion PK66ch5. The reference spectrum, denoted as “A-S model”,
is obtained for each record from Eq. 10.1 with S = 1.0 and Mw and R values listed in
Table 10.1. The reason for choosing S = 1.0 is that the ground motions are recorded on
USGS soil type C which is more compatible with Geomatrix soil type C than type B.
Since the reference spectrum pertains to the average horizontal component, it should be
compared with the average of spectral values for the fault-normal and fault-parallel
components of the ground motions shown in the figures.

A more comprehensive comparison is provided by the ratio of the average spectral values
(fault-normal and fault-parallel components) of each near-fault record to the value of the
corresponding reference spectrum obtained from Eq. 10.1. Such ratios, as a function of
period, are illustrated in Fig. 10.5 for all five near-fault ground motions under
consideration. Superimposed on the graph is the mean of the ratios for the five ground
motions. It can be observed that even though the ratio for individual ground motions
varies in a relatively wide range (from 0.4 to 2.2) and is also period dependent, on
average it is close to 1.0. Drawing strong conclusions is difficult on account of the small
number of ground motions considered in this study, but the results indicate that the
spectral values of the near-fault ground motions, which are recorded during moderate
earthquakes with magnitude 6.1 and smaller, may be represented by the spectral shape of
ordinary ground motions that are not affected by directivity in the near-fault region.

Chapter 10 Near-Fault Ground Motions from Moderate …

266
To provide a comprehensive comparison of the fault-normal and fault-parallel
components of near-fault ground motions from moderate events, Fig. 10.6 illustrates the
ratio of fault-normal to fault-parallel spectral values as a function of period for the five
ground motions. The mean of the ratios for the five ground motions is also shown. It can
be observed that the ratio is period and ground motion dependent, but on average it is not
far from 1.0 (with values generally smaller than 1.0). As also pointed out previously, this
is in contrast to the observations made for near-fault records from large events, in which
case the fault-normal component is significantly more severe than the fault-parallel
component.

Scenario Events:

Figure 10.7 compares elastic acceleration spectra for two scenario events with R = 3 km
and Mw = 6.0 and 7.0, obtained from two different approaches as follows: (1) using the
Abrahamson-Silva attenuation model discussed previously, which does not take
directivity effects into consideration, and (2) using the equivalent pulse model discussed
in Section 9.1, which takes forward directivity into account.

In the first approach, Eqs. 10.1 to 10.3 are utilized with S = 1.0. It should be noted that
for Mw > 6.4, the coefficient a2 in Eq. 10.2 is replaced by the coefficient a4 given in
Abrahamson and Silva (1997). The values obtained from Eq. 10.1 correspond to the
mean acceleration spectrum of the two components. A modification factor proposed by
Somerville et al. (1997b) is utilized here in order to convert the mean spectrum to the
spectrum corresponding to the fault-normal component. The following equation
represents the ratio of fault-normal to average spectral values for Mw ≥ 6:

ln(SaFN/SaAVG) = C1 + C2 ln(R + 1) + C3 (Mw – 6) (10.4)

where C1, C2, and C3 are period dependent coefficients given in Somerville et al. (1997b).

In the second approach, the elastic acceleration spectrum of pulse P2 (see Fig. 5.4) is
utilized, along with Eqs. 9.1 and 9.3, to estimate the pulse period (Tp) and intensity (ag,max
= 4veff/Tp) of the equivalent pulse for the scenario events. In order to be compatible with
the spectrum obtained from the first approach, the pulse spectrum is computed for a 5%
damping ratio. Since the fault-normal component of near-fault ground motions with
forward directivity was used in developing Eqs. 9.1 and 9.3, the spectral values obtained

Chapter 10 Near-Fault Ground Motions from Moderate …

267
from the equivalent pulse model pertain to the fault-normal component and include
directivity effects. The following observations are made from Fig. 10.7:

•= For large earthquakes (Mw = 7.0) the spectral shape of the pulse-type ground
motion, which represents near-fault ground motions with forward directivity,
differs radically from the spectral shape of ordinary ground motions without
directivity effects. In the period range T > 1.0 sec., the spectral values for the
pulse-type ground motion are considerably larger than those for ordinary records.

•= For moderate earthquakes (Mw = 6.0) the spectral shapes are much closer,
implying that the directivity effect in the near-fault region of moderate
earthquakes is less important than that of large earthquakes. In the period range T
> 2.0 sec., the spectral values for the pulse-type ground motion are even smaller
than the corresponding values for ordinary records.

These observations further support the hypothesis that spectral values of the near-fault
ground motions from moderate earthquakes (Mw ≤ 6) can be represented by the spectral
shape of ordinary ground motions not affected by directivity in the near-fault region.

10.3. Story Ductility Demands Over Height

To evaluate MDOF demands for structures subjected to near-fault records from moderate
earthquakes, distributions of story ductility demands over the height of the structure are
presented in this section. Figures 10.8 and 10.9 illustrate the story ductility distributions
for fault-normal and fault-parallel components of ground motions CL79ga3 and
PK66ch8. The generic 20-story frame introduced in Section 3.2 is utilized here with
periods of T = 0.5, 1.0, and 2.0 sec., and base shear strengths defined by the coefficient γ
= Vy/W. The following observations can be made:

•= For short period (T = 0.5 sec.) and strong structures, the distribution of story
ductility is relatively uniform over the height. For weaker structures the
maximum story ductility occurs in the bottom story, and grows rapidly with a
decrease in strength. This pattern is not much different from the observations
made for near-fault records from large earthquakes (see Chapter 4) and pulse-type
ground motions (see Chapter 6).

Chapter 10 Near-Fault Ground Motions from Moderate …

268
 •= For long period structures (T = 2.0 sec.) the story ductility distributions for
CL79ga3 exhibit some pulse-type patterns, but to a lesser extent compared to
records from large events. Similar patterns are observed for both the fault-normal
and fault-parallel components. The critical story (with maximum ductility) is in
the top portion for relatively strong structures, and migrates to the base when the
base shear strength is reduced. The story ductility distributions for PK66ch8 are
less pulse-like and show characteristics similar to those of ordinary ground
motion. This is in agreement with the observations previously made from the
spectra and time history traces.

•= For a given base shear strength value, larger spectral values of the fault-parallel
component of CL79ga3 translate into larger story ductility demands for this
component, especially at T = 0.5 sec.

The conclusion based on the results presented here is that the response of MDOF
structures to ground motions recorded in the near-fault region of moderate earthquakes
does not exhibit strong pulse-type characteristics.

In order to explore the strength dependence of story ductility demands, unrealistically

small values for the base shear strength coefficient γ had to be used in the investigations.
To put the presented results in perspective, a realistic base shear strength is estimated
here based on the design provisions of UBC’97. Assuming that the structure is at a
distance of R = 3 km from the source in Seismic Zone 4 (soil type SD), the base shear
strength of the structure can be estimated from the UBC design spectrum (including the
near-fault factor for Type A sources), considering a strength reduction factor of 8, and an
overstrength factor of 2. As shown in Fig. 9.4 (the curve denotes as “UBC/4, with
NFF”), the strength coefficients for T = 0.5, 1.0, and 2.0 sec. are γ = 0.39, 0.30, and 0.15,
respectively. The results presented in Figs. 10.8 and 10.9 indicate that for these γ values,
the structure subjected to the two records under consideration remains elastic or almost
elastic. This close to elastic behavior signifies minor structural damage, and indicates
that structures subjected to these ground motions are likely to meet standard performance
objectives.

The conclusion to be drawn is that the near-fault effect diminishes with decreasing
magnitude and becomes unimportant for magnitudes of about 6.1 and smaller.

Chapter 10 Near-Fault Ground Motions from Moderate …

269
 Table 10.1 Designation and Properties of Near-Fault Records from
Moderate Earthquakes

Designation Earthquake Station Magnitude Distance

PK66ch5 Parkfield, 1966 Cholame #5 6.1 5.3
PK66ch8 Parkfield, 1966 Cholame #8 6.1 9.2
CL79ga2 Coyote Lake, 1979 Gilroy Array #2 5.7 7.5
CL79ga3 Coyote Lake, 1979 Gilroy Array #3 5.7 6.0
CL79ga4 Coyote Lake, 1979 Gilroy Array #4 5.7 4.5

Chapter 10 Near-Fault Ground Motions from Moderate …

270
 40 40
max = 23.0 PK66ch5 max = 23.9
PK66ch5

20 20

vg (cm / sec)
0 0

-20 -20

-40 -40
20 20
PK66ch8 max = 10.5 PK66ch8 max = 11.6

10 10
vg (cm / sec)

0 0

-10 -10

-20 -20
40 40
CL79ga2 max = 12.3 CL79ga2 max = 25.0

20 20
vg (cm / sec)

0 0

-20 -20

-40 -40
40 40
CL79ga3 max = 18.4 CL79ga3 max = 28.2

20 20
vg (cm / sec)

0 0

-20 -20

-40 -40
40 40
CL79ga4 max = 21.6 CL79ga4 max = 27.5

20 20
vg (cm / sec)

0 0

-20 -20

-40 -40
0 4 8 12 16 20 0 4 8 12 16 20
Time (sec) Time (sec)

(a) Fault-Normal (b) Fault-Parallel

Figure 10.1 Ground Velocity Time Histories for Near-Fault Ground Motions from
Moderate Earthquakes

Chapter 10 Near-Fault Ground Motions from Moderate …

271
 PK66ch5, ξ = 2% PK66ch8, ξ = 2%
2 1.6
Fault-Normal Fault-Normal
Fault-Parallel Fault-Parallel
1.5 1.2

Sa (g)

1 0.8

0.5 0.4

0 0
100 50
Fault-Normal Fault-Normal

80 Fault-Parallel 40 Fault-Parallel
Sv (cm / sec)

60 30

40 20

20 10

0 0
20 16
Fault-Normal Fault-Normal
Fault-Parallel Fault-Parallel
15 12
Sd (cm)

10 8

5 4

0 0
0 0.5 1 1.5 2 2.5 3 3.5 4 0 0.5 1 1.5 2 2.5 3 3.5 4
T (sec) T (sec)

CL79ga2, ξ = 2% CL79ga3, ξ = 2%
1.6 1.6
Fault-Normal Fault-Normal
Fault-Parallel Fault-Parallel
1.2 1.2
Sa (g)

0.8 0.8

0.4 0.4

0 0
100 100
Fault-Normal Fault-Normal

80 Fault-Parallel 80 Fault-Parallel
Sv (cm / sec)

60 60

40 40

20 20

0 0
20 20
Fault-Normal Fault-Normal
Fault-Parallel Fault-Parallel
15 15
Sd (cm)

10 10

5 5

0 0
0 0.5 1 1.5 2 2.5 3 3.5 4 0 0.5 1 1.5 2 2.5 3 3.5 4
T (sec) T (sec)

Figure 10.2 Elastic Acceleration, Velocity, and Displacement Response Spectra of

Near-Fault Ground Motions from Moderate Earthquakes

Chapter 10 Near-Fault Ground Motions from Moderate …

272
 CL79ga4, ξ = 2%
1.6
Fault-Normal
Fault-Parallel
1.2
Sa (g)

0.8

0.4

0
100
Fault-Normal

80 Fault-Parallel
Sv (cm / sec)

60

40

20

0
20
Fault-Normal
Fault-Parallel
15
Sd (cm)

10

0
0 0.5 1 1.5 2 2.5 3 3.5 4
T (sec)

Figure 10.2 (Cont’d) Elastic Acceleration, Velocity, and Displacement Response Spectra
of Near-Fault Ground Motions from Moderate Earthquakes

Chapter 10 Near-Fault Ground Motions from Moderate …

273
 Elastic SDOF Strength Demands
Abrahamson-Silva Model (Mw = 5.7, R = 4.5 km) vs. CL79ga4, ξ = 5%
0.8
A-S Model
Fault-Normal

0.6 Fault-Parallel

Sa (g)
0.4

0.2

0
0 0.5 1 1.5 2 2.5 3 3.5 4
T (sec)

Elastic Pseudo Velocity Demands

Abrahamson-Silva Model (Mw = 5.7, R = 4.5 km) vs. CL79ga4, ξ = 5%
80
A-S Model
Fault-Normal

60 Fault-Parallel
PSV (cm / sec)

40

20

0
0 0.5 1 1.5 2 2.5 3 3.5 4
T (sec)

Elastic Pseudo Displacement Demands

Abrahamson-Silva Model (Mw = 5.7, R = 4.5 km) vs. CL79ga4, ξ = 5%
16
A-S Model
Fault-Normal
Fault-Parallel
12
PSD (cm)

0
0 0.5 1 1.5 2 2.5 3 3.5 4
T (sec)

Figure 10.3 Comparison of Spectra of Horizontal Components of CL79ga4 with

Reference Spectrum for Mw = 5.7 and R = 4.5 km

Chapter 10 Near-Fault Ground Motions from Moderate …

274
 Elastic SDOF Strength Demands
Abrahamson-Silva Model (Mw = 6.1, R = 5.3 km) vs. PK88ch5, ξ = 5%
1.6
A-S Model
Fault-Normal

1.2 Fault-Parallel

Sa (g)
0.8

0.4

0
0 0.5 1 1.5 2 2.5 3 3.5 4
T (sec)

Elastic Pseudo Velocity Demands

Abrahamson-Silva Model (Mw = 6.1, R = 5.3 km) vs. PK88ch5, ξ = 5%
80
A-S Model
Fault-Normal

60 Fault-Parallel
PSV (cm / sec)

40

20

0
0 0.5 1 1.5 2 2.5 3 3.5 4
T (sec)

Elastic Pseudo Displacement Demands

Abrahamson-Silva Model (Mw = 6.1, R = 5.3 km) vs. PK88ch5, ξ = 5%
16
A-S Model
Fault-Normal
Fault-Parallel
12
PSD (cm)

0
0 0.5 1 1.5 2 2.5 3 3.5 4
T (sec)

Figure 10.4 Comparison of Spectra of Horizontal Components of PK88ch5 with

Reference Spectrum for Mw = 6.1 and R = 5.3 km

Chapter 10 Near-Fault Ground Motions from Moderate …

275
 Elastic SDOF Strength Demand Ratios
Near-Fault Records vs. Abrahamson-Silva Model, Avg. Comp., ξ = 5%
2.5
PK66ch5
PK66ch8
CL79ga2
2
CL79ga3
CL79ga4
A-S model

Average
1.5
/ Sa
record

1
Sa

0.5

0
0 1 2 3 4 5
T (sec)

Figure 10.5 Ratio of Average Spectrum of Components of Near-Fault Records to

Corresponding Reference Spectrum

Elastic SDOF Strength Demand Ratios

Comparison of Fault-Normal and Fault-Parallel Componets, ξ = 5%
2.5
PK66ch5
PK66ch8
CL79ga2
2
CL79ga3
CL79ga4
Average
FP

1.5
/ Sa
FN
Sa

0.5

0
0 1 2 3 4 5
T (sec)

Figure 10.6 Ratio of Spectral Values of Fault-Normal to Fault-Parallel Components of

Near-Fault Ground Motions from Moderate Earthquakes

Chapter 10 Near-Fault Ground Motions from Moderate …

276
 Elastic SDOF Strength Demands
Mw = 7.0, R = 3 km, Fault-Normal, Soil, Strike Slip, ξ== 5%
1.6
Abrahamson-Silva model

Equivalent pulse P2
1.2
Sa (g)

0.8

0.4

0
0 1 2 3 4 5
T (sec)

(a) Mw = 7.0

Elastic SDOF Strength Demands

Mw = 6.0, R = 3 km, Fault-Normal, Soil, Strike Slip, ξ== 5%
1
Abrahamson-Silva model

0.8 Equivalent pulse P2

0.6
Sa (g)

0.4

0.2

0
0 1 2 3 4 5
T (sec)

(b) Mw = 6.0

Figure 10.7 Comparison of Elastic Response Spectra for Pulse-Type and Ordinary
Ground Motions for Two Scenario Events with R = 3 km

Chapter 10 Near-Fault Ground Motions from Moderate …

277
 Story Ductility Demands
Story Ductility Demands ∆
CL79ga3, Fault-Parallel, T = 0.5 sec, without P-∆
∆
CL79ga3, Fault-Normal, T = 0.5 sec, without P-∆
1
γ===0.40
γ===0.40
γ===0.30
γ===0.30
γ===0.20
0.8 γ===0.20
γ===0.15
γ===0.15
γ===0.10
γ===0.10
Relative Height

γ===0.08
γ===0.08
0.6 γ===0.06
γ===0.06
γ===0.05
γ===0.05

0.4

0.2

0
0 2 4 6 8 10 0 2 4 6 8 10
Story Ductility Ratio, µi = δmax,i / δy,i Story Ductility Ratio, µi = δmax,i / δy,i

T = 0.5 sec.

Story Ductility Demands Story Ductility Demands

∆
CL79ga3, Fault-Normal, T = 1.0 sec, without P-∆ ∆
CL79ga3, Fault-Parallel, T = 1.0 sec, without P-∆
1
γ===0.40 γ===0.40
γ===0.30 γ===0.30
0.8 γ===0.20 γ===0.20
γ===0.15 γ===0.15
γ===0.10 γ===0.10
Relative Height

γ===0.06 γ===0.06
0.6
γ===0.04 γ===0.04
γ===0.02 γ===0.02

0.4

0.2

0
0 2 4 6 8 10 0 2 4 6 8 10
Story Ductility Ratio, µi = δmax,i / δy,i Story Ductility Ratio, µi = δmax,i / δy,i

T = 1.0 sec.

Story Ductility Demands Story Ductility Demands

∆
CL79ga3, Fault-Normal, T = 2.0 sec, without P-∆ ∆
CL79ga3, Fault-Parallel, T = 2.0 sec, without P-∆
1
γ===0.15 γ===0.15
γ===0.10 γ===0.10
0.8 γ===0.05 γ===0.05
γ===0.04 γ===0.04
γ===0.03 γ===0.03
Relative Height

γ===0.02 γ===0.02
0.6
γ===0.01 γ===0.01
γ===0.008 γ===0.008

0.4

0.2

0
0 2 4 6 8 10 0 2 4 6 8 10
Story Ductility Ratio, µi = δmax,i / δy,i Story Ductility Ratio, µi = δmax,i / δy,i

T = 2.0 sec.
(a) Fault-Normal (b) Fault-Parallel

Figure 10.8 Distributions of Story Ductility Demands Over Height for Fault-Normal and
Fault-Parallel Components of Ground Motion CL79ga3

Chapter 10 Near-Fault Ground Motions from Moderate …

278
 Story Ductility Demands Story Ductility Demands
∆
PK66ch8, Fault-Normal, T = 0.5 sec, without P-∆ ∆
PK66ch8, Fault-Parallel, T = 0.5 sec, without P-∆
1
γ===0.40 γ===0.40
γ===0.25 γ===0.25
0.8 γ===0.15 γ===0.15
γ===0.10 γ===0.10
γ===0.08 γ===0.08
Relative Height

γ===0.05 γ===0.05
0.6
γ===0.04 γ===0.04
γ===0.03 γ===0.03

0.4

0.2

0
0 2 4 6 8 10 0 2 4 6 8 10
Story Ductility Ratio, µi = δmax,i / δy,i Story Ductility Ratio, µi = δmax,i / δy,i

T = 0.5 sec.

Story Ductility Demands Story Ductility Demands

∆
PK66ch8, Fault-Normal, T = 1.0 sec, without P-∆ ∆
PK66ch8, Fault-Parallel, T = 1.0 sec, without P-∆
1
γ===0.30 γ===0.30
γ===0.20 γ===0.20
0.8 γ===0.15 γ===0.15
γ===0.10 γ===0.10
γ===0.06 γ===0.06
Relative Height

γ===0.04 γ===0.04
0.6
γ===0.02 γ===0.02
γ===0.01 γ===0.01

0.4

0.2

0
0 2 4 6 8 10 0 2 4 6 8 10
Story Ductility Ratio, µi = δmax,i / δy,i Story Ductility Ratio, µi = δmax,i / δy,i

T = 1.0 sec.

Story Ductility Demands Story Ductility Demands

∆
PK66ch8, Fault-Normal, T = 2.0 sec, without P-∆ ∆
PK66ch8, Fault-Parallel, T = 2.0 sec, without P-∆
1
γ===0.10 γ===0.10
γ===0.08 γ===0.08
0.8 γ===0.05 γ===0.05
γ===0.02 γ===0.02
γ===0.01 γ===0.01
Relative Height

γ===0.006 γ===0.006
0.6
γ===0.005 γ===0.005
γ===0.004 γ===0.004

0.4

0.2

0
0 2 4 6 8 10 0 2 4 6 8 10
Story Ductility Ratio, µi = δmax,i / δy,i Story Ductility Ratio, µi = δmax,i / δy,i

T = 2.0 sec.
(a) Fault-Normal (b) Fault-Parallel

Figure 10.9 Distributions of Story Ductility Demands Over Height for Fault-Normal and
Fault-Parallel Components of Ground Motion PK66ch8

Chapter 10 Near-Fault Ground Motions from Moderate …

279
 CHAPTER 11

SUMMARY AND CONCLUSIONS

The major objective of the study presented in this report is to develop a basic
understanding of the important attributes that characterize near-fault ground motions and
their effects on the response of elastic and inelastic SDOF and MDOF structural systems.
It is necessary to identify the response characteristics that set near-fault ground motions
apart from “ordinary” ground motions whose effects on the response of structures have
been considered, either explicitly or implicitly, in presently employed design procedures.

A comprehensive evaluation of seismic demands is presented for SDOF systems and

MDOF frame structures subjected to near-fault and pulse-type ground motions. The
SDOF systems employed in this study have bilinear hysteretic characteristics. A generic
20-story frame model is used extensively to evaluate MDOF response characteristics of
frame structures. This frame model is designed according to the “weak beam – strong
column” concept, and member strengths are tuned such that simultaneous yielding occurs
at all beam ends (and the column bases) under an SRSS shear force pattern. Bilinear
moment-rotation relationships are assumed at the plastic hinge locations.

Much effort is devoted to representing near-fault ground motions with forward directivity
using the properties of simple equivalent pulses. The equivalent pulses are taken
advantage of to estimate the response of MDOF structures to given near-fault ground
motions through a relatively simple procedure. Models are developed that relate
equivalent pulse parameters to earthquake magnitude and closest distance to the fault
rupture plane. A design methodology is introduced based on the equivalent pulse concept
to estimate the base shear strength required to limit story ductility ratios to specific target
values. Various distributions of story shear strength over the height are evaluated.
Strengthening techniques are investigated for providing effective protection of structures
subjected to near-fault ground motions.

Chapter 11 Summary and Conclusions

280
The following paragraphs present summary conclusions obtained on near-fault ground
motion characteristics and on the response of elastic and inelastic SDOF and MDOF
systems to recorded and simulated near-fault records and equivalent pulse
representations.

Ground Motion Characteristics

The study of ground time history traces and elastic response spectra for 22 recorded and
18 simulated near-fault ground motions from earthquakes with magnitudes between 6.2
and 7.5, and rupture distances between 0 and 10 km leads to the following conclusions:

•= Near-fault ground motions in the forward direction of a large earthquake, where

the rupture propagates towards the site, have short duration and are characterized
by a large pulse.

•= The fault-normal component of near-fault ground motions with forward

directivity is more severe than the fault-parallel component.

•= The 45° rotated components also exhibit pulse-type characteristics, and at least
one of them is almost as severe as the fault-normal component.

•= Spectral values for individual near-fault ground motions can be several times the
values given by the UBC’97 design spectrum.

To investigate whether near-fault records from moderate earthquakes also exhibit pulse-
type characteristics, the response of SDOF and MDOF systems subjected to ground
motions recorded in the forward direction of two events with magnitudes 6.1 and 5.7 are
evaluated. This part of the study leads to the following conclusions:

•= Near-fault ground motions from moderate earthquakes (Mw < 6.2) do not exhibit
distinct pulse-type characteristics such as those observed for large-magnitude
ground motions.

•= No clear distinction can be made between the intensity of the fault-normal and
fault-parallel components of the ground motions.

Chapter 11 Summary and Conclusions

281
 •= These ground motions may be represented by the spectral shape of ordinary
ground motions that do not include directivity effects.

Response of MDOF Systems to Near-Fault Records and Equivalent Pulses

Elastic Response. Elastic modal and time history analyses are performed for the generic
structure subjected to pulses and the fault-normal component of near-fault ground
motions with forward directivity. The main conclusions are as follows:

•= SRSS modal combinations cannot capture all important response characteristics

of elastic structures when the structure fundamental period (T) is longer than the
(equivalent) pulse period (Tp). A traveling wave effect, which is a fundamental
characteristic of the response of long period MDOF structures to near-fault
ground motions, is not adequately accounted for by standard SRSS spectral
analyses.

•= The results of the time history analyses indicate that for long period structures (T
> Tp) designed to the standard SRSS story shear distribution, the distribution of
elastic story shear forces over the height is sensitive to the ratio T/Tp, and shear
forces in upper stories may be higher than the base shear.

Inelastic Response. When the strength of the generic structure is varied, nonlinear time
history analyses of the generic structure subjected to near-fault and pulse-type ground
motions lead to the following conclusions:

•= The large elastic shear forces in the upper portion of long period structures (T >
Tp) result in early yielding of upper stories when the structure is relatively strong.
When the structure strength is reduced, the story ductility demands stabilize in the
upper portion and the maximum ductility demand migrates to the base, where the
ductility demand grows rapidly with a decrease in the strength.

•= For short period structures (T ≤ Tp) the traveling wave effect is not predominant
and the maximum story ductility demands occur in the bottom portion regardless
of the structure strength.

Chapter 11 Summary and Conclusions

282
 •= Inelastic roof displacements are larger than elastic ones in the range of T/Tp <
0.75, and increase with a decrease in structure strength. The reverse is observed
in the range of T/Tp > 1.0, where inelastic roof displacement demands are smaller
than the elastic ones and decrease when the strength is reduced.

P-Delta Effects. To quantify P-delta effects, a stability coefficient of 10% is assumed in

the first story, which causes a negative post-yield slope in the global pushover curve. A
stability coefficient of 10% is realistic for long period structures but is too large for short
period structures. Nevertheless, this value is consistently used for the generic structure
regardless of the period. Moreover, the effect of interior gravity frames, which are
typically designed with simple connections in US practice, is not considered here.
Incorporating these gravity frames will reduce the magnitude of P-delta effects. Thus,
caution should be taken in the interpretation of the results, especially for short period
structures. The conclusions drawn for dynamic P-delta effects are as follows:

•= For long period structures (T > Tp) subjected to near-fault or pulse-type ground
motions, P-delta effects are not very large when the structure is strong enough to
prevent the migration of the maximum ductilities from the upper portion of the
structure to the base. When the maximum ductility demand occurs at the base, the
effect of P-delta gains much on importance.

•= For short period structures (T ≤ Tp), since the maximum ductility always occurs
close to the base, P-delta effects may be significant even when the structure is
relatively strong.

Sensitivity of Seismic Demands to Structure Configuration. A sensitivity study is

performed to evaluate the extent to which the generic 20-story structure represents global
and story-level demands for generic structures with 3 and 9 stories. Furthermore, to
evaluate the representativeness of the generic structures, a pilot study is performed on
two steel moment resisting frame structures, which were extensively studied in the SAC
project (LA 9- and 3-story). The generic 9- and 3-story models are calibrated using
inelastic static analysis to represent the LA 9- and 3-story structures, respectively, and
their dynamic responses are compared. The conclusions to be drawn are summarized as
follows:

Chapter 11 Summary and Conclusions

283
 •= The seismic demands show various degrees of sensitivity to the number of stories.
Roof displacement and maximum story ductility demands are not very sensitive to
the number of stories, whereas the distribution of story ductility or drift demands
over the height of the structure is rather sensitive to this parameter. The
distribution of story ductility and drift over the height is more uniform for
structures with fewer stories.

•= The generic models used in this study can represent global demands (e.g. roof
displacement demands) for the SAC multi-story frame structures subjected to
near-fault ground motions with good accuracy.

•= The generic model also can be used to estimate story drift demands, provided that
it has the same number of stories as the SAC structure. The representation is
more accurate for the 3-story structure than for the 9-story structure, in which case
the effects of peculiarities caused by subjective design decisions are larger.

Representation of Near-Fault Ground Motions by Equivalent Pulses

Simple pulses of different shapes are utilized to represent near-fault ground motions with
forward directivity. This representation proves to be very efficient for design and an
evaluation of demands. The pulse representation of near-fault ground motions is not
perfect but is believed to be sufficiently accurate for practical purposes. The following
conclusions are drawn:

•= There are clear similarities between the response of frame structures to near-fault
ground motions and the response to pulse-type excitations.

•= Within the approximate period range of 0.375 ≤ T/Tp ≤ 3.0, the salient response
characteristics of near-fault ground motions can be represented by simple
equivalent pulses, which are fully defined by a pulse type, a pulse period, and a
single pulse intensity parameter.

•= The type (shape) of the equivalent pulse for a near-fault record is identified based
on an inspection of time history traces, and on a comparison between ground
motion and pulse spectral shapes. The pulse period is estimated from the location
of a global peak in the velocity response spectrum because the velocity spectra of

Chapter 11 Summary and Conclusions

284
 the pulses investigated have a global hump at T/Tp = 1. This procedure
sometimes requires judgment because near-fault velocity spectra often have more
than one major hump, and a sensitivity analysis may have to be performed to
identify the most relevant hump.

•= The pulse intensity is estimated through an elaborate procedure whose objective is

to minimize the differences between the maximum story ductility demands for the
generic 20-story structure subjected to the near-fault record and the equivalent
pulse. The results indicate that the effective pulse velocity obtained by this
minimization procedure is usually close to the peak ground velocity (PGV) of the
ground motion. Thus, the PGV is a good surrogate for the effective velocity of
the equivalent pulse.

Design Considerations

In the context of seismic design, it is necessary to relate equivalent pulse parameters to

earthquake magnitude and closest distance to the fault rupture plane. Such relationships
are established by means of regression analysis. Since relatively small sets of near-fault
ground motions are available for this purpose, the results of such regression analysis
should be interpreted with caution. Given magnitude and distance, these regression
equations, together with the pulse MDOF strength demand spectra for constant ductility,
can be utilized to develop base shear strength demand spectra for design. Various
distributions of story shear strength over the height of the structure are investigated, and
their advantages and disadvantages are evaluated. As an alternative, strengthening of
frames with walls that are either fixed or hinged at the base is comprehensively
investigated. The conclusions drawn from these efforts are summarized as follows:

•= Preliminary models are developed that relate the pulse period to moment
magnitude, and the pulse intensity to moment magnitude and closest distance to
the fault rupture plane. For a scenario event with Mw = 7 and R = 3 km, a
comparison of the base shear strength demand spectra obtained from the
regression equations with the UBC’97 design spectrum indicates that code-
compliant structures with a medium period will experience excessive ductility
demands.

Chapter 11 Summary and Conclusions

285
 •= Suitable distributions of story shear strength over the height of the structure
depend on the performance objectives and the T/Tp ratio of the structure under
consideration. For long period structures (T/Tp > 1.0) subjected to very severe
ground motions, strengthening of the bottom stories (compared to a standard
SRSS distribution) will significantly reduce the maximum story ductility
demands. For short period structures (T/Tp ≤ 1.0) strengthening of the bottom
stories affects the ductility distribution but has little effect on the maximum story
ductility demand.

•= Strengthening of frame structures with walls that are hinged at the base proves to
be an effective technique that can considerably reduce maximum story drift
demands and cause a more desirable (uniform) distribution of story drifts over the
height of the structure. This technique is effective at various performance levels
and for structures in a wide range of T/Tp ratios, and can provide improved
protection against P-delta effects.

Concluding Remarks

The results of this research shed light on many important response characteristics of near-
fault ground motions, but the conclusions and results presented in this report are
applicable only within the assumptions made in each chapter. It is recognized that much
work remains to be done in order to provide final answers to various aspects of this
complex problem. This study attempts to form a foundation on which to base future
development. There are many issues that were not addressed or need further research.
Examples of such issues are as follows:

1. Evaluation of seismic hazard in the near-fault region of active faults.

2. Development of a systematic procedure for estimating the equivalent pulse type
and period.
3. Improvement of regression equations that relate pulse parameters to magnitude
and distance, as more near-fault ground motions with forward directivity are
recorded, particularly from larger events.
4. Evaluation of P-delta effects for various stability coefficients.
5. Evaluation of alternative strengthening techniques.

Chapter 11 Summary and Conclusions

286
 APPENDIX A

TIME HISTORY TRACES AND ELASTIC SPECTRA

OF NEAR-FAULT GROUND MOTIONS

The set of 22 recorded and 18 simulated near-fault ground motions used in this study is
introduced in Chapter 2. Fifteen of the recorded ground motions, which are utilized in
the response studies, are records with forward directivity. Figure A.1 presents ground
acceleration, velocity, and displacement time histories for the fault-normal component of
these 15 records.

Figure A.2 shows acceleration (elastic strength), velocity, and displacement spectra for
the recorded near-fault ground motions whose fault-normal time histories are presented in
Fig. A.1. Each graph shows the spectra for the fault-normal, fault-parallel, and two 45°
rotated (with respect to the fault direction) components of the ground motion. The
spectral values are computed using a damping ratio of ξ = 2%.

Figures A.3 and A.4 illustrate the magnitude and distance dependence of spectral values
for the simulated records used in this study. These figures show acceleration, velocity,
and displacement spectra for the fault-normal component of the records generated at two
stations f6 and f8. In each graph the spectra for records with various distances (3, 5, and
10 km) are presented.

Appendix A Time History Traces and Elastic Spectra …

287
 LP89lgpc, Fault-Normal LP89lex, Fault-Normal
1000 800
max = 704 max = 673

500 400

ag (cm / sec )
2

0 0

-500 -400

-1000 -800
200 200
max = 173 max = 179

100 100
vg (cm / sec)

0 0

-100 -100

-200 -200
80 60
max = 65.1 max = 56.6

40 30
ug (cm)

0 0

-40 -30

-80 -60
0 4 8 12 16 20 0 2 4 6 8 10 12 14
Time (sec) Time (sec)

EZ92erzi, Fault-Normal LN92lucr, Fault-Normal

600 800
max = 424 max = 700

300 400
ag (cm / sec )
2

0 0

-300 -400

-600 -800
200 150
max = 119 max = 136

100
100
50
vg (cm / sec)

0 0

-50
-100
-100

-200 -150
60 300
max = 42.1 max = 230

200
30
100
ug (cm)

0 0

-100
-30
-200

-60 -300
0 2 4 6 8 10 12 14 0 5 10 15 20 25 30 35 40
Time (sec) Time (sec)

Figure A.1 Ground Acceleration, Velocity, and Displacement Time Histories

Appendix A Time History Traces and Elastic Spectra …

288
 NR94rrs, Fault-Normal NR94sylm, Fault-Normal
1000 800
max = 873 max = 718

500 400

ag (cm / sec )
2

0 0

-500 -400

-1000 -800
200 200
max = 174 max = 122

100 100
vg (cm / sec)

0 0

-100 -100

-200 -200
40 40
max = 38.3 max = 30.7

20 20
ug (cm)

0 0

-20 -20

-40 -40
0 2 4 6 8 10 12 14 0 2 4 6 8 10 12 14
Time (sec) Time (sec)

KB95kobj, Fault-Normal KB95tato, Fault-Normal

1200 800
max = 1067 max = 771

600 400
ag (cm / sec )
2

0 0

-600 -400

-1200 -800
200 200
max = 160 max = 174

100 100
vg (cm / sec)

0 0

-100 -100

-200 -200
40 60
max = 40.1 max = 56.0

20 30
ug (cm)

0 0

-20 -30

-40 -60
0 5 10 15 20 25 0 5 10 15 20 25
Time (sec) Time (sec)

Figure A.1 (Cont’d) Ground Acceleration, Velocity, and Displacement Time Histories

Appendix A Time History Traces and Elastic Spectra …

289
 IV79ar06, Fault-Normal IV79melo, Fault-Normal
600 400
max = 424 max = 372

300 200

ag (cm / sec )
2

0 0

-300 -200

-600 -400
200 200
max = 110 max = 117

100 100
vg (cm / sec)

0 0

-100 -100

-200 -200
60 60
max = 57.9 max = 43.6

30 30
ug (cm)

0 0

-30 -30

-60 -60
0 5 10 15 20 25 0 4 8 12 16 20
Time (sec) Time (sec)

KB95kpi1, Fault-Normal MH84andd, Fault-Normal

600 600
max = 426 max = 436

300 300
ag (cm / sec )
2

0 0

-300 -300

-600 -600
100 40
max = 100 max = 26.6

50 20
vg (cm / sec)

0 0

-50 -20

-100 -40
60 8
max = 49.7 max = 3.72

30 4
ug (cm)

0 0

-30 -4

-60 -8
0 5 10 15 20 25 30 0 2 4 6 8 10 12 14
Time (sec) Time (sec)

Figure A.1 (Cont’d) Ground Acceleration, Velocity, and Displacement Time Histories

Appendix A Time History Traces and Elastic Spectra …

290
 MH84cyld, Fault-Normal NR94newh, Fault-Normal
800 800
max = 712 max = 709

400 400

ag (cm / sec )
2

0 0

-400 -400

-800 -800
80 200
max = 65.3 max = 119

40 100
vg (cm / sec)

0 0

-40 -100

-80 -200
10 40
max = 9.99 max = 34.1

5 20
ug (cm)

0 0

-5 -20

-10 -40
0 2 4 6 8 10 12 14 0 4 8 12 16 20
Time (sec) Time (sec)

NR94spva, Fault-Normal
800
max = 715

400
ag (cm / sec )
2

-400

-800
80
max = 62.6

40
vg (cm / sec)

-40

-80
20
max = 16.0

10
ug (cm)

-10

-20
0 4 8 12 16 20
Time (sec)

Figure A.1 (Cont’d) Ground Acceleration, Velocity, and Displacement Time Histories

Appendix A Time History Traces and Elastic Spectra …

291
 LP89lgpc, ξ = 2% LP89lex, ξ = 2%
5 4
Fault-Normal Fault-Normal
Fault-Parallel Fault-Parallel
4 0.707(FN+FP) 0.707(FN+FP)
3
0.707(FN-FP) 0.707(FN-FP)

3
Sa (g)

2
2

1
1

0 0
600 600
Fault-Normal
Fault-Parallel
500 500
0.707(FN+FP)
0.707(FN-FP)
400 400
Sv (cm / sec)

300 300

200 200
Fault-Normal
Fault-Parallel
100 0.707(FN+FP) 100
0.707(FN-FP)
0 0
250 100
Fault-Normal Fault-Normal
Fault-Parallel Fault-Parallel
200 0.707(FN+FP) 80 0.707(FN+FP)
0.707(FN-FP) 0.707(FN-FP)

150 60
Sd (cm)

100 40

50 20

0 0
0 0.5 1 1.5 2 2.5 3 3.5 4 0 0.5 1 1.5 2 2.5 3 3.5 4
T (sec) T (sec)

EZ92erzi, ξ = 2% LN92lucr, ξ = 2%
2 4
Fault-Normal Fault-Normal
Fault-Parallel Fault-Parallel
0.707(FN+FP) 0.707(FN+FP)
1.5 0.707(FN-FP)
3
0.707(FN-FP)
Sa (g)

1 2

0.5 1

0 0
250 250
Fault-Normal
Fault-Parallel
200 200 0.707(FN+FP)
0.707(FN-FP)
Sv (cm / sec)

150 150

100 100
Fault-Normal

50 Fault-Parallel 50
0.707(FN+FP)
0.707(FN-FP)
0 0
100 160
Fault-Normal Fault-Normal
Fault-Parallel Fault-Parallel
80 0.707(FN+FP) 0.707(FN+FP)
120
0.707(FN-FP) 0.707(FN-FP)

60
Sd (cm)

80
40

40
20

0 0
0 0.5 1 1.5 2 2.5 3 3.5 4 0 0.5 1 1.5 2 2.5 3 3.5 4
T (sec) T (sec)

Figure A.2 Elastic Acceleration, Velocity, and Displacement Spectra, Recorded

Appendix A Time History Traces and Elastic Spectra …

292
 NR94rrs, ξ = 2% NR94sylm, ξ = 2%
3 4
Fault-Normal Fault-Normal
Fault-Parallel Fault-Parallel
2.5
0.707(FN+FP) 0.707(FN+FP)
3
0.707(FN-FP) 0.707(FN-FP)
2
Sa (g)

1.5 2

1
1
0.5

0 0
400 400
Fault-Normal Fault-Normal
Fault-Parallel Fault-Parallel
0.707(FN+FP) 0.707(FN+FP)
300 300
0.707(FN-FP) 0.707(FN-FP)
Sv (cm / sec)

200 200

100 100

0 0
100 120
Fault-Normal Fault-Normal
Fault-Parallel Fault-Parallel
100
80 0.707(FN+FP) 0.707(FN+FP)
0.707(FN-FP) 0.707(FN-FP)
80
60
Sd (cm)

60
40
40

20
20

0 0
0 0.5 1 1.5 2 2.5 3 3.5 4 0 0.5 1 1.5 2 2.5 3 3.5 4
T (sec) T (sec)

KB95kobj, ξ = 2% KB95tato, ξ = 2%
5 4
Fault-Normal Fault-Normal
Fault-Parallel Fault-Parallel
4 0.707(FN+FP) 0.707(FN+FP)
3
0.707(FN-FP) 0.707(FN-FP)

3
Sa (g)

2
2

1
1

0 0
800 800
Fault-Normal Fault-Normal
Fault-Parallel Fault-Parallel
0.707(FN+FP) 0.707(FN+FP)
600 600
0.707(FN-FP) 0.707(FN-FP)
Sv (cm / sec)

400 400

200 200

0 0
120 200
Fault-Normal Fault-Normal
Fault-Parallel Fault-Parallel
100
0.707(FN+FP) 0.707(FN+FP)
150
0.707(FN-FP) 0.707(FN-FP)
80
Sd (cm)

60 100

40
50
20

0 0
0 0.5 1 1.5 2 2.5 3 3.5 4 0 0.5 1 1.5 2 2.5 3 3.5 4
T (sec) T (sec)

Figure A.2 (Cont’d) Elastic Acceleration, Velocity, and Displacement Spectra, Recorded

Appendix A Time History Traces and Elastic Spectra …

293
 IV79ar06, ξ = 2% IV79melo, ξ = 2%
2 1.2
Fault-Normal Fault-Normal
Fault-Parallel Fault-Parallel
1
0.707(FN+FP) 0.707(FN+FP)
1.5
0.707(FN-FP) 0.707(FN-FP)
0.8
Sa (g)

1 0.6

0.4
0.5
0.2

0 0
400 300
Fault-Normal Fault-Normal
Fault-Parallel Fault-Parallel
250
0.707(FN+FP) 0.707(FN+FP)
300
0.707(FN-FP) 0.707(FN-FP)
200
Sv (cm / sec)

200 150

100
100
50

0 0
200 160
Fault-Normal Fault-Normal
Fault-Parallel Fault-Parallel
0.707(FN+FP) 0.707(FN+FP)
150 120
0.707(FN-FP) 0.707(FN-FP)
Sd (cm)

100 80

50 40

0 0
0 0.5 1 1.5 2 2.5 3 3.5 4 0 0.5 1 1.5 2 2.5 3 3.5 4
T (sec) T (sec)

KB95kpi1, ξ = 2% MH84andd, ξ = 2%
1.6 1.6
Fault-Normal Fault-Normal
Fault-Parallel Fault-Parallel
0.707(FN+FP) 0.707(FN+FP)
1.2 1.2
0.707(FN-FP) 0.707(FN-FP)
Sa (g)

0.8 0.8

0.4 0.4

0 0
300 120
Fault-Normal Fault-Normal
Fault-Parallel Fault-Parallel
250 100
0.707(FN+FP) 0.707(FN+FP)
0.707(FN-FP) 0.707(FN-FP)
200 80
Sv (cm / sec)

150 60

100 40

50 20

0 0
120 20
Fault-Normal Fault-Normal
Fault-Parallel Fault-Parallel
100
0.707(FN+FP) 0.707(FN+FP)
15
0.707(FN-FP) 0.707(FN-FP)
80
Sd (cm)

60 10
d

40
5
20

0 0
0 0.5 1 1.5 2 2.5 3 3.5 4 0 0.5 1 1.5 2 2.5 3 3.5 4
T (sec) T (sec)

Figure A.2 (Cont’d) Elastic Acceleration, Velocity, and Displacement Spectra, Recorded

Appendix A Time History Traces and Elastic Spectra …

294
 MH84cyld, ξ = 2% NR94newh, ξ = 2%
3 4
Fault-Normal Fault-Normal
Fault-Parallel Fault-Parallel
2.5
0.707(FN+FP) 0.707(FN+FP)
3
0.707(FN-FP) 0.707(FN-FP)
2
Sa (g)

1.5 2

1
1
0.5

0 0
400 400
Fault-Normal Fault-Normal
Fault-Parallel Fault-Parallel
0.707(FN+FP) 0.707(FN+FP)
300 300
0.707(FN-FP) 0.707(FN-FP)
Sv (cm / sec)

200 200

100 100

0 0
50 80
Fault-Normal Fault-Normal
Fault-Parallel Fault-Parallel
40 0.707(FN+FP) 0.707(FN+FP)
60
0.707(FN-FP) 0.707(FN-FP)

30
Sd (cm)

40
20

20
10

0 0
0 0.5 1 1.5 2 2.5 3 3.5 4 0 0.5 1 1.5 2 2.5 3 3.5 4
T (sec) T (sec)

NR94spva, ξ = 2%
4
Fault-Normal
Fault-Parallel
0.707(FN+FP)
3
0.707(FN-FP)
Sa (g)

0
300
Fault-Normal
Fault-Parallel
250
0.707(FN+FP)
0.707(FN-FP)
200
Sv (cm / sec)

150

100

50

0
80
Fault-Normal
Fault-Parallel
0.707(FN+FP)
60
0.707(FN-FP)
Sd (cm)

40

20

0
0 0.5 1 1.5 2 2.5 3 3.5 4
T (sec)

Figure A.2 (Cont’d) Elastic Acceleration, Velocity, and Displacement Spectra, Recorded

Appendix A Time History Traces and Elastic Spectra …

295
 Simulations, Mw = 6.5, Fault-Normal, Station f6, ξ = 2% Simulations, Mw = 7.0, Fault-Normal, Station f6, ξ = 2%
3
2
R = 3 km
R = 3 km
R = 5 km
R = 5 km 2.5
R = 10 km
R = 10 km
1.5
Sa (g) 2

1 1.5

1
0.5
0.5

0 0
300 400
R = 3 km R = 3 km
R = 5 km R = 5 km
250
R = 10 km R = 10 km
300
200
Sv (cm / sec)

150 200

100
100
50

0 0
80 250
R = 3 km R = 3 km
R = 5 km R = 5 km
R = 10 km 200 R = 10 km
60

150
Sd (cm)

40
100

20
50

0 0
0 0.5 1 1.5 2 2.5 3 3.5 4 0 0.5 1 1.5 2 2.5 3 3.5 4
T (sec) T (sec)

Simulations, Mw = 7.5, Fault-Normal, Station f6, ξ = 2%

4
R = 3 km
R = 5 km
R = 10 km
3
Sa (g)

0
1000
R = 3 km
R = 5 km
800 R = 10 km
Sv (cm / sec)

600

400

200

0
500
R = 3 km
R = 5 km
400 R = 10 km

300
Sd (cm)

200

100

0
0 0.5 1 1.5 2 2.5 3 3.5 4
T (sec)

Figure A.3 Elastic Acceleration, Velocity, and Displacement Spectra, Simulated f6

Appendix A Time History Traces and Elastic Spectra …

296
 Simulations, Mw = 6.5, Fault-Normal, Station f8, ξ = 2% Simulations, Mw = 7.0, Fault-Normal, Station f8, ξ = 2%
2 3
R = 3 km R = 3 km
R = 5 km R = 5 km
2.5
R = 10 km R = 10 km
1.5
2
Sa (g)

1 1.5

1
0.5
0.5

0 0
300 400
R = 3 km R = 3 km
R = 5 km R = 5 km
250
R = 10 km R = 10 km
300
200
Sv (cm / sec)

150 200

100
100
50

0 0
80 250
R = 3 km R = 3 km
R = 5 km R = 5 km
R = 10 km 200 R = 10 km
60

150
Sd (cm)

40
100

20
50

0 0
0 0.5 1 1.5 2 2.5 3 3.5 4 0 0.5 1 1.5 2 2.5 3 3.5 4
T (sec) T (sec)

Simulations, Mw = 7.5, Fault-Normal, Station f8, ξ = 2%

4
R = 3 km
R = 5 km
R = 10 km
3
Sa (g)

0
1000
R = 3 km
R = 5 km
800 R = 10 km
Sv (cm / sec)

600

400

200

0
500
R = 3 km
R = 5 km
400 R = 10 km

300
Sd (cm)

200

100

0
0 0.5 1 1.5 2 2.5 3 3.5 4
T (sec)

Figure A.4 Elastic Acceleration, Velocity, and Displacement Spectra, Simulated f8

Appendix A Time History Traces and Elastic Spectra …

297
 LIST OF REFERENCES

Abrahamson, N., and Silva, W., (1997). “Empirical Response Spectral Attenuation
Relations for Shallow Crustal Earthquakes,” Seismological Research Letters, Vol. 68,
No.1, 94-126.

ACI 318-99 (1999). “Building Code Requirements for Structural Concrete and
Commentary,” American Concrete Institute, Farmington Hills, Michigan, 1999.

Anderson, J., and Bertero, V., (1978). “Uncertainties in Establishing Design

Earthquakes,” ASCE Journal of Structural Engineering, Vol. 113, No. 8, Aug. 1987,
1709-1724.

Anderson, J., Bertero, V., and Bertero, R., (1999). “Performance Improvement of Long
Period Building Structures Subjected to Severe Pulse-Type Ground Motions,” Pacific
Earthquake Engineering Research Center, University of California, Berkeley, Report No.
PEER-1999/09, Oct. 1999.

Baez, J., and Miranda, E., (2000). “Amplification Factors to Estimate Inelastic
Displacement Demands for the Design of Structures in the Near Field,” Proceedings,
12th World Conference on Earthquake Engineering, New Zealand.

Benioff, H., (1955). “Mechanism and Strain Characteristics of the White Wolf Fault as
Indicated by the Aftershock Sequence, Earthquakes in Kern County, California, During
1955 (G.B. Oakeshotte, ed.),” California Division of Mines Bulletin, No. 171, 199-202.

List of References
298
Bertero, V., Mahin, S., and Herrera, R., (1978). “Aseismic Design Implications of Near-
Fault San Fernando Earthquake Records,” Earthquake Engineering and Structural
Dynamics, Vol. 6, Jan.-Feb. 1978, 31-42.

Biggs, J., (1964). “Introduction to Structural Dynamics,” McGraw-Hill, Inc., New York.
Chopra, A., (1995). “Dynamics of Structures: Theory and Applications to Earthquake
Engineering,” Prentice-Hall, Inc., Englewood Cliffs, New Jersey, 131-151.

Graves, R. (1996). “Modification of Lucerne Time History of the 1992 Landers

Earthquake to Include Geodetically Defined Static Displacements,” Unpublished
Manuscript.

Gupta, A., and Krawinkler, H., (1999). “Seismic Demands for Performance Evaluation of
Steel Moment Resisting Frame Structures,” John A. Blume Earthquake Engineering
Center, Report No. 132, Department of Civil Engineering, Stanford University.

Gupta, A., and Krawinkler, H., (2000). “Dynamic P-delta Effects for Flexible Inelastic
Steel Structures,” ASCE Journal of Structural Engineering, Vol. 126, No. 1, Jan. 2000,
145-154.

Hall, J., Heaton, T., Halling, M., Wald, D., (1995). “Near-Source Ground Motion and its
Effects on Flexible Buildings,” Earthquake Spectra, Vol. 11, No. 4, Nov. 1995, 569-605.

Iwan, W., and Chen, X., (1994). “Important Near-Field Ground Motion Data from the
Landers Earthquake,” Proceedings, 10th European Conference on Earthquake
Engineering, Vienna.

Iwan, W., (1997). “Drift Spectrum: Measure of Demand for Earthquake Ground
Motions,” ASCE Journal of Structural Engineering, Vol. 123, No. 4, Apr. 1997, 397-404.

Krawinkler, H., and Gupta, A., (1998). “Story Drift Demands for Steel Moment Frame
Structures in Different Seismic Regions,” Proceedings, Sixth U.S. National Conference
on Earthquake Engineering, Seattle, Washington.

Lawson, R., (1996). “Site-Dependent Inelastic Seismic Demands,” Ph.D. Dissertation,

Department of Civil Engineering, Stanford University.

MacRae, G., Morrow, D., and Roeder, C., (1998). “Near-Field Ground Motion Effects on
Short Structures,” PG&E-PEER Program, University of Washington, Draft Report, Aug.
1998.

Mahin, S., Bertero, V., Chopra, A., and Collins, R., (1976). “Response of the Olive View
Hospital Main Building During the San Fernando Earthquake,” Earthquake Engineering
Research Center, University of California, Berkeley, Report [Link]/EERC-76/22, Oct.
1976.

List of References
299
Makris, N., and Chang, S., (2000). “Effect of Viscous, Viscoplastic and Friction
Damping on the Response of Seismic Isolated Structures,” Earthquake Engineering &
Structural Dynamics, Vol. 29, No.1, Jan. 2000, 85-107.

Nassar, A., and Krawinkler, H., (1991). “Seismic Demands for SDOF and MDOF
systems,” John A. Blume Earthquake Engineering Center, Report No. 95, Department of
Civil Engineering, Stanford University.

NEHRP (1994). “National Earthquake Hazards Reduction Program Recommended

Provisions for Seismic Regulations for New Buildings,” Federal Emergency
Management Agency 223, Washington, DC, 1994.

Prakash, V., Powell, G.H., and Campbell, S., (1993). “DRAIN-2DX Base Program
Description and User Guide, Version 1.10”, Report No. UCB/SEMM-93/17, Department
of Civil Engineering, University of California at Berkeley.

Rahnama, M., and Krawinkler, H., (1993). “Effects of Soft Soils and Hysteresis Model
on Seismic Demands,” John A. Blume Earthquake Engineering Center, Report No. 108,
Department of Civil Engineering, Stanford University.

Ramsey, F., and Schafer, D., (1996). “The Statistical Sleuth: A Course in Methods of
Data Analysis,” Duxbury Press, Belmont, California,181-186.

SEAOC Blue Book (1996). “Recommended Lateral Force Requirements and

Commentary,” Structural Engineers Association of California, sixth edition, Sacramento,
California, 1996.

Seneviratna, G.D.P.K., and Krawinkler, H., (1997). “Evaluation of Inelastic MDOF

Effects for Seismic Design,” John A. Blume Earthquake Engineering Center, Report No.
120, Department of Civil Engineering, Stanford University.

Silva, W., (1999). “Strong Motion Database,” CD-ROM, Pacific Engineering and
Analysis (also available at [Link]

Singh, J., (1985). “Earthquake Ground Motions: Implications for Designing Structures
and Reconciling Structural Damage,” Earthquake Spectra, Feb. 1985, 239-270.

Somerville, P., Saikia C.K., Wald, D., and Graves, R. (1996). “Implications of the
Northridge Earthquake for Strong Ground Motions from Thrust Faults,” Bulletin of the
Seismological Society of America, 86, S115-S125.

Somerville, P., Smith N., Punyamurthula, S., and Sun, J. (1997a). “Development of
Ground Motion Time Histories for Phase 2 of the FEMA/SAC Steel Project,” Report No.
SAC/BD-97-04.

List of References
300
Somerville, P., Smith, N., Graves, R., and Abrahamson, N., (1997b). “Modification of
Empirical Strong Ground Motion Attenuation Relations to Include the Amplitude and
Duration Effects of Rupture Directivity,” Seismological Research Letters, Vol. 68, No. 1,
180-203.

Somerville, P., (1998). “Development of an Improved Ground Motion Representation for

Near Fault Ground Motions,” SMIP98 Seminar on Utilization of Strong-Motion Data,
Oakland, CA, Sept. 1998.

Somerville, P., (2000). “Characterization of Near Fault Ground Motions,” US-Japan

Workshop on Effects of Near-Field Earthquake Shaking, San Francisco, CA, March
2000.

Uniform Building Code (1997). International Conference of Building Officials, Whittier,

California.

Veletsos, A. and Newmark, N., and Chelapati, C., (1965). “Deformation Spectra for
Elastic and Elastoplastic Systems Subjected to Ground Shock and Earthquake Motions,”
Proceeding, Third World Conference on Earthquake Engineering, Vol. II, New Zealand.

Wald, D., and Heaton, T., (1994). “Spatial and Temporal Distribution of Slip for the 1992
Landers, California, Earthquake,” Bulletin of the Seismological Society of America, 84,
668-691.

List of References
301