Introduction to Earthquake

Engineering
Introduction to Earthquake
Engineering

Hector Estrada and Luke S. Lee

MATLAB ® is a trademark of The MathWorks, Inc. and is used with permission. The MathWorks does not warrant the
accuracy of the text or exercises in this book. This book’s use or discussion of MATLAB ® software or related products
does not constitute endorsement or sponsorship by The MathWorks of a particular pedagogical approach or particu-
lar use of the MATLAB ® software.

CRC Press
Taylor & Francis Group
6000 Broken Sound Parkway NW, Suite 300
Boca Raton, FL 33487-2742

© 2017 by Taylor & Francis Group, LLC

CRC Press is an imprint of Taylor & Francis Group, an Informa business

No claim to original U.S. Government works

Printed on acid-free paper

International Standard Book Number-13: 978-1-4987-5826-0 (Hardback)

This book contains information obtained from authentic and highly regarded sources. Reasonable efforts have
been made to publish reliable data and information, but the author and publisher cannot assume responsibility for
the v­ alidity of all materials or the consequences of their use. The authors and publishers have attempted to trace
the ­copyright holders of all material reproduced in this publication and apologize to copyright holders if permission
to publish in this form has not been obtained. If any copyright material has not been acknowledged please write and
let us know so we may rectify in any future reprint.

Except as permitted under U.S. Copyright Law, no part of this book may be reprinted, reproduced, transmitted, or
utilized in any form by any electronic, mechanical, or other means, now known or hereafter invented, including
­photocopying, microfilming, and recording, or in any information storage or retrieval system, without written per-
mission from the publishers.

For permission to photocopy or use material electronically from this work, please access www.copyright.com (http://
www.copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC), 222 Rosewood Drive, Danvers, MA
01923, 978-750-8400. CCC is a not-for-profit organization that provides licenses and registration for a variety of users.
For organizations that have been granted a photocopy license by the CCC, a separate system of payment has been
arranged.

Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only
for identification and explanation without intent to infringe.

Library of Congress Cataloging‑in‑Publication Data

Names: Estrada, Hector, author. | Lee, Luke S., author.

Title: Introduction to earthquake engineering / Hector Estrada and Luke S.
Lee.
Description: Boca Raton : Taylor & Francis, CRC Press, 2017. | Includes
bibliographical references and index.
Identifiers: LCCN 2017018885| ISBN 9781498758260 (hardback : acid-free paper)
| ISBN 9781315171579 (ebook)
Subjects: LCSH: Earthquake engineering.
Classification: LCC TA654.6 .E848 2017 | DDC 624.1/762--dc23
LC record available at https://lccn.loc.gov/2017018885

Visit the Taylor & Francis Web site at

http://www.taylorandfrancis.com
and the CRC Press Web site at
http://www.crcpress.com
Contents
Preface...............................................................................................................................................ix
Acknowledgements............................................................................................................................xi
Authors............................................................................................................................................ xiii

Chapter 1 Introduction...................................................................................................................1
1.1 Structural Effects of Earthquakes......................................................................2
1.1.1 Ground Failures..................................................................................... 2
1.1.2 Indirect Effects of Earthquakes............................................................5
1.1.3 Ground Shaking....................................................................................7
1.2 Types of Earthquakes.........................................................................................8
1.2.1 Man-Made Earthquakes........................................................................ 9
1.2.2 Volcanic Earthquakes.......................................................................... 11
1.2.3 Tectonic Earthquakes.......................................................................... 11
1.3 History of the Development of Mitigation Strategies for the Effects of
Seismic Hazards............................................................................................... 14
References................................................................................................................... 17

Chapter 2 Engineering Seismology Overview............................................................................. 19

2.1 Seismology Terminology.................................................................................. 19
2.1.1 Elastic Rebound Theory: Tectonic Plate Movement........................... 21
2.1.2 Seismic Waves..................................................................................... 21
2.2 Measuring Earthquakes.................................................................................... 23
2.2.1 Earthquake Intensity........................................................................... 23
2.2.2 Earthquake Magnitude........................................................................25
2.3 Effects of Earthquakes on Structures............................................................... 29
2.3.1 Earthquake Characteristics................................................................. 29
2.3.2 Site Characteristics.............................................................................. 31
2.3.3 Structural Characteristics.................................................................... 32
2.4 Earthquake Hazard Assessment.......................................................................34
2.5 Earthquake Prediction...................................................................................... 36
Problems...................................................................................................................... 36
References................................................................................................................... 36

Chapter 3 Single-Degree-of-Freedom Structural Dynamic Analysis.......................................... 37

3.1 Undamped Single-Degree-of-Freedom System............................................... 37
3.1.1 Free Vibration Response of Undamped Systems................................ 38
3.1.2 Structural Weight................................................................................ 42
3.1.3 Structural Stiffness..............................................................................46
3.1.4 Approximate Structural Period...........................................................60
3.1.5 Time-Dependent Forced Undamped Vibration Response.................. 61
3.2 Damped Single-Degree-of-Freedom System...................................................64
3.2.1 Free Vibration Response of Damped Systems.................................... 65
3.2.2 Structural Damping............................................................................. 71
3.2.3 Time-Dependent Forced Damped Vibration Response...................... 73
3.2.4 Time-Dependent Support Accelerations............................................. 78

v
vi Contents

3.3 Base Shears and Stresses Caused by Time-Dependent Forces and

Support Excitations.......................................................................................... 83
Problems...................................................................................................................... 86
References...................................................................................................................90

Chapter 4 Response to General Loading..................................................................................... 91

4.1 Response of an SDOF System to an Impulse................................................... 91
4.2 General Forcing Function.................................................................................94
4.3 Shock Spectra................................................................................................. 102
4.4 Response to Ground Motion........................................................................... 111
4.5 Direct Integration Methods............................................................................ 114
4.5.1 Nigam−Jennings Algorithm (Explicit).............................................. 114
4.5.2 Central Difference Method (Explicit)............................................... 117
4.5.3 Newmark’s Beta Method for Linear Systems (Implicit)................... 120
Problems.................................................................................................................... 121
Reference................................................................................................................... 122

Chapter 5 Response Spectrum Analysis of SDOF System........................................................ 123

5.1 Elastic Response Spectrum............................................................................ 123
5.1.1 Elastic Design Response Spectrum................................................... 132
5.2 Inelastic Response Spectrum.......................................................................... 140
5.2.1 Inelastic Design Response Spectrum................................................ 143
Problems.................................................................................................................... 146
References................................................................................................................. 149

Chapter 6 Generalized SDOF System Analysis........................................................................ 151

6.1 Discrete System (Shear Buildings)................................................................. 152
6.1.1 Equation of Motion........................................................................... 152
6.1.2 Generalized Participation Factor, Frequency, and Natural
Period.......................................................................................... 157
6.1.3 Deflections, Base Shear, and Moments Using Participation
Factors and Response Spectra........................................................... 157
6.2 Generalized SDOF Continuous System......................................................... 162
6.2.1 Equation of Motion........................................................................... 162
6.2.2 Generalized Participation Factor, Frequency, and Natural
Period.......................................................................................... 163
6.2.3 Deflections, Base Shear, and Moments Using Participation
Factors and Response Spectra........................................................... 163
Problems.................................................................................................................... 167
References................................................................................................................. 168

Chapter 7 Multi-Degree-of-Freedom System Analysis............................................................. 169

7.1 Equations of Motion for an MDOF System................................................... 171
7.1.1 Periods and Mode Shapes for an MDOF System.............................. 172
7.1.2 Participation Factors for an MDOF System...................................... 180
7.1.3 Deflections, Base Shear, and Moments Using Participation
Factors and Response Spectra........................................................... 182
Contents vii

7.2 Response Spectrum Analysis Method Summary........................................... 188

Problems.................................................................................................................... 199
References................................................................................................................. 201

Chapter 8 Seismic Code Provisions........................................................................................... 203

8.1 Structural Design Philosophies...................................................................... 203
8.1.1 Allowable Stress Design....................................................................204
8.1.2 Load and Resistance Factor Design..................................................206
8.1.3 Allowable Seismic Force-Resisting System......................................206
8.2 Overview of Seismic Design Codes...............................................................207
8.3 Seismic Load Combinations...........................................................................208
8.3.1 Redundancy Factor, ρ........................................................................ 210
8.3.2 Overstrength Factor Ωo and other Seismic Force-Resisting
System Parameters............................................................................ 211
8.4 Overview of Seismic Load Analysis Procedures........................................... 213
8.4.1 Design Response Spectrum............................................................... 213
8.4.2 Permitted Lateral Analysis Procedures............................................. 219
8.5 Earthquake Loads Based on ASCE-7 Equivalent Lateral Force
Procedure���������������������������������������������������������������������������������������������������220
8.6 Modal Response Spectrum Analysis.............................................................. 229
8.7 Introduction to Seismic Response History Procedures..................................240
Problems.................................................................................................................... 241
References.................................................................................................................246
Index............................................................................................................................................... 247
Preface
Earthquake engineering is an important part of structural engineering. The primary purpose
of Introduction to Earthquake Engineering is to provide students and practitioners with the
­fundamental and practical knowledge of how earthquake engineering is used in the design of lateral
force-resisting systems for buildings and other structures. Special emphasis is placed on obtaining
seismic loads using the most recent code specifications.
The topics and illustrations presented in this text have been carefully chosen for students in
courses dealing with seismic loads as part of an architectural, construction, or civil engineering-
related curriculum: both at the undergraduate and graduate levels. An important and unique
feature of Introduction to Earthquake Engineering is that it ties together the relationship and
interdependence between structural dynamics and seismic loads based on code specifications.
With this in mind, many practical examples are provided throughout the text. Many solutions
to example problems are prepared using computer programing methods (i.e., MATLAB®) to
enable students, and practitioners, to solve complex dynamics problems. Also, considerable
content is available to acquaint students with specifications and reference information found
in the American Society of Civil Engineers’ Minimum Design Loads for Buildings and Other
Structures.
Examples of typical seismic load calculations presented in this text demonstrate the type
of routine technical mathematics an entry-level structural engineer can expect to encounter on
the job. Because Introduction to Earthquake Engineering is first and foremost an introductory
textbook, most complex nonlinear analysis theory has been purposely avoided. However, the
textbook does assume that the reader has had some exposure to rigid body dynamics, differ-
ential equations, and structural analysis coursework. The 2010 edition of the Minimum Design
Loads for Buildings and Other Structures (ASCE-7) is the source of all design code-based
specifications.
This book provides broad coverage of the subject of earthquake engineering, emphasizing the
process of establishing seismic loads for buildings and other structures. To that end, the reader is
introduced to a wide variety of practical dynamics and seismic loading examples and end-of-­chapter
problems. Also, to guide the reader, each chapter begins with a list of learning outcomes. After
carefully studying this textbook, a reader should be able to: (i) demonstrate a basic understanding
of earthquake engineering, (ii) identify and solve fundamental structural dynamics problems, and
(iii) perform equivalent static and dynamic seismic analyses of simple structural systems. The first
objective relates to the student’s ability to operate in the first two cognitive domains of Bloom’s
Taxonomy (viz., the knowledge of seismic engineering terms and the comprehension of the overall
area of earthquake engineering). The second and third objectives primarily concentrate on the next
two cognitive domains (viz., the application of structural dynamics in earthquake engineering and
the analysis of systems subjected to earthquake loading).
Image galleries, PowerPoint presentations, a solutions manual, etc., will be hosted on a
companion website. Visit the book’s CRC Press website for further details:
https://www.crcpress.com/Introduction-to-Earthquake-Engineering/Estrada-Lee/p/
book/9781498758260

Hector Estrada
Luke S. Lee

ix
x Preface

MATLAB® is a registered trademark of The MathWorks, Inc. For product information, please contact:

The MathWorks, Inc.

3 Apple Hill Drive
Natick, MA 01760-2098 USA
Tel: 508-647-7000
Fax: 508-647-7001
E-mail: [email protected]
Web: www.mathworks.com
Acknowledgements
We are grateful to the many students at University of the Pacific whose suggestions and ­reactions
to our earthquake engineering classes have driven the need for us to refine our presentation of
­earthquake engineering as simply and clearly as possible. Finally, we are indebted to Wendy
Estrada for her proofreading and extensive checking of details: both technical and stylistic, and
most ­importantly for all of her exhaustive support during the writing process.

xi
Authors
Hector Estrada, PhD, PE, is currently a professor of civil engineering at University of the
Pacific (Pacific), where he also served as chair of the Civil Engineering Department. Prior to join-
ing Pacific, Dr. Estrada was chair of the Department of Civil and Architectural Engineering at
Texas A&M University-Kingsville. Dr. Estrada has also had visiting research appointments at
NASA, Texas Tech University, and the U.S. Army Engineer Research and Development Center—
Construction Engineering Research Laboratory. His teaching interests include structural engineer-
ing and mechanics, the design of concrete, steel, and timber structures, structural dynamics, and
earthquake engineering. He has published on structural engineering and engineering education in
various peer-reviewed journals, conference proceedings, and presented research work at various
technical conferences. He has published a book on drafting and design of structural steel b­ uildings,
and several book chapters. He has served as a reviewer for a number of journals, conferences,
book publishers, and funding agencies. Dr. Estrada earned his BS (with honors), MS, and PhD all
in Civil Engineering from the University of Illinois at Urbana-Champaign.

Luke S. Lee, PhD, PE, is currently an associate professor of civil engineering at University
of the  Pacific, where he teaches courses in solid mechanics, structural dynamics and health
­monitoring,  structural design, and engineering risk analysis. He has authored and co-authored
numerous journal articles, conference papers, and book chapters in composite materials, structural
health m ­ onitoring, service life estimation, and engineering pedagogy. Dr. Lee earned his BS in
civil ­engineering from  the University of California, Los Angeles, MS in civil engineering from
the University of California, Berkeley, and PhD in structural engineering from the University of
California, San Diego.

xiii
 1 Introduction

After reading this chapter, you will be able to:

1. Explain the general effects of earthquakes on the built environment

2. Describe the main source of earthquakes
3. List the different types of earthquakes
4. Explain tectonic plate movement

“If a tree falls in a forest and no one is around to hear it, does it make a sound?” Since sound
is the sensation excited in the ear when air is set in motion, one could argue that the tree makes
no sound. Likewise, if an earthquake occurs and no one, or no infrastructure system, is there to
experience the resulting motion, we can conclude that the most severe earthquake would cause no
damage. Before humans appeared on earth, numerous earthquakes occurred, as evidenced by the
large shift in the location of the continents with no damage (this will be discussed in more detail
in Section 1.2.3). So we could argue that earthquakes do not cause damage; rather, it is infrastruc-
ture that experiences damage because of an earthquake. That is, the devastating effect of earth-
quakes should be attributed to an inferior built environment. Thus, when infrastructure is designed
to account for the effects of earthquakes, even the largest earthquakes cause little or no damage!
Nevertheless, because of the historical effects of earthquakes on human civilization, they are the
most terrifying of all natural disasters. The principle reason being that they can occur with little or
no notice and can devastate large areas.
This book is an introduction to the essentials of seismic design. The primary focus is on estab-
lishing the forces and deformations experienced by structural systems found in infrastructure. We
begin by providing a broad overview of earthquakes, their effects on structures, sources, and types
(Chapter 1), followed by an introduction to seismology in Chapter 2, the material for which was
selected based on seismology material covered in the California seismic portion of the Professional
Engineering (PE) exam. The book then covers the prerequisite structural dynamics material (based
on single-degree-of-freedom [SDOF] dynamic analyses) needed to develop the ­ fundamental
relationships between structures and ground motion produced by earthquakes (Chapter 3).
­
The material being primarily focused on the development of parameters that are based on free and
forced, damped and un-damped vibrations, with the aim of developing parameters to characterize
­structural system mass, stiffness, and damping, which in turn are used to calculate periods, frequen-
cies, and other relevant vibration properties.
We also explain the computational processes used to develop a critical concept (Chapter 4),
which is the establishment of seismic forces for structures at a particular location: the response
spectrum based on the Newmark–Hall method—both elastic and inelastic spectra (Chapter 5).
Next, we connect the analysis of multi-degree-of-freedom (MDOF) building systems (or multi-
story shear b­ uildings) to fundamental SDOF principles through the development of the general-
ized SDOF equations (Chapter 6). This generalized SDOF formulation can be combined with the
­principles associated with SDOF analyses, including response spectrum analysis, to analyze a
­multistory building as an SDOF system, including finding maximum shear and bending moment in
all members due to a seismic load. The process entails computing approximate participation factors
and using a design response spectrum to determine the story displacements, which are then used to
obtain lateral story seismic forces that are necessary to perform a complete analysis for the internal
loading of a structural system.

1
2 Introduction to Earthquake Engineering

Following the response spectrum analysis of MDOF systems subjected to seismic ground motion
in Chapter 7, we examine code-based analyses—both equivalent-static and dynamic approaches in
Chapter 8. The dynamic analysis will emphasize modal response spectrum analysis. The main goal
of the book is to elucidate code-based formulations such that a practicing engineer or an engineer-
ing student can apply these formulations, and explain how they were developed and describe the
­limitations of these code-based procedures.
The book is intended as a tool for a one-semester course in introduction to earthquake engineer-
ing at the undergraduate or graduate levels, both in civil or architectural engineering.
The material presented in the book is concise and includes both essential structural ­dynamics
and seismic engineering principles, the former being prerequisite to fully grasping the funda-
mental concepts covered in the latter. Thus, students who have a limited background in structural
­dynamics still have the opportunity to cover the essentials of earthquake engineering in a one-
semester course. This is particularly important in an era where information is readily available and
can be overwhelming for a new learner. For completeness, we also include computational tools that
can be used to solve the complex mathematics involved in the solution of even the simplest seismic
­engineering problems.
At the beginning of each chapter, we present a list of learning outcomes in an effort to improve
the delivery of the content and to guide the learner in mastering the material presented. Overall, the
book adheres to the following general learning outcomes:

1. Demonstrate a basic understanding of earthquake engineering

2. Identify and solve basic structural dynamics problems
3. Perform basic equivalent static and dynamic seismic analyses of simple structural systems

The first objective relates to the student’s ability to operate in the first two cognitive domains
of Bloom’s Taxonomy (namely, the knowledge of seismic engineering terms and comprehension of
the overall area of earthquake engineering). The second and third objectives primarily concentrate
on the next two cognitive domains (namely, the application of structural dynamics in earthquake
engineering and the analysis of systems subjected to earthquake loading).

1.1  STRUCTURAL EFFECTS OF EARTHQUAKES

Generally, effects of earthquakes can be divided into several categories. In this section, we present
three categories: ground failure, indirect effects, and ground shaking, the latter being the main topic
covered in the book. Ground failure can be further subdivided into surface rupture, ground subsid-
ence, ground cracking, soil liquefaction, and landslides. Indirect effects include tsunamis (tidal
waves in the oceans), seiches (tidal waves in an enclosed body of water, such as a lake), and fires, all
of which can result in damage comparable to, or even larger than, ground shaking.

1.1.1  Ground Failures

Ground failures are generally considered part of geotechnical earthquake engineering, and they
involve the movement of the ground surface at a location where geological fissures or zones of
weakness in the crust of the earth (faults) slip slowly or suddenly. Usually, sudden movement is
much more damaging because of the associated ground shaking. The terminology and the process
of how earthquakes occur will be discussed in more detail in the rest of this chapter. In this section,
we present an overview of the most severe effects of ground failure on infrastructure.
Surface faulting occurs when the relative movement of rocks on the two sides of a fault takes
place deep within the earth and breaks through to the surface; this can occur as slow movement
in the form of fault creep or suddenly resulting in an earthquake. This type of ground failure
typically follows a preexisting fault line. Figure 1.1a shows the slow movement of the San Andreas
Fault inside the DeRose Winery in the Cienega Valley near Hollister, California. This location
Introduction 3

(a) (b)

1992 Landers earthquake in San Bernardino County

San Andreas Fault creep

in Cienega Valley, inside
the DeRose Winery

FIGURE 1.1  Surface rupture: (a) slow (creep) and (b) caused by the 1992 Landers earthquake in San
Bernardino County. (http://www.conservation.ca.gov/cgs/rghm/ap/Pages/main.aspx)

experienced over 16 inches (400 mm) of fault creep over a 40-year period as indicated by a graph of
fault slip displayed at the winery. Since the winery building straddles the surface of the fault, this
­movement has been splitting it into two, which is clearly evident by the large crack on the concrete
floor. The San Andreas Fault, however, is capable of producing large earthquakes near this site,
such as the 1989 Loma Prieta earthquake, which caused widespread damage throughout the San
Francisco Bay Area, but did not break through to the surface. Figure 1.1b shows the surface rupture
resulting from the 1992 Landers earthquake, in San Bernardino County, California. The sudden slip
of the fault extended for nearly 50 miles with relative displacements from 1 inch to 20 feet.
Ground subsidence occurs as loose soils rearrange and settle into a denser state during vibra-
tions caused by earthquakes. In some cases, the compaction effect may amount to substantial
settlement of the ground surface. Liquefaction, which will be discussed later in this section,
can also ­trigger significant subsidence of unconsolidated sediment. For example, it is believed
that Big  Lake and St.  Francis Lake in Northeast Arkansas formed during the 1811 and 1812
New Madrid e­ arthquakes in areas where liquefaction resulted in significant lowering of the land
surface, ­causing it to flood. Uniform settlement over large tracks of land causes few problems
for infrastructure (except for long horizontal structures, such as roads and pipelines). However,
systems along the boundaries can experience differential settlement that can be extremely dam-
aging, as was experienced in Anchorage, Alaska along Fourth Avenue during the Good Friday
earthquake in 1964; see Figure 1.2.
Ground cracking is usually observed along the edges of ground subsidence as shown in Figure 1.2;
it may also be the result of slope failure or liquefaction, all of which cause the ground to lose its
­support and sink, with the ground surface breaking up into fissures, scarps, horsts, and grabens
(Figure 1.2). The most damaging effect of ground subsidence is differential settlement, which can
severely disrupt the function of any infrastructure system near the vicinity of cracking locations,
particularly those with long foundations that straddle the cracks.
Soil liquefaction occurs when loose, saturated granular soils temporarily change from a solid to
a liquid state, losing their shear strength, which corresponds to a loss in effective stress between soil
particles. Loose saturated (or moderately saturated) sands and nonplastic silts are most susceptible
to this ground failure; however, in rare cases, gravel and clay can also experience liquefaction. In all
cases, poor drainage within the loose soil causes an increase in the pore water pressure as the soil
is compressed by the vibratory effect of seismic waves. As the load is transferred from soil to pore
water pressure, the effective stress between particles is temporarily reduced, or eliminated, causing
4 Introduction to Earthquake Engineering

Scarp

Fissure
Horst Graben

Normal
fault

FIGURE 1.2  Surface subsidence. (Courtesy of USGS.)

a corresponding decrease in shear strength. In some cases, the pore water pressure increases rapidly
and a slurry (soil water mixture) forms that flows vertically to the surface, which can result in craters
and sand boils as shown in Figure 1.3a. Liquefaction is responsible for some of the most spectacular
failures caused by earthquakes as shown in Figure 1.3b. Following the 1964 Niigata, Japan earth-
quake, several apartment buildings experienced severe tilting and settlement. Fortuitously, the tilted
buildings did not suffer major structural damage, and thus minimal human injury.
Landslides caused by earthquakes are uncommon. Consequently, in order for a structure to
­experience damage during the event, it must be located at the top or bottom of the soil mass that
slides down; for this reason, damage resulting from earthquake-induced landslides is rare. Sloped
land that is marginally stable under static conditions is most susceptible to sliding during the intense
shaking of strong earthquakes. For the most severe cases, debris (soil, boulders, and other mate-
rials) flow can move at avalanche speeds and can travel long distances depending on the slope
from which the landslide was formed. Furthermore, earthquake-induced landslides can be sudden

(a) (b)

FIGURE 1.3  (a) Sand boils that erupted following the 2011 Christchurch Earthquake and (b) Tilting and
settlement of buildings caused by liquefaction following the 1964 Niigata Earthquake. (Courtesy of USGS.)
Introduction 5

(a) (b)

FIGURE 1.4  (a) Partial view of Yungay City, Peru in 1966 and (b) after being buried by the 1970 Great
Peruvian earthquake-induced landslide. (Courtesy of EERC, University of California.)

and unpredictable, producing the total destruction of communities in the path of the debris flow.
Some ­devastating landslides have occurred around the world. One of the worst cases happened
in Peru  during the 1970 Ancash (or Great Peruvian) earthquake (magnitude 7.9) off the Pacific
Ocean coast, which produced what is considered the world’s deadliest earthquake-induced land-
slide (20,000 fatalities). During the earthquake, the northern wall of Mount Huascaran, 130 km
(70 miles) from the epicenter, was loosened, mobilizing rock, ice, and snow into a massive landslide
that buried the towns of Ranrahirca and Yongay; see Figure 1.4.

1.1.2 Indirect Effects of Earthquakes

These include some of the most devastating and frightening impacts of earthquakes, such as tsu-
namis and fires. Although these cause major damage to structural systems, they are only briefly
discussed in this book. Designing structural systems to withstand the large magnitude forces result-
ing from water waves generated by a tsunami is challenging. Tsunamis are long-period sea waves
that are generated when an earthquake causes the vertical movement of the seafloor. Tsunamis
travel far, at high speeds (over 500 mph) in the open ocean and are difficult to detect because of
their small crest-to-trough height, and long wavelengths, which typically, are hundreds of miles
long. Unobstructed, these waves can travel around the world and dissipate all their energy without
causing damage. However, as they approach a shore, the water depth decreases causing an increase
in wave speed and wave amplitude (height of wave run-ups). Wave run-ups of 75 feet have been
observed at several locations. Wave run-ups can push water that rushes far inland, and have created
devastating damage to infrastructure and great loss of life. One of the most devastating tsunami
events occurred in the Indian Ocean on December 26, 2004, which killed over 230,000 people
in 14 different countries; the furthest fatality was in Port Elizabeth, South Africa, over 8000 km
away (see Figure 1.5). In some flat areas, such as Aceh on the northwest tip of Sumatra Island in
Indonesia, the wave run-ups pushed water as far as 2 km inland, with wave heights of over 25 m.
The wave forces generated by these types of events are nearly impossible to design against.
Seiches are earthquake-induced waves in an enclosed body of water, such as a lake or a reser-
voir, or one that is partially enclosed, such as a bay. Seiches are caused when long-period seismic
waves resonate with oscillations of the enclosed water and cause standing waves. Earthquakes may
­happen within or far outside the perimeter of the body of water. Although this type of wave has been
observed during most earthquakes (even in swimming pools), related damage to infrastructure has
been minimal.
Fire is probably the most terrifying indirect effect of earthquakes, particularly considering
that people who survived in collapsed buildings, but were trapped in the debris, were burnt alive
6 Introduction to Earthquake Engineering

BANGLADESH

INDIA
MYANMAR
(BURMA)

THAILAND
SOMALIA SRI
LANKA MALAYSIA
MALDIVES
KENYA
INDONESIA
TANZANIA
SEYCHELLES

Indian
ocean
MADAGASCAR

SOUTH
AFRICA

FIGURE 1.5  Countries affected by the 2004 Indian Ocean Tsunami. (From Wikipedia.)

as the fire consumed everything in its path. Traditional firefighting methods are often ineffec-
tive against earthquake-induced fires because most water mains that supply water hydrants break.
Earthquake-induced fires are started by ruptured combustible substance conduits (such as gas
mains) or destroyed combustible substance storage containers (such as oil tanks), and then ignited
by sparks from sources such as downed powerlines. Earthquake-induced fires are most common
in regions where timber construction is prevalent. One of the most devastating earthquake-induced
fires happened after the 1906 San Francisco earthquake. Over a four-day period, the fire burned
nearly 25,000 buildings on 508 city blocks (see Figure 1.6), which constituted about 80% of the
city at the time.

FIGURE 1.6  Extent of San Francisco fire following the 1906 Earthquake. (Courtesy of USGS.)
Introduction 7

1.1.3  Ground Shaking

Ground shaking causes the majority of earthquake damage; additionally most of the aforemen-
tioned effects are caused by shaking. In fact, where the shaking intensity is low, the hazard of other
effects can be minimal. Consequently, shaking is the only effect experienced by everyone within an
afflicted area, and intense shaking can produce widespread damage from various seismic hazards.
For this reason, ground shaking is the main focus of earthquake engineering.
Although earthquakes are caused by numerous natural and human-induced phenomena, the
events posing the highest seismic risks are caused by the relative deformation of crustal tectonic
plates (discussed in Section 1.2.3). Most frequently, earthquakes happen at the boundaries, but in
rare occasions they occur far from the boundaries. The most spectacular examples of these are the
aforementioned New Madrid earthquakes. The boundary interface of adjacent plates is generally
rough, with friction resisting the relative deformation of these plates. This friction induces shear
stress in the rock, which eventually exceeds the inherent rock strength, resulting in slip that releases
elastic strain energy in the form of shock waves. These seismic waves propagate throughout the
earth, and when they reach the earth’s surface move the ground vertically and horizontally. This
ground movement causes accelerations that result in the inertial response of structures, which in
turn, causes significant damage or collapse of countless structures. There are also segments of
the plate boundaries that experience smooth continuous relative deformation, or creep, such as
the aforementioned segment of the San Andreas Fault along the Cienega Valley near Hollister,
California (Figure 1.1a).
Seismic waves that radiate from the location where a fault ruptures (the focus) quickly travel
throughout the earth’s crust, producing ground shaking when they reach the ground surface.
The  intensity and duration of shaking experienced at a particular site during an earthquake
are primarily because of three factors, all of which will be discussed in more detail throughout
this book:

1.
Earthquake size (magnitude): It can be measured objectively or subjectively—larger
earthquakes cause stronger shaking. A strong earthquake can cause ground shaking over
widespread areas, suddenly affecting large numbers of structures. Even relatively small
earthquakes can have a significant impact on large numbers of buildings. The 2014 South
Napa earthquake (magnitude 6), as shown in Figure 1.7, caused extensive damage to
buildings in the town of Napa, as well as to the many wineries in the region, with a total
­estimated loss of approximately one billion US dollars. Fortuitously, the quake struck early
in the morning when businesses were closed and sidewalks were empty; only one fatality
was attributed to the quake.
2. Location (distance from the focus or epicenter): Generally, the closer to the epicenter, the
stronger the shaking. Structures near the epicenter of a strong earthquake can ­experience
extensive damage, in some cases partial or total collapse; see Figure 1.7. Many other
­examples can be found online.
3. The subsurface materials beneath the structure: Soft soils amplify the shaking, while
rocks do not. This is the most insidious of the three factors because the site can be located
at a long distance from the epicenter and still experience extensive ground shaking due to
local soil conditions. Seismic waves travel through rock for most of their trip from the focus
to the surface; however, at many sites, the final part of the trip is through soil, the geologi-
cal characteristics of which have a major influence on the nature of ground shaking. Some
soils act as seismic wave filters, attenuating shaking at some frequencies while amplifying
it at others. The most peculiar example of distant seismic wave amplification is that of the
Valley of Mexico, where Mexico City is located. The origin of these waves (earthquake
focus) is approximately 350–400 km away, along the Pacific cost in Michoacán. Parts of
this valley are situated in a dry lake, Lake Texcoco, which was emptied to make way for
8 Introduction to Earthquake Engineering

CISN ShakeMap: 6.8 km (4.2 mi) NW of American Canyon, CA

Aug 24, 2014 10:20:44 AM UTC M6.0 N38.22 W122.31 Depth: 11.2 km ID:72282711
39°

38.5°
Saintsbury Winery
Sam Kee Laundry Building

km

37.5° 0 50

–123° –122°
Map version 18 processed 2014–08–25 04:21:30 AM UTC
Perceived shaking Not felt Weak Light Moderate Strong Very strong Severe Violent Extreme
Potential damage None None None Very light Light Moderate Mod/heavy Heavy Very heavy
Peak ACC (%g) <0.1 0.5 2.4 6.7 13 24 44 83 >156
Peak vel (cm/s) <0.07 0.4 1.9 5.8 11 22 43 83 >160
Instrumental
intensity I II–III IV V VI VII VIII IX X+

Scale based upon Wald et al. (1999).

FIGURE 1.7  Intensity and extent of 2014 South Napa Earthquake. (Courtesy of USGS.)

the expansion of Mexico City. The near-surface geology of the city center consists of an
­alluvial deposit, a thick layer of a very soft, high-water-content ­mixture of clay and sand
that was deposited there by streams before the lake was drained (so that the Aztecs could
build the island city of Tenochtitlan). As shown in Figure 1.8, rock shakes at the same
frequency as the seismic waves, which attenuate as they move further from the epicen-
ter. However, the unconsolidated alluvial soil has a dominant influence, and magnifies
the intensity of ground shaking (10-fold) near certain periods, resulting in near harmonic
motion with a period of 2 s. During the 1985 Mexico City earthquake (magnitude 8), peak
horizontal accelerations were first attenuated from 150 cm/s2 along the Pacific coast near
the epicenter to 35 cm/s2 at higher elevations with firm subsurface materials; these waves
were then amplified fivefold to 170 cm/s2 near the center of Mexico City. The shaking
intensity was further amplified by resonant vibration properties of certain buildings, those
with natural periods of about 2 s, which correspond to buildings between 10 and 20 sto-
ries, most of which experienced large displacements and extensive structural damage, or
collapse.

1.2  TYPES OF EARTHQUAKES

The majority of damaging earthquakes are produced by surface fault ruptures caused by the
­relative movement of the tectonic plates. However, a number of other sources, including man-made
and volcanic, can produce measurable earthquakes. Volcanic earthquakes can be attributed to
the same forces that are responsible for tectonic earthquakes. In this section, we present a brief
description of these three types of earthquakes. In Chapter 2, a more detailed description of tectonic
­earthquakes is provided.
Introduction 9

USGS ShakeMap : Michoacan, Mexico

Thu Sep 19, 1985 13:17:47 GMT M8.0 N18.42 W102.38 Depth: 15.0 km ID:198509191317
22°

20°

21-story steel building in

18°
the Pino Suarez Complex
170 PGA = 170 cm/s2

–170
SCT 170 PGA = 35 cm/s2
16°
km
–170

0 50 100 170 150 cm/s2 170 18 cm/s2

UNAM

–106° –104° –102° –100° –170 –170

Map version 1.1 processed Sat Nov 8, 2008 11:44:37 AM MST Campos Teacalco
Perceived shaking Not felt Weak Light Moderate Strong Very strong Severe Violent Extreme
Epicenter Mexico
Potential damage None None None Very light Light Moderate Moderate/heavy Heavy Very heavy city
Peak ACC (%g) <.17 .17–1.4 1.4–3.9 3.9–9.2 9.2–18 18–34 34–65 65–124 >124 ~400 km
Peak vel (cm/s) <0.1 0.1–1.1 1.1–3.4 3.4–8.1 8.1–16 16–31 31–60 60–116 >116
Instrumental
intensity I II–III IV V VI VII VIII IX X+

FIGURE 1.8  Intensity of 1985 Mexico City Earthquake. (Courtesy of USGS.)

1.2.1  Man-Made Earthquakes

These generally have a much smaller magnitude than the other two types of earthquakes, and
thus have a lesser impact on infrastructure. However, man-made earthquakes can lead to earlier
fault ruptures (tectonic earthquakes) because the shaking can increase critical stresses at the plate
boundaries. One of the most intense cases is due to explosions, both from conventional and nuclear
­weapons. For example, it is estimated that the Boxcar nuclear bomb explosion in 1968, with a yield
of 1200,000 tons TNT equivalent, excited an earthquake of magnitude 5.0 that lasted for 10–12 s
(Bolt, 2004). This shook buildings in nearby communities, including Las Vegas, NV (30 miles
away), but no serious damage or casualties occurred.
Collapse earthquakes can be generated by landslides (those not generated by tectonic earth-
quakes) as well as collapse of mine/cavern roofs, or buildings. Earthquakes caused by mine roof
collapse can be rather controversial given that tectonic earthquakes can cause mines to collapse.
The  Crandall Canyon, Utah coal mine collapse on August 6, 2007 was blamed on a tectonic
­earthquake by the owner, but seismologists determined that the 3.9 magnitude seismic shock was
caused by the collapse of the mine roof. They reasoned that the coal extraction process being
applied was the primary cause of the mine collapse because the center of the focus (center of the
seismic activity) was in the collapsed mining zone, and most of the previous earthquakes recorded
in that region were attributable to mining activities. No damage from the earthquake was detected
outside the mine; however, six miners lost their lives in the accident. Although not necessarily man-
made, earthquakes caused when underground caverns collapse (which sometimes leave a sinkhole
on the surface) are very similar to those caused when mine roofs collapse, and are very common
in many parts of the world (such as Florida). Building collapse earthquakes are rare, and even the
largest buildings only cause rather small magnitude quakes. For example, the collapse of the North
World Trade Center Tower on September 11, 2001 only caused a 2.3 magnitude earthquake; see
Figure 1.9. This is smaller than a tectonic earthquake recorded just months earlier near the site of
the tragedy.
10 Introduction to Earthquake Engineering

Seconds
0 10 20 30

5 Second WTC Collapse

North Tower, ML = 2.3

11 Sep 2001
14 h 28 m 31 s
0
Ground velocity (µm/s)

–5
5 Station: PAL
17 Jan 2001 Component: E-W
12 h 34 m 22 s

P-wave Earthquake in Manhattan

Magnitude, ML = 2.4
–5
S-wave

FIGURE 1.9  Comparison of building collapse and natural earthquakes. (From Kim et  al. 2001. Seismic
waves  generated by aircraft impacts and building collapses at World Trade Center, New York City.
Eos Transactions American Geophysical Union 82(47): 565–571.)

Seismic activity near large artificial lakes has also been reported. It is believed that the weight
of the water and the extra water pressure in rocks can change the stresses in existing rock frac-
tures. The weight of the water increases the total stress by direct loading, while the extra pore
water pressure decreases the effective stress (pushes the fracture planes apart); this can lead to
sliding along fracture planes where tectonic strain is already present, causing an earthquake.
This is speculated to be most likely during periods when a reservoir is being filled or drained.
One case that continues to be attributed to artificial lake seismicity is the August 1, 1975 Orville,
California earthquake. Bolt (2004), however, presents a compelling argument against this sup-
position. His  ­personal ­experience is that the connection between the earthquake and the fill-
ing of the lake is only circumstantial; however, he does not completely discount the possibility
that the reservoir might have in part triggered the earthquake. There are other earthquakes that
have been conclusively proven to have been triggered by artificial lakes; one such case is the
November 14, 1981 earthquake (­magnitude 5.6) induced by Lake Nasser along the Nile River in
Egypt (Bolt, 2004).
Earthquakes caused by energy production processes have been recorded for over 50 years, though
recently the association of earthquakes and energy production has become contentious. This is
­particularly controversial with respect to the hydraulic fracturing (fracking) process. This ­technique
for extracting oil and gas from low permeability rocks involves fracturing rocks by ­injecting high-
pressure fluid into the rock. The most compelling evidence of fracking-induced seismicity can be
found in Oklahoma and north Texas, regions where seismic activities were once nonexistent, but
now routinely experience small earthquakes; with the only variable being oil and gas production.
In fact, a 2012 report by the U.S. National Research Council concluded that some (not many) pro-
duction activities have induced perceptible seismicity. Geothermal power generation has also been
Introduction 11

conclusively correlated with earthquakes. The process entails injecting water at high pressures to
fracture 500°F solid rock, creating an artificial reservoir of superheated water, the steam of which
is then used to drive electrical turbines. A well-documented case of induced seismicity is from
Geysers Geothermal field in California.

1.2.2  Volcanic Earthquakes

These are caused by the same energy source as tectonic earthquakes, which is the heat from the
earth’s core. Volcanic seismicity affects limited areas near volcanic regions. For example, Hawaii’s
big island has a very active volcano that has caused several earthquakes, none of which have been
felt by the entire island. These occasionally serve as a warning sign of impending volcanic erup-
tions. The movement of magma through tubes below the volcanic vents creates pressure changes
in the surrounding rock that can rupture, releasing elastic strain energy as seismic waves. These
seismic waves have been successfully used to predict eruptions of volcanos such as Mount St. Helen
in 1980 and Pinatubo in 1991. Other seismic waves can be induced by sudden, irregular movement
of magma whose path has been obstructed, or by steady magma movement deep in the mantle.
Damage from all these earthquakes is relatively minor compared with that p­ roduced by tectonic
earthquakes.

1.2.3 Tectonic Earthquakes
As discussed previously, these are caused by a sudden dislocation of segments of the earth’s crust,
the structure of which is composed of plates (large and small) known as tectonic plates that float on
a liquid layer, the mantle. This arrangement resulted from the formation of planet Earth five billion
years ago, when hot gasses cooled into a semi-solid mass. It is estimated that after one to two billion
years of cooling, the crust solidified and cracked forming tectonic plates (different ones than those
that exist today).
From the beginning, the plates have been in constant motion forming and breaking up conti-
nents over time, including the formation of supercontinents that contained most of the landmass.
The  ­latest supercontinent, Pangea, started separating approximately 200 million years ago, and
its parts have drifted apart to the current configuration of the earth’s surface. This process was
­originally proposed by Alfred Wegener in the early 1900s. He noted several different pieces of
evidence to support his theory of the continental drift, including (see Figure 1.10):

1. How the current shape of some continents appear to fit together, particularly the east coast
of South America and the west coast of Africa.
2. The significant similarities between fossil records (both flora and fauna) found in several
continents that could only have occurred if the continents were attached.
3. The similarity in geology across several continents, including grooves carved by glaciers
and the sediments deposited by these glaciers.

Wagner’s theory implies that the earth’s crust (known as the lithosphere) consists of large slabs
of rigid rock, or plates (see Figure 1.11). The lithosphere can be divided into oceanic and continental
parts. The oceanic lithosphere is thinner and forms near mid-oceanic ridges where plates are drift-
ing apart. The thicker continental lithosphere plates consist of rocks that are less dense than oceanic
lithosphere rocks. All plates move at different speeds and different directions, spreading apart along
some of the plate boundaries (spreading or divergent zones), while colliding along other boundaries
(subduction or convergent zones). The spreading zones are continuously replaced with upwelling
lava, creating new seafloor or oceanic lithosphere. At the opposite side of the newly created oceanic
lithosphere, the old part of the plate plunges into the earth’s interior at the subduction zones, being
12 Introduction to Earthquake Engineering

AFRICA Fossil evidence

of the Triassic
INDIA land reptile
Lystrosaurus.

SOUTH AMERICA
AUSTRALIA
ANTARTICA

Fossil remains of Fossils of the fern

Cynognathus, a Fossil remains of the Glossopteris, found in
Triassic land reptile freshwater reptile all of the southern
approximately Mesosaurus. continents, show that
3 m long. they were once joined.

FIGURE 1.10  Continental drift evidence; distribution of fossils across the southern continents of Pangea.
(Courtesy of USGS.)

absorbed into the mantel. At some locations along the mid-oceanic ridge, horizontal slip occurs
normal to the divergent zones as a new s­ urface is created or transformed; this type of boundary is
known as a transformed fault (see Figure 1.11).
Wagner originally proposed that the forces driving the continental drift were the result of g­ ravity,
and the same forces were responsible for ocean tides. This theory was rejected because these forces
are too small to drive the continents through the seafloor. It was later proposed that seafloor spread-
ing at the mid-oceanic ridges was responsible for moving the seafloor and continents over the
­mantle, which is a softer rock immediately below. Although this theory was also rejected, it helped
establish the theory of plate tectonics, which with respect to the shape and size of the plates is
widely accepted today; see Figure 1.11. For a while, it was believed that the movement of the plates
was driven by convection currents in the mantle; this transfers heat from deep within the earth (the
core) to the surface. The temperature difference drives molten lava currents (similar to how heat
from the sun drives wind currents on the earth’s surface). The currents are produced as heated rock
near the core (which has a lower density than that near the bottom of the plates) rises toward the
surface as molted lava; over the course of the trip, lava cools and becomes denser, some of which
after reaching the surface plunges back toward the core.
This theory, however, does not fully explain all phenomena associated with lithospheric plate
movement. The current view is that plate movement is driven by additional forces, known as ridge
push and slab pull, which also derive their energy from the dissipation of heat from the earth’s core
through convection currents (this theory is still being debated). These forces are generated at the
mid-oceanic ridge (divergent zones) and at the subduction zones. At the divergent zones, lava cools
at the surface creating new oceanic ridge; as this new material moves away from the divergent
Introduction 13

Transform Divergent Convergent

Plate
Plate
Asthenosphere
North American
Eurasian plate
plate Eurasian
plate

Juan de Fuca
plate
Caribbean
Filipino plate Arabian
plate plate Indian
Cocos
plate plate
EQUATOR

Nazca African
Pacific
plate plate
Australian plate South American
plate plate
Australian
plate

Scotia plate

Antarctic
plate

FIGURE 1.11  Tectonic plates and types of plate boundaries. (Courtesy of USGS.)

zones, it continues to cool, becoming denser. At the other end of the plate, the oceanic lithosphere
becomes even denser with age as it continues to cool and thicken. This causes the lithospheric plate
to tilt toward the subduction zones, like an unbalanced boat in water. The mid-oceanic ridge then
rises above the rest of the seafloor, and as the new material rises to the top of the ridge, gravity pulls
it down; this is called ridge push. Also, the greater density of old oceanic lithosphere at the subduc-
tion zones, relative to the underlying asthenosphere, allows gravity to pull these rocks (known as
slabs) down into the mantle. This is believed to be the source of the driving force for the majority
of plate movement.
Although the precise force mechanisms causing movement of the lithospheric plates con-
tinue to be debated, there is substantial agreement that heat from the earth’s core serves as the
primary source of energy for these forces. Also, the general outlines of the major plates have
almost conclusively been established. In fact, this outline is clearly illustrated when the loca-
tion of the epicenters of recent earthquakes is plotted on a global map as shown in Figure 1.12.
At the boundaries of the plates, rocks fracture, usually at many locations, creating a web of
smaller plates with edges that rub and push relative to each other; these edges are called faults.
On average, these faults have the potential to displace approximately 2 inches per year. When
the rubbing and pushing is prevented, elastic energy accumulates along the edges of the plates.
When this energy is released with a s­ udden movement (slip), it causes brief strong ground vibra-
tions. The specific location (generally a volume of rock) where the movement or energy release
occurs is known as the focus, or hypocenter. The point on the earth’s surface directly above
the hypocenter is called the epicenter. Usually, the vibrations cause the rocks near the focus to
become unstable; and as these rocks settle into a new equilibrium state they cause aftershocks.
The discipline that studies seismic activity is known as seismology, and will be discussed further
in Chapter 2.
14 Introduction to Earthquake Engineering

African plate

FIGURE 1.12  Epicenters of 358,214 earthquakes, 1963–1998. (Courtesy of NASA.)

1.3 HISTORY OF THE DEVELOPMENT OF MITIGATION

STRATEGIES FOR THE EFFECTS OF SEISMIC HAZARDS
The primary goal of an earthquake engineer is to mitigate hazards resulting from seismic events.
As  previously discussed, it is difficult to mitigate the effects of tsunamis, landslides, and fires;
however, mitigating the effects of ground shaking on structural systems using sound engineering
judgment has made great advances in the last century. This is known as earthquake-resistant design
and has been successfully applied to reduce human and economic losses. The remainder of the book
focuses on earthquake-resistant design for structures. In this section, we briefly discuss the histori-
cal development of earthquake-resistant design, which chronologically has occurred in five general
stages:

1. Development of construction practices to mitigate the effects of shaking.

2. The first half of the twentieth century focused on characterizing the effect of shaking using
lateral forces.
3. In the 1940s, the development of the response spectrum theory marked a major step
forward.
4. In the 1960s, concepts of structural dynamics were incorporated into design practice.
5. The last stage has implicitly been included in building codes since the 1970s and deals with
performance-based design.

While humans did not fully understand the process that causes earthquakes until the middle
of the last century, many civilizations developed construction practices to reduce the s­ tructural dam-
age caused by ground shaking. From the study of prehistoric buildings in seismic-prone regions, it
appears that two approaches to seismic-resistant construction were followed: increase the strength
or change the stiffness of the lateral force-resisting systems. This is particularly evident for the con-
struction of important structures, which received special seismic-resistance detailing, while most
other buildings did not.
In many parts of the world, buildings were built stronger after they collapsed in an earthquake,
while other parts of the world addressed weaknesses in the construction by specifying different
construction detailing. For example, the Incas in Peru were master builders that created ingenious
earthquake-resistant structures. The hilltop city of Machu Picchu (classified as a Civil Engineering
Introduction 15

FIGURE 1.13  Seismic-resistant construction in Machu Picchu, Peru.

International Historic Monument by the American Society of Civil Engineers [ASCE]) has numer-
ous structures constructed of interlocking, mortar-free stonework with perfectly mated joints (these
are known as dry-stone walls or ashlar masonry); see Figure 1.13. Clearly, this type of construc-
tion was deliberate in order to prevent sliding of the stone blocks during earthquakes. The intricate
process was far more complex than the perfectly regular array of stonework used in other parts
of the world, indicating that the extra effort was purposefully made for structural reasons and not
for aesthetics, as some have suggested. In fact, there are buildings in Machu Picchu with masonry
laid in parallel courses (see Figure 1.13); the most important structures, such as temples and the
king’s quarters, were built to withstand large earthquakes using masonry laid in nonparallel courses
(see Figure 1.13).
Ashlar masonry is more seismic resistant than using mortar because higher friction forces
develop between mating surfaces. This along with the interlocking masonry allows the stones to
move slightly and resettle without excessive movement of the walls. This type of construction allows
for passive energy dissipation, a form of passive structural control. Inca construction was based on
designs that follow prescriptive specifications. This type of design philosophy is based on experi-
ence with the construction of proven designs, and is still widely used for design today. More rational
construction procedures based on engineering principles (engineered construction) have only been
available for the past century.
Engineered seismic-resistant design was developed in America, Asia, and Europe; not neces-
sarily independently, but there is no consensus on the order. It is believed that the first rational
attempts to understand the effects of earthquakes on structures were made in Europe following
the ­devastating 1755 Lisbon earthquake (over 50,000 casualties). After this event, the regional
government enacted building regulations that required seismic-resistant detailing to be followed
in the reconstruction of the city. A few years later, scientists thoroughly documented the effects of
the 1783 earthquake in the Calabria region of southern Italy (nearly 50,000 casualties). Although
the results from these studies did not lead to substantive design guidelines, they did inspire future
­scientists to study the effects of earthquakes on structures, including a number of British and Italian
engineers who went on to develop the foundation for seismology.
After the 1881 Nobi earthquake (7000 casualties), the Japanese government formed the first-ever
earthquake investigation committee, and directed the committee members to research earthquake-
resistant structural design. It is believed that this committee conceived the idea of characterizing
the effects of earthquakes on structures by applying a lateral force equal to a fraction of the total
weight of the building, which is based on Newton’s second law (force equals mass times accelera-
tion). Following the 1908 Messina-Reggio earthquake (58,000 casualties), the Italian government
also appointed a committee to develop practical recommendations for the seismic-resistant design
of buildings. This committee also proposed characterizing the effects of earthquakes on structures
by applying lateral forces: first, they proposed a force equal to 1/12 of the building weight. Three
16 Introduction to Earthquake Engineering

years later, this specification was modified as follows: the first story must resist a force equal to 1/12
of the building weight above, while the second and third floors must resist a lateral force of 1/8 of
the building weight above. This specification limited the height of buildings to three stories. These
two committees provided a first iteration to the equivalent static procedure that is widely used in
seismic-resistant design codes around the world today.
In America, the first attempt to characterize seismic effects on buildings also followed a major
seismic event, the 1906 San Francisco earthquake. Resistance to earthquake forces was prescribed
in code provisions indirectly by requiring buildings to resist a wind lateral load of 30 lb/ft2, which
was intended to account for both wind and seismic effects. Following another devastating earth-
quake in California (Long Beach in 1933), the Division of Architecture was given the responsibility
for public-school building safety. Their proposed regulations required that public-school buildings
be designed to resist a lateral force equal to a fraction of the dead load and a portion of the design
live load, entailing 0.1 for masonry buildings without frames and 0.02–0.05 for other types of
buildings. This appears to be the first explicit design specification incorporating the importance of
construction regarding the type of lateral force-resisting system. This also marked a very impor-
tant point in code enforcement; the California Legislature passed the Field Act and the Riley Act.
The Field Act made the seismic-resistant design and construction of public-school buildings manda-
tory, and the Riley Act required that most buildings in the state be designed to resist a lateral load
equal to 2% of the design loads dead plus live loads.
The 1943 City of Los Angeles building code is believed to be the first to recognize that lateral
forces depend on building mass and height, an indirect recognition of building flexibility on these
forces. The lateral forces at different stories of a building were determined by the product of a grav-
ity load (equal to dead load plus half of the live load) and a lateral force coefficient that varied with
the number of stories above the story being analyzed. The 1947 San Francisco building code also
recognized the influence of building flexibility on the magnitude of seismic forces, as well as the
influence of site soil conditions on these forces.
Although many city building codes have been in existence around the world for decades (even
millennia, such as the Hammurabi Code from Babylon, which did not include specifications, rather
promoted good construction by prescribing penalties for failures), regional model codes were not
developed until the introduction of the Uniform Building Code (UBC) in 1927. The second edition
of the UBC published in 1930 incorporated seismic provisions and suggested these be incorporated
by cities in areas prone to earthquakes. The UBC seismic provisions were based primarily on two
observations of building performance during earthquakes: first, most damage was caused by lat-
eral shaking, and second, structural damage was more extensive for buildings on soft soil deposits
(likely the first code to recognize soil effects). The code specified a lateral force equal to 10% of
the dead plus design live loads, reducible to 3% for buildings on firm soil. In the ensuing editions
of the UBC, maps of the United States depicting zones of different levels of seismic risk were
­published. These maps recognized the hazard of large magnitude earthquakes for different areas
of the country.
In 1948, a joint committee of the Structural Engineers Association of Northern California and
the San Francisco Section of ASCE proposed model lateral-force provisions for California building
codes. A major advance was the introduction of the earthquake-response spectrum concept, which
was first developed by Biot in 1943 and later proposed as a design tool by Housner. This theory
combines ground motion and the dynamic properties of the structure (namely the period). This led
to the building period becoming an explicit factor in the determination of seismic design forces.
The response spectrum work was made possible by recordings of strong-motion earthquake accel-
erations using recently deployed accelerographs by the US Coast and Geodetic Survey. In 1952,
a second joint committee of ASCE and now the Structural Engineers Association of California
(SEAOC) developed a comprehensive guide, the Recommended Lateral Force Requirements and
Commentary (the so-called SEAOC Blue Book), first published in 1959. This explicitly related lat-
eral forces to building periods based on the design response spectrum.
Introduction 17

Structural dynamic concepts were further developed in the 1960s and incorporated in seismic
codes after the 1971 San Fernando earthquake. In particular, structural dynamic specifications
explicitly incorporating the response spectrum theory were required for the design of hospitals
and other critical facilities. Since the incorporation of structural dynamics in code specifications,
significant advances have been made in the development of analytical procedures; much of this has
been in parallel with the development of computers because solutions based on structural dynamics
are computationally intensive. Even with the great advances in the development of mitigation strate-
gies for ground shaking, structures designed using modern seismic-resistant design principles have
been observed to experience much larger forces during major earthquakes than those prescribed by
design codes, leading to major damage, even collapse. These observations have led researchers to
conclude that structures subjected to larger imposed loadings from earthquakes can survive intense
shaking with damage; however, will not collapse as long as lateral force-resisting systems are tough
(i.e., able to absorb the seismic energy) and include a continuous path for lateral load to be trans-
ferred to the ground. Also, it is important for engineers to understand that the structural engineering
principles used for the analysis of structures subjected to gravity static loads do not result in the
greatest dynamic forces a structure may experience during an earthquake. These dynamic forces
are only equivalent to code-specified lateral forces in that a structure designed to resist these forces
has the capability of deforming without overstressing from load reversals, and provide adequate
member ductility, as well as provide connections with sufficient strength and resiliency to accom-
plish the following performance goals:

• Resist minor earthquakes without damage.

• Resist moderate earthquakes without structural damage, but with some nonstructural
damage.
• Resist major earthquakes without collapse, but with both structural and nonstructural
damage.

These performance criteria have generally remained unchanged since the 1970s, but have
recently taken on a broader approach in order to allow the real estate owner to decide the desired
level of structural performance and safety; both in terms of protecting occupants and capital invest-
ment. This new philosophy is known as performance-based design and is awaiting to appear in stan-
dards such as the 2010 ANSI/ASCE “Minimum Design Loads for Buildings and Other Structures”
standard; see Section 1.3.1.3 in this document.

REFERENCES
Bolt, B. A., Earthquakes, 5th edition, W.H. Freeman and Co., New York, 2004.
Wald, D., Quitoriano, V., Heaton, T., and Kanamori, H., Relationships between peak ground accelera-
tion, peak ground velocity and modified mercalli intensity in California, Earthquake Spectra, 15(3),
557–­564, 1999.
 2 Engineering Seismology
Overview
After reading this chapter, you will be able to:

1. Define seismology terminology

2. Describe tectonic plate movement based on the elastic rebound theory
3. Describe different types of seismic waves
4. Describe how earthquakes are measured based on intensity and magnitude
5. Describe major effects of earthquakes on structures
6. Describe earthquake characteristics
7. Describe basic site characteristics
8. Describe structural characteristics
9. Compute the probability of occurrence of an earthquake

Seismology is the scientific discipline that studies seismic activity from various sources
(­tectonic, volcanic, etc.), in particular, the propagation of seismic waves through the interior of the
earth and their environmental effects, such as ground shaking and tsunamis. This discipline grew out
of the need to understand how earthquakes occur, and to try and predict their occurrence. Engineering
seismology is the applied branch of seismology that focuses on characterizing seismic hazards at a site
or region in order to evaluate risks to various vulnerable infrastructure systems. This is accomplished
by studying the seismic history of a region (including the frequency and ­intensity of earthquakes)
to establish characteristics (such as the intensity of shaking) and frequency of occurrence of future
earthquakes. In this chapter, we briefly cover the most important concepts of engineering seismology.

2.1  SEISMOLOGY TERMINOLOGY

Just like any other discipline, seismology has developed a jargon as a means to communicate
­succinctly. In this section, we introduce and define terminology related to earthquake engineering.
One of the most recognized terms related to seismology is epicenter, which is the geographical
point on the ground surface where an earthquake is estimated to be centered. This point along with
the focal depth gives the location where the rock ruptures at a fault (fault rupture) that generates
the main earthquake, the focus or hypocenter. This is an area (not a point) that can extend for many
miles along a fault. The plane along which the rock ruptures and slips is known as the fault plane,
and may be at an angle with respect to the ground surface, the dip angle. The angle the fault plane
makes with respect to the north direction along the surface is known as the strike angle. The ­relative
displacement between the two sides of the fault plane is known as the fault slip. The radiating
­seismic waves (discussed in Section 2.1.2) are recorded using a seismometer at an observation sta-
tion located at a distance known as the epicentral distance. These terms are used to characterize
earthquakes and are illustrated in Figure 2.1.
The focal depth is used to classify earthquakes as: shallow for focal depths less than 70 km
(43 miles), intermediate for focal depths between 70 km (43 miles) and 300 km (186 miles), and
deep for focal depths greater than 300 km (186 miles). Focal depths have been known to be as large
as 720 km (450 miles). Shallow earthquakes are more destructive than others because the mass of
the rock above deeper earthquakes attenuates their shock waves. Shallow earthquakes generally
occur near ocean trenches, like coastal California earthquakes, which have focal depths of less than

19
20 Introduction to Earthquake Engineering

Epicentral distance
North Epicenter
Strike angle
Observing station

Dip angle
Fault line

Fault plane Focal depth

Hypocenter (focus)

FIGURE 2.1  Earthquake terminology.

16 km (10 miles). Most moderate-to-large shallow earthquakes are preceded by smaller quakes,
called foreshocks, and followed by smaller quakes called aftershocks. Foreshocks are precursors of
the impending fault rupture, while aftershocks result from adjustments to the stress imbalance in
the rocks produced during the rupture.
The fault slip is used to classify faults depending on the direction of the movement of rocks on one
side relative to the other side; vertical movement is known as dip-slip, while horizontal movement
is known as strike-slip. These terms correspond to the definition of the angles used to describe the
direction of fault movement. As discussed in Section 1.2.3, faults occur along the edges of tectonic
plates (or their interior, which is much less common). Dip-slip faults are further subdivided based on
the relative vertical movement of the tectonic plates: normal faults occur at plate boundaries where
tectonic plates spread apart at divergent zones causing the hanging wall to move down relative to the
footwall, while reverse faults occur at plate boundaries where tectonic plates collide at convergent
zones causing the hanging wall to move up relative to the footwall; see Figure 2.2. Thrust faults
are reverse faults with a small dip angle. Faults along oceanic ridges are normal, resulting in the
lengthening of the crust, as shown previously in Figure 1.11, whereas those along subduction zones
are reverse, resulting in the shortening of the crust. When these faults break through to the surface,
surface rupture, they produce an exposed steep slope known as the fault’s scarp; see Figure 2.2.
At transform zones, tectonic plates move horizontal relative to each other; the faults associated
with this motion are strike-slip, which can be further subdivided into right-lateral or left-lateral
faults. Figure 2.3 depicts a left-lateral fault, which can be described by an observer standing in front
of the fault line as watching the land on the other side moving to the left. Similarly, an observer
­facing a right-lateral fault line would see the land on the other side moving to the right. The San
Andres fault discussed in Chapter 1 is a type of right-lateral fault. This is demonstrated by the fault
extending from left to right through the drainage culvert on the south side of the DeRose Winery
building in the Cienega Valley; see Figure 2.3. Therefore, when these faults break through to the
surface, any pipes, streams, fences, etc. that straddle the surface rupture are offset. It should be
noted that dip-slip and strike-slip faults can occur simultaneously, in which case the fault is clas-
sified as an oblique-slip. Also, fault rupture does not reach the surface in the case of many earth-
quakes, particularly for intermediate and deep cases.

Scarp
Hanging
Footwall wall
Hanging Footwall
wall
Normal fault Reverse fault

FIGURE 2.2  Vertical movement, dip-slip faults.

Engineering Seismology Overview 21

Right-lateral fault

Left-lateral fault

FIGURE 2.3  Horizontal movement, strike-slip fault.

2.1.1 Elastic Rebound Theory: Tectonic Plate Movement

As discussed in Section 1.2.3, tectonic plates are forced to move by several mechanisms, includ-
ing ridge push and slab pull, convection currents in the semimolten rock of the mantle, etc. At the
boundaries, relative movement between plates tends to occur; however, shear stresses develop along
the plane that separates the fault because of friction between the plate boundary surfaces. As these
shear stresses reach the shear strength of the rock along the fault, the rock fails and elastic strain
built-up in the rock is released as the rock fractures and slips. The rate at which the stored strain
energy is released is dependent on the geometry and material properties along the fault, and it
­dictates whether the resulting movement is aseismic (in the form of creep; see Figure 1.1) or seismic
(in the form of an earthquake). Along relatively straight stretches of fault where the rock is weak
and ductile, the rock tends to rupture slowly releasing the strain energy gradually, which results in
creep. Where the rock is strong and brittle, it ruptures fast, releasing the strain energy abruptly—
partly in the form of heat and partly as radiating stress waves that cause earthquakes on the ground
surface. For this case, after the strain energy is released, there is a period of no seismic activity
until the s­ ubsequent buildup of strain energy occurs followed by rupture. Therefore, earthquakes
are more likely to occur along parts of a fault line that have not ruptured for some time; these are
known as seismic gaps, which can be identified by plotting seismic and creep activity along a fault
line (Kramer, 1996). This process of successive buildup and release of strain energy is described
by elastic rebound theory and can be described graphically, as shown in Figure 2.4. The figure
depicts the gradual buildup of strain resulting from the relative motion of two plate boundaries
when the motion is being resisted by friction forces. The fault slips when the stresses overcome the
shear strength of the rock, and the two sides of the fault rebound to an unstressed state, resulting in
radiating seismic waves.

2.1.2 Seismic Waves
Seismic waves radiate from the focus and travel in every direction, as shown in Figure 2.5. The ­portion
of the energy released from a fault rupture as shaking first travels through the interior of the earth
as body waves, following the shortest path to the ground surface where they are transformed into
surface waves (see Figure 2.5). Thus, body waves arrive at an observation station (or site) before any
surface waves. An understanding of these different waves is necessary to establish the epicenter and
to characterize the size of the earthquake (earthquake magnitude and ground acceleration).

Unstrained rock Strain buildup Rebound to unstrained state

FIGURE 2.4  Elastic rebound theory.

22 Introduction to Earthquake Engineering

Surface
waves
S Body waves
Earthquake
P
Vibration
direction

FIGURE 2.5  Seismic waves.

Body waves naturally divide into two different types, P-waves (primary or pressure) and
S-waves (secondary or shear), which are slower than the P-waves; see Figure 2.5. The stiffness of
the ­material through which the waves travel controls their speeds. P-waves are longitudinal waves
that move through earth’s interior in successive compression and rarefaction (i.e., P-waves move
material ­particles back and forth parallel to their direction). Their average speed through solids
is approximately 6 km/s and slows down when traveling through liquid, for example, speeds in
water are only about 2 km/s. Also, given that liquid medium (such as water) has no shear stiffness,
S-waves can only travel through solids, at about 3 km/s. Because of the slower speeds of S-waves,
they travel at a lower frequency, but larger amplitude, which makes them more destructive than
P-waves. This speed differential between P- and S-waves can be recorded using a seismograph as a
time lapse, Δtp−s (see Figure 2.6), and allows the seismologist to predict an impending earthquake
at a site, albeit with a very short warning. This was the basis for an earthquake early warning
system implemented in Mexico City a few years after the 1985 earthquake. Because of the large
distances (approximately 400 km; see Figure 1.8) from the epicenter to the city, it is estimated that
the ­residents of the city have 60–90 s (the time it takes between the arrival of the P-waves and the
intense shaking from S-waves) to prepare for the impending earthquake.
When body waves reach the ground surface, they are classified as surface waves. These have
lower frequency than their parent body waves and are obviously responsible for the damage associ-
ated with earthquakes. The intensity of these waves diminishes approximately exponentially as the
focal depth increases. These waves are also divided into two types: Rayleigh and Love. Rayleigh
waves combine effects of both P- and S-waves on the earth’s surface and move disturbed material
vertically and horizontally (in the direction of wave propagation) simultaneously, similar to ocean
waves. These waves have low frequencies and large amplitudes, and travel at average speeds of

Δtp–s Peak Ground Acceleration (PGA)

Acceleration

Time

P-wave arrival
S-wave arrival

FIGURE 2.6  Seismograph—a strong-motion recording using a seismometer.

Engineering Seismology Overview 23

3 km/s. Love waves move the ground side to side on the surface plane, perpendicular to the direc-
tion of the wave propagation, similar in character and speed to S-waves. The shear motion of these
waves is primarily responsible for the ground shaking experienced during earthquakes.
As shown in Figure 2.6, seismic waves can be recorded using a seismograph, which can be used
to decipher the four different types of waves since each travels at different speeds, with P-waves
being the fastest, followed by S-waves, then Love waves, and finally Rayleigh waves. Along with
the direction of shaking on the ground surface (vertical and horizontal), the different speeds relative
to one another can be used to separate these waves from one another in a predictable pattern; this
is particularly clear for P- and S-waves in Figure 2.6. For surface waves, the vertical and horizontal
(north–south and east–west) recordings are needed to decipher the two surface waves. The different
arrival times of these waves can be used to estimate the epicentral distance (and the hypocentral
distance—the distance from the observation station to the focus). With recordings from three differ-
ent observation stations, an estimate of the geographic location of the epicenter can be triangulated.
That is, the epicenter is located where three circles drawn on a map intersect, each circle of radius
equal to the epicentral distance, and centered at their corresponding observation station location.

2.2  MEASURING EARTHQUAKES

The duration and size of an earthquake can be measured using several objective and subjective
methods, including relative strength (shaking intensity), amount of energy radiated during the fault
rupture, and the magnitude of the largest acceleration imparted to a site or a structure. The shaking
intensity measures earthquake size using a single quantitative measure, either the intensity (subjec-
tive) or magnitude (objective); see Figure 2.7. These two will be discussed further in the next two
sections and are typically used to compare events of different sizes quantitatively. Energy released
is measured objectively using one of a number of methods to be discussed later. The acceleration
is quantified at a site using the peak ground acceleration (PGA), which is defined as the maximum
ground motion (see Figure 2.6).

2.2.1 Earthquake Intensity
The intensity of an earthquake is a subjective, nonempirical approach for estimating the size of an
earthquake based on the subjective assessment of human observations of the effects of earthquake
shaking on buildings (amount of damage sustained by structures and land surface) and on people.
Different buildings (based on type and construction material) respond differently to shaking; that
is, the performance of poorly designed and constructed buildings is more sensitive to shaking than

1933
Observation (subjective) Measured motion (objective)

Timeline

First seismograph Richter magnitude scale

Rossi–Forel (RF) (Long Beach, CA)
Body and surface waves scale
Modified Mercalli (MM)
Moment magnitude scale
Earthquake energy
Comparison of intensity of MMI and RF:

- MM I II III IV V VI VII VIII IX X XI XII

- RF I II III IV V VI VII VIII IX X

FIGURE 2.7  Scales used to estimate the relative size of earthquakes.

24 Introduction to Earthquake Engineering

well-designed and/or well-constructed structures. Also, people have varying perceptions of shaking,
that is, psychologically some people are more sensitive to shaking than others. This subjective measure
of earthquake intensity was developed before any instruments were capable of recording earthquake
shaking, and thus its value has diminished since the development of instrumental earthquake size
measurements. However, the intensity methodology is still useful for assessing historic events from
descriptions of damage to buildings and land, particularly given criteria that account for construction
quality, g­ eologic conditions, and potential discrepancies between observers of the same earthquake.
One of the earliest criteria (or scales) was a 10-degree scale proposed in 1883 by two scientists
for whom the scale was named the Rossi–Forel (RF) scale. This is an arbitrary scale that character-
izes shaking numerically in Roman numerals from 1 (RF-I) for barely noticeable to 10 (RF-X) for
an extremely intense earthquake; see Figure 2.7. This scale was not widely accepted; however, it
is still occasionally used in some parts of the world. In 1902, Giuseppe Mercalli proposed a modi-
fied ­version of the RF scale that found wide consensus amongst seismologists. The main issue with
the RF scale was that the 10th degree was too broad (see Figure 2.7). The new scale still had 10
degrees, but intensities were distributed more evenly, with more detailed and explicit definitions
of each degree. This new scale was later modified by a number of seismologists, including Richter
who modified it to better account for contemporary construction in California, and was published in
1931 as the modified Mercalli (MM) scale (Richter, 1958). This is still widely used today, and it has
12 points, ranging from 1 (MM-I) for imperceptible shaking to 12 (MM-XII) for total destruction;
see Table 2.1. It is important to note that even this scale has many limitations because it is based

TABLE 2.1
Modified Mercalli Intensity Scale, and Comparison to Magnitude and PGA
Degree Intensity Description Magnitude PGA ( g)
I Not felt, except by some in rare circumstances. <3.0 <0.03
II Felt only by a few persons at rest on upper floors of buildings. 3.0–3.9
III Felt indoors, but many people do not recognize it as an earthquake.
IV During the day felt indoors by many, outdoors by few, while at night some 4.0–4.9
awakened.
Dishes, windows, doors are disturbed and walls make creaking sounds.
V Felt by nearly everyone during the day and many awakened at night. Some dishes, 0.03–0.08
windows, etc. are broken; cracked plaster in a few places; unstable objects
overturned.
VI Felt by all, many frightened and run outdoors. Some heavy furniture moved; a few 5.0–5.9 0.08–0.15
instances of fallen plaster and damaged chimneys. Damage slight.
VII Everybody runs outdoors. Damage negligible in buildings of good design and 0.15–0.25
construction; slight to moderate in well-built ordinary structures; considerable in
poorly built or badly designed structures; some chimneys damaged.
VIII Damage slight in specially designed structures; considerable in ordinary buildings 6.0–6.9 0.25–0.45
(some with partial collapse); great in poorly built structures. Chimneys, factory
stack, columns, monuments, walls toppled. Heavy furniture overturned.
IX Damage considerable in specially designed structures; well-designed frame 7.0–7.9 0.45–0.6
structures thrown out of plumb; great damage for others, including partial
collapse. Buildings shifted off foundations. Ground cracked conspicuously.
Underground pipes broken.
X Some well-built wooden structures destroyed; most masonry and frame structures 0.6–0.8
destroyed. Ground badly cracked.
XI Few masonry structures remain standing. Bridges destroyed. Broad fissures in >8.0 0.8–0.9
ground. Underground pipelines completely out of service.
XII Damage total. Waves seen on ground surface. Objects thrown into the air. >0.90
Engineering Seismology Overview 25

on a subjective assessment of observations (observer’s previous experience with earthquakes), local

design and construction practices, and population density. Furthermore, the MM scale considers the
response of a building to ground shaking, not the ground shaking alone. Shaking intensity is highly
affected by regional and near-surface geology (soil/rock type, etc.), and it is difficult to compare
deep and shallow event intensities because deep events produce smaller levels of shaking for the
same size earthquake. Table 2.1 also shows a comparison of approximate magnitudes and PGA
ranges that would be expected for shaking near the epicenter.
Using either scale, various sites can be assigned intensity values and can also be used to develop
an isoseismal map for a particular event. That is, sites with equal intensity values are connected
using contour lines, similar to those that represent data on a topographic map. This type of map is
useful in estimating the general location of the epicenter of an earthquake and its magnitude (as
shown in Table 2.1 for MM), particularly earthquakes that predate instruments. Additional infor-
mation that can be inferred from these maps include the ground conditions at a particular site, the
underlying geology, radiation pattern of seismic waves, and response of different types of buildings.

2.2.2 Earthquake Magnitude
Whereas the intensity of a given earthquake varies from one observation point to another, earth-
quakes can be associated with a single value of magnitude. The consensus measure of magnitude is
based on the Richter scale, which quantifies the size of an earthquake with an index of the amount
of energy released, and while this approach is an improvement as compared to the intensity scales,
the Richter scale does not accurately account for all factors that contribute to the actual size of an
earthquake. This scale, however, does appropriately measure the relative strength of an earthquake
and remains an important parameter in earthquake hazard analysis. In addition to being used for
earthquake hazard analysis by seismologists and engineers, it is the preferred scale used to inform
the public of the size of an earthquake.
Because it was originally developed to quantify the strength of Southern California earthquakes,
the Richter scale is also known as the local magnitude scale, M L . Richter defined M L using the
base-10 logarithm of the peak trace amplitude (in micrometers, µm) of a standard Wood–Anderson
seismograph (which has a magnification factor of 2800, a natural period of 0.8 s, and damping of
80%), located on firm ground at a distance of 100 km from the epicenter. The following relationship
gives the local magnitude:

A
M L = log10
Ao

where:
A is the peak amplitude in µm, measured from a seismogram
Ao is the peak amplitude of a zero-magnitude earthquake in µm, which is used to adjust the
variation of ground motion amplitude for epicentral distances other than 100 km

This equation is typically not used directly to determine the magnitude of an earthquake; instead,
a correction nomogram provided by Richter is used (see Bolt 2004). The nomogram was convenient
when digital calculators were not available. The equation used to develop the nomogram is

M L = log10 A + 3 log10 (8∆t p−s ) − 2.92 (2.1)

where:
A is now given in millimeters (mm)
Δtp−s is the time between the arrival of P- and S-waves in seconds (see Figure 2.6); this
­indirectly measures the epicentral distance
26 Introduction to Earthquake Engineering

The Richter scale has a practical range from 0 to 9.0 (theoretically, the scale has no upper or
lower limits). Also, the base-10 logarithmic scale indicates that each unit increase in M L corresponds
to a 10-fold increase of the earthquake wave amplitude. For example, a 7 M L earthquake is 100
times stronger than a 5 M L event (10 × 10 = 100).

EXAMPLE 2.1
Estimate the local magnitude of a southern California earthquake recorded in two perpendicular
directions at several stations using standard Wood–Anderson seismographs. The trace amplitudes
and epicentral distances are as follows (Richter, 1958):

Amplitude on the Seismograph, A (µm)

Epicentral Distance, Δ
Station N–S Component E–W Component (km)
1 8400 6000 114
2 7900 8500 179
3 24,500 30,000 90
4 8100 7000 246

Solution
Equation 2.1 can be rewritten in a slightly different form to include the epicentral distance rather
than Δt p–s, ML = log10 A + 2.56 log Δ − 5.12 (this type of equation is dependent on the seismologi-
cal station and still yields significant scatter of the data). This equation can be used to determine
ML for the two perpendicular directions, which can then be averaged to obtain a sufficiently
­accurate result without having to combine the results vectorially. The calculations for Station 1 in
the north–south and east–west directions are as follows:

MLN−S = log10 AN−S + 2.56 log10 ΔN−S − 5.12

MLN−S = log10 (8.4 ×103 μm) + 2.56 log10 (114 km) − 5.12 = 4.07

MLE−W = log10 (6 ×103 μm) + 2.56 log10 (114 km) − 5.12 = 4.92

The average of these two values gives ML at Station 1. That is,

4.07 + 3.92
ML1 = = 4 .0
2

All other stations are treated similarly and the results are summarized as follows:

Local Magnitude

Station MLN– S MLE–W Average ML

1 4.07 3.92 4.0
2 4.54 4.58 4.6
3 4.27 4.36 4.3
4 4.91 4.84 4.9

Therefore, the magnitude of this particular earthquake can be determined as the average of the
averages of each station:

4 . 0 + 4 . 6 + 4 . 3 + 4 .9
ML = = 4 .4
4
Engineering Seismology Overview 27

The local energy associated with earthquakes during fault fracture growth is primarily trans-
formed into heat, with only 1%–10% being released as seismic waves. The relationship between the
fraction of energy radiated as waves and the local magnitude, M L, is given as follows:
log10 E = 11.8 + 1.5ML (2.2)

where:
E is in ergs, which is a relatively small unit, 1 ft lb = 1.356 × 107 ergs

In Equation 2.2, the logarithmic scale coupled with the other terms results in a 32-fold increase
in energy radiated for each unit increase in M L . For example, a 7 M L earthquake radiates 1000 times
more energy than a 5 M L event (32 × 32 = 1024). Since this relationship was calibrated using vari-
ous types of seismic waves, it is also applicable to other magnitude scales, including those discussed
in the following paragraphs.
The local magnitude scale does not accurately account for deep focal depths (>45 km) or long
epicentral distances (>600 km) because seismic waves attenuate with distance (whether the distance
is depth or length). The wave train recorded on a seismograph for a deep earthquake is very differ-
ent than for a shallow one, leading to two different values of M L even for events that release equal
amounts of total energy. Consequently, more accurate measures of magnitude had to be developed
in order to improve the uniform coverage of earthquake size. Richter working with his colleague
Gutenberg addressed the shortcomings of the local magnitude scale by developing the body-wave
scale, mb, to handle deep-focus earthquakes and the surface-wave scale, Ms, to handle distant earth-
quakes. The three scales are related by the following empirical relationships:

M s = 1.27( ML − 1) − 0.016 ML2 (2.3)

and

mb = 0.63M s + 2.5 (2.4)

Unfortunately, even these scales do not provide an accurate estimate for the largest magnitude
(>8M L ), most devastating earthquakes. The limitations of the scales are purported to be the result
of ground-shaking characteristics (such as the natural period) not increasing proportionally with
increasing amount of the total energy radiated by very large earthquakes. M L, mb, and Ms are valid
for natural periods of up to 0.8 s (for a Wood−Anderson seismograph), 1.0 s, and 20 s, respectively.
Because of these ground-shaking limitations, earthquakes larger than certain sizes will have a con-
stant magnitude. This is known as magnitude saturation. For instance, M L begins to saturate at
about 6.5, mb at about 7.0, and Ms at about 8.0.
The strength of an earthquake can be more accurately measured using the moment magnitude
scale, MW, which accurately measures a wide range of earthquake sizes and is applicable globally.
This scale is a function of the total moment release by an earthquake. The moment is a measure of
the total energy released. The concept of moment is adopted from mechanics and is defined as the
product of the fault displacement and the force causing the displacement. A simple derivation of the
moment, M0, is presented in Villaverde and is based on the size of the fault rupture, the slip amount,
and the stiffness of the fractured rocks. That is,

M 0 = GAf Ds (2.5)

where:
M0 is in dyne cm, which is a relatively small unit, 1 dyne cm = 1 × 10−7 N m
G is the shear modulus of the rocks included in the fault in dyne/cm2, which ranges from
3.2 × 1011 dyne/cm2 in the crust to 7.5 × 1011 dyne/cm2 in the mantle
28 Introduction to Earthquake Engineering

Af is the area of the fault rupture in cm2

Ds is the average fault slip or displacement in cm

All of which can be estimated relatively accurately. Alternatively, the moment can be directly
estimated from the amplitudes of long-period waves at large distances, with corrections for attenu-
ation and directional effects. Although moment is an effective way to establish the size of an earth-
quake, it is customarily convenient to convert it into a magnitude quantity so that it can be compared
to M L, mb, and Ms. This can be accomplished by relating the seismic moment M0 to the radiated
energy E. That is,

∆σ
E= M0 (2.6)
2G

where:
Δσ is the static stress drop in the earthquake, which ranges from 30 to 60 bars (1 bar = 100 kPa)
G is the shear modulus of the medium near the fault, the same as used in Equation 2.5

The moment magnitude, Mw, is then derived by substituting Equation 2.6 into Equation 2.2 and
replacing M L with Mw since the two scales give similar values for earthquakes of magnitude ranges
from 3 to 5. For average values of Δσ and G (Δσ/G ≅ 10 –4), the relationship between M0 and Mw is
then given as

2
Mw = log10 M 0 − 10.7 (2.7)
3

where:
M0 is in dyne cm

The resulting Mw does not saturate at large magnitude values. Mw is also known as the Kanamori
wave energy, after the scientist who developed this relationship (Villaverde, 2009).

EXAMPLE 2.2
Estimate the seismic moment and moment magnitude of the January 12, 2010 Haiti earthquake.
It is estimated that the blind thrust fault (the slip plane ends before reaching the earth’s surface)
caused an average strike-slip displacement of 2 m over an area equal to 30 km long by 15 km
deep (Eberhard et al. 2010). Assume that the rock along the fault has an average shear rigidity of
3.2 × 1011 dyne/cm2.

Solution
1. Determine the fault’s rupture area A f and fault slip, Ds, in consistent units.

Af = (30 ×105 cm)(15×105 cm) = 4.5×1012 cm2

Ds = 2 m = 20 cm

2. Determine the seismic moment M0 using Equation 2.5.

M0 = GAf Ds
= (3.2×1011 dyne /cm2 )( 4.5 ×1012 cm2 )( 200 cm)
= 2.88 ×10 26 dyne cm

Engineering Seismology Overview 29

3. Determine the moment magnitude, Mw, using Equation 2.7.

Mw = 2/ 3log10M0 – 10.7
= 2/ 3log10 (2.88 ×10 26 dyne cm) – 10.7
= 6.94

It is worth noting that a magnitude 7 Mw has been widely reported in the media.

There are other magnitude scales in use that are not covered in this section; thus, it is important
to specify the type of magnitude scale being used. In this book, M L is used unless otherwise noted.

2.3  EFFECTS OF EARTHQUAKES ON STRUCTURES

To design structures that pose low risk to the public, engineers must understand the behavior of
these systems when subjected to various loads, including seismic loads. Strong shaking during an
earthquake is responsible for most damage to structures except in cases where structures are built
directly on a fault line, in which case the shaking and ground rupture contribute to the damage.
The severity of shaking at a site is controlled primarily by two factors: earthquake characteristics
and attenuation of ground motion. Attenuation is the dissipation of seismic energy as seismic waves
move through layers of varying soil and rock strata. Also, once at a site, seismic waves can be
­modified by the site characteristics. Furthermore, during an earthquake, different structures will
behave differently, so characteristics of the structural system have a major influence on seismic
structural performance. Therefore, damage to a structure is primarily controlled by the characteris-
tics of the earthquake, site, and structure.

2.3.1 Earthquake Characteristics
Although the primary factor controlling an earthquake’s effect on structural behavior at a particular
site is the PGA, the largest PGA earthquakes do not always cause the most damage from shaking.
Shaking is rather complex and it is difficult to characterize its full effect using a single parameter.
Strong shaking also significantly depends on duration and frequency content, and to a lesser extent
on factors such as length of fault rupture, focal depth, orientation of the fault, speed of rupture, and
whether or not fault rupture reaches the ground surface.
As discussed earlier, the size of an earthquake can be measured using various parameters
(intensity, magnitude, etc.), the most important of which is PGA when it comes to seismic-resistant
design. This is obtained from the acceleration time history recorded with an accelerometer (which
is a recording of the variation of acceleration amplitude with time) and is typically given as a
­fraction of the acceleration due to gravity. PGA is generally largest near the epicenter and tends to
decrease away from it because of attenuation. However, PGA can increase significantly depending
on a number of factors, especially the geology of the path of the waves to a specific site and local
soil ­conditions (to be discussed in the next section). Ground acceleration is generally resolved into
a vertical and two horizontal components. The vertical component is approximately one-third of
either of the horizontal components; thus, the majority of the damage is caused by the horizontal
components of acceleration. Nonetheless, the vertical component of acceleration is also included in
modern seismic design codes, as discussed in Chapter 8.
As mentioned earlier, large PGA earthquakes do not always create the most damage. In fact, a
large PGA earthquake would produce little damage to some structures if it only occurs for a short
time, whereas a relatively small PGA earthquake that continues for several seconds can be devas-
tating to some structures. This can be exemplified using the first two earthquake records in Figure
2.8; whereas the PGA for the 1940 El Centro earthquake (PGA = 0.32g, M = 7.0) is smaller than
30 Introduction to Earthquake Engineering

that from the 1966 Parkfield earthquake (PGA = 0.5g, M = 6.2), the resulting damage from the El
Centro earthquake was much more extensive than that associated with the Parkfield earthquake.
This can largely be attributed to the much longer strong-motion duration of the El Centro earth-
quake. The duration of strong motion (or shaking) is important because the longer the strong motion,
the more energy is imparted to the structure; and because of the limited capacity of a structure to
absorb energy, there is a greater potential for inelastic response for longer strong shaking. Since
inelastic damage accumulates from load reversals, an earthquake with a large number of load rever-
sals would result in more damage compared to an earthquake with only a few larger load reversals.
It is important to note that the duration of strong shaking is not the duration of the entire earthquake
event, which is much longer, as can be seen in acceleration time history data in Figure 2.8. While
there are a number of procedures that have been proposed to define the strong-shaking portion of
an acceleration time history or accelerogram, none have garnered the consensus of the seismologi-
cal or engineering communities. One such procedure for determining strong-motion duration is the

PGA = 0.32g PGA = 0.5g

Parkfield 1966
El Centro 1940 - S00E
No. 2 - N65E

PGA = 1.25g Lytle Creek 1970 - S65E San Francisco 1957

State Building - S09E

San Fernando 1971

Pacoima Dam - N76W
Managua 1972 - East
1.0

Olympia 1949 - N86E 0.5

Acceleration, g

0.0
Helena 1935 - S90W Stone Canyon 1972
Melendy Ranch - N29W
Koyna 1967 - Long –0.5
PGA = 0.84g
PGA = 0.65g –1.0

Loma Prieta 1989

Northridge 1994
Corralitos - CHAN1 : 90 deg
Sylmar County hospital
Parking Lot - CHAN3 : 360 deg

Chile 1985, Llolleo - N10E

Mexico City, 1985 - SCT S00E

0 10 20 30 40 50 60 70 80
Time (s)

FIGURE 2.8  Ground motions plotted using same horizontal and vertical scales. (Reproduced from Chopra,
A. K., Dynamics of Structures: Theory and Applications to Earthquake Engineering, 4th edition, Prentice-
Hall, Upper Saddle River, NJ, 2012. By permission of Prentice-Hall, Upper Saddle River, NJ.)
Engineering Seismology Overview 31

bracketed duration, which quantifies strong-motion duration as the time elapsed between the first
and last peak acceleration amplitudes that exceed a specified value (Bolt proposed 0.05g).
The dominant period at which the strong shaking occurs is often complementary and interrelated
to both PGA amplitude and strong-motion duration. Small PGA amplitude earthquakes cause little
damage to some structures because stresses may not be large enough, and even large PGA amplitude
earthquakes with short strong-motion duration may not cause enough stress reversals to cause any
significant structural damage. However, a relatively small PGA that continues for several seconds at a
uniform frequency can be devastating to some structures, particularly if the dominant period coincides
with the natural period of the structure, which results in a resonant condition, leading to severe damage.
This is the reason for the devastating damage experienced by some buildings during the 1985 Mexico
City earthquake, which had an epicenter approximately 400 km from Mexico City. The shaking in
poor soil areas was devastating to 10–20-story buildings because the frequency content was filtered as
the seismic waves entered these areas, leaving a dominant period of shaking of approximately 2 s (see
Figure 2.8). The dominant period of shaking coincided with the natural period of 10–20-story build-
ings, leading to a resonant condition; interestingly, other nearby buildings suffered little or no damage.

2.3.2 Site Characteristics
As discussed earlier, the strength of shaking usually diminishes as waves travel away from the focus
because of attenuation. Surface waves usually experience the same effect; this results in smaller
shaking for sites at greater distances from the epicenter. However, shaking can also strengthen
depending on the geologic conditions along the path of the wave between the focus and the site, as
well as the soil conditions and topography at the site. Soil condition and topography, in particular,
can greatly affect the characteristics of the input motion in terms of the amplitude and natural
period of the shaking at the site. A soft subsurface deposit or steep topography (ridge or hill) causes
the ground to vibrate like a flexible system, while a stiff bedrock causes the ground to vibrate
as a rigid body. That is, hard rock vibrates with the same frequency and amplitude as the input
motion, while the vibration of a flexible underlying deposit would depend on its inherent stiffness
and damping characteristics (in the same way as the dynamic response of a structure is affected by
its dynamic characteristics, as discussed in Section 2.3.3). For this case, the stiffness of the under-
lying material has the greatest effect on the characteristics of the input motion at a site (assuming
effects of damping to be negligible). Low-frequency (long-period) input motions are amplified more
at sites underlain by soft soils than those underlain by stiff ones, whereas stiff soil deposits amplify
high-frequency (short-period) input motions more than soft ones.
The inherent stiffness of the ground beneath a site is a function of its shear-wave velocity, and
thickness of the sediment above the bedrock. Shear-wave velocity is faster for hard rock than for soft
soil; consequently, seismic waves travel faster through hard rock than through soft soil. So when seis-
mic waves pass from rock into soil, they slow down. This slower speed must be accompanied by an
increase in wave amplitude (amplifying the ground shaking) in order for the flow of energy to remain
constant, which is required for the conservation of elastic wave energy. An increase in soft sediment
thickness has a similar effect (increased amplitude) because of a decrease in material density.
The most spectacular example of a deep flexible soil deposit amplifying low-frequency motions
occurred during the 1985 Mexico City earthquake. As discussed earlier (Section 1.1.3), a relatively
small rock acceleration caused a high PGA at the ground surface underlain by a very soft, high-
water-content mixture of clay and sand. The most severe damage was limited to an area near the city
center (the so-called Lake Zone), with little or no damage to buildings built on solid volcanic rock
just a few kilometers away (the so-called Foothill Zone); see Figure 1.8. The PGA for the Lake Zone
was about five times greater than that in the Foothill Zone. The frequency content for the two zones
was also very different; the Lake Zone had predominant periods of about 2 s, which was about 10
times larger than that in the Foothill Zone. Also, as shown in Figure 2.8, the strong shaking in the
Lake Zone continued for a long time. The duration of strong shaking along with the 2 s site period
32 Introduction to Earthquake Engineering

contributed to a peculiar damage pattern. Minor damage occurred to buildings of less than five sto-
ries and those of more than 30 stories; most other buildings collapsed or sustained extensive dam-
age. A simple rule of thumb for estimating the fundamental period of an N-story building is N/10 s
(Equation 3.19), which is the reason that buildings of about 20 stories collapsed; their fundamental
period matched that of the site ground motion period, creating a resonance condition. In effect,
these buildings suffered a double resonance condition: the amplification of the bedrock motion by
the soil deposit and the amplification of the soil motion by the structure.
During the 1989 Loma Prieta earthquake, the San Francisco Bay Area (some 100 km north
of the epicenter) experienced more damage than the epicentral region. The reason for the exten-
sive damage along the margins of the bay is because of the amplification of the PGA at soft-soil
sites. The bay basin is filled with an alluvial soil deposit of silts and clays, along with some layers
of sandy and gravelly soils. The upper deposit, or young mud, is composed of silty clay, known
as San Francisco Bay Mud. This material is of loose-to-medium density, and very compressible.
As shown in Figure 2.9, the ground stratum in the vicinity of the bay is divided into three zones
based on underlying material seismic response, from soft mud to hard rock. From Figure 2.9, it is
clear that the attenuation of seismic waves varied considerably in different regions of the bay due
to the ­varying subsurface soil conditions. The rock material attenuated the seismic waves relatively
quickly compared to the soft young bay mud found at the margins of the bay. The areas underlain by
this bay mud experienced the most damage because the soft mud greatly amplified ground motions.
These two examples clearly show that local site characteristics, particularly different underlying
ground types, greatly influence the seismic wave frequency content and amplitude, as will be shown
in Chapter 5. The National Earthquake Hazards Reduction Program (NEHRP) defines six ground
types based on shear-wave velocity in order to establish amplification effects for design purposes:

1. Type A, hard rock (igneous rock)

2. Type B, rock (volcanic rock)
3. Type C, very dense soil and soft rock (sandstone)
4. Type D, stiff soil (mud)—it is conservative to assume this when no soil properties are
known
5. Type E, soft soil (artificial fill)
6. Type F, soils requiring site-specific evaluations

Type A has the largest stiffness and generates the smallest amplifications, while Type E is the
softest and generates the largest amplifications.

2.3.3 Structural Characteristics
Once seismic waves have been modified for ground conditions at the site of a structure, the struc-
ture responds to ground excitation based on its inherent characteristics, as well as the age of the
structure and the quality of its construction. The characteristics of a structure that have the largest
influence on its response include its weight and stiffness, and they will be covered in later chapters.
Damping can play a significant role when structures remain within the elastic limit, but remains a
much smaller effect for structures that deform into inelastic behavior as will be covered in Chapter
5. Although all these characteristics affect the displacement amplitude and natural period of a struc-
ture, stiffness has the greatest effect on structural response. A stiff structure vibrates with the same
frequency and amplitude as the input motion, while a flexible structure may or may not vibrate in
sync with the input motion, depending on how similar its natural period is to the input excitation
period. For  instance, long-period (low-frequency) input ground excitations amplify the deforma-
tions of flexible structures more than those of stiff ones, whereas short-period (high-frequency)
input ground excitations amplify deformations of stiff structures more than those of soft ones. This
structural response will be discussed in detail in later chapters.
Engineering Seismology Overview 33

.11 .08 USGS

.06 .16 .26 .12 CSMIP

.07 .29 HAYWARD FAULT
.24 .08 .26
.18 Bay Mud
.16
.21 .13 .09 .04
Alluvium
.11 .06
.09 Rock and shallow residual soil
.12
.11 .10
.14
.09
.08
.15 .16 N
.33 .06
.10 .06
.12
.29 .20 CALAVERAS FAULT
.09
.13
.11
.28
.16 .21
.27
.10 .12 .13

.08 .14 .09

.29
.22
.38 .17 .18
.11 .13
SAN ANDREAS .13
FAULT

.53 .33

.45
SAN GREGORIO FAULT

APPROX. ZONE .26 .19

OF RUPTURE
.49
ZAYANTE FAULT

.64 .33
Corralitos - CHAN1 : 0 deg .28
.47 EPICENTER .55
.50
.37
.54H
.60V .39H
Scale .56V .2

0 10 20 Miles .15 .25

FIGURE 2.9  Horizontal PGA for the 1989 Loma Prieta earthquake (CSMIP = California Strong Motion
Instrumentation Program). (Reproduced from Chopra, A. K., Dynamics of Structures: Theory and Applications
to Earthquake Engineering, 4th edition, Prentice-Hall, Upper Saddle River, NJ, 2012. By ­permission of
Prentice-Hall, Upper Saddle River, NJ.)

The age and construction of structures play a major role in their performance during an
e­ arthquake. Older and/or poorly constructed structures are at a greater risk of failure during an
­earthquake because of inadequate ductility, which is the ability of a structure to absorb the seismic
energy ­without failure. A structure can include seismic detailing to satisfy the ductility demand
during an earthquake, and to provide an adequate path for the seismic load to the foundation.
Addressing ­ductility demand and load path through proper detailing can allow a structure to per-
form adequately during even major earthquakes. The ductility demand can be addressed by allow-
ing the structure to behave in the inelastic range, provided that the stability of the overall system is
not compromised. An alternative approach to resisting seismic loads is to permit the movement of
structures; these systems are known as compliant systems.
34 Introduction to Earthquake Engineering

Seismic loads on structures result from inertial forces created by the ground accelerations. Based
on Newton’s second law (force equals mass times acceleration), the magnitude of these forces is a
function of the structure’s mass (or weight). Also, the acceleration is a function of the characteristics
of the input ground motion, particularly intensity, duration, and frequency content. The hazard a
given earthquake risk possesses to structures of different functions is accounted for by specifying
higher loads for structures of increased importance. Importance is classified in a range from unoc-
cupied buildings to critical structures such as hospitals and schools; as discussed in Section 8.4.2.

2.4  EARTHQUAKE HAZARD ASSESSMENT

The intensity and magnitude scales are simplified methods for reporting the relative strength of an
earthquake; however, these quantities do not necessarily provide a measure of seismic hazard at a
particular site because not all regions are susceptible to strong earthquakes (small earthquakes occur
all the time along active faults and generally do not pose a hazard to structures or human life). Also,
intensity and magnitude are the measurements of past earthquakes and might not properly represent
future ones. However, past regional seismicity (rate of occurrence of earthquakes) can be used to
establish active seismic faults and forecast future seismic activity at a site. There are two quantitative
methods that have been developed to characterize seismic hazard at a particular site: deterministic,
where a particular earthquake is assumed, and probabilistic, where statistical methods are explicitly
used to account for uncertainties in earthquake strength, location, and time of occurrence.
Even in highly seismic regions, the movement of the ground during an earthquake does not
present a significant hazard to people (as discussed in Chapter 1, secondary effects such as fires,
landslides, and tsunamis can). The highest risk to people in seismically active regions is shaking
of buildings and other structures, particularly when structures collapse. The two methods used to
characterize seismic hazards (deterministic and probabilistic) primarily focus on shaking effects.
The deterministic approach can be used in areas where seismic activity is frequent and its sources
are well defined. The process is relatively simple and entails the following:

1. Identifying nearby seismic sources

2. Identifying the distance to the structure site from nearby seismic sources
3. Determining the magnitude and characteristics of nearby seismic sources
4. Establishing the structural response to the effects from all nearby seismic sources
5. Selecting the case that produces the largest structural response.

This method is too conservative because it does not properly account for uncertainties. For this
reason, the probabilistic method was developed.
The probabilistic approach is similar, but includes the uncertainties in each step. The probabilis-
tic approach, although quite complicated, has been incorporated in the various seismic hazard maps
used in contemporary seismic design codes. These codes have incorporated probabilistic seismic
hazard zone maps based on past geographic seismology to minimize risk to human life. These maps
depict where damage might occur, based on the probability of the occurrence of strong shaking and
the local geology (areas underlain by deep, loose sediments are at a greater risk for damage than
those directly over solid rock); see Figure 2.10. These maps can be used to draw probabilistic design
spectra, which will be discussed in detail in Chapters 5 and 8. With this information, structural
engineers can plan and design structures at a particular site using the methods covered in Chapter 8.
The key issue in the probabilistic approach found in the seismic design codes is the probability
of exceedance of a particular event for a specified intensity, over a certain period of time. The codes
consider three different earthquake levels for design: the maximum design earthquake (MDE),
which has a 10% probability of being exceeded over a period of 50 years; one with 5% probability
of being exceeded over a period of 50 years; and the maximum considered earthquake (MCE),
which has a 2% probability of being exceeded over a period of 50 years. Also, an earthquake
Engineering Seismology Overview 35

50° –120° 50°

–70° %g
–110° –80°
–100° –90°

60

30
40° 40°
20

10

5
30° 30°
2

–120° –70°
–110° –80°
–100° –90°
Horizontal ground acceleration (%g)
with 2% probability of exceedance in 50 years

FIGURE 2.10  Probabilistic seismic hazard map showing ground shaking hazard zones. (From USGS
National Seismic Hazard Mapping Project website: geohazards.cr.usgs.gov.)

that has a significant chance of occurring during the life of a structure is known as the maximum
­probable earthquake (MPE), and has a 50% probability of being exceeded over a period of 50 years.
Designing for the MCE provides the highest level of protection and is reserved only for the most
important or critical structures. These earthquakes can also be characterized in terms of the return
period (RP) using the following relationship:

−T
RP = c (2.8)
ln(1 − P )

where:
P is the probability of exceedance in T years

That is, the RPs in years for the three earthquake levels included in codes and the MPE are
as follows:

−T −50 years
MCE → RP = = = 2475 years
ln(1 − P ) ln(1 − 0.02)
−T −50 years
RP = = = 975 years
ln(1 − P ) ln(1 − 0.05)

−T −50 years
MDE → RP = = = 475 years
ln(1 − P ) ln(1 − 0.10)

−T −50 years
MPE → RP = = = 72 years
ln(1 − P ) ln(1 − 0.50)
36 Introduction to Earthquake Engineering

2.5  EARTHQUAKE PREDICTION

The temporal prediction of earthquakes has been investigated for many years without success. Also,
all signs, such as radon detection and unusual animal behavior, that have been claimed as portend-
ing earthquakes have been proven wrong. However, sensitive instruments can detect the arrival of
P-waves that are followed by the devastating S-waves. This is the current basis for proposed early
earthquake detection systems. In some parts of the world, these systems have already been deployed
and have successfully provided warnings to residents of afflicted areas, albeit short warnings.
Although it is unlikely that we will develop the ability to predict when earthquakes will occur, it
is generally possible to predict where earthquakes can occur. Great advances have been made in the
past century to map the most seismic-prone regions of the world, since past seismic activity in these
regions has provided the information needed to map many faults. It is now possible to predict the
spatial occurrence of earthquakes relatively accurately, including the possible strength of the event,
as well as forecasting the expected average rates of damaging earthquakes. Also, because of the
instrumental records from past earthquakes, we have been able to map subsurface conditions, which
have allowed us to predict the characteristics of the ground motions that may occur at a particular
site from a statistical stand point.

PROBLEMS
2.1 A Richter magnitude earthquake of 6.8 would radiate how much more energy than an
M L 5.8 earthquake?
2.2 A Richter magnitude earthquake of 5.0 is how much stronger than an M L 3.0
earthquake?
2.3 Estimate the local magnitude of a southern California earthquake recorded using a
­standard Wood–Anderson seismograph that shows a trace amplitude of 23 mm, and the
P- and S-wave arrival times at the recording station of 07:19:45 and 07:20:09, respectively.
2.4 An earthquake at a transformed fault caused an average strike-slip displacement of
2.5 m over an area equal to 80 km long by 23 km deep. Assuming the rock along the
fault has the average shear stiffness of 175 kPa, estimate the seismic moment and
moment ­magnitude of the earthquake.
2.5 Estimate the seismic moment and moment magnitude of the February 27, 2010 Chile
earthquake. It is estimated that the thrust fault (the slip plane ends before reaching the
earth’s surface) caused an average strike-slip displacement of 5 m over an area equal
to 600 km long by 150 km deep. Assume the rock along the fault has the average shear
rigidity of 3 × 1011 dyne/cm2.
2.6 What is the energy difference between the Haiti and Chile earthquakes from
Example 2.2 and Problem 2.5, respectively?

REFERENCES
Bolt, B. A., Earthquakes, 5th edition, W.H. Freeman and Co., New York, 2004.
Chopra, A. K., Dynamics of Structures: Theory and Applications to Earthquake Engineering, 4th edition,
Prentice-Hall, Upper Saddle River, NJ, 2012.
Eberhard, M. O., S. Baldridge, J. Marshall, W. Mooney, and G. J. Rix, The MW 7.0 Haiti earthquake of
January 12, 2010; USGS/EERI Advance Reconnaissance Team Report: U.S. Geological Survey Open-
File Report 2010-1048, 58 p, 2010.
Kramer, S. L., Geotechnical Earthquake Engineering, Prentice-Hall, Upper Saddle River, NJ, 1996.
Richter, C. F., Elementary Seismology, W.H. Freeman and Co., San Francisco, CA, 1958.
Villaverde, R., Fundamental Concepts of Earthquake Engineering, CRC Press, Boca Raton, FL,
2009.
 3 Single-Degree-of-Freedom
Structural Dynamic Analysis
After reading this chapter, you will be able to:

1. Determine the weight of a structure

2. Determine the stiffness of a structure
3. Determine the natural angular frequency of a structure
4. Determine the natural period of a structure
5. Determine the natural cyclic frequency of a structure
6. Perform a dynamic analyses of damped and undamped single-degree-of-freedom systems
7. Perform a dynamic analyses of damped and undamped single-degree-of-freedom systems
subjected to harmonic time-dependent forces
8. Determine deflections and stresses induced in a structure by motions of its supports

Traditional seismic design follows two approaches: one based on static analysis and the other
one based on dynamic analysis (see Chapter 8). The static analysis is formulated by estimating an
­equivalent lateral seismic load as a fraction of the total system weight; the fraction is determined from
a notional earthquake, which is based on the risk posed by the seismicity of the site and is captured in
a single graph known as the response spectrum (to be discussed in Chapters 4 and 5). The dynamic
analysis procedure models the load effect of earthquakes more accurately by using either a set of spe-
cific earthquake ground motion time-history records (response time-history analysis procedure) or the
effect of the same notional earthquake as the equivalent static analysis (the design response spectrum),
but using a modal analysis (modal response spectrum analysis procedure). In the response time-­history
analysis procedure, a linear or nonlinear mathematical model of the system is used to determine the
system response (displacements and accelerations, which are used to determine the internal loading in
each member) at each increment of time for a suite of ground motion acceleration time histories. With
the full history of the response, the absolute maximum internal loads can be determined by combin-
ing the results of the set of ground motion acceleration time histories; these loads can then be used to
design each member. The modal response spectrum analysis procedure uses established methods of
structural dynamics to determine system vibration mode shapes and their associated natural periods,
which along with the response spectrum are used to establish the maximum structural response. With
this maximum response, the maximum internal loads can be determined and used to design each
member. In this chapter, we explain the essential concepts of structural dynamics, also known as
vibration theory, needed to understand both dynamic seismic analyses procedures.

3.1  UNDAMPED SINGLE-DEGREE-OF-FREEDOM SYSTEM

The concept of degree-of-freedom (DOF) is central to the understanding of vibration theory.
A DOF in the context of vibration theory is an independent displacement or rotation of a point on
a structure, which means that even the simplest case could potentially have an infinite number of
DOFs. However, determination of the number of DOFs depends on the idealization of the structural
system. The ends of a member are often points of interest; for example, consider a portal frame with
rigid connections between columns and beam and fixed supports at the base of each column. The
ends of the column at the supports would have zero DOFs, but three DOFs at each of the connec-
tions between beam and column (see Figure 3.1). These connections must both move and rotate in

37
38 Introduction to Earthquake Engineering

u
u
m

FIGURE 3.1  Idealized single-degree-of-freedom system for a portal frame.

sync, resulting in six DOFs. However, for our discussion in this chapter, the beam will be assumed
to be much stiffer than the columns, preventing rotations of the connections. Also, the two columns
move in parallel, and if we assume that the axial deformations of the columns and beam are rela-
tively small, we can reduce the entire system to a single-degree-of-freedom (SDOF) case as shown
in Figure 3.1. Finally, note that there is no damping in the system, the lack of which would cause the
system to move in perpetuity once set in motion. We will cover damping in Section 3.2.

3.1.1  Free Vibration Response of Undamped Systems

Free vibration is caused by initial conditions (displacement and velocity) and not by applied force.
To characterize the free vibration motion of the portal frame idealized as an SDOF system, we first
need to formulate a mathematical model in terms of the displacement as a function of time. This can
be accomplished by establishing an equation of motion for the system. This equation can be deter-
mined by applying equilibrium to a free-body diagram (FBD) of the mass using D’Alembert’s
­principle, where the inertial force developed by the mass (from Newton’s second law, F = ma)
is included in the FBD as shown in Figure 3.2. Note that this is not a complete FBD because the
­internal axial loading and the internal moment in the column stem are missing.
Horizontal equilibrium of the FBD shown in Figure 3.2 yields the equation of motion,

+→ ∑ F = 0;
x − mu − ku = 0 ⇒ mu + ku = 0

where:
m is the mass of the system, discussed in more detail in Section 3.1.2
k is the lateral stiffness, discussed in more detail in Section 3.1.3

The double dot over the u indicates differentiation with respect to time, that is,

d 2u
u =
dt 2

The equation of motion, a second-order, linear, and homogeneous differential equation with
constant coefficients, can be written in the following form:

u + ωn2u = 0 (3.1)

m mü

V=ku

FIGURE 3.2  Free-body diagram of a portal frame SDOF model.

Single-Degree-of-Freedom Structural Dynamic Analysis 39

where for convenience a new parameter is introduced, the natural circular frequency (from now on
called natural frequency) with units of radians per second (rad/s),

k
ωn = (3.2)
m

The solution to the second-order, linear, and homogeneous differential equation of motion is of
the form

u(t ) = eλt

This equation can be differentiated with respect to time to obtain the velocity,

u(t ) = λ eλt

and the velocity can be differentiated with respect to time to obtain the acceleration,

u(t ) = λ 2eλt

substituting the equations for displacement and acceleration into the equation of motion yields

λ 2eλt + ωn2eλt = 0

which can be rearranged as

(λ2 + ωn2 )eλt = 0
In general, the exponential function is not zero, thus the quantity within parenthesis must be zero
in order to determine a valid solution to the differential equation,

(λ2 + ωn2 ) = 0
We then solve for the unknown parameter λ,

λ = ±ωni

where i is the imaginary unit of a complex number given as i = −1. As the equation of motion is a
second-order differential equation, two constants of integration are needed, A1 and A2,

u(t ) = A1eωn it + A2e−ωn it

This equation can be expressed in a polar form (in terms of sines and cosines) by making use of
Euler’s identities:

eiθ = cos θ + i sin θ and e−iθ = cos θ − i sin θ

40 Introduction to Earthquake Engineering

The result is

u(t ) = A1 (cos ωnt + i sin ωnt ) + A2 (cos ωnt − i sin ωnt )

which can be rewritten as

u(t ) = ( A1 + A2 ) cos ωnt + i( A1 − A2 )sin ωnt

As the two trigonometric functions are real-valued solutions to the equation of motion, we can
express the general solution in real form as

u(t ) = A cos ωnt + B sin ωnt (3.3)

where A and B are arbitrary constants.

We now use initial conditions (at time t = 0) to solve for constants A and B. First, we substitute
the initial displacement (at time t = 0), u(0), into Equation 3.3 in order to determine the constant A,

u(0) = A cos(0) + B sin(0)

which yields A = u(0). Next, we take a time derivative of Equation 3.3 and substitute the initial
 0), in order to determine the constant B,
velocity (at time t = 0), u(

u(0) = − Aωn sin(0) + Bωn cos(0)

which yields B = (u(0) /ωn).

The complete solution is the free vibration of an undamped SDOF system, which describes the
position of the mass as a function of time,

u(0)
u(t ) = u(0) cos ωnt + sin ωnt (3.4)
ωn

A graph of the solution is shown in Figure 3.3. This motion is described as harmonic (and there-
fore periodic) because it is a function of sine and cosine of the same frequency, ωn. The time required
to complete a full cycle (2π) is known as the natural period of vibration, Tn (s), and is determined as
(see Figure 3.3)

2π
ωnTn = 2π ⇒ Tn = (3.5)
ωn

. Tn = 2π /ωn
u u(0)
uo uo
b
a b c d e
uo
u(0)
a c e
t

β/ωn

FIGURE 3.3  Free vibration of an undamped SDOF system.

Single-Degree-of-Freedom Structural Dynamic Analysis 41

The natural cyclic frequency, f n (hertz or cycles/s), is defined as the reciprocal of the natural
period of vibration, Tn (s), and is proportional to the natural circular frequency, ωn (rad/s), that is,

1 ω
fn = = n (3.6)
Tn 2π

Note that these three quantities are related and essentially represent the same physical quantity
expressed in different units.
An alternate formulation for the free vibration response of an undamped SDOF leads to the defi-
nition of two additional parameters introduced in Figure 3.3: the phase angle, β, and the amplitude,
uo. We start with Equation 3.3 and assume that constants A and B are related to new variables C and
β as follows:

A = C cos β and B = C sin β (3.7)

Squaring these two relationships and adding the results yields

A2 + B2 = C 2 (cos2 β + sin 2 β )

The quantity in parentheses is equal to 1.0 (trigonometric identity); thus,

C = A2 + B2 = u(0)2 + (u (0) /ωn )2 = uo

where uo is the maximum amplitude shown in Figure 3.3.

Also, taking the ratio of B to A, we obtain the phase angle, β,

B C sin β
= = tan β
A C cos β

or

 B  u(0) 
β = tan−1   = tan−1  
 A   u(0)ωn 

Substituting Equation 3.7 into Equation 3.3 yields the equation of motion in terms of the ampli-
tude and phase angle, that is,

u(t ) = A cos ωnt + B sin ωnt = C cos β cos ωnt + C sin β sin ωnt

which can be rearranged using the following trigonometric identities:

1 1
cos β cos ωt = cos(ωt − β ) + cos(ωt + β )
2 2

1 1
sin β sin ωt = cos(ωt − β ) − cos(ωt + β )
2 2
42 Introduction to Earthquake Engineering

.
C = u(0)2 + (u(0)/ωn)2
β u(0)
α
.
u(0)/ωn

FIGURE 3.4  Relationship between phase angles and amplitude.

The equation in the new form is

u(t ) = C cos(ωnt − β )

Similarly,

u(t ) = C sin(ωnt + α)

where a new phase angle, α, is introduced,

 u(0)ωn 
α = tan−1  
 u(0) 

The two phase angles and the amplitude are related as shown in Figure 3.4.

3.1.2 Structural Weight
As noted earlier, the circular frequency is proportional to the mass, which in turn is directly pro-
portional to the weight. For seismic analysis of structures, the weight is defined as the total effective
weight W, which is specified in ASCE/SEI (2010) “Minimum Design Loads for Buildings and Other
Structures” standard, (from here on referred to as ASCE-7) as

W = DL + 0.25StL + larger of PL or 10-psf + WPE + 0.2SL + LaL (3.8)

where:
DL is the dead load of the structural system that is tributary to each floor
StL is the storage load, which is a live load in areas used for storage, such as a warehouse
PL is the partition load, when applicable
WPE is the operating weight of permanent equipment
SL is the flat roof snow load when it exceeds 30 psf, regardless of roof slope
LaL is the landscape loads associated with roof and balcony gardens

For partition loading, ASCE-7 requires the given partition load or 10 psf when determining the
effective weight. This is only applicable where the location of partitions is subject to change (such as
buildings designated as “office building” occupancy), where ASCE-7 requires a partition live load
of 15 psf for designing individual floor members for vertical loads, if the floor live load is <80 psf.
For information about estimating the other loads, see ASCE-7 specifications.
Once the dead load is estimated, one can determine the tributary weight of a floor or roof as
shown in Figure 3.5 for a multistory building. The weight of each floor and the tributary wall
loads halfway between adjacent floors is assumed to be concentrated or lumped at each floor level
(see Figure 3.6). The story weight, Wx, for this diagram is given as

Wx = walls (A, B, C) + floor (D) + equipment (E) = WA + WB + WC + WD + WE

Single-Degree-of-Freedom Structural Dynamic Analysis 43

Parapet

Roof W2
Wall and floor DL
tributary to the 2nd floor
of the building
h2
h2
2
1st level
h1 W1
2
h1 Tributary DL at this
level is assumed to be
concentrated at the
Base floor level and is
known as lumped
Wall and foundation DL
act directly on ground

FIGURE 3.5  Building structure and lumped mass simplified model.

Mid-height of C
Normal B
story above wall
A
WC
Shear
WA E wall
WD
WE D WB
Floor

Shear
wall
ke f
ua o
hq on
rt ti

Mid-height of
ea irec

story below
D

FIGURE 3.6  Story weight, Wx for calculating lateral forces.

Here WD includes weight of the structure floor, suspended ceiling, mechanical equipment (unless
taken separately as WE ), and when applicable 10 psf for partitions or the actual partition weight.
In general, foundation weight and half of the first-story wall weight are assumed to act directly on
the ground and are commonly omitted in seismic calculations (though common practice, this is not
an explicit provision in seismic design specifications).

EXAMPLE 3.1
The load pressures for dead, live, and snow loads for the rectangular (30 ft × 40 ft) warehouse
structure shown are provided below. Determine the effective weight at the roof.

Roof DL = 45 psf
Floor DL = 65 psf
Wall DL = 25 psf
Floor LL = 250 psf (storage)
SL = 50 psf
44 Introduction to Earthquake Engineering

Solution
1. Identify the effective height of the roof in order to determine the tributary height of the
walls needed to calculate the effective weight of the walls.
  In this case, the tributary height of the walls is the sum of the height from the roof to
midway to the base floor and the entire height of a parapet wall above the roof level as
shown in Figure E3.1:

h1 16 ft
htributary = + hparapet = + 4 ft =12 ft
2 2

2. Calculate individual weights of components associated with the floor of interest.

  The components contributing to the effective weight of the roof include the walls
­parallel to north–south (N–S) and east–west (E–W) directions, roof load, and snow load.
Note that the storage live load and floor dead load do not contribute to the seismic
weight of the roof because they are associated with the base level and act directly on
the ground.
The weight of the walls is calculated by multiplying the area of the walls with the
wall dead load.
Weight of walls parallel to the N–S direction:

( 40 ft ×12 ft )
WN– S,Walls = AreaN– S,Walls × Wall DL = ( 2 walls) ×25 psf = 24,000 lb
wall

Similarly, for the weight of walls parallel to the E–W direction:

(30 ft ×12 ft )
WE– W,Walls = ( 2 walls) × 25 psf =18,000 lb
wall

The weight of the roof level is calculated by multiplying the roof area with the roof
dead load.

WRoof Level = AreaRoof Level ×Roof DL = (30 ft × 40 ft ) × 45 psf = 54,000 lb

This problem specifies a snow load pressure which is necessary for the seismic weight
calculation of the roof level. Based on Equation 3.7, snow loads exceeding 30 psf are
factored by 0.2 in the seismic weight calculation.

WSnow = AreaRoof Level × (0.2× SL)= (30 ft × 40 ft ) × (0.2× 50 psf) =1

12,000 lb

3. The effective seismic weight for the roof can be determined by summing the weight
contribution from each of the components.
  As the roof is the only level above the base, the effective seismic weight of the roof is
denoted as W1:

4 ft
Roof htributary W

16 ft

Base

FIGURE E3.1  One story schematic for effective weight calculations for an idealized structural model.
Single-Degree-of-Freedom Structural Dynamic Analysis 45

W1 = WN-S,Walls +WE-W,Walls +WRoof Floor +WSnow

W1 = 24,000 lb +18,000 lb + 54,000 lb +12,000 lb =108,000 lb=108 kips

EXAMPLE 3.2
Given the following rectangular (30 ft × 40 ft) two-story warehouse and loading shown, deter-
mine the effective weight at the roof and first floor:
Roof DL = 45 psf
Floor DL = 65 psf
Wall DL = 25 psf
Floor LL = 250 psf (storage)
SL = 50 psf

Solution
1. Identify the effective height of the first floor and the roof in order to calculate the effec-
tive weight of the walls at each level.
  The effective height of the first level is the sum of the heights halfway to the adjacent
levels as shown in Figure E3.2:
h1 h2 16 ft 14 ft
heffective,1 = + = + =15 ft
2 2 2 2
The effective height of the roof is the sum of the heights from the roof midway to
the first level and the entire height of the parapet wall above the roof floor as shown in
Figure E3.2:

h2 14 ft
heffective,roof = + hparapet = + 2 ft = 9 ft
2 2
2. Calculate individual weights of components associated with each level of interest.
First level component weights:
For the first level, the seismic weight includes dead loads from the walls and floor,
whereas the live load is a result of storage load.
Weight of walls parallel to the N–S direction,
( 40 ft ×15 ft )
WN– S,Walls,1 = AreaN– S,Walls,1 × Wall DL = ( 2walls) × 25 psf = 30,000 lb
wall

2 ft
Roof W2
heffective, roof

14 ft

heffective, 1
1st level
W1

16 ft

Base

FIGURE E3.2  Two story schematic for effective weight calculations for an idealized structural model.
46 Introduction to Earthquake Engineering

Similarly, for the weight of walls parallel to the E–W direction,

(30 ft ×15 ft )
WE– W,Walls,1 = ( 2 walls) × 25 psf = 22,500 lb
wall
The floor load components on the first level include the floor dead load and the stor-
age live load,

WFloor DL,1= AreaFloor,1×FloorDL = (30 ft × 40 ft ) × 65 psf =78,000 lb

This problem specifies a storage live load pressure which is necessary for the seis-
mic weight calculation of the first level. Based on Equation 3.7, storage live load, StL, is
­factored by 0.25 in the seismic weight calculation.

WFloor LL,1 = AreaFloor,1× 0.25× StL = (30 ft × 40 ft ) × 0.25× 250 psf =75,000 lb

Roof level component weights:

For the roof level, the seismic weight includes dead loads from the walls and floor
and the snow load. The process for calculating the effective seismic weight is detailed in
Example 3.1.
Weight of walls parallel to the N–S direction,

( 40 ft × 9 ft )
WN– S,Walls,roof = ( 2 walls) × 25 psf =18,000 lb
wall

Weight of walls parallel to the E–W direction,

(30 ft × 9 ft )
WE– W,Walls,roof = ( 2 walls) × 25 psf =13,500 lb
wall

The roof floor load components include the floor dead load and the snow load,

WFloor DL, roof = AreaRoof Floor ×Roof DL = (30 ft × 40 ft ) × 45 psf = 54,000 lb

WSnow = AreaRoof Floor × (0.2× SL) = (30 ft × 40 ft ) × (0.2× 50 psf) =1

12,000 lb

3. Determine the effective seismic weight for each level.

  The effective seismic weight for each level can be determined by summing the weight
of the components. Level 1 is denoted as W1 and the roof level is denoted as W2:

W1 =WN– S,Walls,1 +WE – W,Walls,1 +WFloor DL,1 +WFloor LL,,1

W1 = 30,000 lb + 22,500 lb +78,000 lb + 75,000 lb = 205,500 lb = 205.5 kiips

W2 = WRoof = WN-S,Walls,Roof + WE-W,Walls,Roof + WFloor DL,Roof + WSnnow

W2 = 18,000 lb + 13,500 lb + 54,000 lb + 12,000 lb = 97,500 lb = 97.5 kips

3.1.3 Structural Stiffness
To completely characterize the equation of motion for a linear elastic SDOF system, we need the
lateral stiffness, k. This variable, along with the mass, is needed to determine the circular frequency,
Single-Degree-of-Freedom Structural Dynamic Analysis 47

which once obtained can be used to determine the natural period of vibration of the sys­tem. The
stiffness is defined as the force or moment that results in a unit displacement or rotation at a DOF.

EXAMPLE 3.3
Determine the stiffness for a simply supported beam with a concentrated load F applied at mid-
span. This loading can be considered lateral with respect to the axis of bending.

Solution
1. Determine the deflection at point of interest (or DOF) for the given loading.
  The deflection of the beam can be determined with one of the many structural analy-
ses methods (e.g., double integration, conjugate beam, virtual work, and moment-area)
used to c­ ompute deflections. Also, beam deflections can be obtained from deflection
tables such as those found in Part 4 of the American Institute of Steel Construction
Manual. The midspan deflection for a simply supported beam with a concentrated load
F is shown in Figure E3.3.
2. Determine the stiffness for the given loading and associated DOF.
  The stiffness k is the ratio of force (or moment) to displacement (or rotation). The mid-
span deflection can be rearranged to determine the stiffness,

48EI
k=
L3

where the quantity EI is the flexural stiffness given by the product of the modulus of elas-
ticity, E, and the second moment of the cross-section about the axis of bending (moment
of inertia), I.

Following the approach of Example 3.3, we can derive equations for k for a number of simple
structural systems that can be modeled as SDOF (see Table 3.1 for a list of cases).

EXAMPLE 3.4
A fixed–fixed beam of length, L, with a concentrated weight, W, applied at midspan has a constant
flexural rigidity, EI. Idealize the beam and weight shown in Figure E3.4 into an SDOF system and
determine its natural and cyclic frequencies.

Solution
1. Idealize the structural system.
In order to idealize or model this structural system as a SDOF system, we assume that
the weight of the beam is small relative to the concentrated weight, W. In addition, we
limit the motion of the weight to vertical displacements, u, resulting in the lumped mass
model shown in Figure E3.5.
2. Determine the stiffness parameters.
Considering the displaced shape of the SDOF s­ystem as shown in Figure E3.6, the
static load–displacement relationship can be used to ­determine the stiffness associated
with the DOF.

F
FL3
Δ ∆=
48EI

FIGURE E3.3  Maximum deflection of a simply supported beam with concentrated force at
midspan.
48 Introduction to Earthquake Engineering

TABLE 3.1
Equivalent Stiffness Constants, k
Case Max Deflection, Δ Stiffness, k
Case 5wL4 384 EI
w 384 EI 5L3
Δ

F FL3 192 EI
Δ 192 EI L3

w wL4 384 EI
384 EI L3
Δ
Axially loaded bar F Fh AE
Δ
AE h
h

Fixed column F Δ Fh3 3EI

subjected to a lateral 3EI h3
load, F h

Fixed–fixed column F Δ Fh3 12 EI

subjected to a lateral 12 EI h3
load, F
h

Springs in parallel F k1 + k2
k1 + k2
F

Springs in series F F 1 1 kk
+ 1/  +  = 1 2
k1 k2 k k k + k2
F  1 2  1

FIGURE E3.4  Fixed–fixed beam with concentrated weight at midspan.

m
u

FIGURE E3.5  Idealized lumped mass model of a fixed–fixed beam with concentrated weight at midspan.

FIGURE E3.6  Displaced shape of a fixed–fixed beam with concentrated force at midspan.
Single-Degree-of-Freedom Structural Dynamic Analysis 49

From Table 3.1, the force–displacement relationship is given by

FL3
∆=
192EI

The stiffness of the SDOF system is determined by applying the same process as in
Example 3.3:

F 192EI
k= = 3
∆ L

3. Determine the natural and cyclic frequencies.

The equation of motion for this SDOF system under free vibration is described by
Equation 3.1:

u + ωn2u = 0

where the natural frequency is determined using Equation 3.2. Substituting for stiffness
and mass in terms of known quantities yields

192(EI /L3 ) 3EIg

ωn = =8 rad/s
(W /g ) WL3

The natural frequency, ωn, results in units of radians per second. In order to determine
the cyclic frequency, f, the units are converted from radians per second into cycles per
second by applying the conversion of 1 cycle = 2π radians or using Equation 3.6:

 1cycle  4 3EIg
f = ωn  = cycles/s or hertz
 2π rad  π WL3

EXAMPLE 3.5
Given a water tank supported on a slender column as shown in Figure E3.7, determine the natural
frequency and period of the system.

Solution
1. Idealize the structural system.
  Similar to Example 3.4, the weight of the column is ­considered negligible with respect
to the weight of the water tank. Considering only the h
­ orizontal displacement of the
water tank, the lumped mass model is shown in Figure E3.7.

W = 10 kips Go
Tigers! W

E = 3 × 107 psi 50 ft
I = 20,000 in4

FIGURE E3.7  Water tank supported on a slender column and an idealized structural model.
50 Introduction to Earthquake Engineering

2. Determine the stiffness parameters.

  From Table 3.1, the stiffness of the SDOF system is determined by applying the same
process described in Examples 3.3 and 3.4:

3EI 3(3×107 psi)( 20, 000 in4 )

k= = = 8333 lb/in
L3 (50 ft ×12 in/ft )3

3. Determine the natural frequency and period of the system.

  As the water tank is modeled as an SDOF system, the equation of motion is described
by Equation 3.1. The natural f­requency of water tank is determined using Equation 3.2:

k 8333 lb/in
ωn = = = 17.94 rad/s
m (10, 000 lb/ (32.2 ft/s2 ×12 in/ft))

  The natural period of vibration is determined using Equation 3.5:

2π 2π
Tn = = = 0.35 s
ωn 17.94 rad/s

EXAMPLE 3.6
The space-grid roof structure may be considered rigid, and it has a dead load of 20 psf. The side
sheathing has a dead load of 10 psf. All steel columns are W10 × 30 (I = 170 in4). Determine the
dynamic properties, natural frequency, and period, in the E–W direction. Note that the first pair of
columns in Figure E3.8 is fixed–pinned, the second is pinned–pinned, and the last pair is fixed–
fixed. A plan view of the structure is shown in Figure E3.9.

Solution
1. Idealize the structural system into an SDOF system.
  The frame can be modeled as an SDOF system assuming an rigid roof and the total
stiffness of the SDOF system is the sum of the individual column stiffnesses.

A B C

4′

12′

FIGURE E3.8  Elevation view of a space grid roof structure.

A B C
1

N
40′

2 @ 40′

FIGURE E3.9  Plan view of a space grid roof structure.

Single-Degree-of-Freedom Structural Dynamic Analysis 51

2. Determine the weight (proportional to the mass) of the SDOF system.

  The weight of the SDOF system includes the roof dead load of 20 psf and the wall
sheathing weight around the perimeter of the building (240 ft). The weight due to sheath-
ing includes the tributary height of the roof, 10 ft, in the area calculation:

W = 20 lb/ft 2( 40 ft × 80 ft ) + 10 lb/ft 2(10 ft × 240 ft ) = 88, 000 lb = 88 kips

3. Determine stiffness parameters.

  The lateral stiffness of the columns depends on the boundary conditions and connec-
tions to the roof beams. In this case, columns A1 and A2 are pinned and fixed at their
ends, columns B1 and B2 are pinned–pinned ­connected at their ends, and columns C1
and C2 are fixed–fixed connected at their ends.
A1 and A2 column stiffnesses,

3EI
kA1 = kA 2 =
h3

B1 and B2 column stiffnesses,

kB1 = kB2 = 0

C1 and C2 column stiffnesses

12EI
kC1 = kC 2 =
h3

The total stiffness of the SDOF system is the sum of column stiffnesses in the E–W
direction.

3EI 12EI 30EI 30( 29, 000 ksi)(170 in4 )

KE−W = 2× 3
+ 2× 3 = 3 = = 49.5 kips/in
h h h (144 in)3

4. Determine the natural frequency and period of the SDOF system.

  As the frame is modeled as an SDOF system, the equation of motion is described by
Equation 3.1. The natural ­frequency of the roof structure is determined using Equation
3.2:

KE−W 49.5 kips/in

ωn = = = 14.74 rad/s
m (88 kips/(32.2 ft/s2 ×12 in/ft))

The natural period of vibration is determined using Equation 3.5:

2π 2π
Tn = = = 0.43 s
ωn 14.74 rad/s

Lateral force-resisting systems are used to resist lateral seismic (or wind) loads and are generally
categorized into one of three different categories: unbraced frames, braced frames, or shear walls.
Following are simplified derivations of the equivalent lateral stiffness constants for single-story
systems of each of these three cases.
Unbraced frames have rigid connections between columns and beams and carry lateral load by
developing moments in the columns and beams. If we consider the portal frame introduced earlier
52 Introduction to Earthquake Engineering

in this chapter (see Figure 3.1), we can derive the equivalent stiffness of the frame by assuming that
each column contributes equally to the frame stiffness. For example, for a portal frame with fixed–
fixed columns (such as the one listed in the sixth row of Table 3.1), which would be equivalent to a
frame with a very stiff beam, the total stiffness is given as

12 EI 24 EI (3.9)
K = (2 ) = 3
h3 h
where h is the height of the frame and EI is the flexural stiffness of the columns, which was defined
in the sixth row of Table 3.1. Silva and Badie (2008) gave the following ­formula for a general case
with variable beam-to-column stiffness ratio of α = Ib/Ic (assuming variable moments of inertia, but
constant modulus of elasticity) and variable beam-span-to-column-height ratio of κ = Lb/h:

24 EI  6α + κ 
K=   (3.10)
h3  6α + 4κ 

For a portal frame case with pinned supports and a very stiff beam, K is given by the contribution
of two pinned–fixed columns (such as the one listed in the fifth row of Table 3.1) as

3EI 6 EI
K = (2 ) = 3 (3.11)
h3 h

Silva and Badie (2008) also gave a general case with variable beam-to-column stiffness ratio of
α and variable beam-span-to-column-height ratio of κ:

24 EI  α 
K=   (3.12)
h3  4α + 2κ 

EXAMPLE 3.7
The bridge systems depicted below has a rigid deck that weighs 500 kips. The lateral force-resist-
ing system consists of three concrete (E = 3000 ksi) columns as shown in Figure E3.10. Determine
the natural period of the system.

Solution
1. Idealize the structural system into a SDOF system.
  The frame can be modeled as an SDOF system, and the total stiffness of the SDOF
system is the sum of the individual column stiffnesses.

Fixed–fixed Fixed–fixed
I = 15,000 in4 I = 15,000 in4
20 ft
Fixed–pinned
I = 15,000 in4

Soft soil
Rock

FIGURE E3.10  Schematic of the bridge system.

Single-Degree-of-Freedom Structural Dynamic Analysis 53

2. Determine the mass of the SDOF system

m = W /g = 500 kips/ (386.4 in/s2 ) = 1294

. kips ⋅ s2 /in

3. Determine the stiffness.

  The lateral stiffness of the columns depends on the boundary conditions and connec-
tions to the bridge deck. In this case, the end conditions of the columns are specified in
the problem statement, and the stiffness parameters for each column can be obtained
from Table 3.1:

12EI 3EI 12EI 27EI 27(3000)(15, 000)

K= 3
+ 3 + 3 = 3 = = 87.9 kips/in
h h h h ( 20 ⋅ 12)3

4. Determine the natural period of the SDOF system.

  The natural period of the bridge is determined using Equation 3.5:

m 1.294
Tn = 2π = 2π = 0.762 s
K 87.9

Braced frames can be assumed to behave as cantilever beams that carry the lateral load by devel-
oping an internal shear force (known as panel shear) and a bending moment (known as overturning
moment). For portal braced frames similar to the one shown in Figure 3.7, we can assume that the
panel shear is carried by the diagonal members (one axial compression and the other axial tension),
whereas the overturning moment is carried by the axial forces in the columns. This braced frame
can be modeled as a statically indeterminate truss in which case the number of unknown reactions,
r, and unknown member forces, b, exceeds the number of equations of equilibrium (two times the
number of joints), 2j. To solve for the additional unknown in Figure 3.7, we can use compatibil-
ity of displacements. However, because the member sizes are initially unknown, we cannot apply

h F j = 4 ⇒ 2j = 8
2j < b + r
b=5 b+r=5
r=4

h
sin θ =
h2 + d2
θ d
cos θ =
h + d2
2

FIGURE 3.7  Braced portal frame model.

54 Introduction to Earthquake Engineering

displacement compatibility. Therefore, we must make some simplifying assumptions to render the
truss statically determinate and perform an approximate analysis. First, the diagonal members can
be assumed to be stiff (in which case the diagonals can carry both tension and compression) or they
can be assumed to be slender (in which case the member in axial compression cannot support axial
force because it buckles). In the case of stiff diagonal members, we assume that the tension and
compression diagonals each carry half of the panel shear, whereas in the case of slender ­diagonals,
we assume that the panel shear is resisted entirely by the tension diagonal. The assumption of slen-
der diagonal members is used to derive the equation for the ­stiffness of a braced portal frame using
the principle of virtual work in Example 3.8.
The principle of virtual forces can be used to establish the displacement of a truss joint in any
direction. The process entails determining the internal forces in each ­member caused by the applied
loads and determining the virtual internal forces caused by a unit virtual load applied at the joint in
the direction in question. With these internal loads, the contribution from each member to the truss
deflection is summed to determine the total displacement,
m

∑n E A
N i Li
∆= i (3.13)
i i
i =1

where:
Ni is the axial force in each member caused by the actual loading
ni is the axial force in each member caused by a unit virtual load applied at the joint in the
­direction in question
Li is the length of each member
Ai is the area of each member
Ei is the modulus of elasticity of each member

EXAMPLE 3.8
Derive an equation for the stiffness for a braced frame (see Figure 3.7) when the diagonal
­members are slender.

Solution
1. Apply equilibrium to determine internal forces for actual and virtual loading.
Slender diagonals cannot carry compression load; thus, from statics we can deduce
that the left-hand column and beam are zero-force members, leaving us with the cases
shown in Figure E3.11 for both the actual and virtual loading.
Use the method of joints to determine the internal forces in the remaining members
by first drawing an FBD of the upper right-hand joints and applying equilibrium to the
resulting concurrent force systems as shown in Figure E3.12.
2. Apply virtual work (Equation 3.13) to determine the displacement at the top of the frame,
Δ. As the axial displacement of the column contributes little to the lateral d
­ isplacement
of the frame, it’s customary to ignore it:

NL (1/cos θ )(F /cos θ ) h2 + d 2  (h2 + d 2 )3 / 2 

∆= ∑ n
EA
=
E1A1
=F
 d 2E1A1 




3. Determine the stiffness (ratio of force to displacement) of the frame. The frame lateral
deflection equation can be rearranged to determine the stiffness as

d 2E1A1 (3.14)
k=
(h + d 2 )3 / 2
2
Single-Degree-of-Freedom Structural Dynamic Analysis 55

Actual loading Unit virtual load

Δ Δ

F 1

θ θ

FIGURE E3.11  Actual and virtual loading with slender diagonal elements assumed.

FBD of joint: FBD of joint:

N1 F 1
n1

N2 n2

Equilibrium: Equilibrium:
ΣFx = 0; F – N1 cosθ = 0 ΣFx = 0; 1 – n1 cosθ = 0
N1 = F/cosθ n1 = 1/cosθ
ΣFy = 0; – N2 – N1 sinθ = 0 ΣFy = 0; – n2 – n1 sinθ = 0
N2 = – F tanθ n2 = – tanθ

FIGURE E3.12  FBDs of joints due to the actual loading (left) and virtual loading (right).

where the quantity EA is the axial stiffness given by the product of the modulus of elas-
ticity, E, and the cross-sectional area, A.

EXAMPLE 3.9
The platform shown in Figure E3.13 is used at a stadium to film football games and is experiencing
large dynamic motions. Preliminary investigations indicate that the natural period is 0.9 s. Camera
personnel ­recommend the period to be limited to 0.3 s. Using steel (E = 29,000 ksi), determine the
required diameter of a system of diagonal ties (wires) to retrofit the system to the new specifications.

Solution
The system can be idealized as an SDOF system. Also, as the platform is square, only one
­direction of motion needs to be considered, which is retrofitted with two diagonal ties (one along
each side—dash lines in Figure E3.13).

1. Determine the existing stiffness of the system using the observed natural period. In terms
of mass,

m = W /g = 4.5 kips/ (386.4in/s2 ) = 0.01164 kips ⋅ s2 /in

56 Introduction to Earthquake Engineering

8 ft

8 ft
4.5 kips

6 ft

FIGURE E3.13  Schematic of a structural platform.

and the existing natural period of the SDOF system (Tn = 0.9 s), we can determine the
available stiffness using Equation 3.5:
2
m  2π 
Tn = 2π ⇒ k = m  
k  Tn 

That is,
2
 2π   2π 2
k = m   = 0.01164  = 0.5676 kips/in
 Tn   0.9 

2. Determine the required stiffness necessary to correspond to the limiting period. We now
determine the required stiffness to deduce the natural period to 0.3 s.
2
 2π   2π 2
knew = m   = 0.01164  = 5.108 kips/in
 Tn   0.3 

3. Determine the required diameter of diagonal ties needed to increase the stiffness of
the system. Determine the difference between the available and the new required
­stiffnesses. This must be provided by the remedial diagonal ties. The change in s­ tiffness
can be applied to Equation 3.14 to determine the required radius of the diagonal ties:

d 2EA
knew − k = 2
(h2 + d 2 )3 / 2

where d = 8 ft, h = 6 ft, and A = πr2, so

(8 ×12)2( 29, 000)πr 2

5.108 × 0.5676 = 2
((6 ×12)2 + (8 ×12)2 )3/ 2

Solving for the radius of the wire yields

(5.108 − 0.5676)((6 ×12)2 + (8 ×12)2 )3/ 2

r= = 0.0685 in
2(8 ×12)2( 29, 000)π

Thus, the required diameter is 2r = 0.137 in, which can be provided by a 3/16-inch wire.
Single-Degree-of-Freedom Structural Dynamic Analysis 57

Shear walls can be assumed to behave as deep cantilever beams that carry the l­ ateral load by
developing internal shear force and bending moment as shown in Figure E3.14. That is, a shear
wall has deflections that are affected by both moment and shear (unlike regular beams, where
most of the deflection is caused by flexure). We can once again use the principle of virtual work
to establish the lateral displacement of the top of a shear wall. The process entails determin-
ing the internal shear and moment functions (diagrams) caused by the applied lateral load and
determining the virtual internal shear force and moment functions (diagrams) caused by a lateral
unit virtual load also applied at the top of the wall. With the internal shear force and moment
functions, the deflection of the top of the wall can be determined using the following virtual
work equation:

L L
M  vV 
∆=
∫ m f f dx +
EI ∫ κ  GA  dx (3.15)
0 0

where:
mf is the internal virtual moment caused by an external virtual unit load
Mf is the internal moment caused by the lateral load
v is the internal shear force caused by an external virtual unit load
V is the internal shear force caused by the lateral load
L is the length of the member
I is the moment of inertial of the member
E is the modulus of elasticity of the member
A is the area of the member
G is the shear modulus of elasticity of the member
κ is a constant that takes the following values depending on the cross-section of the shear wall:
κ = 1.2 for rectangular beams
κ = 10/9 for circular sections
κ = A/Aweb (ratio of total area to area of web) for I or box sections

EXAMPLE 3.10
Derive an equation for the stiffness for the cantilever shear wall shown in Figure E3.14.

Δ
F

FIGURE E3.14  Cantilever shear wall.

58 Introduction to Earthquake Engineering

Actual load x x Virtual load x x

F 1

V(x) F Mf(x) Fh v(x) 1 mf(x) h

FIGURE E3.15  Internal shear and bending moment diagrams for actual loading and virtual loading.

Solution
1. Apply equilibrium to determine shear force and bending moment functions for actual
and virtual loading.
  Use structural analysis to determine the shear force and bending moment diagrams for
a cantilever beam, for both real load and virtual loading; see Figure E3.15.
2. Determine the displacement, Δ, at the top of the wall.
  Apply virtual work (Equation 3.15) to determine the displacement at the top of the
wall, Δ. As the moments vary l­ineally and the shears are constant, we get the following
result:
L L
1 κ 1 κ
∆=
EI ∫ 0
mf ( x )Mf ( x ) dx +
GA ∫ v(x)V (x) dx = EI3 (h)(Fh)h + GA (1)(FF) h
0
3
Fh κFh
= +
3EI GA
3. Determine the stiffness (ratio of force to displacement) of the wall.
  The wall lateral deflection equation can be rearranged to determine the stiffness,
which for a rectangular (κ = 1.2) shear wall is
3EI AG
K= + (3.16)
h3 1.2h

where the quantity EI is the flexural stiffness given by the product of the modulus of elasticity, E,
and the moment of inertia of the area, I, and AG is the shear stiffness given by the product of the
shear modulus of elasticity, G, and the area, A.

Figure 3.8 illustrates the relationship between the shear wall aspect ratio, d/h, and relative effect
of shear displacement, Δ/ΔM. ΔM is the displacement cause by moment only. Note that as the shear
wall gets taller, the effect of the shear force decreases rather quickly. For d/h < 1/5, the shear

Δ
Δ/ΔM

h
2
Δ d
= 1 + 0.94
ΔM h

1 d/h
0

FIGURE 3.8  Effect of shear force in the lateral deflection of a shear wall.
Single-Degree-of-Freedom Structural Dynamic Analysis 59

FIGURE 3.9  Fixed top and bottom shear wall.

effect represents <4% so it may be neglected; h­ owever, for d/h > 1/3, the shear effect is important
as it is >10%.
We can also derive the equation for a shear wall that is fixed at the top and bottom (see Figure 3.9),
12 EI AG
K= + (3.17)
h3 1.2h

EXAMPLE 3.11
The structural system depicted in Example 3.6 (repeated here as Figure E3.16 for convenience)
has a roof weighing 20 psf and side sheeting weighing 10 psf. Assuming the lateral force-resisting
system in the N–S direction (one on each side) consists of ½-inch plywood (E = 1000 ksi and
G = 500 ksi) shear walls, determine the building’s natural period in this direction.

Solution
1. Determine the weight (proportional to the mass) of the SDOF system (same as
Example 3.6).
The weight of the SDOF system includes the roof dead load of 20 psf and the wall
sheathing weight around the perimeter of the building (240 ft). The weight due to
­sheathing includes the tributary height of the roof, 10 ft, in the area calculation:
W = 20 psf (40′ × 80′) + 10 psf (10′ × 240′) = 88, 000 lb = 88 kips

A B C
1

N
40′

2 @ 40′

4′

12′

E–W elevation N–S elevation

FIGURE E3.16  Plan and elevation views of a structural system.

60 Introduction to Earthquake Engineering

2. Determine the stiffness parameters.

The lateral stiffness of the shear walls depends on the moment of inertia and area of
the footprint of the wall.
Moment of inertia of shear wall,

I = bh3 /12 = 0.5 in( 20 ft ×12 in/ft )3 /12 = 576, 000 in4

Area of shear wall,

A = bh = 0.5 in ( 20 ft ×12 in/ft ) = 120 in2

The total stiffness of the SDOF system is given by Equation 3.16:

 3EI AG   3(1000 ksi)(576, 000 in4 ) 120 in2(500 ksi) 

K = 2  3 + = 2  +  = 1324 k/in
h 1.2 h  
 (14 ⋅ 12)3 1.2(14 ⋅ 12) 

3. Determine the natural period of vibration.
As the frame is modeled as an SDOF system, the equation of motion is described by
Equation 3.1. The natural period of the system is determined using Equation 3.5:

m W /g 88 kips/ 386.4 in/s2

Tn = 2π = 2π = 2π = 0.083 s
k k 1324 kips/in

3.1.4 Approximate Structural Period

For the equivalent static method of analysis, the ASCE-7 specifications provide approximate rela-
tions to determine the period of a structural system. The following relationships are used to deter-
mine an approximate period, Ta:

Ta = Ct hn x (3.18)

where:
hn is the height in feet above the base to the highest level of the structure
Ct and x are determined from ASCE-7 (Table 12.8.2; see Table 3.2)

An even more approximate relation that can be used to quickly check the values determined
using Equation 3.18 or the full dynamic analysis period is

Ta = 0.1N (3.19)

TABLE 3.2
Values of Approximate Period Parameters
Ct and x (ASCE-7 Table 12.8.2)
Structure Type Ct x
Steel moment-resisting frames 0.028 0.8
Concrete moment-resisting frames 0.016 0.9
Eccentrically braced steel frames 0.03 0.75
All other structural systems 0.02 0.75
Single-Degree-of-Freedom Structural Dynamic Analysis 61

where N is the number of stories.

This approximate period can be considered valid for buildings of up to 12 stories and story
heights uniformly distributed at about 10 ft.
Approximate period for shear walls:

0.0019
Ta = hn (3.20)
Cw

where:
100
x
 hn 2
∑   Ai
CW =
AB  hi  [1 + 0.83(hi /Di )2 ]
i =1

AB  is the  area of the base of the structure in square feet

Ai is the web area of shear wall i in square feet
Di  is the  length of shear wall i in feet
hi  is the  height of shear wall i in feet

3.1.5 Time-Dependent Forced Undamped Vibration Response

We now characterize the forced vibration motion of the portal frame model introduced in
Section  3.1.1 (see Figure 3.10). In this case, the mathematical model includes a time-dependent
excitation force, p(t) (or displacement). We can again apply equilibrium to an FBD of the mass
(which now includes the time-dependent force as shown in Figure 3.10) using D’Alembert’s prin-
ciple. Horizontal ­equilibrium of the FBD shown in Figure 3.10 yields the equation of motion,

∑ F = 0;x − mu − ku + p(t ) = 0 ⇒ mu + ku = p(t )

(3.21)

For the purpose of this analysis, we use a periodic, harmonic excitation force, which is a force
with magnitudes represented by sine or cosine as functions of time. (This force can also represented
a time-dependent displacement excitation, such as one produced by rotating machinery.) Sine and
cosine functions can also be applied to nonharmonic loadings with the response being obtained
using a Fourier method—a superposition of individual responses to the harmonic components of the
external excitations. First, assume a forcing function of the following form:

p(t ) = po sin ωt (3.22)

where:
po is the peak magnitude of the force
ω is the frequency of the force in radians per second

u Free-body diagram
p(t)
m p(t) m u

k V=ku

FIGURE 3.10  Free-body diagram of the SDOF model including a time-dependent force.
62 Introduction to Earthquake Engineering

This forcing function produces bound displacements provided that ω ≠ ωn = k /m , the resonant
condition, which will be discussed in more detail later in this section.
With a forcing function, we can solve the equation of motion following standard methods used in
the solution of nonhomogeneous differential equations. Equation 3.21 can be rewritten as

mu + ku = po sin ωt

The general solution to this equation can be expressed as a combination of the particular and
complementary solutions,

u(t ) = uc (t ) + uP (t ) (3.23)

where:
uc is the complementary solution, which is the solution to the homogeneous equation (the free
vibration solution) given by Equation 3.3 and repeated here for convenience,

uc (t ) = A cos ωnt + B sin ωnt

uP is the particular solution, which usually takes the same form as the forcing function, that is,

u p (t ) = C sin ωt (3.24)

where C is the peak value, which is obtained by substituting Equation 3.24 into the original ­equation
of motion,

−mω 2C sin ωt + kC sin ωt = po sin ωt

After factoring the sine function that appears in each term, and recognizing that this function is
not always zero, we can write this equation as

−mω 2C + kC = po

Solving for C,

po po (u )
C= = = st o
k − mω 2 k (1 − m /k ω 2 ) 1 − r 2

where:
r is the frequency ratio, that is,

ω
r=
ωn

(ust)o is the static displacement caused by po; that is, (ust)o = po/k

We can rewrite the particular solution to include the phase angle, φ, as follows:

uP (t ) = uo sin(ωt − φ )

Single-Degree-of-Freedom Structural Dynamic Analysis 63

where:
uo is the maximum steady-state displacement response caused by the time-dependent
­excitation force;
or

uP (t ) = (ust )o Rd sin (ωt − φ ) (3.25)

where:
Rd is the deformation response factor, which is the ratio of the maximum steady-state
­displacement, uo, and the equivalent static displacement, (ust)o,

uo 1
Rd = = (3.26)
(ust )o 1 − r 2

and the phase angle, φ, ranges as follows:

 0 ω < ωn
φ =  
180 ω > ωn


Figure 3.11 shows graphs depicting the deformation response factor, Rd, and phase angle, φ, as
functions of frequency ratio, r. First, note that the displacement grows unbounded as r approaches
one, both from the right and the left. In actual practice, however, the system would come apart
or yield (changing the stiffness, thus changing the natural frequency). This point on the graph
corresponds to the resonant frequency, and as indicated earlier, the solution to the equation of
motion is not valid here. Also, the excitation force changes from being in phase with the natural

5
Deformation response factor, Rd

2
r= 2

0
180
uo is out of phase
Phase angle, φ

with force
degrees

90
uo is in phase
with force
0
0 1 2 3
Frequency ratio, r

FIGURE 3.11  Deformation response factor, Rd, and phase angle, φ, as a function of frequency ratio, r.
64 Introduction to Earthquake Engineering

vibration of the system to being out of phase; that is, the forcing function is additive up to r = 1;
and it begins to counter the force after r = 1. Furthermore, note that for small values of r (slowly
varying force), Rd is ∼1; and for large values of r (rapidly varying force), Rd approaches zero,
which corresponds to no displacement. An additional point of interest along the abscissa of these
graphs is r = 2, beyond which all values of Rd are ≤ 1. The point can be obtained by setting
Rd = 1.0. This implies that the dynamic displacement is always less than the static displacement
beyond r = 2 .
We now obtain the total dynamic response of the system by combining the complementary and
particular solutions,

po /k
u(t ) = A cos ωnt + B sin ωnt + sin ωt (3.27)
1− r 2

The two constants A and B are arbitrary and can be obtained by evaluating the equation at time
t = 0 (initial conditions). Again, the displacement at time t = 0 is u(0), that is,

po /k
u(0) = A cos(0) + B sin(0) + sin(0)
1− r 2
which again yields A = u(0). Next, differentiate the displacement equation with respect to time,

po /k
u (t ) = − Aωn sin ωnt + Bωn cos ωnt + ω cos ωt (3.28)
1− r 2

Evaluating the velocity at time t = 0,

po /k
u (0) = − Aωn sin(0) + Bωn cos(0) + ω cos(0)
1− r 2

yields

u (0) po /k
B= − r
ωn 1− r 2

The complete solution is the forced vibration of an undamped SDOF system, which describes the
position of the mass as a function of time,

 u (0) po /k  p /k
u(t ) = u(0) cos ωnt +  − r  sin ωnt + o 2 sin ωt
2 
(3.29)
 ωn 1− r  1− r

A graph of which is given in Figure 3.12. In general, the resulting motion is the sum of periodic
motions of two different frequencies, ω and ωn, with different amplitudes. As noted earlier, when ω
approaches ωn, the motion becomes unbounded as t goes to infinity. However, when damping (dis-
cussed in Section 3.2) is included, the motion remains bounded.

3.2  DAMPED SINGLE-DEGREE-OF-FREEDOM SYSTEM

An undamped system, once excited, will oscillate indefinitely with constant amplitude at its natu-
ral frequency; however, from experience we know that no systems can oscillate indefinitely. Any
system undergoing motion always experiences internal damping forces that dissipate mechanical
Single-Degree-of-Freedom Structural Dynamic Analysis 65

p
Amplitude, po

Period, T = 2π/ω

2
Total response

Steady-state
response
1
u(t)/(ust)o

–1

–2
0 1 2
t/T

FIGURE 3.12  Forcing function and resulting response for r = 0.2, u(0) = 0.5po/k, and u (0) = ωn po /k .

energy (kinetic and potential) by transforming it into other forms of energy, such as heat. This type
of damping is inherent to a system. There may be cases where intentional damping is introduced
in a system, such as the shock absorbers in an automobile. Similar types of shock absorbers can
also be used for building systems, such as the fluid viscous dampers from Taylor Devices shown in
Figure 3.13. Similar to inertial and stiffness effects, damping effects can be characterized using a
force, which in this case is assumed to be proportional to the magnitude of the velocity and opposite
to the direction of motion. This type of damping is known as “viscous damping” and is the type of
damping produced in a body when restrained by surrounding viscous fluid. While viscous damp-
ing is not inherent in structural systems, this type of damping is much easier to manipulate in the
­equation of motion than the more realistic damping from internal friction and is accurate enough
for most practical purposes.

3.2.1  Free Vibration Response of Damped Systems

We can include the damping effects in the free vibration response of the portal frame model
­discussed in Section 3.1.1 by first applying equilibrium to an FBD of the mass using D’Alembert’s
66 Introduction to Earthquake Engineering

FIGURE 3.13  Fluid viscous dampers (photographs courtesy of Taylor Devices, Inc.).

principle, where, in addition to the stiffness and inertial forces, a damping force proportional to the
velocity is included, as shown in Figure 3.14. In this case, we are using an oscillator model, which
is commonly used in structural dynamics textbooks (Chopra, 2012) because it is easier to visualize
the combined effects of damping, stiffness, and inertia.
Horizontal equilibrium of the FBD shown in Figure 3.14 yields the equation of motion,

mu + cu + ku = 0 (3.30)

Here, c is the damping coefficient, which is discussed in more detail in the next section.
Again, this is a second-order, linear, and homogeneous differential equation with constant
­coefficients, the solution to which is also of exponential form, that is,

u(t ) = e ρ t

This equation can be differentiated once to determine the velocity,

u(t ) = ρeρ t

and the velocity can be differentiated to determine the acceleration,

u(t ) = ρ 2eρ t

u u
k ku
m u
m
c cu
Viscous
damper
FBD

FIGURE 3.14  Idealized SDOF system and FBD for a portal frame.
Single-Degree-of-Freedom Structural Dynamic Analysis 67

substituting the equations for the displacement, velocity, and the acceleration into Equation 3.30
yields

mρ 2eρ t + cρeρ t + keρ t = 0

which can be rearranged as follows:

(mρ 2 + cρ + k )eρ t = 0

In general, the exponential function is not zero, thus the quantity within parenthesis (also known
as the characteristic equation) must be zero,

mρ 2 + cρ + k = 0

Solving for the unknown parameter ρ yields


 −c + c 2 − 4mk


2 ρ1 =
−c ± c − 4mk  2m
ρ= ⇒
2m  c 2 − 4mk
ρ = −c −
 2 2m

Since the equation of motion (Equation 3.30) is a second-order differential equation, we need
two constants of integration, C1 and C2, which are determined from initial conditions (displacement
and velocity).

u(t ) = C1eρ 1t + C2eρ 2 t

Substituting the ρ parameters into this equation yields

−c  c2 − 4 mk − c2 − 4 mk 
t 
u(t ) = e 2 m C1e
t t
2m + C2 e 2m

 


This result leads to three different solutions depending on the value of the quantity under the
square root operator (see Figure 3.15):

1. If c2–4 mk > 0, exponents are real numbers. The motion for this case decays exponentially;
thus, no oscillation occurs. For this case, the system is classified as overdamped.
2. If c2–4 mk < 0, exponents are imaginary values, and oscillatory motion occurs. For this
case, the system is classified as underdamped.
3. If c2–4 mk = 0, then there is a boundary between oscillatory and nonoscillatory motion.
For this case, the system is classified as critically damped.

Structural systems are typically underdamped, so our discussion on damping will be focused
primarily on this type of system. However, critically damped systems are important because they
are used to define the damping ratio, ζ, which is typically used to describe damping in structural
systems. For a critically damped system, c2−4 mk = 0, which leads to the definition of the critical
damping coefficient, ccr:
68 Introduction to Earthquake Engineering

1.5

Critically damped, ζ = 1
1
Overdamped, ζ = 2
0.5
u(t)/u(0)

0
0 1 2 3
t/T
–0.5
Underdamped, ζ = 0.1
–1

–1.5

FIGURE 3.15  Free vibration response of a damped SDOF system.

2k
ccr = 2 km = 2mωn = (3.31)
ωn

The other parameters were defined earlier. With the critical damping coefficient, ccr, we now
define the damping ratio or damping factor (which is an expedient way to define damping for practi-
cal structural systems and is usually expressed in a percentage form) as
c
ζ= (3.32)
ccr

We can also use this value to define the three different general states of damping: overdamped
systems have ζ > 1, critically damped systems have ζ = 1, and underdamped systems have ζ < 1, as
shown in Figure 3.15. Since most structural systems are underdamped, in this book we only provide
the solution for this case. Rewriting the unknown parameters ρ1 and ρ2 as

(
ρ 1, ρ 2= −ζ ± ζ 2 −1 ωn )
This will be complex as ζ < 1. We can write this in complex number form as

(
ρ 1, ρ 2= −ζ ± i 1 − ζ 2 ωn )

Substituting these parameters into the solution yields

 1−ζ 2 ωn t 1−ζ 2 ωn t 
u(t ) = e−ζωn t C1ei + C2e−i 
 

This equation can be expressed in polar form (in terms of sines and cosines) by making use of
Euler’s identities; and because trigonometric functions are real-value solutions to the equation of
motion, we can express the general solution in real form as

u(t ) = e−ζωn t [ B sin ωD t + A cos ωD t ] (3.33)

Single-Degree-of-Freedom Structural Dynamic Analysis 69

where A and B are arbitrary constants. While this motion is periodic, it is no longer of constant
frequency.  In fact, the frequency (and the period) is a function of the damping ratio—the damped
frequency,

ω D = ωn 1 − ζ 2 (3.34)

The damped period,

2π 2π Tn
TD = = = (3.35)
ω D ωn 1 − ζ 2
1− ζ 2

We now use initial conditions (at time t = 0) to solve for constants A and B. First, the initial
­displacement (at time t = 0) is u(0), that is,

u(0) = e0 [ B sin(0) + A cos(0)]

which yields A = u(0). Next, differentiating the displacement equation with respect to time we
obtain

u(t ) = −ζωne−ζωn t [ B sin ωD t + A cos ωD t ] + e−ζωn t [ BωD cos ωDt + AωD sin ωD t]

Evaluating the velocity at time t = 0 yields

u(0) = −ζωne0 [ B sin(0) + A cos(0)] + e0 [ BωD cos(0) + AωD sin(0)]

where:

u(0) + ζωnu(0)
B=
ωD

Again, the complete solution is the free vibration response of a damped SDOF system, which
describes the position of the mass as a function of time,

 u(0) + ζωnu(0)  
u(t ) = e−ζωn t   sin ω t + u(0) cos ω t  (3.36)
 ωD  D D

 

A graph of which is given in Figure 3.16. The damped structure clearly shows motion decay
over time. The decay in the motion is mathematically produced by the exponential part of the
solution.
This oscillation decay can be used to experimentally determine the inherent damping in a ­system.
The process entails setting the structure into free vibration from an initial displacement and obtain-
ing a record of the motion. The rate of decay of amplitude of motion is characterized by a ratio
of the value of two successive peak displacement amplitudes, which is known as the logarithmic
­decrement, δ. The values of two successive peak displacement amplitudes are

u1 = ιe−ζωn t1 u2 = ιe−ζωn ( t1 +TD )

70 Introduction to Earthquake Engineering

u(0)
1 Undamped
u1
ρe–ζωnt structure

Damped
u2 structure
u(0)
u(t)

–ρe–ζωnt Tn = 2π/ωn

TD = 2π/ωD

FIGURE 3.16  Comparison of free vibration response of undamped and damped SDOF systems.

where ι is a constant that depends on the initial conditions. The ratio of these displacements is

u1 ιe−ζωn t1
= −ζωn ( t1 +TD ) = eζωn ( t1 +TD )−ζωn t1 = eζωnTD
u2 ιe

Taking the natural log of both sides, we can define the logarithmic decrement, δ,

u 
δ = ln  1  = ζωnTD (3.37)
 u2 

The logarithmic decrement can also be defined in terms of acceleration as

u1
δ = ln = ζωnTD (3.38)
u2

In both cases, δ can be defined in terms of ζ only. This is accomplished by replacing the damped
period as follows:

ζωn 2π 2πζ
δ = ζωTD = = (3.39)
2
ωn 1 − ζ 1− ζ 2

For small damping ratio values (ζ < 0.2), the value under the square root operator is close to
unity and the logarithmic decrement can be approximated as

δ ≈ 2πζ (3.40)

Single-Degree-of-Freedom Structural Dynamic Analysis 71

Furthermore, for lightly damped systems where successive oscillation peaks have similar ordi-
nates, two nonconsecutive amplitudes ui and ui+n (where n is any integer) can be used to determine δ,

ui
ln = nζωnTD = nδ (3.41)
ui +n

Also, note that the equation of motion can be rewritten in the following form:

u + 2ζωnu + ωn2u = 0 (3.42)

3.2.2 Structural Damping
As noted earlier, in addition to attenuating vibration amplitudes, damping has the effect of lower-
ing the natural frequency, and thus lengthening the period from the natural state. Most structural
systems possess small inherent damping, <20% as listed in Table 3.3; and as shown in Figure 3.17,
damping effects on the period are negligible when the damping ratio is low. Thus, the damped and
undamped periods are practically equal, and in practice, TD is usually taken as Tn (except when
supplemental damping is added to a structural system).
The damping ratio characterizes how fast oscillations decay from one cycle of motion to the
next and has a significant effect on the response of elastic systems. However, the inelastic response
is dominated by ductility, and damping has little attenuation effect relative to ductility—this will
be described in Chapter 5. For this reason, seismic design guidelines (which assume inelastic
behavior) assume a uniform 5% damping for all building structural systems. Furthermore, current
­k nowledge of damping mechanisms in structural systems, particularly buildings, is rather limited,
with l­ ittle ongoing research in the area. It is believed that damping in building systems is the result
of the f­ ollowing damping mechanisms: friction in structural elements (friction in bolted and nailed
­connections, and cracking of concrete), intrinsic material damping, friction in nonstructural com-
ponents and their connections to the structure, and soil–structure interaction (radiation of seismic
waves back into the soil and intrinsic damping in the soil) (Villaverde, 2009). All of the aforemen-
tioned damping mechanisms are difficult to characterize, thus making it difficult to determine
damping analytically for actual structures. Damping ratios tabulated in Table 3.3 (Lindeburg and

TABLE 3.3
Typical Damping Ratios
Type of Construction ζ
Steel frame with welded connections and flexible walls 0.02
Steel frame with welded connections, normal floors, and exterior cladding 0.05
Steel frame with bolted connections, normal floors, and exterior cladding 0.10
Concrete frame with flexible interior walls 0.05
Concrete frame with flexible interior walls and exterior cladding 0.07
Concrete frame with concrete or masonry shear walls 0.10
Concrete or masonry shear walls 0.10
Wood frame and shear walls 0.15

Source: Lindeburg, M. R. and K. M. McMullin, Seismic Design of Building Structures:

A Professional Introduction to Earthquake Forces and Design Details,
9th ­edition, Professional Publications, Inc., Belmont, CA, 2008.
72 Introduction to Earthquake Engineering

1
ωD Tn
= = 1 – ζ2
ωn TD

Tn
TD

0 ζ
0 1

FIGURE 3.17  Effects of damping on vibration period.

McMullin 2008) are only provided to illustrate that real structures do not possess inherent damp-
ing >15%. From the given data, it should also be clear that the damping ratio depends on the type
of building construction.

EXAMPLE 3.12
A free vibration test is conducted on an empty elevated water tank. A cable attached to the tank
applies a lateral force of 16.4 kips and pulls the tank horizontally by 2 in. The cable is suddenly cut
and the resulting free vibration is recorded. At the end of four complete cycles, the time is 2.0 s,
and the amplitude is 1 in. From the given information, compute the damping ratio, natural period,
effective stiffness, effective weight, damping coefficient, and number of cycles required for the
displacement to decrease to 0.2 in.

Solution
1. Determine the dampin
As ui = u1 = 2 in at t = 0 and ui+n = u5 = 1 in at n = 4, from Equation 3.41

u   2
ln  1  = ln   = nζ 2π
 u5   1

1  2
Therefore, ζ = ln   = 0.0276 or 2.76%.
4 × 2π  1
The low damping ratio allows for a small damping assumption.
2. The damped period is determined by dividing the lapse time by the number of cycles,

2s
TD = = 0.5 s /cycle
4 cycles

In addition, due to the small damping assumption, the natural period can be approxi-
mated as the damped period.

Tn ≅ TD = 0.5 s

Single-Degree-of-Freedom Structural Dynamic Analysis 73

3. Determine the equivalent stiffness of the water tank assuming a linear force–­displacement
relationship

F 16.4 kips
k= = = 8.2 kips/in
u 2 in

4. Determine the effective weight using the period and the stiffness

2π 2π
ωn = = = 12.57 rad/s
Tn 0 .5 s

k 8.2 kips/in
m= = = 0.0519 kip ⋅ s2 /in
ωn2 (12.57 rad/s)2

W = mg = (0.0519 kip ⋅ s2 /in)(386.4 in/s2 ) = 20.05 kips

5. The damping coefficient

( )
c = ς ( 2 Km ) = 0.0276 2 (8.2kips/in)(0.0519 kip ⋅ s2 /in) = 0.036 kip ⋅ s/in

6. Determine the number of cycles for displacement to reduce to 0.2 inches using the
­logarithmic decrement δ

 u 
ln  i  = nζ 2π ⇒
 ui +n 
1  u  1  2 in 
n= ln  i  = ln   = 13.3 cycles or 13 cycles

 276)2π  0.2 in 
ζ 2π  ui +n  (0.02

3.2.3 Time-Dependent Forced Damped Vibration Response

We now characterize the forced damped vibration motion of the portal frame discussed in Section
3.1.5, shown in Figure 3.10, including the damping force in the mathematical model. The equilib-
rium equation in this case yields an equation of motion of the form

mu + cu + ku = p(t ) (3.43)

For the purpose of this analysis, we again use a periodic, harmonic excitation force. The force
excitation can also be represented as a time-dependent displacement excitation as will be shown in
Section 3.2.4. We begin again by assuming a forcing function of the form

p(t ) = po sin ω t

where:
po is the peak magnitude of the force
ω is the frequency of the force in radians per second

This forcing function produces bound displacements; even for the resonant condition, pro-
vided damping is nonzero. Again, with this forcing function, we can solve the equation of motion,
74 Introduction to Earthquake Engineering

mu + cu + ku = po sin ω t , following standard methods used in the solution of nonhomogeneous
differential equations. The general solution to this equation is of the same form as that for the
undamped case,

u(t ) = uc (t ) + uP (t ) (3.44)

where uc is the complementary solution, which is the solution to the homogeneous equation for
the underdamped case (the free vibration solution) given by Equation 3.33, and repeated here for
convenience,

uc (t ) = e−ζωn t [ A cos ωD t + B sin ωD t ]

up is the particular solution, which usually takes the same form as the forcing function, that is,

uP (t ) = C sin ω t + D cos ω t (3.45)

where C and D are constants that are obtained by substituting the particular solution into the
­original equation of motion,

po (1 − r 2 ) po 2ζ r
C= and D =−
k (1 − r 2 )2 + (2ζ r )2 k (1 − r 2 )2 + (2ζ r )2

Substituting the constants into the particular solution yields

po /k
uP (t ) = ((1 − r 2 )sin ωt − 2ζ r cos ωt ) (3.46)
(1 − r 2 )2 + (2ζ r )2

We can rewrite the particular solution to include the phase angle, φ, as follows:

uP (t ) = uo sin(ω t − φ)

where:

(ust )o
uo = C 2 + D 2 =
(1 − r 2 )2 + (2ζ r )2

is the maximum steady-state displacement response and

 2ζ r 
φ = tan−1 (−D /C ) = tan−1 
 1 − r 2 

(ust)o = po/k is the equivalent static displacement

And as given earlier, the frequency ratio is r = ω/ωn.
We can again define the deformation response factor (also known as the dynamic magnifica-
tion factor), Rd, as the ratio of the maximum steady-state displacement, uo, to the equivalent static
displacement, (ust)o:

uo 1
Rd = = (3.47)
ust (1 − r ) + (2rζ )2
2 2

Single-Degree-of-Freedom Structural Dynamic Analysis 75

Note that this ratio is a function of r and ζ only. Maximizing Rd in terms of r yields several
­interesting conditions. First, take the partial derivative of Rd with respect to r is

∂Rd r (1 − r 2 − 2ζ 2 )
= =0
∂r [(1 − r 2 )2 + (2rζ )2 ]3/2

This leads to four conditions:

r = 0 leads to 1–2ζ2 = 0, or ζ = 0.5 ,

1.
2.
r = ∞ does not yield any useful information,
r = 1 leads to resonance, Rd res = 1/ 2ζ , so, as ζ → 0, Rd res → ∞ ,
3.
r other than 0, 1, or infinity yields 1 − r 2 − 2ζ 2 = 0 ⇒ r = 1 − 2ζ 2 .
4.

Figure 3.18 shows graphs depicting the deformation response factor, Rd, as a function of
­f requency ratio, r. This graph is the same as that shown in Figure 3.11, but includes damp-
ing. Notice that in this case, the displacement does not grow unbounded as r approaches 1.
In fact, as  damping  increases the displacement at the resonant frequency decreases, and for
­damping ratios exceeding 70.7%, the dynamic displacement is always less than the static dis-
placement. Again, for small values of r (slowly varying force), Rd is approximately equal to 1;
and for large values of r (rapidly varying force), Rd approaches zero, which corresponds to no
displacement.
We now obtain the total dynamic response of the system by combining the complementary
­solution (transient response) and particular solution (steady-state response),

u(t ) = e−ζωn t ( A cos ωD t + B sin ωD t )

po /k (3.48)
+ ((1 − r 2 )sin ωt − 2ζ r cos ωt )
(1 − r 2 )2 + (2ζ r )2

ζ = 0.01
5

4 ζ = 0.1

3
Rd

ζ = 0.2
2

1
ζ = 0.7
ζ=1
0
0 1 2 3
r

FIGURE 3.18  Damped Rd as a function of r.

76 Introduction to Earthquake Engineering

The two constants A and B are arbitrary and can be obtained using the initial (at time t = 0) con-
ditions. Again, the displacement at time t = 0 is u(0), which yields the following when substituted
into Equation 3.48:
(ust )o sin(−φ)
A = u(0) −
(1 − r 2 )2 + (2ζ r )2

Evaluating the velocity at time t = 0 yields

 
1  (ust )o (ω cos(−φ) + ζω sin(−φ)) 
B= u(0) + u(0)ζω − 
ωD  (1 − r 2 )2 + (2ζ r )2 
 

A graph of the complete solution that describes the position of the mass as a function of time is
given in Figure 3.19 (similar to Figure 3.12, but with damping). In general, the resulting motion is
the sum of periodic motions of two different frequencies (ω and ωn) and amplitudes. Note that the
transient response (the one due to the inherent motion in the system) attenuates as t goes to infinity.
Also, as noted earlier for undamped systems, the motion becomes unbounded when ω approaches
ωn, as t goes to infinity. However, for damped cases, motion remains bounded to a maximum of 0.5ζ,
as shown in Figure 3.20.
The effect of a harmonic force excitation on the foundation of a supporting system can be deter-
mined as the magnitude of the force transmitted to the support through the spring (stiffness) and
the damping elements,

( fT )o = ku p + cu p = (ust )o Rd [ k sin(ωt − φ) + cω cos(ωt − φ)]

The maximum value of this force over t is

( fT )o = (ust )o Rd k 2 + c 2ω 2

2 Total response

Steady-state
response
1
u(t)/(ust )o

–1

–2
0 1 2
t/T

FIGURE 3.19  Harmonic force response for r = 0.2, ζ = 0.05, u(0) = 0, and u (0) = ωn po /k .
Single-Degree-of-Freedom Structural Dynamic Analysis 77

15
Steady-state
amplitude
1
10
2ζ
Envelope
curves
5
u(t)/(ust )o

–5

–10 – 1
2ζ

–15
0 2 4 6 8 10
t/T

FIGURE 3.20  Resonant response (r = 1) to harmonic force for ζ = 0.05.

Rewriting this as a fraction of equivalent static force, po, we get the transmissibility, Tr,

( fT )o
= Rd 1 + (2ζ r )2
po

or

( fT )o 1 + (2rζ )2
Tr = = (3.49)
po (1 − r 2 )2 + (2rζ )2

EXAMPLE 3.13
A sensitive, 50-lb instrument that requires insulation from vibration is being installed in a building
where a reciprocating machine is in use. The machine causes the building floor to vibrate in a
harmonic motion at frequency of 1000 cycles per minute. The equipment is going to be installed
on four springs of equal stiffness and negligible damping. Determine the stiffness of each spring if
the amplitude of the transmitted vibration is to be limited to <15% of the floor vibration.

Solution
After idealizing the system as an SDOF system, use transmissibility with ζ = 0 to determine the
required stiffness to limit the vibration transmitted.

1+ ( 2rζ )2 1
Tr = =
2 2
(1− r ) + ( 2rζ ) 2 ±(1− r 2 )

1. Determine the maximum required frequency ratio, r. Since Tr < 1, r > 2, so, apply the
negative sign throughout the denominator to obtain

1
Tr =
(r 2 − 1)
78 Introduction to Earthquake Engineering

and establish an inequality such that the transmissibility ratio is <0.15, that is,

1
< 0.15
(r 2 − 1)

solving this inequality for r, we determine that frequency ratio, r, must be >2.77.
2. Determine the required natural cyclical frequency. Use the frequency ratio in the follow-
ing form:

f
r= > 2.77
fn

where; f = 1000 cycles/min (1 min/60 s) = 16.67 Hz.

The frequency ratio equation then gives a natural cyclical frequency of

16.67 Hz
> 2.77 ⇒ fn < 6.02Hz
fn
3. Determine the mass

m = W /g = 50 lbs/(386.4in/s2 ) = 0.1294 lbs ⋅ s2 /in

4. We now determine the total required stiffness to limit the vibration transmitted.

1 kt 1
fn = < 6.02
2π m s

Solving for the total stiffness, kt < 185 lbs/in.

As the instrument is supported on four springs, the stiffness of each spring must be
<46.3 lbs/in in order to limit the transmitted vibration to <15%.

3.2.4 Time-Dependent Support Accelerations

There are a number of cases where the support (or foundation) of a system is subjected to time-
varying displacements, such as seismic wave excitations, which is the main topic of this book.
The dynamic response of a system to support motion is similar to that of the same system subjected
to a time-varying force as discussed in the last section. Let us characterize the excitation motion of
the portal frame discussed in Section 3.1.5, shown in Figure 3.21; including the damping force in the
mathematical model. Here, we have three different displacements: ground, ug(t), relative, u(t), and
total with respect to the reference axis, ut(t). This can be accomplished by establishing the equation

ut
u
u ut
k fs
m fI
c fD

ug
ut (t) = ug (t) + u(t) FBD
ug

FIGURE 3.21  Idealized SDOF system and free-body diagram for a portal frame.
Single-Degree-of-Freedom Structural Dynamic Analysis 79

of motion for the system by applying equilibrium to the FBD of the mass using D’Alembert’s prin-
ciple, where the stiffness and damping forces are proportional to the relative displacement and the
inertial force to the total mass displacement. Again, we are using an oscillator model because it is
easier to visualize the combined effects of damping, stiffness, and inertia.
Horizontal equilibrium of the FBD shown in Figure 3.21 yields the equation of motion,

∑ F = 0;
x f I (t ) + f D (t ) + f S (t ) = 0

where:
fI (t ) = mut (t ) is the inertial force
f D (t ) = cu(t ) is the damping force and fS (t ) = ku(t ) is the stiffness force

So the equation of motion can be rewritten as

mut (t ) + cu(t ) + ku(t ) = 0

This equation can be further rewritten in terms of the relative and ground motion by substituting
the definition of total acceleration as ut (t ) = u(t ) + ug (t ) into the equation of motion,

mu(t ) + cu(t ) + ku(t ) = −mug (t ) (3.50)

The right-hand side is the effective support excitation loading that opposes the sense of ground
acceleration and is similar to the time-varying force discussed in Section 3.1.5.
Alternatively, we can write this equation in terms of the total displacement, in which case the
right-hand side is the effective loading; the resulting response is the total displacement of the mass
from a fixed reference, rather than displacement relative to the moving base:

mut + cu t + kut = kug (t ) + cug (t ) (3.51)

Before we consider an earthquake excitation, let us analyze a system subjected to a periodic,

harmonic base excitation (such as the dynamic action of machinery) given by

ug (t ) = ugo sin ωt

where:
ugo is the peak amplitude of the ground displacement
ω is the frequency of the support motion in this case

Substituting this displacement relation into Equation 3.51 yields

mut + cu t + kut = kugo sin ωt + cωugo cos ωt (3.52)

The two terms on the right-hand side can be combined into an equivalent form by performing
a trigonometric transformation such as the one performed in Section 3.1.1 to obtain u(t) = C sin
(ωnt + α),

mut + cu t + kut = Fo sin(ω t + α) (3.53)

80 Introduction to Earthquake Engineering

where:

Fo = (kugo )2 + (cω ugo )2 = ugo k 2 + (cω )2 = ugo k 1 + (2rζ )2

cω
tan α = = 2rζ
k

Equation 3.53 is the differential equation for the oscillator excited by the harmonic force Fo sin
(ωt + α); the solution to which is of the same form as that for Equation 3.43. That is, following stan-
dard methods used in the solution of nonhomogeneous differential equations such as the solution in
Equation 3.44, which includes the complementary (transient) and particular (steady state) solutions.
The steady-state solution describes the relative transmission of support motion to the mass,

Fo /k sin(ωt + α − φ)
u t (t ) =
(1 − r 2 )2 + (2rζ )2

or

u t (t ) 1 + (2rζ )2 sin(ωt + α − φ)
= (3.54)
ugo (1 − r 2 )2 + (2rζ )2

Maximizing this relationship leads to the expression for transmissibility, which was discussed
earlier and given by Equation 3.49. That is, transmissibility of motion from the foundation to the
structure is given by the same function as the transmissibility of force from the structure to the
­foundation. This is an important expression in vibration isolation; in this context, it is the degree
of relative isolation defined as the ratio of the amplitude of motion of the system to the static
displacement:

uot 1 + (2rζ )2 ut
Tr = = = o (3.55)
ust (1 − r 2 )2 + (2rζ )2 ugo

where:
ust = Fo/k is the static displacement

Again, we note that this is a function of the frequency ratio, r, and the damping ratio, ζ, only.
This can also be expressed in terms of acceleration,

uot 1 + (2rζ )2
Tr = = (3.56)
ugo (1 − r 2 )2 + (2rζ )2

Figure 3.22 shows graphs depicting the transmissibility as a function of frequency ratio,
r, for ­various values of ζ. This graph is similar to that shown in Figure 3.18.
From this graph, the following observations are noted:

1. All curves pass through the same point at r = 2 .

2. Increasing damping when r < 2 increases the effectiveness of the vibration isolation
system.
Single-Degree-of-Freedom Structural Dynamic Analysis 81

3
ζ=0
2.5
ζ = 1/5

Transmissibility, Tr
2
ζ = 1/4
1.5
ζ = 1/3
ζ = 1/2
1
ζ = 1/2
ζ = 1/3
0.5
ζ = 1/4
ζ = 1/5
0
0 0.5 1 2 1.5 2 2.5 3
Frequency ratio, r

FIGURE 3.22  Tr versus r for various values of ζ.

3. As r approaches zero (slowly varying displacements), uot = ugo; that is, the mass moves
rigidly with the ground motion.
4. r = 1 again leads to resonance; as ζ→0, Tr→∞.
5. Increasing damping when r > 2 decreases the effectiveness of the vibration isolation
system.
6. As r approaches infinity (rapidly varying displacements), uot = 0; that is, the mass stays still
while the ground beneath it moves.

These observations lead to two important conclusions regarding vibration isolation:

1. Since Tr > 1 for r < 2 and Tr < 1 for r > 2, vibration isolation can be achieved only in
the range of r > 2 .
2. Since damping increases transmissibility in the isolation range (r > 2 ), the most effective
vibration absorbers consist of spring elements having little or no damping. This implies a
tradeoff when selecting a soft spring to reduce the transmitted force but increases the static
displacement or vice versa with a stiff spring.

Displacement transmitted to foundations can also be determined by considering the relative

­displacement, u = ut – ug, which can be written as

uo r2
= (3.57)
ugo (1 − r 2 )2 + (2rζ )2

EXAMPLE 3.14
A 50-lb package is suspended in a box (damping of 2%) by two springs each having a stiffness
k of 250 lb/in. The box is transported on the bed of a truck, which due to the suspension motions
­experiences vertical harmonic excitations of amplitude u(t) = 1.5 in sin (2 rad/s ⋅ t). The owner
of the box has specified that the maximum relative displacement of the package be <0.05 in.
Does the system satisfy this requirement?

Solution
There are two ways to solve this problem; we can directly determine the relative displacement or
use transmissibility to determine the total displacement and then compute the relative displace-
ment. In both cases, we first need the frequency ratio.
82 Introduction to Earthquake Engineering

1. Determine the frequency ratio of the idealized SDOF system. Mass,

m = W /g = 50 lbs/ (386.4 in/s2 ) = 0.1294 lbs ⋅ s 2 /in

Stiffness,

kt = 2k = 2(250 lbs/in) = 500 lbs/in

Natural frequency,

kt 500
ωn = = = 62.16 rad/s
m 0.1294

Also, the forcing frequency is given in the displacement equation as

ω = 2 rad/s

So the frequency ratio is

ω 2
r= = = 0.0322
ωn 62.16

2. Determine the relative displacement of the mass. Using the transmissibility ratio defi-
nition of Equation 3.57, the relative displacement of the system can be determined.
ugo is the amplitude of the system displacement given in the displacement equation as
1.5 in,

uo r2
=
ug o (1− r )2 + ( 2rζ )2
2

ug o r 2 (1.5)(0.0322)2
⇒ uo = = = 0.00155′′ < 0.05′′
(1− r 2 )2 + ( 2rζ )2 (1− (0.0322)2 )2 + (2(0.0322)(0.05))
2

The displacement limit is satisfied!

The second approach utilizes the transmissibility ­relationship in Equation 3.55, where
ugo = 1.5 in again!

uot 1+ ( 2rζ )2
Tr = =
ug o (1− r 2 )2 + ( 2rζ )2
2
ug o 1+ ( 2rζ )2 (1.5) 1+ (2(0.0322)(0.05))
⇒ uot = = = 1.50155′′
(1− r 2 )2 + ( 2rζ)2 (1− (0.0322)2 )2 + ( 2(0.0322)(0.05))2

Here the relative displacement is given by

uo = uot − ug o = 1.50155′′ − 1.5′′ = 0.00155′′

Same as above!
Single-Degree-of-Freedom Structural Dynamic Analysis 83

3.3 BASE SHEARS AND STRESSES CAUSED BY ­TIME-DEPENDE­NT​

FORCES AND SUPPORT EXCITATIONS
After determining the maximum dynamic displacement, either from a dynamic excitation or force,
we can conduct a static structural analysis to determine element forces (bending moment, shear
force, and axial force) and stresses needed for design; no additional dynamic analysis is necessary.
For the simple portal frame system that we have considered in this chapter, we can follow one of two
different approaches to conduct this structural analysis:

1. The dynamic effect can be replaced with a slowly applied force that produces deforma-
tion umax = uo (or acceleration ümax = üo), which is determined by a dynamic analysis. As
shown in the prior two sections, this displacement can equally be produced by a dynami-
cally applied force or a support excitation. Displacements can be used to determine an
equivalent, external static force, Fs,max, as shown in Figure 3.23, which can in turn be used
to determine all internal forces.
2. With known dynamic displacements at column ends, the forces (moment and shear) can
be determined using individual element stiffness via fixed-end moment analyses, see a
structural analysis textbook. Figure 3.24 depicts two cases we have used in this chapter.
The first case shows the base shear and moment, which are also equal to the maximum
internal shear and moment. The second case has maximum shear and moment at the top
and bottom of the column.

In both cases, maximum stresses (shear and flexural) can be determined using standard strength
of materials procedures. For example, the maximum shear stress is

VQ
τ=
It

uo
Fs,max

Tributary mass
Fs,max = k uo = m üo

Lateral stiffness

FIGURE 3.23  Equivalent static force.

uo
uo

3EI V 12EI
Vb = kuo = uo M V = kuo = uo
h3 h3
EI EI
Mb = hVb = 3 uo M=6 uo
h2 h2

M
Vb V
Mb

FIGURE 3.24  Internal shear force and moment for pinned–fixed and fixed–fixed columns.
84 Introduction to Earthquake Engineering

where:
Q is the first moment of the area above the plane in question
I is the moment of inertia of the cross section
t is the width of the section at the plane in question

This equation yields the following maximum stresses for rectangular and I-shaped sections:
Rectangular section,

Vmax
τ max = 1.5 (3.58)
A

I-shaped section,
Vmax
τ max ≈ (3.59)
Aweb

where:
A is the area of the entire rectangular cross section
Aweb is the area of the web; in practice, this is equal to the product of the depth of section and
thickness of the web

The maximum flexural stress can be determined as

M max
σmax = (3.60)
S

where:
S is the section elastic modulus

EXAMPLE 3.15
Given the building frame shown in Figure E3.17 with damping of 5% and load p(t) = 200 lbs sin 5.3 t,
determine (a) maximum displacement u, (b) the maximum base shear, and (c) the maximum normal
stresses in the columns (Ix = 75 in4 and Sx = 18 in3 for W8 × 21). Let us assume that the beam is rigid.

Solution
1. Determine mass, stiffness, and natural frequency of the SDOF system.
The building frame can be modeled as an SDOF system assuming that the horizontal
beam is rigid and only lateral deformations of the columns occur. The stiffness of the
system is then the sum of column lateral stiffnesses. The mass and stiffness of the SDOF
system are calculated as follows:

u
p(t) 15 kips

15 ft

FIGURE E3.17  Structural frame and loading.

Single-Degree-of-Freedom Structural Dynamic Analysis 85

Mass,

W 15 kips(1000 lb/kip)
m= = = 38.86 lb ⋅ s2 /in
g 386 in/s 2

Stiffness,

 4 
k=
3EI 3EI
+ = 2 3( 29, 000, 000 psi)(75 in )  = 2237.6 lb/in
L3 L3  3 
 (15 ft ×12 in/ft ) 

The natural frequency of the SDOF system is determined using Equation 3.2:

k 2237.6 lb/in
ωn = = = 7.59 rad/s
m 38.86 lb ⋅ s2 /in

2. Determine the maximum steady-state displacement.

The maximum steady-state ­displacement, uo, due to the applied forcing function can
be determined using the d­ efinition of the deformation response factor in Equation 3.47.
The equivalent static displacement can be calculated using the basic force–­
displacement relationship for the SDOF system as follows:

po 200 lb
(ust )o = = = 0.08938 in
k 2237.6 lb/in

Also, given forcing frequency of ω = 5.3 rad/s, the ratio of forcing frequency to natu-
ral ­frequency can be determined as follows:

ω 5 .3
r= = = 0.6984
ωn 7.59
Now substituting this value into Equation 3.47, the maximum steady-state
displacement is

(ust )o 0.08938 in
uo = = = 0.1729 in
(1− r 2 )2 + ( 2rζ )2 (1− 0.6982 )2 + ( 2(0.698)(0.05))2

3. Maximum base shear.
Calculate maximum base shear at the support for each column using the force–dis-
placement relationship:

 
 3EI   3(29, 000, 000 psi)(75 in4 ) 
Vmax = kuo =  3  uo =   0.1729 in = 193.5 lb
 L   (15 ft ×12 in/ft )
3

 
4. Maximum normal stress due to bending.
The maximum bending moment at the top of the column is calculated from equilib-
rium of the column member. As the column has a pin support, summing moments about
the top of the column will result in the ­following ­maximum internal bending moment at
the top of the column (see Figure 3.24):

Mmax = VmaxL = 193.5 lb(15 ft ×12 in/ft ) = 34, 822 lb ⋅ in

Applying the flexure formula from mechanics of materials (Equation 3.60), the normal
stress in the column due to the bending moment is determined as
86 Introduction to Earthquake Engineering

Mmax 34, 822 lb ⋅ in

σmax = = = 2048 psi
S 17 in3

PROBLEMS
3.1 Determine the effective weight at the roof and floor levels for a 60 ft by 150 ft, three-
story office building with equal story heights of 10 ft and the following loading:
Roof DL  = 20 psf
Roof LL  = 20 psf
Floor DL  = 20 psf
Floor LL  = 60 psf
Wall DL  = 10 psf
Also, the partition load is 10 psf as the occupancy is designated as office building
and the floor live load is <80 psf.
3.2 The bridge beam depicted below supports a rigid deck that weighs 200 kips. Assuming
that the weight is lumped at midspan, determine the natural frequency and period of the
system (E = 29,000 kip/in2 and Ix = 13,000 in4).

50 ft

3.3 Given a water tank supported on a slender (I = 10,000 in4), concrete (E = 3 × 107 psi)

column as shown, determine the natural frequency and period of the system.

W = 5 kips

30 ft

3.4 Given the following steel (E = 29,000 kip/in2) building frame, determine the period.
The columns have Ix = 75 in4 and the beam has Ix = 150 in4.

15 kips

15 ft

30 ft
Single-Degree-of-Freedom Structural Dynamic Analysis 87

3.5 The structural system depicted below has a roof weighing 22.5 psf and side sheathing
weighing 10 psf. The total weight is assumed to be concentrated at the bottom of the
roof trusses. The lateral force-resisting system in the North–South direction (one on
each side) consists of bracing with ½-inch steel rods (E = 29,000 ksi). Determine the
natural period of the building.

1 2
4′

12′

A B C
1

N
40′

2 @ 40′

3.6 Given the following structural system and properties, determine the total stiffness and
structural period.

W = 750-kips

I1 = I3 = 25,000 in4, I2 = 60,000 in4

15 ft I3
20 ft I1 E = 3,000 ksi
30 ft I2

3.7 Consider the following building frame with a rigid beam that is pin connected

to one c­olumn and rigidly connected to the other as shown; each column has
EI = 40,000,000 kip-in2. Determine the natural frequency and period of the system.
p(t) 200 kips

10 ft pin

15 ft

3.8 After constructing a bridge pier, the owner decides to determine the dynamic proper-
ties of the system. This type of system can be modeled as a concentrated mass atop a
weightless tower. To determine the properties, a cable is attached to the top and pulled
with a crane with a force of 20 kips, which causes a horizontal displacement of 1 in.
The cable then is suddenly cut and the resulting free vibration is recorded. At the end of
10 complete cycles, the time is 2 s and the amplitude is 0.14 in. From this information,
88 Introduction to Earthquake Engineering

compute: (a) undamped natural period, (b) effective stiffness, (c) effective weight, and
(d) effective damping coefficient.

Pier cap

3.9 A wind turbine can be modeled as a concentrated mass atop a weightless tower.
To ­determine the dynamic properties of the system, a cable is attached to the turbine
and a crane. A lateral force of 200 pounds is applied to the turbine causing a horizontal
displacement of 1.0 in. When the cable is suddenly cut, the resulting free vibration is
recorded. At the end of four complete cycles, the time is 1.25 s and the amplitude is
0.64 in. From this information, compute the: (a) undamped natural period, (b) effective
stiffness, (c) ­effective weight, and (d) effective damping coefficient.

3.10 To determine the dynamic properties of a simply supported bridge girder, a cable is
attached to its midspan and pulled with a crane. To produce a 4.0-inch deflection at
midspan, the crane exerts a vertical force of 10 kips. When the cable is suddenly cut,
Single-Degree-of-Freedom Structural Dynamic Analysis 89

the resulting free vibration is recorded. At the end of five complete cycles, the time
is 0.25 s and the amplitude is 2.7 in. From this information, compute: (a) undamped
natural period, (b) effective stiffness, (c) effective weight, and (d) effective damping
coefficient.

10 kip, which is released when the cable is cut

3.11 A vibration isolation block weighing 2000 lbs is to be installed in a laboratory on four

springs so that the vibration from adjacent factory operations (as high as 1500 cycles
per minute) will not disturb certain experiments. Determine the stiffness of system,
k, such that the motion of the block is limited to 10% of the floor vibration; neglect
damping.
3.12 A sensitive instrument that weighs 250 lb is mounted in the New Engineering Building
in one of the labs. The instrument currently experiences a vibration amplitude of 0.4 in
when an air compressor runs at a frequency of 35 Hz. Neglecting damping, determine
(a) the stiffness necessary for a mount in order to achieve 80% isolation and (b) the
resulting vibration amplitude of the instrument.

Instrument

3.13 A 500-pound tiger that was rescued will be placed inside a box and transported on
a flatbed truck. The box will be supported by four springs, each having stiffness of
250 lb/in. The damping for the box supporting system has been estimated at 2%. During
the transport process, it is estimated that the box will experience vertical displace-
ments of u(t) = 1.5 in ⋅ sin 4 rad/s ⋅ t. For the tiger to have a comfortable trip, maximum
­relative displacement and velocity of the box should be <0.05 in and 1.0 in/s, respec-
tively. Does the spring damping system satisfy these requirements?
3.14 Given the following building frame with damping of 5% and load p(t) = 200 lbs

sin(5.3 rad/s ⋅ t), determine (a) maximum displacement u, (b) the maximum base shear,
and (c) the maximum normal stresses in the columns (Ix = 75 in4 and Sx = 18 in3 for
W8 × 21). Assume that the beam is rigid.
p(t) 250 kips

pin
10 ft

15 ft
90 Introduction to Earthquake Engineering

3.15 The following frame has a rigid beam, a diagonal 1/2-in diameter steel rod (E = 29,000
ksi) brace, and two steel columns (E = 29,000 ksi and Ix = 82.7 in4). Determine the maxi-
mum force in the diagonal member and the maximum shear force in each of the c­ olumns
when the frame is subjected to force p(t) = 900 lbs sin 8 rad/s ⋅ t. Damping is 5%.
p(t) 600 kips pin

10 ft
15 ft
pin

20 ft

REFERENCES
American Society of Civil Engineers, ASCE/SEI Minimum Design Loads for Buildings and Other Structures,
Edition, American Society of Civil Engineers, Reston, VA, 2010.
Chopra, A. K., Dynamics of Structures: Theory and Applications to Earthquake Engineering, 4th edition,
Prentice-Hall, Upper Saddle River, NJ, 2012.
Lindeburg, M. R. and K. M. McMullin, Seismic Design of Building Structures: A Professional Introduction
to Earthquake Forces and Design Details, 9th edition, Professional Publications, Inc., Belmont,
CA, 2008.
Silva, P. and S. S. Badie, Optimum beam-to-column stiffness ratio of portal frames under later loads.
Structure Magazine, August 2008.
Villaverde, R., Fundamental Concepts of Earthquake Engineering, CRC Press, Boca Raton, FL, 2009.
 4 Response to General Loading

After reading this chapter, you will be able to:

1. Determine the response of a single degree of freedom (SDOF) system to a general load by
the direct and numerical integration of Duhamel’s integral
2. Use shock spectra to solve general dynamics load problems
3. Perform the direct numerical solution of equation of motion to solve for the response of an
SDOF system subjected to a general dynamic load

Thus far, we covered the response of structures to harmonic loadings; however, the ground vibra-
tion of a structure during an earthquake is nonperiodic, and it can be of short or long duration. The
response of an SDOF system subjected to a nonperiodic force can be determined using a convo-
lution integral, Laplace transform, or numerical methods. The convolution integral and Laplace
­transform approaches are analytical methods that can be used to evaluate the response of the
­system in a closed form when the nonperiodic forcing function can be described analytically. The
­analytical approaches are useful in describing the response of an SDOF system and characterizing
the ­influence of parameters on system response, as well as informing system design. Numerical
methods are often employed in cases where an analytical solution is difficult or not available under
a general forcing function.
In this chapter, we first focus on the use of a convolution integral known as Duhamel’s integral
in the context of structural dynamics to analyze an SDOF system subjected to a general forcing
­function. Duhamel’s integral is based on the response of the system due to an impulse, which is a
force that abruptly jumps from zero to a constant value. Several examples using Duhamel’s inte-
gral are ­outlined in order to demonstrate the analytical solution to nonperiodic, general ­forcing
functions; these include step, rectangular pulse, and ramp forcing functions. These analyti-
­
cal ­examples are used to introduce the concept of the shock spectra for a given forcing function
and  how  shock  ­spectra  can  be used to determine the maximum response of structures to such
loadings.
In addition, we demonstrate the use of numerical methods to evaluate the response of an SDOF
system to a seismic ground motion. In particular, we provide examples of how an earthquake
response spectrum is generated for a given earthquake event.

4.1  RESPONSE OF AN SDOF SYSTEM TO AN IMPULSE

In this section, we determine the response of an SDOF system to a nonperiodic forcing function
using Duhamel’s integral. A nonperiodic forcing function, or a general forcing function, is charac-
terized by the magnitude of the force, which varies with time and occurs during a time period of
specified length. An impulsive force represents the most basic nonperiodic forcing function. An
impulse is the effect of a force on a structure over a very short time span. Figure 4.1 illustrates an
impulse loading function, p(t), that is applied to an SDOF system and shows the impulse calculated
as the area of the shaded rectangle. The impulse represents a force that acts on the mass of a struc-
ture for a specific interval of time, Δτ.
Using Newton’s second law of motion, the equation of motion can be written in terms of velocity
 Notice that we only include the effects of mass because damping and stiffness have negligible
u.
effect on the maximum response for impulsive loadings since there is little time for mass to experi-
ence a displacement.

91
92 Introduction to Earthquake Engineering

p(t)
p(t) m
p Impulse = p∆τ

0
τ τ + ∆τ
u(t) p∆τ ⋅ g(t − τ)
Undamped
Damped

0 ⋅ t
u(τ)

FIGURE 4.1  Impulse loading applied to an SDOF system and response.

For the equilibrium of mass, we obtain:

du

∑ F = ma : p(t ) = m dt (4.1)

where:
p(t) represents the general loading function

Rearranging Equation 4.1 and integrating the result over the limits of t1 = τ to t2 = τ + Δτ result
in the principle of linear impulse and momentum (the product of mass and velocity), which directly
relates a change in momentum mdu to the impulse applied to the SDOF system p(t)dt, that is, the
magnitude of an impulse is equal to a change in momentum,

mdu = p(t ) dt (4.2)

Integrating,
t2 τ +Δτ

mdu =
∫ p(t ) dt = ∫ τ
p(t ) dt = pΔτ = impulse (4.3)
t1

Solving for the change in velocity yields

p∆τ
du = (4.4)
m

Since the impulse acts for an infinitesimally short duration of time (Δτ is very small), the spring
and damper have no effect during the time the impulse is applied. The impulse essentially imparts a
velocity to the mass of the system. The change in velocity described in Equation 4.4 is equal to the
initial velocity of the system, since prior to time t = τ, the system was at rest.
Considering an underdamped SDOF system, subjected to the impulse at time t = τ as shown in
Figure 4.1, the equation of motion is given by

mu + cu + ku = 0 (4.5)

Response to General Loading 93

where: (like in Chapter 3)

m is the mass
c is the damping coefficient
k is the stiffness of the system

The initial conditions can now be defined using the initial velocity on the mass at time t = τ,
based on Equation 4.4. The initial velocity is given by

pΔτ
u(τ ) = (4.6)
m

and displacement before and up to the impulse at time t = τ is zero,

u(τ ) = 0 (4.7)

The solution to the equation of motion is given by the free vibration response of an underdamped
SDOF system shown in Equation 3.36 and rewritten in the following for an impulse applied at time
t = τ :

 u(τ ) + ζωnu(τ )  
u(t ) = e−ζωn ( t−τ )  sin ω (t − τ ) + u(τ )cos ω (t − τ ) (4.8)
 ωD 

D D

 

Substituting the initial conditions from Equations 4.6 and 4.7 into Equation 4.8, the response of
an underdamped SDOF subjected to an impulse at time t = τ is as follows:

 pΔτ  
u(t ) = e−ζωn ( t−τ )  sin ω (t − τ )
 mωD  D
 (4.9)
 

Neglecting damping in the system (because the short impulse does not allow the damping
­mechanism to absorb much energy), the response function is given by

 pΔτ 
u(t ) =   sin ωn (t − τ ) (4.10)
 mωn 

The impulse response function of a damped and undamped SDOF system is shown in Figure 4.1.
If a unit impulse is applied to the system, such that pΔτ = 1, then the response of the SDOF
s­ ystem becomes

 1  
g(t − τ ) ≡ u(t ) = e−ζωn ( t−τ )  sin ω (t − τ )
 mωD  D
 (4.11)
 

Here, we define g(t − τ) as the unit impulse response function. The unit impulse response can
be multiplied by the magnitude of the impulse in order to ascertain the response of the system for
damped and undamped cases, as shown in Figure 4.1, u(t) = pΔτ ⋅ g(t − τ).
94 Introduction to Earthquake Engineering

4.2  GENERAL FORCING FUNCTION

A general loading function given by p(t), as shown in Figure 4.2, can be represented by a series of
short impulses having varying magnitudes at successive incremental times, one after another. At
some time, τ the force p(τ) acts on the system for a short duration of time Δτ. The impulse acting on
the system at time t = τ is given as p(τ)Δτ, and at any time t, the elapsed time since the pulse is t − τ.
Each impulse causes an incremental displacement response Δu(t) of the SDOF system. The incre-
mental displacement response to an impulse at time t = τ is given by multiplying the impulse p(τ)
Δτ and the unit impulse response function g(t − τ), as shown in Figure 4.1, and can be rewritten as

∆u(t ) = [ p(τ )∆τ ]g(t − τ ) for t > τ (4.12)

By summing the incremental displacement responses, the total displacement response from all
the incremental impulses can be determined,

u(t ) = ∑ ∆u(t ) ≅ ∑[ p(τ )∆τ ]g(t − τ ) (4.13)

As Δτ becomes infinitesimally small or approaches zero, the summation can be replaced with
integration, yielding the following result:

u(t ) =
∫ p(τ )g(t − τ ) dτ (4.14)
0

Then substituting the definition of the impulse response function from Equation 4.11 into
Equation 4.14, we can determine the total response of an underdamped SDOF system at any time t
as the superposition of the response to all the impulses occurring before t, that is,

p(t)
2
1

t
τ τ + ∆τ
du(t)
Response to impulse at τ

t
τ
u(t)
Total response

FIGURE 4.2  Discretization of a general loading function and response.

Response to General Loading 95

t
1
u(t ) =
mωD ∫ p(τ )e −ζωn ( t −τ )
[sin ω D (t − τ )] d τ (4.15)
0

Equation 4.15 is a convolution integral and is known as Duhamel’s integral. It allows for the deter-
mination of the response of an underdamped SDOF system to any general forcing function, and is par-
ticularly useful for determining the effect of random excitations associated with ground motion during
earthquakes. Also, if the forcing function p(t) cannot be expressed by an analytical equation, which is
often the case in earthquake-induced ground motion, the integral of Equation 4.15 can be evaluated
using numerical methods. It is important to recognize that Equation 4.15 assumes that the system is at
rest before the application of the forcing function p(t). Also, Equation 4.15 accounts for the particular
solution of the response up(t) as described in Chapter 3, with the total solution given by Equation 3.23.
If an initial displacement u(0) and initial velocity u( 0) were present, and the complementary (free
vibration response) part of the solution given by Equation 3.36 would be added to Equation 4.15, then
the total response of an underdamped case including initial conditions is

 u(0) + ζωnu(0)  
u(t ) = e−ζ ωnt   sin ω t + u(0)cos ω t 
 ωD 

D D

 
t
1
+
mωD ∫ p(τ )e −ζωn ( t −τ ) sin ω D (t − τ ) d τ
  (4.16)
0

If damping in the system is neglected, then Equations 4.15 and 4.16 can be expressed as follows:
t
1
u(t ) =
mωn ∫ p(τ )[sin ω (t − τ )] dτ
n (4.17)
0

and
t
 u(0)  1
u(t ) = 
 ωn 
sin ωnt + u(0)cos ωnt +
mωn ∫ p(τ )[sin ω (t − τ )] dτ
n (4.18)
0

Again, it is important to note that Equation 4.17 assumes that the system is initially at rest,
whereas Equation 4.18 is given to show how the initial displacement, u(0), and the initial velocity,
 0), would be included in the response.
u(

EXAMPLE 4.1
Consider the undamped SDOF system, shown in Figure 4.1, subjected to the constant force shown
in Figure E4.1. Assuming that the SDOF system is initially at rest, use Duhamel’s integral to deter-
mine the displacement response, u(t), of the system.

p(t)

po

FIGURE E4.1  Step or constant force loading function.

96 Introduction to Earthquake Engineering

Solution
1. Describe the forcing function and equation of motion to be solved. The forcing function
p(t) can be described analytically as

p(t ) = po

This forcing function can be considered a step force, or a dynamically applied con-
stant force, because it is abruptly applied to the mass of the system as opposed to a force
that is applied very slowly inducing no vibration on the system and can be considered
­statically applied.
In this problem, damping is neglected and so the equation of motion to be solved is
given as

mu(t ) + ku(t ) = po

2. Apply Duhamel’s integral. Since we are neglecting damping effects in the system (ζ = 0)
 0) = 0), Equation 4.17 can be applied
and the system is initially at rest (u(0) = 0 and u(
directly. Substituting the forcing function into the equation and integrating results in the
response of the system,

t
1 po t
u(t ) =
mωn ∫ p sin ω (t − τ )dτ = mω
o n 2
n
cos ωn (t − τ )
0
0

or

po
u(t ) = [1− cos ωnt ]
mωn2

Recalling that stiffness k = ω n2m , the response can be written as follows:

po
u(t ) = [1− cos ωnt ] = ust[1− cos ωnt ]
k

where:
ust represents the static displacement of the SDOF system; that is, if the force po were
applied very slowly such that no vibrations were induced, then the displacement
would be  ust = po /k
The term [1− cos ωnt ] is the dynamic load factor (DLF) and represents the ampli-
fication of the displacement above the static displacement; that is, u(t)/ust.
3. Plot the response of the SDOF system. By rearranging the response in terms of the
DLF and substituting for ωn = 2π/Tn, the following equation can be used to represent the
response of the SDOF system:

u(t )  t 
= DLF = 1− cos 2π 
ust  Tn 

We can use this equation to obtain the displacement response of the SDOF system
subjected to a step force (Figure E4.2).The response of an SDOF system to the step force
p(t) = po is similar to the free vibration response of an undamped system except that it
is shifted up by the static displacement ust. The maximum displacement (umax = 2ust) of
an SDOF system subjected to a suddenly applied force po is twice the displacement of
Response to General Loading 97

1.5
u(t)/ust

0.5

0
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
t/Tn

FIGURE E4.2  Response of an undamped SDOF to step force.

the same SDOF system when an equivalent static force po is applied very slowly, and it
occurs at ωt = π or when t = T/2.

EXAMPLE 4.2
Use Duhamel’s integral to determine the response of an undamped SDOF system subjected to a
rectangular impulse that is applied from 0 < t < td. The rectangular impulse is a suddenly applied
constant force, po, which is applied for a specified duration of time, td, as shown in Figure E4.3.
Assume that the SDOF system is initially at rest.

Solution
1. Describe the forcing function and equation of motion to be solved. The forcing function
p(t) can be described analytically as

 po t ≤ td
p(t ) = 
0 t > td

By superposition, the given forcing function can be described as a sum of step func-
tions where the first forcing function p1(t), with a magnitude of po, starts at t = 0, and the
second forcing function p2(t), with a magnitude of –po, starts at t = td. This is analytically
described as

p1(t ) = po

 0 t ≤ td
p2(t ) = 
− po t > td

and graphically illustrated using superposition as the sum of step forces shown in
Figure E4.4.

p(t)

po

t
td

FIGURE E4.3  Rectangular impulse loading function.

98 Introduction to Earthquake Engineering

p(t) p1(t) p2(t)

po po

= + td
t t t
td

–po

FIGURE E4.4  Illustration of the rectangular impulse as the sum of two step force loading functions.
2. Apply Duhamel’s integral. The response of the SDOF system to the step force p1(t) = po
can be determined by applying Duhamel’s integral and substituting the forcing function.
This result is the same as the response determined in Example 4.1 and rewritten in the
following for convenience:

u1(t ) = ust[1− cos ωnt ]

where:
u1(t) is the response of the SDOF system to the first forcing function p1(t)

The second forcing function p2(t) is a time-delayed step force with a magnitude of
–po, and the response to this second forcing function can also be determined by substi-
tuting the forcing function and evaluating Duhamel’s integral as follows:

t
1 − po t
u2 (t ) =
mωn ∫ − p sin ω (t − τ ) dτ = mω
td
o n 2
n
cos ωn (t − τ )
td

−po
u2(t ) = [1− cos ωn (t − td )] = ust[cos ωn (t − td ) − 1]
k

There are two response regions due to the rectangular pulse: during the pulse when
t ≤ td, the total response of the system is given by u1(t), and after the pulse when t > td, the
total response of the system is given by the two responses, u1(t) + u2(t), that is,

 u1(t ) t ≤ td  ust[1− cos ωnt ] t ≤ td

u(t ) =  or u(t ) = 
u1(t ) + u2(t ) t > td ust[cos ωn (t − td ) − cos ωnt ] t > td

which can be written in terms of the DLF = u(t)/ust and a dimensionless time parameter,
the ratio of duration of the loading to the natural period of the structure, td/Tn:

  t 
 1− cos 2π  t ≤ td
  Tn 
u(t ) 
DLF = =
ust cos 2π  t − td  − cos 2π t 
   
t > td
  Tn Tn  Tn 

3. Plot the response of the SDOF system. We can use these equations for different values of
td /Tn to obtain the displacement response of the SDOF system subjected to a rectangular
pulse force as shown in Figure E4.5.

Although we can find closed-form solutions to some general forcing function problems, it is more
efficient to determine the results using computer software, such as MATLAB®. In this chapter, we
use MATLAB to develop short scripts to solve many of the problems encountered in structural
Response to General Loading 99

1.5 2
td td
1 = 0.25 = 0.5
Tn Tn
1
0.5
u(t)/ust

u(t)/ust
0 0

–0.5
–1
–1

–1.5 –2
0 1 2 3 4 5 0 1 2 3 4 5
t/Tn t/Tn

2 2
td td
=1 =2
1.5 Tn 1.5 Tn

1 1
u(t)/ust

0.5 u(t)/ust 0.5

0 0

–0.5 –0.5
0 1 2 3 4 5 0 1 2 3 4 5
t/Tn t/Tn

FIGURE E4.5  Response of an SDOF system due to a rectangular pulse and varying pulse durations.

dynamics. And in later chapters, we will use these MATLAB scripts to solve earthquake engineer-
ing problems. These programs have powerful algorithms that can perform complex calculations
efficiently. In fact, modern analysis and design have become highly dependent on these computa-
tional tools.
MATLAB (MATrix LABoratory) was originally designed as an interactive software system
for matrix operations, such as solving systems of linear equations, and computing eigenvalues and
eigenvectors. More recent versions of the program include extensive graphics capabilities that can
be used to easily plot results, such as those shown in Examples 4.1 and 4.2. It also has an extensive
library of functions, including the one that performs convolution integrals numerically (the conv
function), as well as differentiation and integration. To facilitate the development of a script to
perform the convolution operation, we first rewrite Equation 4.17 (or Equation 4.15) in terms of two
functions that are then discretized into dimensionless vectors. The first function includes the shape
of the pulse and is written as a generalized forcing function p(t) = pox(t) that can be substituted into
Equation 4.17 (as p(τ) = pox(τ)),
t
p
u(t ) = o
mωn ∫ x(τ ) ⋅ sin ω (t − τ ) dτ
n (4.19)
0

We also write this equation in terms of a dimensionless time factor, the ratio of time to the natu-
ral period of the structure, t/Tn. That is, substituting ωn = 2π/Tn and ωn = k /m , we obtain
t t
ω p 2π po 2π
u(t ) = n o
ωn mωn ∫
0
x(τ ) ⋅ sin ωn (t − τ ) d τ =
Tn m
( k /m ) ∫
x(τ ) ⋅ sin
2
0
T n
(t − τ ) d τ

100 Introduction to Earthquake Engineering

or
t
u(t ) τ  t τ 
ust
= 2π
∫ x   ⋅ sin  2π  −  d τ
 Tn    Tn Tn 
 
(4.20)
0
where ust = po/k. Equation 4.20 can be written in convolution form as
t
u(t ) τ   t τ 
DLF =
ust
= 2π
∫ x   ⋅ h  2π  −  dτ
 Tn    Tn Tn 
 
(4.21)
0
where the function h(ξ) = sin(ξ), and ξ is the argument of the function, which is the quantity in the
brackets in this case.

EXAMPLE 4.3
Use the convolution function in MATLAB and Equation 4.21 to solve for the response due to the
rectangular impulse shown in Figure E4.3.

Solution
1 t ≤ td
x(t ) = 
0 t > td

1. The MATLAB script is as follows:
clear all
clc
% specify the duration of the pulse td/T
tdT = input ('Enter the value for td/T ratio: ');
% create a vector with n equally spaced points to represent t/Tn
n = 500;
tT = linspace(0,5,n);% loop over t/Tn to create vector of input pulse
for i = 1:length(tT)
if tT(i) <= tdT
p(i) = 1;
else
p(i) = 0;
end
end
% integrate pulse function to get response
dt=tT(2)-tT(1);
p=p*dt;
h=sin(2*pi*tT);
uust=2*pi*conv(p,h);
%create plot
plot (tT, uust(1:length(tT)), 'LineWidth',2, 'Color',[0 0 0]);
xlabel ('t/T_n', 'FontSize', 12, 'FontName', 'Times New Roman',…
'FontAngle', 'italic');
ylabel ('u(t)/u_{st}', 'FontSize', 12, 'FontName', 'Times New Roman', …
'FontAngle', 'italic');

This script can be used to generate graphs of the normalized displacement response
of SDOF systems subjected to a rectangular impulse, identical to those obtained using
the results of Example 2.

EXAMPLE 4.4
Use the convolution function of MATLAB and Equation 4.21 to determine graphically the dynamic
response of a tower subjected to the load shown in Figure E4.6.
Response to General Loading 101

p(t)
p(t)
W = 38.6 kips
120k

k = 100 kips/in

t(s)
0.02 0.04 0.06

FIGURE E4.6  An SDOF system and trapezoidal impulse loading function.

Solution
1. Mass, stiffness, and natural frequency of the SDOF system. The tower can be modeled
as an SDOF system. The stiffness of the system is given as 100 kips/in and the mass can
be calculated as follows:

W 38.6 kips (1000 lb/kip)

m= = = 100 lb s 2 /in
g 386.4 in/s 2

The natural frequency of the SDOF system is determined using Equation 3.2,

k 100, 000
ωn = = = 31.6 rad/s
m 100

2. Obtain function x(t):

 50t t ≤ 0.02 s
 0.02 s < t ≤ 0.04 s
 1
x(t ) = 
3 − 50t 0.04 s < t ≤ 0.06 s
 0 t > 0.06 s

Also, po = 120 kips
3. Displacement response as a function of time script:

clear all
clc
% input data
omega = 31.6; % natural frequency in rad/sec
po=120000; % peak force in lbs
m=100; % mass in lbs per square second
dt=0.002; % time increment for calculations
% loop over t to create vector based on pulse function values
for i = 1:500;
t(i)=dt*i;
if t(i) <= 0.02
p(i) = 50*t(i);
elseif t(i)<= 0.04
p(i) = 1;
elseif t(i)<= 0.06
p(i) = 3-50*t(i);
else
p(i) = 0;
end
end
102 Introduction to Earthquake Engineering

u(in)
0

–1

–2
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
t(s)

FIGURE E4.7  Response of an SDOF system to trapezoidal loading function.

p=p*dt;
h=sin(omega*t);
uust=po*conv(p,h)/(m*omega);
uustr = max(abs(uust)) % maximum displacement
%create plot
plot (t, uust(1:500), 'LineWidth',2, 'Color',[0 0 0]);
xlabel ('t(sec)', 'FontSize', 12, 'FontName', 'Times New Roman',...
'FontAngle', 'italic');
ylabel ('u(in)', 'FontSize', 12, 'FontName', 'Times New Roman', ...
'FontAngle', 'italic');

The maximum dynamic displacement is 1.3968 in versus the static displacement of

po /k = 1.2 in (Figure E4.7).

4.3  SHOCK SPECTRA

The results of Example 4.2 (Figure E4.5) show that each time we specify a different natural
period of the structure relative to the duration of the pulse (t d /Tn), we obtain a new response, each
of which has a maximum value. In design, only these peak values are necessary and not the entire
history of the response. The envelope of the maximum response as a function of the loading dura-
tion relative to the period of the structure can be graphically represented as a shock spectrum.
The shape of the envelope, or the shock spectrum, is also dependent on the shape of the pulse.
The response used to obtain shock spectra can be expressed in terms of displacement, velocity,
or acceleration. This is useful in ­determining the values of the peak (absolute) response that are
used in design.
To obtain shock spectra, we can solve Duhamel’s integral analytically or numerically, both
of which were introduced in the previous section. Numerical integration can also be performed
using one of the standard numerical integration techniques such as Euler (forward rectangular),
Trapezoidal (linear), or Simpson’s (parabolic) methods, all of which can be found in a standard
calculus textbook. In this book, we use the numerical integration operators available in MATLAB;
for Duhamel’s integral, we apply the conv function used in Example 4.3. Following are some exam-
ples used to generate the shock spectra using both analytical solutions and convolution numerical
­integration. Also, we demonstrate the use of these spectra in determining the maximum response
of structural systems.

EXAMPLE 4.5
Use the results from Example 4.2 to draw the shock spectrum for the rectangular pulse force, as
shown in Figure E4.2.
Response to General Loading 103

Solution
1. Use the DLF versus td/Tn results obtained in part 2 of Solution 4.2; that is,

  t 
 1− cos 2π  t ≤ td
u(t )  
 Tn 
DLF = =
ust  t t  t 
cos 2π  − d  − cos 2π  t > td

 
 Tn Tn  Tn 

The maximum values of the DLF, DLFmax, for various values of td/Tn can be obtained
more efficiently using MATLAB. The results of DLFmax versus td/Tn can then be graphed
to obtain the shock spectrum of a pulse loading, as shown in the Figure E4.8. The script
is as follows:

clear all
clc
% create two vectors of equal n length to represent td/Tn and t/Tn
n = 500;
tdT = linspace(0,5,n);
tT = linspace(0,5,n);
% loop over td/Tn and t/Tn to create the shock spectrum
for j = 1:n
for i = 1:n
if tT(i) <= tdT(j)
uust(i) = 1 - cos(2*pi*tT(i));
else
uust(i) = cos(2*pi*(tT(i)-tdT(j)))- cos(2*pi*tT(i));
end
end
uustr(j)=max(abs(uust)); % Select the max values from each response
end
% create figure
figure1=figure;
axes1 = axes('Parent',figure1);
xlim (axes1,[0 5]);
ylim (axes1,[-2.5 2.5]);
box (axes1, 'on');

p(t)
1.5
po
DLFmax

1
td t

0.5

0
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
td/Tn

FIGURE E4.8  Shock spectrum of an SDOF system subject to a rectangular impulse load.
104 Introduction to Earthquake Engineering

grid (axes1, 'on');

hold (axes1, 'all');
%create plot of the shock spectrum
plot (tdT,uustr,'LineWidth',2,'Color',[0 0 0]);
xlabel ('t_d /T_n','FontSize',12,'FontName','Times New Roman',...
'FontAngle','italic');
ylabel ('DLF_{max}','FontSize',12,'FontName','Times New Roman',...
'FontAngle','italic');

EXAMPLE 4.6
Write a MATLAB script using the conv function to plot the shock spectrum for the rectangular
pulse in Example 4.2.

Solution
clear all
clc
% create two vectors of equal n length to represent td/Tn and t/Tn
n = 500;
tdT = linspace(0,5,n);
tT = linspace(0,5,n);
% loop over td/Tn and t/Tn to create the shock spectrum
for j = 1:n
for i = 1:n
if tT(i) <= tdT(j)
p(i) = 1;
else
p(i) = 0;
end
end
% integrate pulse function to get response
dt=tT(2)-tT(1);
p=p*dt;
h=sin(2*pi*tT);
uust=2*pi*conv(p,h);
uustr(j)=max(abs(uust)); % Select the max values from each response
end
% create figure
figure1=figure;
axes1 = axes('Parent',figure1);
xlim (axes1,[0 5]);
ylim (axes1,[0 2.2]);
box (axes1, 'on');
grid (axes1, 'on');
hold (axes1, 'all');
%create plot of the shock spectrum
plot (tdT,uustr,'LineWidth',2,'Color',[0 0 0]);
xlabel ('t_d /T_n','FontSize',12,'FontName','Times New Roman',...
'FontAngle','italic');
ylabel ('DLF_{max}','FontSize',12,'FontName','Times New Roman',...
'FontAngle','italic');

This script creates a graph identical to the graph shown in Figure E4.8. The difference in the
two scripts is in the way they establish the two arrays that are plotted to generate the figure.
Example 4.5 uses the analytical solution, while Example 4.6 applies the conv function to obtain
the response array.

EXAMPLE 4.7
Draw the shock spectrum for the response of an SDOF system subjected to the symmetric trian-
gular pulse loading function shown in Figure E4.9.
Response to General Loading 105

p(t)

po

t
td/2 td

FIGURE E4.9  Symmetric triangular impulse loading function.

Solution
1. Obtain function x(t):

 2t /td t ≤ td / 2

x(t ) = 2 − 2t /td td / 2 < t ≤ td
 0 t > td


2. Displacement response as a function of time script: (Figure E4.10)

clear all
clc
% create two vectors of equal n length to represent td/Tn and t/Tn
n = 500;
tdT = linspace(0,5,n);
tT = linspace(0,5,n);
% loop over td/Tn and t/Tn to create the shock spectrum
for j = 1:n
for i = 1:n
if tT(i) <= tdT(j)/2
p(i) = 2*tT(i)/tdT(j);
elseif tT(i) <= tdT(j)
p(i) = 2-2*tT(i)/tdT(j);

1.6

1.4

1.2

1
DLFmax

p(t)
0.8
po
0.6

0.4

0.2
td/2 td t
0
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
td /Tn

FIGURE E4.10  Shock spectrum of an SDOF system subject to a symmetric triangular impulse load.
106 Introduction to Earthquake Engineering

else
p(i) = 0;
end
end
% integrate pulse function to get response
dt=tT(2)-tT(1);
p=p*dt;
h=sin(2*pi*tT);
uust=2*pi*conv(p,h);
uustr(j)=max(abs(uust)); % Select the max values from each response
end
% create figure
figure1=figure;
axes1 = axes('Parent',figure1);
xlim (axes1,[0 5]);
ylim (axes1);
box (axes1, 'on');
grid (axes1, 'on');
hold (axes1, 'all');
%create plot of the shock spectrum
plot (tdT,uustr,'LineWidth',2,'Color',[0 0 0]);
xlabel ('t_d /T_n','FontSize',12,'FontName','Times New Roman',...
'FontAngle','italic');
ylabel ('DLF_{max}','FontSize',12,'FontName','Times New Roman',...
'FontAngle','italic');

EXAMPLE 4.8

Draw the shock spectrum for the response of an SDOF system subjected to the triangular pulse
loading function shown in Figure E4.11.

Solution
1. Obtain function x(t):

1− t /td t ≤ td
x(t ) = 
 0 t > td

2. Displacement response as a function of time script (Figure E4.12):

clear all
clc
% create two vectors of equal n length to represent td/Tn and t/Tn
n = 500;

p(t)

po

t
td

FIGURE E4.11  Triangular impulse loading function.

Response to General Loading 107

1.8

1.6

1.4

DLFmax 1.2
p(t)
1
po
0.8

0.6

0.4
td t
0.2

0
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
td /Tn

FIGURE E4.12  Shock spectrum of an SDOF system subject to a triangular impulse load.

tdT = linspace(0,5,n);
tT = linspace(0,5,n);
% loop over td/Tn and t/Tn to create the shock spectrum
for j = 1:n
for i = 1:n
if tT(i) <= tdT(j)
p(i) = 1-tT(i)/tdT(j);
else
p(i) = 0;
end
end
% integrate pulse function to get response
dt=tT(2)-tT(1);
p=p*dt;
h=sin(2*pi*tT);
uust=2*pi*conv(p,h);
uustr(j)=max(abs(uust)); % Select the max values from each response
end
% create figure
figure1=figure;
axes1 = axes('Parent',figure1);
xlim (axes1,[0 5]);
ylim (axes1);
box (axes1, 'on');
grid (axes1, 'on');
hold (axes1, 'all');
%create plot of the shock spectrum
plot (tdT,uustr,'LineWidth',2,'Color',[0 0 0]);
xlabel ('t_d /T_n','FontSize',12,'FontName','Times New Roman',…
'FontAngle','italic');
ylabel ('DLF_{max}','FontSize',12,'FontName','Times New Roman',…
'FontAngle','italic');

EXAMPLE 4.9
Draw the shock spectrum for the response of an SDOF system subjected to the half-cycle sine
pulse force shown in Figure E4.13.
108 Introduction to Earthquake Engineering

p(t)

po

t
td

FIGURE E4.13  Half-cycle sine impulse loading function.

Solution
1. Obtain function x(t):

sin(πt /td ) t ≤ td
x(t ) = 
 0 t > td

2. Displacement response as a function of time script (Figure E4.14):

clear all
clc
% create two vectors of equal n length to represent td/Tn and t/Tn
n = 500;
tdT = linspace(0,5,n);
tT = linspace(0,5,n);
% loop over td/Tn and t/Tn to create the shock spectrum
for j = 1:n
for i = 1:n
if tT(i) <= tdT(j)
p(i) = sin(pi*tT(i)/tdT(j));
else
p(i) = 0;
end
end
% integrate pulse function to get response
dt=tT(2)-tT(1);

1.8

1.6

1.4

1.2

1
DLFmax

0.8
p(t)
0.6 po
0.4

0.2
td t
0
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
td/Tn

FIGURE E4.14  Shock spectrum of an SDOF system subject to a half-cycle sine impulse load.
Response to General Loading 109

p=p*dt;
h=sin(2*pi*tT);
uust=2*pi*conv(p,h);
uustr(j)=max(abs(uust)); % Select the max values from each response
end
% create figure
figure1=figure;
axes1 = axes('Parent',figure1);
xlim (axes1,[0 5]);
ylim (axes1);
box (axes1, 'on');
grid (axes1, 'on');
hold (axes1, 'all');
%create plot of the shock spectrum
plot (tdT,uustr,'LineWidth',2,'Color',[0 0 0]);
xlabel ('t_d /T_n','FontSize',12,'FontName','Times New Roman',...
'FontAngle','italic');
ylabel ('DLF_{max}','FontSize',12,'FontName','Times New Roman',...
'FontAngle','italic')

EXAMPLE 4.10
Given the building frame shown in Figure E4.15, which is subjected to a triangular impulse force
(see Figure E4.11) of amplitude po = 5 kips and duration td = 0.6 s, determine (1) maximum dis-
placement at the top, (2) the maximum base shear, and (3) the ­maximum bending stresses in the
columns (Ix = 82.7 in4 and Sx = 20.9 in3 for W8 × 24). Assume that the beam is rigid.

Solution
1. Mass, stiffness, and natural period and frequency of the SDOF system. The building
frame can be modeled as an SDOF system, assuming that only lateral deformations of
the columns occur. The stiffness of the system is the sum of column lateral stiffnesses.
The mass and stiffness of the SDOF system are calculated as follows:
Mass:

W 20 kips (1000 lb/kip)

m= = = 51.8 lb s 2 /in
g 386.4 in/s 2

  Stiffness (see Table 3.1); the 15 ft column is fixed–fixed, while the 20 ft column is
fixed–pinned,

12EI 3EI 12( 29, 000, 000 psi)(82.7 in4 ) 3( 29, 000, 000 psi)(82.7 in4 )
k= + 3 = + = 5455 lb/in

h13 h2 (15 ft × 12 in/ft )3 (20 ft × 12 in/ft )3

u
20 kips
F(t)

15 ft W8 × 24

20 ft

FIGURE E4.15  SDOF building frame geometry.

110 Introduction to Earthquake Engineering

The natural frequency of the SDOF system is determined using Equation 3.2,

k 5455
ωn = = = 10.3 rad/s
m 51.8

The natural period of the SDOF system is determined using Equation 3.5,

2π 2π
Tn = = = 0.61 s
ωn 10.3 rad/s

2. Determine the maximum displacement. The maximum displacement due to the applied
triangular pulse forcing function can be determined using the shock spectrum obtained
in Figure E4.12. This figure provides the maximum dynamic load factor, DLFmax = uo /ust
as a function of the ratio of pulse duration to natural period, td /Tn, where the equivalent
static displacement ust can be calculated using the basic force−­displacement relationship
for the SDOF system as follows:

po 5 kips (1000 lb/kip)

ust = = = 0.92 in
k 5455 lb/in

Also, the ratio of pulse duration to natural period can be determined as follows,

td 0 .6
= = 0.98
Tn 0.61 s

Finally, from the shock spectrum obtained in Figure E4.12, we can ­

determine
DFLmax ≅ 1.55. Now, solve for umax,

uo = (FDL)max ⋅ ust = 1.55(0.92 in) = 1.42 in

3. Maximum base shear. Calculate the maximum base shear at the support for each column
using the force−displacement relationship or the equilibrium of the column member as
described in Figure 3.24:
The 15-foot column:

 12EI   12( 29, 000 ksi)(82.7 in4 ) 

V1max = k1uo =  3  uo =   1.42 in = 7.01kips

 h1   (15 ft × 12 inn/ft )3 

The 20-foot column:

 3EI   3( 29, 000 ksi)(82.7 in4 ) 

V2max = k2uo =  3  uo =   1.42 in = 0.74 kips
 h2   ( 20 ft × 12 in/ftt )3 

4. Maximum normal stress due to bending. The maximum bending moment at the top of
each column is calculated from the equilibrium of the column member as described in
Figure 3.24.
The 15-foot column:

 6EI   6( 29, 000 ksi)(82.7 in4 ) 

M1max =  2  uo =   1.42 in = 631kip ⋅ in

 h1   (15 ft × 12 in/ft )2 
Response to General Loading 111

The 20-foot column:

 3EI   3( 29, 000 ksi)(82.7 in4 ) 

M2max =  2  uo =   1.42 in = 177 kip ⋅ in
 h2   ( 20 ft × 12 in/ft )2 

The maximum moment yields the maximum bending stress; thus, the 15-foot column
experiences the maximum stress. Applying the flexure formula from mechanics of mate-
rials (Equation 3.60), the normal stress in this column due to the bending moment is

M1max 631kip ⋅ in
σ max = = = 30.2 ksi
S 20.9 in3

4.4  RESPONSE TO GROUND MOTION

The equation of motion for ground excitation was derived in Chapter 3 (Equation 3.50) and is shown
here for convenience:

mu(t ) + cu (t ) + ku(t ) = − mug (t ) (4.22)

The total response of an underdamped SDOF system including initial conditions subjected to a
ground excitation, ug (t ), can be determined using Equation 4.16 with a new forcing function

p(t ) = −mug (t ) (4.23)

When the system starts from rest, we can determine the response due to a ground excitation for
an underdamped system in terms of the relative displacement of the mass with respect to the ground
by substituting Equation 4.22 into Duhamel’s integral (Equation 4.15). That is,

t
1
u(t ) = −
ωD ∫ u (t )e
g
−ζ ωn ( t −τ )
[sin ωD (t − τ )] d τ (4.24)
0

EXAMPLE 4.11
Draw the response of the frame shown in Figure E4.15, when subjected to the ground motion of
the North−South component of the horizontal ground acceleration recorded at the Imperial Valley
Irrigation District substation, El Centro, California, during the Imperial Valley Earthquake of May
18, 1940, from here on referred to as the El Centro earthquake. Assume 2% damping. The input
ground acceleration was obtained from http://www.vibrationdata.com/elcentro.dat, and is graphi-
cally shown in Figure E4.16.

Solution
1. The natural frequency of the frame ωn was obtained in Example 4.10 as 10.3 rad/s.
Using the conv function introduced in Section 4.2, we can write the following script to
perform the analysis:

clear all
clc
dam = 0.02; % given damping of 2%
wn = 10.3; % rad/sec, determined in Example 10
load elcentro.mat; % load El Centro data
N = length(elcentro); % number of points in the ground acceleration file
112 Introduction to Earthquake Engineering

üg(g)
0.4
0.3
0.2
0.1
0
–0.1 0 5 10 15 20 25 30
–0.2
–0.3
t(s)
–0.4

FIGURE E4.16  Ground motion acceleration time history for El Centro earthquake.

acc = elcentro; % ground acceleration data

DT = 0.02; % sampling rate
% create the time vector using the sampling rate
for i=N:-1:1
t(i)=i*DT;
end
% Integrate the ground acceleration using convolution
p=acc*DT;
h=exp(-dam*wn*t).*sin(wn*t);
u=conv(p,h)*386.4/wn;
uo=max(abs(u)) % determine maximum displacement
% create a new time vector adjusted to the conv length to graph response
for j=1:2*N-1
t2(j)=j*DT/2;
end
% create figure
figure1=figure;
axes1 = axes('Parent',figure1);
xlim (axes1,[0 20]);
ylim (axes1);
hold (axes1, 'all');
%create plot
plot (t2, u, 'LineWidth',2, 'Color',[0 0 0]);
xlabel ('t(sec)','FontName','Times New Roman','FontAngle','italic');
ylabel ('u(in)','FontName','Times New Roman','FontAngle', 'italic');

This gives a maximum displacement of 3.12 in, and the graphical response history
results are shown in Figure E4.17.

2
u(in)

–2
–3.12 in
–4
0 2 4 6 8 10 12 14 16 18 20
t(s)

FIGURE E4.17  Displacement response of the SDOF building frame due to El Centro earthquake.
Response to General Loading 113

Equation 4.23 can be rewritten as the DLF in terms of the relative displacement uo/ust or
the  total  accelerations uot /ugo . Additionally, we can normalize time with respect to the natural
period as a dimensionless time parameter, t/Tn. First, let ust = peff/k; assuming peff = mugo leads to
ust = mugo /k = ugo /ωn2, since k = ωn2 m and ugo is the peak value of the ground acceleration time his-
tory, ug . The DLF can then be written as
u ω 2u ut
DLF = o = n o = o (4.25)
ust ugo ugo

where uot is the total acceleration, which is related to the relative displacement by ωn2uo. From
Equation 4.21, we determine the convolution integral component of Equation 4.24,
t
τ   t τ 
DLF = 2 π
∫ x   ⋅ h  2π  −  d τ
 Tn    Tn Tn 
 
(4.26)
0

where x(ξ) is a function that describes the applied displacement (or acceleration) history function.
This implies that all of the shock spectra derived thus far are applicable to ground excitation forcing
functions as well.
However, generating a shock spectrum using the conv function in MATLAB is relatively simple.
For instance, in the following example, we use the El Centro earthquake ground acceleration pre-
sented in Example 4.11. This data will be used extensively in the remainder of the book. Also, the
terminology changes when dealing with seismic response; rather than shock spectrum, it is custom-
ary (and more appropriate) to refer to the shock spectrum as the response spectrum. The response
spectrum is the plot of the maximum response as a function of the SDOF system period. This topic
will be covered extensively in Chapter 5, including the development of the design response spec-
trum for elastic and inelastic structural systems.

EXAMPLE 4.12
Draw the response spectrum for the response of an SDOF system subjected to the El Centro
­earthquake ground acceleration.

Solution
The input ground acceleration can be obtained from http://www.vibrationdata.com/elcentro.dat
as a fraction of the acceleration due to gravity, g. In this case, we derive the response based on
Equation 4.23 and the natural period since we do not have a lapse time for the input forcing
­function. Thus, the displacement response as a function of the natural period script is:

clear all
clc
dam = 0.02; % damping; can changed to generate different spectra
load elcentro.mat; % load El Centro data
N = length(elcentro); % number of points in the ground acceleration file
acc = elcentro; % ground acceleration data
DT = 0.02; % sampling rate
for i=N:-1:1
t(i)=i*DT;
end
for j = 500:-1:1
period(j)=j*DT/2; % natural period array
wn(j)=2*pi/period(j); % natural frequency array
p=acc*DT;
h=exp(-dam*wn(j)*t).*sin(wn(j)*t);
u=conv(p,h);
uustr(j)=max(abs(u))/wn(j); % maximum displacement
114 Introduction to Earthquake Engineering

end
%create plot
plot (period, uustr*386.4, 'LineWidth',2, 'Color',[0 0 0]);
xlabel ('T_n(sec)','FontName','Times New Roman','FontAngle','italic');
ylabel ('u(in)','FontName','Times New Roman','FontAngle','italic');

As mentioned previously, seismic ground accelerations cannot be described analytically and

the evaluation of the integral in Equation 4.23 typically requires the use of numerical methods
in order to characterize the response of the SDOF system subjected to ground accelerations.
Numerical integration using the conv function in MATLAB, or other numerical integration
methods, can be difficult to implement for complex cases, and are not applicable in the case of
nonlinear behavior. Therefore, seismic ground excitation loadings are typically treated using the
direct integration of the equation of motion. Following is a discussion of three direct integration
methods.

4.5  DIRECT INTEGRATION METHODS

Since the numerical integration of the convolution (Duhamel’s) integral is difficult to implement
for complex loading functions, and not applicable for nonlinear response, time-stepping numerical
methods for the direct integration of the differential equation of motion were developed. First, to
derive any direct integration numerical algorithm, we assume an SDOF system characterized by
the natural frequency ωn and the damping ratio ζ. Also, we divide the acceleration excitation a(t)
­function into N equal intervals of Δt.
Time-stepping methods are divided into two categories: explicit and implicit. Explicit meth-
ods use the state of the system at the current time to calculate the state of the system at the next
time interval, while implicit methods use relationships involving both current and later states of
the ­system to find a solution. Consequently, implicit methods require iteration, making them more
difficult to implement. However, implicit methods are more practical in cases where the use of an
explicit method requires impractically small time steps, Δt, to keep the error in the result bounded.
Thus, for a desired level of accuracy, implicit methods can take much less computational time
because they can use longer time steps. Summaries of three of the most commonly applied methods
follow.

4.5.1 Nigam−Jennings Algorithm (Explicit)

For each interval Δt, the response is calculated using the conditions at the beginning of the time
interval—initial conditions. These initial conditions are given by the displacement and velocity
at the end of the preceding time interval. We also assume the forcing function to be linear and
­piecewise continuous (see Figure 4.3),

 ∆a 
a(τ ) = ai +  i  τ , ti ≤ τ ≤ ti+1 (4.27)
 ∆t 

a(t) ai+1
ai
a(τ )

∆ ti

ti ti+1 t
τ

FIGURE 4.3  Discretization of acceleration time history function.

Response to General Loading 115

where:
∆ai = ai +1 − ai
ti = i ⋅ ∆t
and
∆t = ti +1 − ti
for i = 1, 2, 3, … ,N

The equation of motion for the time interval can be written as

 ∆a 
u + 2ωnζυ + ωn2u = −ai −  i  τ , ti ≤ τ ≤ ti+1 (4.28)
 ∆t 

The solution to this linear differential equation with constant coefficients is given by the
­complementary and particular components,

u(t ) = uc (t ) + u p (t ) (4.29)

where, over the interval Δt, the complementary solution is

uc (τ ) = e−ζ ωn (τ −ti ) (Ci cos ωD (τ − ti ) + Di sin ω D (τ − ti ))

and over the same interval, the particular solution is

u p (τ ) = Bi + Ai (τ − ti )

Here, Ai and Bi are the constants of integration, determined by substituting up back into the equa-
tion of motion. After applying boundary conditions, the formulas to calculate the displacement,
velocity, and acceleration at time step ti+1 = ti + Δt are

ui +1 = a11ui + a12ui + b11ai + b12 ai +1 (4.30)

ui +1 = a21ui + a22ui + b21ai + b22 ai +1 (4.31)

ui+1 = −ωn2ui+1 − 2ζ ωnui+1 (4.32)

respectively. Where the coefficients of the displacements and velocities in the first two equations
only need to be computed once and are given as

 
 ζ
a11 = e−ζωn∆t  sin ωD∆t + cos ωD∆t 
 1 − ζ 2 


sin ωD∆t
a12 = e−ζ ωn∆t
ωD
116 Introduction to Earthquake Engineering

ωn
a21 = − e−ζωn∆t sin ωD∆t
2
1− ζ

 
 ζ
a22 = e−ζωn∆t cos ωD∆t − sin ωD∆t 
 1− ζ 2 


 2ζ 2 −1 ζ  sin ω ∆t  2ζ 1  2ζ
b11 = e−ζ ωn∆t  2 +  D
−  3 + 2  cosωD∆t  − 3
  
 ωn ∆t ωn  ω D  ωn∆t ωn   ωn∆t

 2ζ 2 −1 sin ω ∆t 2ζ  1 2ζ
b12 = −e−ζωn∆t  2  D
+ 3 cos ωD∆t  − 2 + 3
 

 ωn ∆t  ωD ωn∆t  ωn ωn∆t

 2  
 2ζ −1 ζ  ζ
b21 = e−ζωn ∆t  2 + cos ωD∆t − sin ωD∆t 
 ωn ∆t ω n  1− ζ 2 


 2ζ 1  1
− 3 + 2 (ωD sin ωD∆t + ζ ωn cos ωD∆t ) + 2
 ωn∆t ωn   ωn ∆t


 
e−ζωn∆t  ζ  − 1
b22 =  sin ω D ∆t + cos ω D ∆t  ω 2∆t
2 
ωn ∆t  1 − ζ 2
 n

EXAMPLE 4.13
Write a MATLAB script for the Nigam−Jennings algorithm; then use the script to draw displace-
ment, velocity, and acceleration response spectra for the response of the SDOF system subjected
to the El Centro earthquake ground acceleration presented in Example 4.11.

Solution
Again, the input ground acceleration can be obtained from http://www.vibrationdata.com/
elcentro.dat as a fraction of the acceleration due to gravity g. In this example, we write a MATLAB
script based on Equations 4.30 and 4.31. The maximum displacement, velocity, and acceleration
response as a function of the natural period script is:

clear all
clc
d(1) = 0;% initial displacement
v(1) = 0;% initial velocity
dam = 0.02; % damping; can be changed to generate different response spectra
DT = 0.02; % sampling rate
load elcentro.mat; % load El Centro data
N = length(elcentro); % number of points in the ground acceleration file
acc = elcentro; % ground acceleration data
% generate the data to graph a response spectrum by changing Omega
for j = 250:-1:1
period(j) = j*DT; % natural period
wd = 2*pi/period(j)*sqrt(1-dam^2); % damped frequency
wn = 2*pi/period(j); % natural frequency
Response to General Loading 117

%%% compute values of the response and choose the maximum %%%
% first calculate the elements of the matrices A and B
a11=exp(−dam*wn*DT)*(dam/sqrt(1-dam^2)*sin(wd*DT)+cos(wd*DT));
a12=exp(−dam*wn*DT)/wd*sin(wd*DT);
a21=-wn/sqrt(1-dam^2)*exp(-dam*wn*DT)*sin(wd*DT);
a22=exp(−dam*wn*DT)*(cos(wd*DT)−dam/sqrt(1−dam^2)*sin(wd*DT));
b11=exp(−dam*wn*DT)*(((2*dam^2−1)/(wn^2*DT)+dam/wn)*sin(wd*DT)/wd+...
(2*dam/(wn^3*DT)+1/wn^2)*cos(wd*DT))−2*dam/(wn^3*DT);
b12=-exp(−dam*wn*DT)*((2*dam^2−1)/(wn^2*DT)*sin(wd*DT)/wd+...
2*dam/(wn^3*DT)*cos(wd*DT))−1/wn^2+2*dam/(wn^3*DT);
b21=exp(−dam*wn*DT)*(((2*dam^2−1)/(wn^2*DT)+dam/wn)*(cos(wd*DT)-...
dam/sqrt(1−dam^2)*sin(wd*DT))−(2*dam/(wn^3*DT)+1/wn^2)*(wd*sin(wd*DT)+...
dam*wn*cos(wd*DT)))+1/(wn^2*DT);
b22=(−1+exp(-dam*wn*DT)*(dam/sqrt(1-dam^2)*sin(wd*DT)+cos(wd*DT)))/(wn^2*DT);
%loop to find response
for i = 1:N-1
d(i+1) = a11*d(i)+a12*v(i)+b11*acc(i)+b12*acc(i+1);
v(i+1) = a21*d(i)+a22*v(i)+b21*acc(i)+b22*acc(i+1);
Responseaccelation(i) = -2*wn*dam*v(i+1)-(wn^2)*d(i+1);
end
% compute the value of the largest response
max_acc(j) = max(abs(Responseaccelation));
max_vel(j) = max(abs(v));
max_dis(j) = max(abs(d));
end
subplot(3,1,1), plot(period,max_dis*386.4)
ylabel('Displacement(in)','FontName','Times New Roman','FontAngle','italic')
subplot(3,1,2), plot(period,max_vel*386.4)
ylabel('Velocity(in/s)','FontName','Times New Roman','FontAngle','italic')
subplot(3,1,3), plot(period,max_acc)
xlabel('T_n(sec)','FontName','Times New Roman','FontAngle', 'italic')
ylabel('Acceleration(g)','FontName','Times New Roman','FontAngle','italic')

The results for the displacement response spectrum obtained using this method and the conv
f­ unction in MATLAB, shown in Figure E4.18, are virtually identical. The other spectrum are shown
in Figure E4.19.

4.5.2 Central Difference Method (Explicit)

With this method, the response is calculated using the constant time interval Δt, based on a finite
difference approximation of the time derivatives of the displacement in the equation of motion.
We also assume the excitation forcing function to be linear and piecewise continuous, with approxi-
mate time derivatives for time step i of duration Δt shown in Figure 4.4. Substituting the velocity

20

15
u(in)

10

0
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
Tn(s)

FIGURE E4.18  Displacement response spectrum for an SDOF system due to El Centro earthquake.
118 Introduction to Earthquake Engineering

20

Displacement (in)
15

10

0
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5

50

40
Velocity (in/s)

30

20

10

0
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5

1.5
Acceleration (g)

0.5

0
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
Tn(s)

FIGURE E4.19  Displacement, velocity, and acceleration response spectra for the SDOF system
­subjected to El Centro earthquake.

and acceleration as shown in this figure into the equation of motion, Equation 3.49, we obtain the
­equation of motion for the time interval 2Δt,

ui+1 − 2ui + ui−1 u −u

2
+ 2ωnζ i+1 i−1 + ωn2ui = −ugi (4.33)
(Δt ) 2Δt

where ugi is the ground acceleration at time step i. Also, ui and ui –1 are known from the preceding
step. Rearranging Equation 4.33 to have these known quantities on the right,
 1 ω ζ  1 ω ζ  2 
 + n u = −ugi −  − n u − ω 2 − u (4.34)
 (Δt )2 Δt  i+1  (Δt )2 Δt  i−1  n (Δt )2  i

u(t) ui+1
ui+1 – ui–1
u⋅ i =
2∆t
ui–1

u – 2ui + ui–1 ∆t ∆t
u⋅⋅i = i+1
(∆t)2
t
ti – ∆t ti ti + ∆t

FIGURE 4.4  Finite difference approximation of velocity and acceleration.

Response to General Loading 119

or
ˆ
kui +1 = a
ˆi (4.35)

where:

1 ωζ
kˆ = 2
+ n
(Δt ) Δt

 1 ω ζ  2 
aˆi = −ugi −  − n  ui−1 − ωn2 −  ui = −ugi − aui−1 − bui
 (Δt ) 2
Δt   (Δt )2 

Determining the initial velocity uo and displacement uo from the specified initial conditions,
we can use the following algorithm for the central difference method:

1. Initial calculations based on the known quantities:

uo = −ugo − 2ωnζ uo − ω 2nuo

(∆t )2
u−1 = uo − ∆t ⋅ uo + ⋅ uo
2

1 2ω ζ
kˆ = 2
+ n
(Δt ) 2Δt

1 ωζ
a= 2
− n
(Δt ) Δt

2
b = ω 2n−
(Δt )2

2. Calculations for step i, where i = 0, 1, 2, 3, …

âi = −ugi − aui −1 − bui

aˆi
ui +1 =
kˆ

ui +1 − ui −1
ui =
2∆t
ui +1 − 2ui + ui −1
ui =
(∆t )2

This method is conditionally stable. It requires the following time step length for stability.

Tn (4.36)
Δt <
π
120 Introduction to Earthquake Engineering

This is rarely a limitation in seismic response analyses since Δt is typically specified as 0.02 s or
smaller to accurately define the ground acceleration ug .

4.5.3 Newmark’s Beta Method for Linear Systems (Implicit)

Professor Nathan Newmark developed a general method that approximates the variation of the
acceleration response within each time interval. In this method, the response can be calculated
using nonconstant time intervals Δti following approximations of the time derivatives of the
­displacement in the equation of motion as

ui+1 = ui + [(1 − γ )Δt ]ui + [γΔt ]ui+1 (4.37)

ui+1 = ui + Δt ⋅ ui + [(0.5 − β )(Δt )2 ]ui + [β (Δt )2 ]ui+1 (4.38)

Where parameters β and γ  define the variation of the acceleration over nonconstant time inter-
vals Δti. The values of these parameters must be chosen carefully in order to obtain stable results.
Two sets of values that produce stable results include β = 1/2 and γ  = 1/4 (constant average accelera-
tion) and β = ½ and γ  = 1/6 (linear acceleration).
After obtaining the initial velocity uo and displacement uo from the specified initial conditions,
we can use the following algorithm for the Newmark Beta method:

1. Initial calculations based on the known quantities and time intervals Δti:

uo = −ugo − 2ωnζ uo − ω 2nuo

1 2ω ζγ
a1 = + n
β (Δt )2 β ⋅ Δt

1 γ 
a2 = + 2ωnζ  −1
β ⋅ Δt 
 β 

1 γ 
a3 = −1 + 2ωnζ  −1Δt
2β 
 2β 

k̂ = ω 2n+ a1

2. Calculations for step i, where i = 0, 1, 2, 3, …,

âi +1 = −ugi +1 + a1ui + a2ui + a3ui

aˆi +1
ui +1 =
kˆ

γ  γ  γ 
ui +1 = (ui +1 − ui ) + 1 −  ui + Δt 1 −  ui
β ⋅ Δt 
 β  
 2β 

Response to General Loading 121

(ui +1 − ui ) u  1 
ui +1 = − i −  −1 ui
β (Δt ) 2 
β ⋅ Δt  2β 

This method is also conditionally stable. It requires the following time step for stability:

Δt ≤ Tn π 2(γ − 2β ) (4.39)

which for β = 1/2 and γ  = 1/4 is unconditionally stable.

This particular method is useful in the analysis of nonlinear systems; however, details of the
algorithm are slightly different than those presented here. For complete details of the analysis of
nonlinear systems, see the textbook by Chopra (2012).

PROBLEMS
4.1 Use Duhamel’s integral to determine the response of an undamped SDOF system
­subjected to a half-sine impulse that is applied from 0 < t < t d. Assume that the SDOF
­system is initially at rest and t d /T ≠ 0.5. Hint, this can be solved by using superposition
of two sinusoidal excitations.
4.2 Use Duhamel’s integral to determine the dynamic response of a tower subjected to the
load shown.

p(t) p(t)
W = 38.6 kips
120

k = 100 kips/in

0.04 t(s)
0.02

4.3 Draw the shock spectrum for an SDOF system for the triangular impulse loading shown.

p(t)

po

td t

4.4 Assume that Burn’s tower (a water tank weighing 4200 kips) on the campus of University
of the Pacific is supported on a 120 ft tall cantilever tower with stiffness of 2000 kips/in.,
determine the design values for lateral deformation and base shear for a symmetric tri-
angular impulse of amplitude 100 kips and duration t d = 0.1 s.
4.5 The bridge beam depicted below supports a rigid deck that weighs 200 kips. The beam
is subjected to the blast load described by the triangular impulse loading shown.
Determine the maximum stress in the beam. (E = 29,000 kip/in2, Ix = 13,000 in4, and
Sx = 830 in3).
122 Introduction to Earthquake Engineering

p(t)

50 ft
p(t)

100 kips

t (s)
0.5

4.6 Consider the following building frame with a rigid beam that is pinned con-

nected to one column and rigidly connected to the other as shown; each column has
EI = 40,000,000 kip in2. The frame is subjected to the blast load shown at the beam
level. Determine the maximum stress in each column given Sx = 2000 in3.

p(t) 200 kips

p(t)

10 ft Pin
1000 kips
15 ft
t(s)
0.13

4.7 Write a MATLAB script for the central difference algorithm; then use the script to
draw displacement, velocity, and acceleration response spectra for the response of the
SDOF system subjected to the El Centro earthquake ground acceleration presented in
Example 4.11.
4.8 Write a MATLAB script for the Newmark Beta algorithm; then use the script to
draw displacement, velocity, and acceleration response spectra for the response of the
SDOF system subjected to the El Centro earthquake ground acceleration presented in
Example 4.11.
4.9 Write a MATLAB script to solve Problem 4.8 using the lsim function in MATLAB,
which simulates time response of dynamic systems to arbitrary inputs.

REFERENCE
Chopra, A. K., Dynamics of Structures: Theory and Applications to Earthquake Engineering, 4th edition,
Prentice-Hall, Upper Saddle River, NJ, 2012.
 5 Response Spectrum
Analysis of SDOF System
After reading this chapter, you will be able to:

1. Use response spectra to calculate the maximum response of a structure, specifically dis-
placements and base shear
2. Obtain a design elastic response spectrum
3. Use design response spectra to calculate maximum response of a structure, specifically
displacements and base shear
4. Develop an inelastic response spectrum
5. Use an inelastic response spectrum to calculate the maximum response of a structure,
specifically displacements and base shear

Given the complexity of seismic ground acceleration records, numerical methods such as those
introduced in Chapter 4 must be applied to obtain the response of a structure to an input ground
excitation (see Figure 5.1). The response depicted in Figure 5.1 is analogous to the analysis per-
formed in Example 4.11 of Chapter 4. For seismic-resistant design, the entire response history of
the system is not necessary, only the absolute maximum value of the response is required. However,
it is necessary to process a large portion of the response history to identify this maximum response
value. Without the aid of computer software, this process can be time consuming, even for the
simple impulsive loadings presented in Chapter 4. Also, the response (displacement, velocity, or
acceleration) of a structure to a seismic excitation depends on system dynamic properties (mass,
stiffness, and damping) and ground excitation characteristics. As discussed in Chapter 4, it is more
convenient to use a plot of the maximum response to a specified loading for all SDOF systems
based on their inherent natural period (or frequency) and damping; the so-called shock spectrum
for impulse loading and response spectrum for seismic loading. Here, we review the theoretical con-
cepts as well as a practical approach for developing a response spectrum that can be used for design;
this is the precursor to the contemporary design response spectrum used in current design codes
(see Chapter 8). Since most systems are allowed to deform inelastically during a seismic event, we
also present details of the design response spectrum for systems in the inelastic range.

5.1  ELASTIC RESPONSE SPECTRUM

Response spectrum analysis is an effective approach to assess the behavior of an SDOF ­system when
it is subjected to a particular peak ground acceleration (PGA) input. And as shown in Example 4.12
of Chapter 4, developing an elastic response spectrum is a straightforward process, provided we can
use a computer program to perform the tedious calculations. The elastic response ­spectrum only
depends on the system natural period and damping because these are the only dynamic properties
that influence the maximum values of the response. The response may be expressed as displacement
(relative or total), velocity (relative or total), or acceleration (­relative or total), as shown by Example
4.13 in Chapter 4; however, the convention is to show relative d­ isplacement, relative velocity, and
total acceleration.
Figures 5.2 and 5.3 illustrate the general process for developing an elastic response spectrum.
First, we select an earthquake accelerogram (ground acceleration vs. time); also known as a ­seismic
time-history. The time-history for various earthquakes (at different sites) can be obtained from a

123
124 Introduction to Earthquake Engineering

ut
Response action
Tn = 1.0 s and ζ = 0.05 u u (in)
5 Max value
m

–5

k üg(g) Ground excitation

0.5

ug(t)
–0.5

FIGURE 5.1  Response of a building frame to a seismic ground excitation.

Vary period Response

u
A
Tn= 0.2 s t
umaxA

Ground excitation input u

üg Tn = 0.5 s B
t
umaxB
t
u umaxC
Tn = 1 s C
t
....

....

u umaxD

Tn = 3 s t D

ζ is constant Max displacement

FIGURE 5.2  Displacement response to a seismic ground excitation.

umax = Sd = D
D

A
Tn (s)
0.2 0.5 1 3

FIGURE 5.3  Displacement response spectrum from Figure 5.2 results.

Response Spectrum Analysis of SDOF System 125

Fs (t) = ku(t) = müt(t) = mω n2u(t)

Vb(t) Mb(t)

FIGURE 5.4  Equivalent static force, base shear, and overturning moment.

number of websites, particularly http://ngawest2.berkeley.edu/. For a specific seismic time-­history, nat-

ural period, Tn, and damping ratio, ζ, we generate a complete response history (relative d­ isplacement,
relative velocity, or total acceleration) using one of the numerical methods discussed in Chapter 4.
Figure 5.2 illustrates the process for generating relative displacement responses. We then scan each
response history to identify the maximum value, which corresponds to a s­ ingle point on the response
spectrum plot, as shown in Figure 5.3. For instance, the maximum displacement response for case A in
Figure 5.2 corresponds to point A on the response spectrum in Figure 5.3. We then repeat the process
for different natural periods (with constant damping) until the response spectrum plot is completed.
Finally, we repeat the entire process for different values of damping to generate other response spectra.
The response quantities graphed on these spectra are known as spectral quantities; that is,
Sd = D = spectral displacement = maximum relative displacement of an SDOF system.
Sv = spectral velocity = maximum relative velocity of an SDOF system.
Sa = spectral acceleration = maximum total acceleration of an SDOF system.
As discussed in Section 3.3, we only need the relative displacement to conduct a structural ­analysis
(see Figure 5.4). That is, the base shear, Vb, and the overturning moment, Mb, can be ­determined
using the static equilibrium analysis of the structure when subjected to an equivalent static force,
Fs = ku(t); Vb = mωn2u(t ); and Mb = hVb. For design purposes, we are primarily interested in the
maximum response so we apply the peak values of the relative displacement or total acceleration.
But the total acceleration is related to the displacement as shown in Figure 5.4. So it would appear
that only the spectral displacement would be necessary to compute the peak values of ­deformations
and internal forces needed for seismic design. However, the velocity and the acceleration response
spectra are needed in order to construct design spectra and to demonstrate how these design spectra
relate to modern building code specifications.

EXAMPLE 5.1
Use the MATLAB® script provided in Chapter 4, Example 4.12 to draw response spectrum of the
displacement, velocity, and acceleration for an SDOF system subjected to the El Centro earth-
quake ground acceleration for 2%, 5%, and 10% damping.

Solution
Input ground acceleration is given in Figure E5.1.

200
üg (in/s/s)

–200
0 5 10 15 20 25 30
t (s)

FIGURE E5.1  Seismic (acceleration) time-history for El Centro earthquake.

126 Introduction to Earthquake Engineering

20
ζ = 0.02
15
ζ = 0.05
Sd (in)

10 ζ = 0.1

0
0 0.5 1 1.5 2 2.5 3
60

40
Sv (in/s)

20

0
0 0.5 1 1.5 2 2.5 3
1.5

1
Sa (g)

0.5

0
0 0.5 1 1.5 2 2.5 3
Tn (s)

FIGURE E5.2  Displacement, velocity, and acceleration response spectra for El Centro earthquake.

The response spectra results are shown in Figure E5.2; for damping factors of 0.02, 0.05, and
0.1, top to bottom.

EXAMPLE 5.2
Use the MATLAB script provided in Chapter 4, Example 4.13 to draw the response spectrum of the
displacement, velocity, and acceleration for an SDOF system subjected to the 1985 Mexico City
earthquake ground acceleration for 2%, 5%, and 10% damping.

Solution
Input ground acceleration is given in Figure E5.3.

0.2
Acceleration (g)

0.1

–0.1

–0.2
0 10 20 30 40 50 60 70
t (s)

FIGURE E5.3  Seismic (acceleration) time-history for 1985 Mexico City earthquake, SCT1 (EW).
Response Spectrum Analysis of SDOF System 127

Sa (g)
1

0
0 0.5 1 1.5 2 2.5 3

300

200 ζ = 0.02
Sv (in/s)

ζ = 0.05
ζ = 0.1
100

0
0 0.5 1 1.5 2 2.5 3

100
Sd (in)

50

0
0 0.5 1 1.5 2 2.5 3
Tn (s)

FIGURE E5.4  Seismic ground motion and response spectra for Mexico City earthquake.

The response spectra results are shown in Figure E5.4; for damping factors of 0.02, 0.05, and
0.1, top to bottom.
As can be seen from these spectra in Figure E5.4, the response is largest for a natural period of
∼2 s, which is the dominant period in Mexico City as discussed in Chapter 2.

Note that the relationship between the relative displacement and total acceleration mut (t ) = mωn2u(t )
is only valid for systems without damping. For this reason, the acceleration determined from this
relationship is known as the pseudo-acceleration, A, in order to distinguish A from the actual accel-
eration. That is, consider the equation of motion in terms of the relative quantities,

u + 2ζωnu + ωn2u = −ug (5.1)

Since most structures have small damping (see Table 3.3), we can assume that the velocity term
is negligible (2ζωnu  ωn2u ); thus, Equation 5.1 can be rewritten as

u + ug = −ωn2u (5.2)

Substituting the value of total acceleration, u + ug = ut , we obtain

ut = −ωn2u (5.3)

This results in the definition of the spectral pseudo-acceleration, A, in terms of the spectral
displacement, D, as

A = ωn2 D (5.4)

128 Introduction to Earthquake Engineering

0.8 ζ = 0.05

Pseudo-acceleration
0.6
Actual acceleration
Sa (g)

0.4

0.2

0
0 0.5 1 1.5 2 2.5 3
Tn (s)

FIGURE 5.5  Comparison of actual and pseudo-acceleration response spectra for El Centro earthquake.

Note that in general A ≠ umax

t
, but it is adequate for most practical purposes, as shown in Figure
5.5, for 5% damping where pseudo-acceleration and actual acceleration response spectra are plotted
for the El Centro earthquake. A 5% damping ratio was selected in order to match the damping level
assumed in building code design spectra.
Similarly, we can derive a quantity for the relative velocity in terms of the displacement, the
pseudo-velocity, V. For this case, we use an energy–balance relationship. Recall that the strain
energy stored in the system during an earthquake is approximately given by

1 2 1
kumax = kD 2 (5.5)
2 2
And if there are no other processes dissipating energy, such as damping, and so on (although not
entirely correct, we assume damping is negligible because it is small), then all the strain energy is
transferred into kinetic energy:

1 2 1
mumax = mSv2 (5.6)
2 2

From the principle of conservation of energy, these energy quantities must be equal; that is,

1 2 1
kD = mSv2 (5.7)
2 2

This results in the definition of the spectral pseudo-velocity, V, in terms of the spectral displace-
ment, D, as

k
D = V ⇒ ωn D = V (5.8)
m

Note that in general V ≠ umax, but it is sufficiently close for most practical purposes. Figure 5.6
shows the pseudo-velocity and actual velocity response spectra due to El Centro with 5% damping,
though the agreement between the response spectra is not as accurate as the comparison between
­acceleration response spectra.
From Equations 5.4 and 5.8 it can be seen that

A
ωn D = V = (5.9)
ωn
Response Spectrum Analysis of SDOF System 129

40

30
Sv (in/s)

20
ζ = 0.05
10 Pseudo-velocity
Actual velocity
0
0 0.5 1 1.5 2 2.5 3
Tn (s)

FIGURE 5.6  Comparison of actual and pseudo-velocity for El Centro quake.

or,

2π T
D =V = n A (5.10)
Tn 2π

which means that D, V, and A contain the same information; thus, we can combine the three
­quantities into a single three-plot (tripartite) graph as a function of natural period, Tn, or natural
frequency, ωn.
Consider separately the relationships between D and V and that between V and A in Equation 5.9,
and now take the logarithms of both sides of each equation,

logV = logωn + logD (5.11)

log V = − log ωn + log A (5.12)

These represent equations of straight lines: the displacement with a slope of +1 and the accel-
eration with a slope of –1. That is, lines of constant D and A are inclined at +45° and –45°, respec-
tively. Thus, to construct a tripartite graph, we need to plot V versus ωn on vertical and horizontal
­logarithmic scales and add logarithmic scales inclined at +45° and –45° with respect to the vertical
axis. So, V is read on the vertical scale, D is read on the scale at –45°, and A is read on the scale
at +45°. When the argument is Tn, D is read from the scale at +45° and A is read from the scale at
–45° (see Figure 5.7).

EXAMPLE 5.3
Given the structure shown in Figure E5.5, determine the response (D, V, and A) for the El Centro
ground motion using the tripartite spectra in Figure 5.7.

Tn = 0.5 s
ζ = 2%
h

FIGURE E5.5  Schematic of frame structure in Example 5.3.

130 Introduction to Earthquake Engineering

100

50 V = 35 in/s

10
10

0
50
20

5
Pseudo-velocity V, in/s

10
g
1.2

10
1

5
=
5 A
Ps

n.
eud

,i
D

D
1
o- =

0.

n
ac 3.

io
ce 0

at
in

0.
2 le

m
5
ra

r
fo
tio

01 D e
n
A,
1 g
0.

0.
1

0.5

1
0.

00
01

0.
0.2
0.02 0.05 0.1 0.2 0.5 1 2 5 10 20 50
Natural vibration period Tn (s)

FIGURE 5.7  Response spectra for El Centro quake for ζ = 0%, 2%, 5%, 10%, and 20%, top to bot-
tom. (Reproduced from Chopra, A. K., Dynamics of Structures: Theory and Applications to Earthquake
Engineering, 4th edition, Prentice-Hall, Upper Saddle River, NJ, 2012. By permission of Prentice-Hall,
Upper Saddle River, NJ.)

Solution
Enter the 2%-damping response spectrum in Figure 5.7 with natural period of 0.5 s, and read
V = 35 in/s from the vertical scale, D = 7.5 in from the scale at +45°, and A = 1.2g from the scale
at –45°. The lines showing these values are depicted in Figure 5.7.

EXAMPLE 5.4
The reinforced concrete bridge structure shown in Figure E5.6 is subjected to the El Centro earth-
quake; determine the total stiffness, structural period, deflection of the deck, and base shear. The
deck weighs 750 kips and the substructure has 2% damping, modulus of elasticity, E = 3000 ksi,
and the other geometric properties shown in Figure E5.6.

Fixed–pinned Fixed–pinned 15 ft
20 ft
I = 25,000 in4 I = 20,000 in4
Fixed–fixed
30 ft
I = 60,000 in4

FIGURE E5.6  Schematic of reinforced concrete bridge structure.

Response Spectrum Analysis of SDOF System 131

Solution
1. Mass, stiffness, and natural period of the system: The bridge can be modeled as an SDOF
system assuming a rigid deck and only lateral deformations of the columns occur. The
stiffness of the system is then the sum of column lateral stiffnesses. The mass and stiffness
of the SDOF system are calculated as follows:
Mass,

w 750 kips
m= = = 1.94 kips s2 /in
g 386.4 in /s2

Stiffness (see Table 3.1), the 30-foot column is fixed–fixed, while the other two are
fixed–pinned,

3EI1 12EI2 3EI3

k= + + 3
h 13 h 32 h3
 3( 25, 000 in4 ) 12(60, 000 in4 ) 3( 20, 000 in4 ) 
= 3000 ksi  + + = 93.4 kips /in
 ( 20 ft ×12 in /ft )
3
(30 ft ×12 in /ft )3
(15 ft ×12 in /ft )3 

The natural period of the system is

m 1.94 kips s2
Tn = 2π = 2π = 0.906 s
k 93.4 kips /in

2. Determine the maximum displacement: The maximum displacement can be deter-

mined using the response spectra depicted in Figure 5.7. Entering the 2%-damping
response spectrum with Tn = 0.91 s, we can read the deck displacement from the scale
at +45°,

D ≅ 6 in

3. Determine the maximum base shear: Calculate the maximum base shear at the support
using the force–displacement relationship or equilibrium as described in Figure 3.24:

Vbmax = kD = 93.4 kips /in (6 in) = 565 kips

A careful examination of the response spectra shown in Figure 5.7 shows that there are regions
where each of the response parameters (D, V, and A) are nearly constant over a wide range of values
for Tn. For example, for very flexible systems (those with large periods), the displacement is equal
to the peak ground displacement for all damping ratios and experience negligible accelerations.
However, stiff systems (those with small periods) accelerate at the PGA for all damping ratios and
experience negligible relative displacements (only ground displacement). Also, damping has the
greatest effect over the center region. This analysis of Figure 5.7 is specific to a set of response spec-
tra for a ­specific earthquake. In order to compare the results of different earthquakes, we ­normalize
each response parameter (spectral displacement, velocity, and acceleration) with respect to its cor-
responding peak ground value (peak ground spectral d­ isplacement, peak ground velocity, and PGA,
respectively) as shown in Figure 5.8.
Figure 5.8 also depicts an idealized response spectrum, shown by a dashed line. We can observe
that for the first portion (Tn < 0.5 s) of the response spectra, the ­acceleration is largest (acceleration
sensitive), before decreasing precipitously; for the middle portion (0.5 s < Tn < 3 s), the velocity is
largest (velocity sensitive); and for the last portion (Tn > 3 s), the displacement is largest (displace-
ment sensitive). This ­phenomenon is one of the reasons for determining V and A. Furthermore, as
132 Introduction to Earthquake Engineering

10
Acceleration Velocity Displacement
0 sensitive sensitive sensitive
5 10

10
2 c d
10
1

Td = 3.0 s
Tc = 0.5 s
e

1
b
0.5
V/ugo

1
.

Ta = 0.035 s

0.
A/

1
..
0.2

ug

go
1

/u
Tb = 0.125 s
0.

Te = 10 s
0.1
0.

Tf = 15 s
a 01

01
0.05

0.
0.
00
1

0.02
0.02 0.05 0.1 0.2 0.5 1 2 5 10 20 50
Tn (s)

FIGURE 5.8  Spectral regions for El Centro quake; ζ = 5%. (Reproduced from Chopra, A. K., Dynamics of
Structures: Theory and Applications to Earthquake Engineering, 4th edition, Prentice-Hall, Upper Saddle
River, NJ, 2012. By permission of Prentice-Hall, Upper Saddle River, NJ.)

mentioned ­previously, delineating regions along the spectra where the different response quantities
are dominant (or maximum) is useful in establishing design spectra, as discussed in Section 5.1.1.

5.1.1 Elastic Design Response Spectrum

As discussed in Chapter 2, seismic ground accelerations vary greatly in magnitude of their peak
amplitude and frequency content, both in space and time. For instance, two earthquake events
of the same magnitude that strike a site at two different times will result in two different ground
acceleration time histories even if the sites are at equal distances from the epicenter. For sites at
different distances, seismic wave attenuation typically causes the seismic energy to diminish (in
some cases, however, local soil conditions can amplify the effects for certain frequencies as was
the case for the 1985 Mexico City earthquake discussed in Chapter 2). This makes it challenging to
design for future buildings relying on past earthquake records to estimate the probable earthquake
response. Therefore, we must assess the response of future structures to seismic loadings based
on the p­ robable PGAs (PPGA), with only secondary considerations to the frequency content. The
PPGA can be determined from statistical analyses of several similar seismic events at sites with
similar geological conditions.
An effective strategy to estimating accurate PPGA is to take averages of past earthquakes with
­similar characteristics, such as soil condition, epicentral distance, magnitude, and source mecha-
nism. The statistical analysis uses normalized response spectra, assuming a Gaussian distribution
for each point along the period axis as shown in Figure 5.9. The general procedure entails the
­following steps:

1. Select N earthquakes (10 for Figure 5.9) with similar characteristics.

2. Obtain D, V, and A for different Tn’s for each earthquake; and normalize with respect to
­average peak ground displacement, ugo, peak ground velocity, u go , and PGA, ügo, respectively.
Response Spectrum Analysis of SDOF System 133

10

5 0
10
Mean + 1σ

10
2

go
/u
c d

D
10
A/
..
1 ug o Mean

Tf = 33 s
Tc = 0.349 s

1
b
0.5
V/ugo

1
.

e
Ta = 1/33 s

Td = 3.135 s
0.
0.2

1
1
0.
Tb = 1/8 s

Te = 10 s
0.1 f
0.
01

a
0.05

01
0.
0.
00

0.02
1

0.02 0.05 0.1 0.2 0.5 1 2 5 10 20 50

Tn (s)

FIGURE 5.9  Mean and mean + 1σ spectra for a group of 10 quakes; ζ = 5%. (Reproduced from Chopra,
A. K., Dynamics of Structures: Theory and Applications to Earthquake Engineering, 4th edition, Prentice-
Hall, Upper Saddle River, NJ, 2012. By permission of Prentice-Hall, Upper Saddle River, NJ.)

3. Determine the mean and standard deviation (σ) at each period. Figure 5.9 schematically
shows the probability distributions for V at Tn = 0.25, 1, and 4 s, which also show that the
coefficient of variation (COV = σ/mean) changes with Tn.

Figure 5.9 also clearly shows that the graphs of the mean and the mean + 1σ are smoother than
any individual parent spectrum, one of which is Figure 5.7.
The elastic design spectrum shown in Figure 5.10 is then obtained by smoothing out into a
series of straight lines the spectra graphs. The dashed lines in Figures 5.9 and 5.10 represent the
means of the response spectra. This simplified empirical approach to constructing elastic response
spectra was first developed by Professors Nathan Newmark and William Hall, and is known as
the Newmark–Hall method. The key parameters are the structural amplification factors, α, which
represent the spectral amplification above ground motion. A summary of structural amplification
factors is listed in Table 5.1. These empirical values were computed using records from firm ground,
such as rock, soft rock, and competent sediments. Though limited in its applicability because of the
small number of records and narrow site characteristics, many pieces of the general shape can be
found in current code spectra, as will be demonstrated in Chapter 8.
After selecting a damping ratio, ζ, and a mean or mean + 1σ plot, the general procedure for
drawing a design response spectrum entails the following steps:

1. Mark the key periods delineating the transition regions, Ta, Tb, Te, and Tf. The periods
­delineating the velocity, acceleration, and displacement-controlled regions are obtained
from the intersection of A and V and the intersection of V and D.
2. Plot lines corresponding to values of ugo, u go , and ügo for the specified design ground
motion. This peak ground motion graph represents the backbone of the response spectrum
and can be considered the lower bound of the response.
134 Introduction to Earthquake Engineering

Tc αV ugo Td
α
Du
go

Te

go
Au
ugo

α
Tb u
go
V (log scale)

go
u
Tf

Ta

1/33 s 1/8 s 10 s 33 s
Tn (log scale)

FIGURE 5.10  Newmark–Hall elastic design response spectrum.

TABLE 5.1
Structural Amplification Factors for Elastic Design Spectra
Mean (50th Percentile) σ (84.1th Percentile)

Damping (%) αA αV αD αA αV αD
2 2.74 2.03 1.63 3.66 2.92 2.42
5 2.12 1.65 1.39 2.71 2.30 2.01
10 1.64 1.37 1.20 1.99 1.84 1.69
ζ 3.21–0.68 lnζ 2.31–0.41 lnζ 1.82–0.27 lnζ 4.38–1.04 lnζ 3.38–0.67 lnζ 2.73–0.45 lnζ

Source: Newmark, N. M. and W.J. Hall, Earthquake Spectra and Design, Earthquake Engineering Research Institute,
Oakland, CA, 1982.

3. Obtain values of α A, αV, and αD for a specified damping ratio, ζ, from Table 5.1; where the
mean or 50th percentile represents a 50% chance of being exceeded (50% exceedance).
4. Amplify the ground peak values (backbone) using these parameters to get: α Aügo, αV  u go ,
and αDugo.
5. Finally, connect the transition lines from a to b and from e to f.

The axes of the Newmark–Hall elastic design spectrum were normalized with respect to ügo = 1g,
u go  = 48 in/s, and ugo = 36 in. Values of ügo can be obtained from deterministic or probabilistic site
hazard analysis, whereas ugo and u go can be determined from site hazard analyses or as empirical
functions of ügo when only ügo is available. The following empirical relationships are recommended
for estimating u go and ugo in this case:

ugo ugo ⋅ ugo

= 48g in /s and 2
=6 (5.13)
ugo ugo

Response Spectrum Analysis of SDOF System 135

A number of other researchers have conducted statistical analyses and determined values
that vary considerably because these values depend on magnitude, epicentral distance, and soil
­characteristics at the recording station site.

EXAMPLE 5.5
Draw the 84.1th percentile response spectrum for 5% damping using the tripartite graph shown
in Figure 5.11. The ground motion is anticipated to have a PGA of ügo = 0.5g.

Solution
1. Determine the peak ground velocity, u go , and peak ground displacement, ugo: Use the
PGA and Equation 5.13 to determine these quantities.
Peak ground velocity,

 in /s   in /s 
u go = 48 ugo = 48

 ⋅ 0.5 g = 24 in /s
 g    g 

Peak ground displacement,

ugo = 6u go
2
/ugo = 6( 24 in /s)2 / 0.5(386 in /s2 ) = 18 in

Plot the backbone quantities on the tripartite graph paper provided in Figure 5.11, as
shown in Figure E5.7.

2. Determine the structural amplification factors, αA, αV, and αD: For 84.1th percentile spec-
trum and 5% damping, these quantities are obtained from Table 5.1 as

αA = 2.71, αV = 2.30, and αD = 2.01

100

50
0
.0
10

20
10
0.
00

10
00
1.

10
V (in/s)

5
,i

.0
D

0
10
A,

0.
g

2
1.
00

1
01
0.
0.

0.5
10

00
0.

0.2
0.
01

0.02 0.05 0.1 0.2 0.5 1 2 5 10 20 50

Tn (s)

FIGURE 5.11  Tripartite graph paper.

136 Introduction to Earthquake Engineering

100
55.2 in/s
50

36
.2
0

in
24 in/s

.0
10

g
36
20

1.

10
18

0.
in

00
10

5g
00
0.
1.

10
V (in/s)

n
5

.0
,i

0
D
10
A,

0.
2 g

1.
00
1

01
0.
0.
10
0.5

00
0.
0.2
0.
01

0.02 0.05 0.1 0.2 0.5 1 2 5 10 20 50

Tn (s)

FIGURE E5.7  Elastic design response spectrum for peak ground acceleration of 0.5g and 5% damping.

3. Determine spectral quantities, A, V, and D: Amplify the backbone quantities to obtain A,

V, and D.

A = ügoαA = 0.5g (2.71) = 1.36g
V = u go αV = 24 in/s (2.3) = 55.2 in/s
D = ugoαD = 18 in (2.01) = 36.2 in

Plot these quantities in the same graph as the backbone quantities, Figure E5.7.
4. Complete the response spectrum by connecting the transition lines. Alternatively,
the spectral graph can be obtained using a conventional graphing computer program
such as MATLAB.

Note that the results of this example would be the same if we had derived the response spectrum
for the standard peak ground values of ügo = 1g, u go  = 48 in/s, and ugo = 36 in and scaled the result-
ing spectrum by a factor of 0.5. Therefore, constructing an elastic design response spectrum using
the Newmark–Hall method only requires the damping ratio and expected PGA, ügo = ηg; The scal-
ing factor, η, can easily be applied to the ordinates of the design spectra.
We have used the tripartite graph to delineate the three main regions of elastic design spectra
(acceleration, velocity, and displacement sensitive); however, for design purposes, it is more conve-
nient to return to the single acceleration spectra. This can be accomplished by plotting Equation 5.10
as shown in Figure 5.12 for the Newmark–Hall spectrum, ügo = 1g, u go  = 48 in/s, and ugo = 36 in
and ζ = 5%. The axes of this spectrum are also plotted using linear scales in Figure 5.13. This figure
is similar in shape to contemporary code-based design spectra.
In cases where only a single value of the pseudo-acceleration, A, is needed, it may not be ­necessary
to construct the entire elastic design response spectrum; particularly if the period is between 1/8 
and 10 s (constant acceleration, velocity, and displacement regions). For such a case, the governing
design acceleration is given by the smallest of the following quantities:
Response Spectrum Analysis of SDOF System 137

2.71

11.7Tn0.704
1 1.8Tn–1
A (g)

7.4Tn–2

0.1

0.01
0.01 1/33 0.1 1/8 0.66 1 4.12 10
Tn (s)

FIGURE 5.12  84.1th percentile acceleration response spectrum for ζ = 5%, plotted using logarithmic scales.

A = ügoα A
A = ωnV = 2πu go αV/Tn
A = ωn2D = (2π/Tn)2ugoαD

Also, as noted earlier, spectral quantities not only depend on PGA magnitude, but also on soil
characteristics at the recording station site and epicentral distance. The values for α A, αV, and αD in
the Newmark–Hall method were derived based on a limited number of records that were obtained
on a firm ground. More comprehensive studies have revealed that local site soil conditions have

3
2.71

2.5

2
11.7Tn0.704
A (g)

1.5 1.8Tn–1

0.5

0
0 1/8 0.5 0.66 1 1.5 2 2.5 3
1/33
Tn (s)

FIGURE 5.13  84.1th percentile acceleration response spectrum for ζ = 5%, plotted using linear scales.
138 Introduction to Earthquake Engineering

3
A B Soft to medium clay and
sand—15 records

2 Deep cohesionless soils

(>250 ft)—30 records
A/ü

C Stiff soils
(<200 ft)—31 records
1
D
Rock – 28 records
0
0 0.5 1.0 1.5 2.0 2.5 3.0
Tn

FIGURE 5.14  Influence of local soil conditions on response spectra for 5% damping. (Based in part on Seed,
H. B. and I. M. Idriss, Ground Motions and Soil Liquefactions during Earthquakes, Earthquake Engineering
Research Institute, Oakland, CA, 1982.)

a significant impact on spectra shapes. As shown in Figure 5.14, for soft soils, ügo remains the
same or decreases relative to firm soil, but u go and ugo increase. Also, as the epicentral distance
increases, the geology over the distance tends to filter the acceleration effects of certain frequencies.
This results in the short-period portion of the design spectra being controlled by nearby earthquakes
and the long-period portion of the design spectra controlled by earthquakes at further distances
from the structure.
Contemporary design spectral analysis is based on a uniform hazard spectrum that accounts
for the effects of all earthquakes expected at a site over a period of time using probability-based
­relations known as attenuation relations. These account for PGA magnitude, local site soil con-
ditions, and epicentral distance all together; and along with probabilistic models of earthquake
­occurrence, these relations are used to obtain site-specific design spectra. Usually, ­however, the
entire spectra are not computed following this procedure; rather, only accelerations correspond-
ing to a couple of natural periods (0.2 and 1 s) are determined, which are then used to gener-
ate full design spectra using approximate relations. This approach will be discussed further in
Chapter 8.

EXAMPLE 5.6
For the concrete bridge structure of Example 5.4, determine the deflection of the deck and base
shear when the bridge is subjected to ground acceleration due to an earthquake characterized by
the 84.1% design spectrum scaled to 0.32g PGA. The deck weighs 750 kips and the substructure
has 5% damping, modulus of elasticity, E = 3000 ksi, and the other geometric properties are
given in Example 5.4.

Solution
1. Mass, stiffness, and natural period of the system: The bridge can be modeled as an SDOF
system assuming a rigid deck and considering only lateral deformations of the columns.
The stiffness of the system is then the sum of column lateral stiffnesses. The mass and
stiffness of the SDOF system are calculated in Example 5.4; the resulting quantities are

m = 1.94 kips s2 /in

k = 93.4 kips /in

Tn = 0.906 s

Response Spectrum Analysis of SDOF System 139

The natural frequency of the system is

k 93.4 kips /in

ωn = = = 6.94 rad /s
m 1.94 kips s2 /in

2. Determine the maximum displacement, D: To determine D due to an 84.1% design

spectrum, scaled to 0.32g PGA, first, enter the response spectrum in Figure 5.12 with
Tn = 0.91 s > 0.66 s, which is in the constant velocity region, so

V = η · u goαV = 0.32( 48 in /s)( 2.3) = 35.3 in / s

The maximum spectral displacement is

V 35.3 in /s
D= = = 5.1in
ωn 18.38 rad /s

3. Determine the maximum base shear: Calculate maximum base shear at the support
using the force–displacement relationship or e­ quilibrium as described in Figure 3.24:

Vbmax = kD = 93.4 kips /in(5.1in) = 476 kips

EXAMPLE 5.7
The building frame shown in Figure E5.8, is subjected to a ground acceleration from an earth-
quake ­characterized by the 84.1% design spectrum scaled to 0.25g PGA, determine (i) maxi-
mum ­displacement u, (ii) the maximum base shear, and (iii) the maximum normal stresses in the
­columns (Ix = 75 in4 and Sx = 18 in3 for W8x21). Let us assume that the beam is rigid and system
damping is 5%.

Solution
1. Determine mass, stiffness, and natural period and frequency of the SDOF system: Since we
have assumed the beam to be rigid, the stiffness of the system is given by the sum of column
lateral stiffnesses. The mass and stiffness of the SDOF system are calculated as follows:
Mass,

W 15 kips(1000 lb /kip)
m= = = 38.86 lb s2 /in
g 386.4 in /s2

Stiffness (see Table 3.1), when both column are fixed–pinned,

3EI 3EI  3( 29, 000, 000 psi)(75 in4 ) 

k= + 3 = 2  = 2237.6 lb /in
L 3
L  (15 ft ×12 in /ft )3 

15 kips

15 ft

FIGURE E5.8  Schematic of frame structure in Example 5.7.

140 Introduction to Earthquake Engineering

The natural period of the system is

m 38.86 lb s2 /in
Tn = 2π = 2π = 0.828 s
k 2237.6 lb /in

The natural frequency of the system is

k 2237.6 lb /in
ωn = = = 7.59 rad /s
m 38.86 lb s2 /in

2. Determine the maximum displacement, D: To determine D due to an 84.1% design
spectrum, scaled to 0.25g PGA, first, enter the response spectrum in Figure 5.12 with
Tn = 0.83 s > 0.66 s, so velocity is constant,

V = η · u goαV = 0.25( 48 in / s)( 2.3) = 27.6 in /s

The maximum spectral displacement is

V 27.6 in/s
D= = = 3.6 in
ωn 7.59 rad /s

3. Determine the maximum base shear: Calculate maximum base shear at the support for
each column using the force–­displacement relationship:

 3EI   
 3( 29, 000 ksi)(75 in4 ) 
Vmax = kuo =  3  uo =   3.6 in = 4.1kips
 L   (15 ft ×12 in /ft )3 
 
4. Maximum normal stress due to bending: The maximum bending moment at the top of
each column is calculated from equilibrium of the column member. Since the column has
a pin support, summing moments about the top of the column will result in the following
maximum internal bending moment (see Figure 3.24):

Mmax = VmaxL = 4.1lb(15 ft ×12 in /ft ) = 732 kip ⋅ in

Applying the flexure formula from mechanics of materials (Equation 3.60), the normal
stress in the column due to the bending moment is determined as

Mmax 732 kip ⋅ in

σmax = = = 43 ksi
S 17 in3

5.2  INELASTIC RESPONSE SPECTRUM

It is not practical or cost effective to design structures to behave in the elastic range for all possible
seismic events. For the most severe cases, structural systems should be designed to resist seismic
loads that cause strains beyond the elastic limit. Newmark and Hall also developed a procedure
for accounting for inelastic behavior in a structural system response due to an earthquake—an
inelastic response spectrum. To simplify the analysis, they assumed the behavior of the structure
to be elastic-perfectly plastic (or elastoplastic) (see Figure 5.15). This assumption yields relatively
accurate results and was also used to develop the inelastic design response spectrum. In this case,
we solve the following nonlinear equation of motion:

mu + cu + FS (u) = −mug (t ) (5.14)

Response Spectrum Analysis of SDOF System 141

Fs
Corresponding linear system
Fo Fo = kuo

Fy Elastoplastic system

Fy = kuy

u
uy uo um

–Fy

FIGURE 5.15  Elastoplastic and corresponding linear system behavior.

where Fs(u) is the inelastic stiffness force.

All other parameters are the same as for elastic systems. The solution to this system requires
numerical integration of the equation and the results can be presented in the same fashion as those
for the elastic case, with one major difference, the ductility of the system, which is defined in terms
of the ductility ratio μ. This parameter is defined as
um
μ= ⇒ um = μuy (5.15)
uy

where:
um represents the maximum displacement of the SDOF system
uy is the yield displacement of the SDOF system

We can also define a parameter based on the fictitious corresponding linear system and the yield
force, known as the yield strength reduction factor, Ry:
Fo kuo uo
Ry = = = (5.16)
Fy kuy uy

where uo represents the maximum displacement of the corresponding linear elastic system.
Note that when u ≤ uy, Equation 5.16 is related to a real elastic system, whereas for u > uy, the
system behaves inelastically and may not return to its equilibrium position after the shaking has
stopped—it may experience permanent deformation.
In addition to μ, we can define a second displacement ratio between maximum displacement, um,
and the maximum displacement of the fictitious corresponding linear system, uo, which relates Ry and μ,

um um uy μuy μ
= = = (5.17)
uo uo uy uo Ry

This gives the ductility demand imposed on an elastoplastic system by ground motion:
um
μ= Ry (5.18)
uo
142 Introduction to Earthquake Engineering

50

10

10
ζ = 2, 5, 10%

0
20
6g
0.1
=
10 1 Ay

10
5
Vy (in/s)

1 µ=1
0.

1
2

01
1

0.

0.
1
µ=4

Ay
0.5

,g
n
y,i

1
00
0.

µ=8
u
01

0.
0.2
0.02 0.05 0.1 0.2 0.5 1 2 5 10 20 50
Tn (s)

FIGURE 5.16  Elastoplastic response spectra for El Centro quake. (Reproduced from Chopra, A. K.,
Dynamics of Structures: Theory and Applications to Earthquake Engineering, 4th edition, Prentice-Hall,
Upper Saddle River, NJ, 2012. By permission of Prentice-Hall, Upper Saddle River, NJ.)

For the purpose of designing a system, ductility capacity must be larger than the ductility
demand, that is, the system must have the ability to deform beyond the elastic limit.
The comparison between the elastoplastic and the corresponding linear systems is also useful in
establishing a procedure for using the elastic design response spectrum in Figure 5.10 to c­ reate the
inelastic design response spectrum. Data presented by Chopra (2012) shows that in the ­displacement
and velocity-sensitive regions, the maximum displacement is nearly equal to the ground ­displacement
and the strength reduction factor is nearly equal to the ductility factor, that is, the result is ­independent
of inelastic response. However, in the acceleration-sensitive region, the ductility demand is much
greater than the strength reduction factor; short-period systems are stiffer and tend to attract more load,
requiring a larger stiffness. Thus, we now have the response ­quantities (yield deformation, pseudo-
velocity, and pseudo-acceleration) being functions of three different variables (period, damping ratio,
and ductility ratio) (see Figure 5.16). This figure clearly indicates that d­ amping has a decreasing effect
as ductility increases and is much less effective than d­ uctility in reducing the maximum response.
Furthermore, to increase a system’s earthquake r­esistance, it can be designed by either making it
strong or ductile (or a combination of the two), because i­ ncreasing ductility greatly decreases response.
Given yield displacement, uy, the relationships between elastic (Equations 5.9) and inelastic
quantities are as follows:
Dy = uy
Vy = ωnDy = ωnuy
Ay = ωn2Dy = ωn2uy = (2π/Tn)2Dy
And since um = μuy or um = μDy,
 T 2
um = µ  n  Ay (5.19)
 2π 

EXAMPLE 5.8
Consider the building frame and characteristics given in Example 5.7. If the building frame is
subjected to the El Centro earthquake, determine the maximum required column strength for a
ductility ratio of 4.
Response Spectrum Analysis of SDOF System 143

Solution
1. Mass, stiffness, and natural period and frequency of the SDOF system were deter-
mined in Example 5.7. That is, m = 38.86 lb s2/in, k = 2237.6 lb/in, Tn = 0.828 s, and
ωn = 7.59 rad/s.
2. Determine the required column strength for a ductility demand of 4, Fy: To determine Fy,
first, enter the response spectrum in Figure 5.16 with Tn = 0.83 s and read the pseudo-
acceleration, Ay,

Ay ≅ 0.16g

With the yield displacement, we obtain Fy,

Fy = mAy = (38.86 (lb − s2 ) / in)(0.16)(386.4 in /s2 ) = 2400 lbs

5.2.1 Inelastic Design Response Spectrum

The elastic design response spectrum is constructed using a backbone of peak ground properties
and adjusting these values with the structural amplification factors. In a similar fashion, we can
construct the inelastic design response spectrum using a new backbone of elastic spectral response
properties and a new set of adjustment parameters that account for inelastic behavior. These adjust-
ment parameters are functions of system ductility. Data analyses for elastic and inelastic constant
spectral displacement and velocity regions show that the maximum inelastic displacement of the
elastoplastic system is approximately equal to the maximum displacement of a linear system
with the same period (see left-hand-side graph of Figure 5.17). First, we write the ductility ratio
(Equation 5.15) in terms of forces using the system stiffness. That is,

um kFe F
μ= = = e (5.20)
uy kFy Fy

Now we can rewrite this relationship in terms of accelerations using the system mass,

Fe Fe muy uy
μ= = = = (5.21)
Fy Fi mui ui

Fs
Fs
Fe Fe

Fy Fi

Fy Fi

u u
uy um uy ue ui

FIGURE 5.17  Force deformation curves for elastic and elastoplastic systems.
144 Introduction to Earthquake Engineering

Since Fi = Fy. This results in the maximum inelastic acceleration being equal to the maximum
­elastic acceleration divided by the ductility factor,

uy A
ui = ⇒ Ay = (5.22)
μ μ

The result presented in Equation 5.22 agrees relatively well with experimental data for the
c­ onstant displacement and velocity regions, but not for the constant acceleration region. For  the
constant acceleration region, the relationship is obtained using the equivalence of energy in
­
the  ­elastic (AREAe) and inelastic (AREAi) systems. AREAe is the hatched area, while AREAi is
the solid area in the right-hand-side graph of Figure 5.17. That is,

1 1 u2
AREAe = ue Fe = Fy e (5.23)
2 2 uy

Since k = Fe/ue = Fy/uy or Fe = Fy ue/uy,

1
AREAi = uy Fy + (ui − uy )Fy (5.24)
2

Setting Equations 5.23 and 5.24 equal, we obtain

1 ue2 1
Fy = uy Fy + (ui − uy )Fy (5.25)
2 uy 2

Using the definition of the ductility ratio, μ = ui/uy, we can now solve for the inelastic
­displacement in terms of the elastic displacement,

μ μ
ui = ue ⇒ Dy = D (5.26)
2μ − 1 2μ − 1

In terms of the accelerations,

ue A
ui = ⇒ Ay = (5.27)
2μ − 1 2μ − 1

since Fe = Fy ue/uy, Fy = müy = müi, and Fe = müe.

With the adjustment parameters, we can easily construct the inelastic design response spectrum
as shown in Figure 5.18. Constructing the inelastic design response spectrum follows similar pro-
cedures to the elastic case. That is, after identifying the ductility demand, μ, the value of the damp-
ing ratio, ζ, and choosing a mean or mean + 1σ plot, we can follow the details depicted in Figure
5.18, where the quantities Ay, Vy, and Dy are the inelastic acceleration, velocity, and displacement,
respectively. Note that stiff systems having periods <1/33 s do not experience any adjustment—as
it is for the elastic case.
As shown in the last section, elastic response spectra are typically given in terms of accel-
eration, which is how contemporary codes express elastic and inelastic seismic response.
Response Spectrum Analysis of SDOF System 145

Tc V = αV ugo Td
D
=
α
Du
Tc′ Vi = V/µ Td′ go
u go
Te

Di
αA

) 0.5
=

=
A

D/
µ–

µ
(2
A/
Tb

=
A
i
V or Vi (log scale)

Tf

D/
µ

u go
D/
Ta µ

u go

1/33 s 1/8 s 10 s 33 s

Tn (log scale)

FIGURE 5.18  Newmark–Hall inelastic design response spectrum.

Thus, to obtain the maximum displacement, we use the relationship between displacement and
pseudo-acceleration:

 T 2
uo = µ  n  Ay (5.28)
 2π 

Also, the maximum force (required yield strength) can be determined using the relations between
mass and acceleration:
Fo = mAy (5.29)

EXAMPLE 5.9

Consider the building frame shown in Figure E5.9 with ζ = 5% and Ix = 75 in4 for the two columns
subjected to a ground acceleration due to an earthquake characterized by the 84.1% design spec-
trum scaled to 0.5g PGA. Assume elastoplastic behavior and determine the lateral deformation
and force for which the frame should be designed if μ = 4.

1500 lbs

15 ft

FIGURE E5.9  Building frame schematic.

146 Introduction to Earthquake Engineering

Solution
1. Mass, stiffness, and natural period of the SDOF system: Since we assumed the beam to
be rigid, the stiffness of the system is given by the sum of column lateral stiffnesses. The
mass and stiffness of the SDOF system are calculated as follows:
Mass,

W 1500 lbs
m= = = 3.88 lb ⋅ s2 /in
g 386.4 in /s2

Stiffness (see Table 3.1), when both columns are fixed–pinned,

3EI 3EI  3( 29, 000, 000 psi)(75 in4 ) 

k= + 3 = 2  = 2237.6 lb /in
L 3
L  (15 ft ×12 in /ft )3 

The natural period of the system is

m 3.88 lb s2 /in
Tn = 2π = 2π = 0.262 s
k 2237.6 lb /in

2. Determine the maximum displacement for ductility of 4, uo: First determine A due to an
84.1% design spectrum, scaled to 0.5g PGA, by entering the response spectrum in Figure
5.12 with Tn = 0.26 s, which is in the acceleration-sensitive region,

A = η · 2.71g = 0.5( 2.71g ) = 1.335g

The maximum displacement is computed using Equation 5.28 and Figure 5.18,

2
 T 2  T 2 A  0.262 s  1.335(386.4 in /s2 )
uo = µ  n  Ay = µ  n  = 4   = 1.353 in
 2π   2π  2µ − 1  2π  2(4) − 1

3. Determine the required column strength for a ductility of 4, Fy: To determine Fy, use
A = 1.335g and Equation 5.29.

mA 3.88 lb s2 /in(1.355)(386.4 in /s2 ) A

Fy = mA y = = = 770 lbs
2μ − 1 2(4) − 1

PROBLEMS
5.1 Use the MATLAB script provided in Chapter 4, Example 4.13 to draw response
­spectra for the displacement, velocity, and acceleration for an SDOF system sub-
jected to the 1994 Northridge earthquake ground acceleration for 2%, 5%, and 10%
damping.
5.2 The building frame shown is subjected to the El Centro earthquake; determine the
total stiffness, structural period, deflection of the beam, base shear, and bending
stresses in each of the columns. The roof weighs 10 kips and the substructure has 5%
damping, modulus of elasticity, E = 29,000 ksi, and the other geometric properties
shown.
Response Spectrum Analysis of SDOF System 147

10 kips

Pin connection I1 = 200 in4, I2 = 400 in4

14 ft
1 2 S1 = 40 in3, S2 = 80 in3

5.3 The following structural system is subjected to the El Centro earthquake; determine
the total stiffness, structural period, deflection of the beam, and base shear. The sub-
structure has 2% damping, modulus of ­elasticity, E = 3000 ksi, and the other geometric
properties as shown.

W = 750 kips u

I1 = I3 = 25,000 in4, I2 = 60,000 in4

15 ft I3
20 ft I1
30 ft I2

5.4 A 25 kip sensitive piece of equipment will be installed on the roof of the following
frame shown with a rigid slab and four columns with EIx = 125,000 kip in2 and 5%
damping. The frame is located in Mexico City at a site with response spectrum depicted
in Example 5.2. Determine approximately the magnitude of the shear force (Vreq’d)
needed to design the equipment anchorage.

25 kips

Vreq’d

12 ft

5.5 Use a tripartite graph to draw the elastic 50th percentile design response spectrum with
a maximum ground acceleration of 0.25g and a damping ratio of 2%.
5.6 For the frame of Problem 5.2, determine the deflection of the beam, base shear, and
­bending stresses when the frame is subjected to ground acceleration due to an ­earthquake
characterized by the 84.1% design spectrum scaled to 0.5g peak ground acceleration. The
substructure has 5% damping and the other geometric properties given in Problem 5.2.
5.7 Burn’s tower (a water tank weighing 4200 kips) on the campus of University of the
Pacific is supported on a 120-ft high cantilever tower with a stiffness of 2000 kips/in
and 5% damping. Determine the design values for lateral deformation and base shear if
the tower is subjected to ground acceleration due to an earthquake characterized by the
84.1% design spectrum scaled to 0.3g peak ground acceleration.
148 Introduction to Earthquake Engineering

5.8 Use a tripartite graph to draw the inelastic 50th percentile design response spectrum
with a maximum ground acceleration of 0.25g, a damping ratio of 2%, and a ductility
ratio of 5.
5.9 Given the following 14-ft. steel tower subjected to a ground acceleration due to an earth-
quake characterized by the 84.1% design spectrum scaled to 0.25g peak ground accel-
eration, determine (i) maximum lateral displacement, and (ii) the maximum base shear.
Assume system damping of 5% and ductility of 5.

W = 38.6 kips

Ix = 75 in4

Sx = 18 in3

5.10 Consider the following building frame with a rigid beam that is pinned con-

nected to one column and rigidly connected to the other as shown; each column has
EI = 40,000,000 kip in2. The frame is subjected to ground acceleration due to an
earthquake characterized by the 84.1% design spectrum scaled to 0.33g peak ground
acceleration. Determine maximum stress in each column given that Sx = 2,000 in3 and
damping is 5%.

200 kips

10 ft Pinned

15 ft

5.11 Solve for the maximum base shear in the frame shown in problem 5.10 assuming a duc-
tility ratio of 4.
5.12 Consider a one-story frame with lumped weight w and natural vibration period,

T = 0.25 s, in the elastic range. Determine the lateral deformation and force in terms of
w for which the frame should be designed if μ = 6. Assume that ζ = 5%, elastoplastic
behavior, and an 84.1% design earthquake with ügo = 0.25g.

uo
Fo
w

k/2 k/2

5.13 Two identical power transformers to be erected adjacent to each other are supported
by steel (E = 29,000 ksi) pedestals and can be idealized as SDOF systems. The ped-
estals are designed to withstand inelastic deformations of up to six times their yield
Response Spectrum Analysis of SDOF System 149

values. Using an inelastic 50th percentile design response spectrum with a maximum
ground ­acceleration of 0.25g, estimate the minimum clear distance (Δ) that must be left
between the two transformers to avoid collision. I = 1380 in4 and W = 1.5 kips.

W W
∆

20 ft

REFERENCES
Chopra, A. K., Dynamics of Structures: Theory and Applications to Earthquake Engineering, 4th edition,
Prentice-Hall, Upper Saddle River, NJ, 2012.
Seed, H. B. and I. M. Idriss, Ground Motions and Soil Liquefactions during Earthquakes, Earthquake
Engineering Research Institute, Oakland, CA, 1982.
 6 Generalized SDOF
System Analysis
After reading this chapter, you will be able to:

1. Setup the equation of motion for generalized SDOF systems

2. Determine the generalized properties, including natural period and participation factor for
generalized shear building systems and distributed mass/elasticity systems
3. Use response spectra and participation factors to calculate the maximum base shear and
overturning moment response of generalized shear building systems and distributed mass/
elasticity systems

Before presenting the formulation for a system discretized into multiple degrees of freedom, it
is instructive to study the development of an approach that condenses these MDOF systems into a
SDOF case. Unlike the MDOF system formulation, this approach does not require simultaneous
manipulation of multiple algebraic equations. The key to obtaining relatively accurate results with
a generalized SDOF system analysis is selecting an appropriate shape function for the deformed
system during vibration.
In this chapter, we also describe the analysis of systems with distributed mass and stiffness as
generalized SDOF systems. While this analysis approach is not as precise as modeling a system
with an infinite number of degrees of freedom, modeling systems as generalized SDOF systems
is more accurate than approximating the entire system as a simple SDOF system as discussed in
Chapter 3. Furthermore, utilizing an infinite number of degrees of freedom to model the distributed
mass and stiffness as a function of position would require formulation in terms of partial differential
equations in both position and time.
The generalized SDOF system analysis presented in this chapter is based on replacing the
required partial differential equation in time and space with an approximate differential equation in
time only. The space variation is tracked using an approximate equation of the mode of vibration for
the system. The process entails selecting a point of interest along the distributed system and track-
ing its motion with respect to time using an SDOF analysis; in order to obtain the response of all
other points, we scale the magnitude of the point in question based on the approximate equation of
the mode of vibration. The solution to the differential equation representing the generalized SDOF
system follows the standard procedures presented in Chapter 3; and we can apply all the solution
techniques discussed thus far, including the response spectrum analysis presented in Chapter 5.
Replacing the distributed dynamic properties (mass, damping, and stiffness) with lumped prop-
erties at the point being tracked requires integrating the variation in these properties along the
length of the system. For discrete MDOF systems, the process involves combining the discrete
properties associated with each DOF into a set of generalized properties. These generalized proper-
ties are most useful in developing the participation factor, which is critical in applying the results
from an SDOF system analysis to an MDOF system analysis.
The formulation of the equation of motion of a generalized SDOF system entails condensing
all degrees of freedom into one. In the following sections, we present analysis procedures for a
multistory shear building frame (discrete system) and a cantilever tower with distributed properties
(continuous system), both subjected to earthquake ground motions. In each case, we assume axial
deformations to be negligible during the dynamic excitation, which leaves the lateral translation
(or deflection) as the only possible deformation of the system.

151
152 Introduction to Earthquake Engineering

6.1  DISCRETE SYSTEM (SHEAR BUILDINGS)

Multistory building systems can be modeled as having a discrete number of degrees of freedom, the
equation of motion for which can be formulated by assuming the building mass to be distributed as
concentrated masses at various levels, as in the simplified lumped mass tower shown in Figure 6.1;
see Section 3.1.2 for a discussion on how to determine the lumped mass at each level. The deformed
shape of the system along with the virtual deformation used in the formulation of the equation of
motion is also depicted in the figure. The actual deformation of this system is a function of position
and time. The time effect is only tracked at the top of the building using z(t), whereas the posi-
tion of the entire shear building is followed using an assumed shape vector, ψj, which is typically
determined by evaluating a shape function at each level. Also shown in Figure 6.1 is an equivalent
generalized SDOF system.

6.1.1 Equation of Motion
The actual motion of the shear building shown in Figure 6.1 can be obtained by determining the
motion of the top of the building, z(t), and scaling it to the other levels using an appropriate shape
vector, ψj. That is, at an arbitrary level, j, the relative displacement is

u j (t ) = ψ j ⋅ z (t ) (6.1)

where:
j = 1, 2,… , N

The motion at the top of the building, z(t), can be obtained using a generalized equation of motion
for the equivalent SDOF system shown in Figure 6.1. This equation of motion can be obtained
using D’Alembert’s principle as discussed in Chapter 3 or the principle of virtual work, which
is the approach we follow in this chapter. The resulting equation is of the same form as the one

ut
mN z
N uN δz
z(t)
kN
~
m

mj
j uj δuj
kj
~
k
m2
u2
2
k2
m1
1 u1
k1

ug(t)

ug(t)

FIGURE 6.1  MDOF shear building and equivalent generalized SDOF system.
Generalized SDOF System Analysis 153

that c­ haracterizes the vibration of the SDOF oscillator introduced in Chapter 3; thus, all analysis
­procedures covered so far are applicable to this case.
The shape vector, ψj, can be obtained from any function, ψ(x), provided it satisfies the geometric
boundary conditions. However, the accuracy of the results depends on how close the assumed shape
function approximates the actual shape of the deformed frame. Also, the function ψ(x) must be
­normalized at the top of the building such that ψ(H) = 1.0, where H is the height of the building.
Figure 6.2 presents several possible shape functions based on building elevation aspect ratio.
To account for the ground excitation loading in the formulation of the equation of motion, we
also need the total displacement at level j, which is a combination of the relative displacement, uj(t),
and the ground level lateral displacement, ug(t):

utj (t ) = u j (t ) + ug (t ) (6.2)

We can now use the principle of virtual work to derive the generalized equation of motion.
This principle states that the virtual work of all internal forces, δWI, equals the virtual work of all
­external forces, δWE, for a given virtual displacement, δu. That is,

δWI = δWE (6.3)

where δWI includes the contribution from stiffness and damping forces:

δWI = δWstiffness + δWdamping (6.4)

And δWE includes the contribution from inertial forces of each mass:

δWE = δWinertial (6.5)

The virtual work of each of these groups of forces is given by summation of the product of the
magnitude of the force and the corresponding virtual displacement. The stiffness and damping
forces can be viewed as effective shear forces in each story, whereas the inertial forces can be
regarded as effective lateral loads applied at each level. The stiffness forces are functions of story

D
1

D
1
H

D
1
H

ψ(x)
H

ψ(x) ψ(x)

Low-rise, H/D < 1.5 Mid-rise, 1.5 < H/D < 3 High-rise, H/D > 3
πx πx
ψ(x) = sin ( 2H ) ψ(x) = ( Hx ) ψ(x) =1–cos ( 2H )

FIGURE 6.2  Possible shape functions based on building elevation aspect ratio. (With kind permission from
Springer Science+Business Media: Naeim, F. 2001. The Seismic Design Handbook, 2nd edition, Dordrecht,
The Netherlands.)
154 Introduction to Earthquake Engineering

drift (relative displacement from one floor to the next); the damping forces are functions of relative
velocity from one floor to the next; and the inertial forces are functions of the total acceleration at
each level.
First, determine the contribution of the stiffness forces to the virtual work equation. The stiffness
force at level j, fsj, is equal to the corresponding story stiffness times story drift,

fsj = k j (u j − u j −1 ) (6.6)

where kj is the sum of all the lateral stiffness systems (columns, shear walls, bracing, etc.) in the
direction in question at story j. Substituting Equation 6.1 evaluated at stories j and j–1 into this
equation yields

fsj = k j ( z(t )ψ j − z(t )ψ j−1 ) (6.7)

We can factor out z(t) and replace the shape vector values with the change in values between two
stories, (Δψj = ψj − ψj−1),

fsj = k j z(t ) ⋅ Δψ j (6.8)

The virtual work of this force is

δWstiffness j = fsj (δu j − δu j−1 ) (6.9)

where δuj represents the internal virtual displacement of the structure. Substituting the value of fsj
from Equation 6.8 and virtual values of Equation 6.1 into Equation 6.9 results in the internal virtual
work due to stiffness,

δWstiffness j = k j z(t ) ⋅ Δψ j (δ zψ j − δ zψ j−1 ) (6.10)

We can factor out δz and substitute Δψj = ψj − ψj−1,

δWstiffness j = δ z ⋅ z(t )k j ⋅ Δψ 2j (6.11)

We can then add the contribution of all N stories to obtain the total internal work from the stiff-
ness forces in each story,

δWstiffness = δ z ⋅ z(t ) ∑ k ⋅ Δψ
j =1
j
2
j (6.12)

Similarly, we can determine the contribution of the damping forces to the virtual work equation.
The damping force at level j, fdj, is equal to the corresponding story damping times story relative
velocity,

fdj = c j (u j − u j −1 ) (6.13)

where cj is the sum of all the damping systems in the direction in question at story j.
Generalized SDOF System Analysis 155

Substituting Equation 6.1 evaluated at stories j and j–1 into this equation yields

fdj = c j ( z(t )ψ j − z(t )ψ j−1 ) (6.14)

We can factor out the velocity and substitute Δψj = ψj − ψj−1,

fdj = c j z(t ) ⋅ Δψ j (6.15)

The virtual work of this force is

δWdamping j = fdj (δu j − δu j−1 ) (6.16)

Substituting the value of fdj from Equation 6.15 and virtual values of Equation 6.1 into Equation
6.16 results in the internal virtual work due to damping,

δWdamping j = c j z(t ) ⋅ Δψ j (δ zψ j − δ zψ j−1 ) (6.17)

We can factor out δz and substitute Δψj = ψj − ψj−1,

δWdamping j = δ z ⋅ z(t )c j ⋅ Δψ 2j (6.18)

We can then add the contribution of all N stories to obtain the total internal work from the damp-
ing forces in each story,

δWdamping = δ z ⋅ z(t ) ∑ c ⋅ Δψ
j =1
j
2
j (6.19)

Substituting Equations 6.12 and 6.19 into Equation 6.4, we obtain the total internal virtual work,

N N

δWI = δ z ⋅ z(t ) ∑ j =1
k j ⋅ Δψ 2j + δ z ⋅ z(t ) ∑ c ⋅ Δψ
j =1
j
2
j (6.20)

Finally, we determine the contribution of the external forces to the virtual work equation. The
inertial force at level j, f Ij, is equal to the corresponding mass times story total acceleration,

fIj = m jutj (6.21)

Substituting the second derivative with respect to time of Equation 6.2 into this equation yields

fIj = m j (ug + uj ) (6.22)

Now taking the second time derivative of Equation 6.1 evaluated at story j and substituting into
Equation 6.22 gives,

fIj = m j (ug + 
z (t )ψ j ) (6.23)

156 Introduction to Earthquake Engineering

The external virtual work of the inertial force is

δWinertial j = − fIjδu j (6.24)

This is negative because the sense of the force is opposite to the direction of the virtual displace-
ment. Substituting the value of f Ij from Equation 6.23 and virtual values of Equation 6.1 yields

δWinertial j = −m j (ug + 
z (t )ψ j )δzψ j (6.25)

We can now sum the contribution of all N stories to obtain the total external work from the
­inertial forces at each level j,

N N

δWE = −δz ⋅ ug ∑j =1

m j ψ j − δz ⋅ 
z (t ) ∑m ψ
j =1
j
2
j (6.26)

We can now apply the principle of virtual work (Equation 6.3) by setting the internal virtual
work, Equation 6.20, equal to the external virtual work, Equation 6.26,

N N N N

δ z ⋅ z (t ) ∑ j =1
k j ⋅ ∆ψ 2j + δ z ⋅ z(t ) ∑j =1
c j ⋅ ∆ψ 2j = −δ z ⋅ ug ∑ j =1
m j ψ j − δ z ⋅ (
z t) ∑m ψ
j =1
j
2
j (6.27)

After eliminating the virtual displacement, δz, since it appears in each term, and rearranging this
equation we obtain

N N N N

z (t )∑ j =1
∑
k j ⋅ ∆ψ 2j + z(t )
j =1
c j ⋅ ∆ψ 2j + 
z (t ) ∑ j =1
m j ψ 2j = −ug ∑m ψ
j =1
j j (6.28)

This is the equation of motion of a generalized SDOF system, which can be written in the same
form as the equation of motion derived in Chapter 3,

  + cz
mz  = − Lu
 + kz  g (t ) (6.29)

where:

∑
N
m = m j ⋅ ψ 2j is the generalized mass
j =1

k = ∑
N
k j ⋅ (∆ψ j )2 is the generalized stiffness
i =1

c = ∑
N
c j ⋅ (∆ψ j )2 is the generalized damping
i =1

L = ∑
N
m j ⋅ ψ j is the generalized force
i =1
N in the summation is the number of stories in the building
Δψj is the relative value of the shape function of two consecutive levels, that is, ∆ψ j = ψ j − ψ j−1
mj is the mass at level j
kj is the stiffness for the jth story
cj is the damping for the jth story
Generalized SDOF System Analysis 157

6.1.2  Generalized Participation Factor, Frequency, and Natural Period

Rearranging the generalized equation of motion (Equation 6.29) into the form of Equation 3.42, we
obtain

z + 2ζωn z + ωn2 z = −Γ ug (t )

 (6.30)

where:
the generalized natural frequency is given as

k (6.31)
ωn =
m

the generalized participation factor is given as

L
Γ = (6.32)
m

ζ is an estimate of the damping ratio.

We can also establish the generalized natural period,

m
Tn = 2π (6.33)
k

6.1.3 Deflections, Base Shear, and Moments Using

Participation Factors and Response Spectra
After finding the period, Tn, using Equation 6.33, we can determine the maximum dynamic response
caused by any dynamic excitation (or force) using the SDOF system analysis discussed in Chapters 3
and 4. Also, we can use the response spectrum analysis method discussed in Chapter 5 to ­determine
the peak response to an earthquake characterized by a design response spectrum with spectral dis-
placement, velocity, and acceleration (D, V, and A, respectively) for an appropriate damping ratio. For
 Thus, the
this case, the original ground acceleration, ug(t), is multiplied by the participation f­ actor, Γ.
solution to the equation of motion based on response spectrum analysis using the spectral displace-
ment, D, gives the maximum displacement at the top of the building, relative to the ground, as

z = Γ ⋅ D (6.34)

This maximum displacement can be distributed to the other floors using the shape vector, ψj
(Equation 6.1), as shown in Figure 6.3,

u jo = ψ j ⋅ z = ψ j ⋅ Γ ⋅ D (6.35)

This can also be written in terms of the spectral acceleration, A as,

Γ
u jo = ψ j A (6.36)
ωn2
158 Introduction to Earthquake Engineering

mN z mN
fNo uNo fNo
kN kN

mj mj
fjo ujo fjo
kj kj

hj fio
Vmax i
m1 Mmax i
f1o u1o hi
k1

Mmax b Vmax b

FIGURE 6.3  Maximum dynamic displacements and associated equivalent static forces.

The acceleration is proportional to the displacement; that is, A = ωn2 D.

The equivalent static forces associated with these floor displacements can be obtained using the
product of the maximum floor displacement and associated stiffness or the product of the associated
story mass and maximum story acceleration; that is,

f jo = Γ m j ψ j A (6.37)

Also, as discussed in Section 3.3, with these forces we can conduct a static structural analysis to
determine element forces (bending moment, shear force, and axial force) and stresses needed for
design of structural elements; no additional dynamic analysis is necessary. Figure 6.3 shows these
equivalent static forces along with the base shear and overturning moment. The internal story shear
force, Vmax i, and internal story moment, Mmax i, at an arbitrary level i can be obtained by applying
static equilibrium to the right-hand-side free-body diagram,

Vmax i = ∑f
j =i
jo
(6.38)

N

M max i = ∑ (h − h ) f
j =i
j i jo
(6.39)

Setting i equal to 1 in Equations 6.38 and 6.39 results in the base shear force, Vmax b, and overturn-
ing moment, Mmax b,
N

Vmax b = ∑f
j =1
jo
(6.40)

N

M max b = ∑h f
j =1
j oj
(6.41)

In the derivation of Equations 6.39 through 6.41, it is assumed that the floor weights are directly
in the center of the building frame; thus, making no contribution to the moment equilibrium
Generalized SDOF System Analysis 159

e­ quation. However, for buildings with off-center weights, a full moment static equilibrium analysis
should be conducted to account for the contribution of these off-center forces.

EXAMPLE 6.1
Given the three-story building shown in Figure E6.1 frame with damping of 5% and subjected to
a ground acceleration due to a 0.25g peak ground acceleration (PGA) earthquake characterized
by an 84.1% design spectrum, determine (i) peak displacements, (ii) maximum base shear, and
(iii) maximum floor overturning moments. Let us assume that beams are rigid. Each story has a
stiffness k = 326.3 kips/in and a shape function appropriate for a mid-rise building is to be used.

Solution
1. Determine the shape vector, ψj. This can be obtained using Figure 6.2 for a mid-rise
building function, ψj = x/H, where H = 36 ft.

Level, j x (ft) ψ (x) ψj

3 36 36/H 1
2 24 24/H 2/3
1 12 12/H 1/3

ψj in the vector form,

1/ 3
 
ψ j = 2/ 3
 
 1 

2. Determine the generalized properties. The building frame can be modeled as a gen-
eralized SDOF system assuming that the beams are rigid and only lateral deformations
of columns occur. The stiffness of each floor is then obtained by summing the column
­lateral stiffnesses at each level. The generalized mass, stiffness, and force of the general-
ized SDOF system are calculated as follows:

Level, j mj kj ψj Δψj m j ⋅ ψj m j ⋅ ψ2j k j ⋅(∆ψj )2

3 m/2 k 1 1/3 m/2 ⋅ (1) m/2 ⋅ (1)2 k ⋅ (1/3)2
2 m k 2/3 1/3 m ⋅ (2/3) m ⋅ (2/3)2 k ⋅ (1/3)2
1 m k 1/3 1/3 m ⋅ (1/3) m ⋅ (1/3)2 k ⋅ (1/3)2

∑
3
= 1.5 m 19/18 m 1/3 k
j =1

Substituting the given values for mass and stiffness, the following properties of the
generalized SDOF are calculated:
Generalized mass,

19 100 kips (1000 lb /kip)

3
19 19 W
 =
m ∑m ⋅ψ
j =1
j
2
j =
18
m=
18 g
=
18 386.4 in/s2
= 273.2lb s2 /in

Generalized stiffness,

3
1 1
k = ∑ k ⋅ (∆ψ )
j =1
j j
2
=
3
k = 326.3 kips /in(1000 lb /kip) = 108,767 lb /in
3

160 Introduction to Earthquake Engineering

Generalized force,

3
W 100 kips(1000 lb /kip)
L = ∑ m ⋅ ψ = 1 . 5 m = 1 . 5 g = 1 .5
j =1
j j
386.4 in/s2
= 388.2 lb s2 /in

3. Determine the natural period, frequency, and participation factor of the generalized
SDOF system:
Natural period,


m 273.2 lb s2 /in
Tn = 2π  = 2π = 0.315 s
k 108,767 lb /in

Natural frequency,

k 108,767 lb /in
ωn = = = 19.95rad /s
m 273.2 lb s2 /in

Participation factor,

L 388.2 lb s2 /in
Γ = = = 1.421

m 273.2 lb s2 /in

4. Determine maximum floor displacements: Determine the maximum displacement at the

top of the building, z(t) or D due to an 84.1% design spectrum, scaled to 0.25g PGA;
first, enter the response spectrum in Figure 5.11 with Tn = 0.315 s, which is greater than
1/8 s so the acceleration is constant,

A = η ⋅ 2.71g = 0.25( 2.71)(386.4 in /s2 ) = 261.8 in /s2

where:
η is the PGA scale for a given earthquake; in this case 0.25

The maximum spectral displacement is

A 261.8 in /s2
D= = = 0.658 in
ωn2 (19.95 rad /s)2

We can now determine the maximum floor displacements using Equation 6.35,

1/ 3  0.311

   
u jo = ψ j Γ ⋅ D = 2/ 3 (1.421)(0.658 in) = 0.623 in

   
 1  0.934

5. Determine the equivalent static story forces: Determine the equivalent static story forces
using Equation 6.37,

f jo = Γ m j ψ j A

Generalized SDOF System Analysis 161

50 kips

k 100 kips 12 ft

k 12 ft
100 kips

12 ft
k

FIGURE E6.1  Schematic of three-story building frame.

f3o = Γ m3ψ3 A = 1.42150

( kips/g )(1)(0.25 ⋅ 2.71⋅ g ) = 48.1kips

f2o = Γ m2ψ2 A = 1.421(100 kips/g )( 2/ 3)(0.25 ⋅ 2.71⋅ g ) = 64.2 kips

f1o = Γ m1ψ1A = 1.421100

( kips/g )(1/ 3)(0.25 ⋅ 2.71⋅ g ) = 32.1kips

6. Determine base shear and overturning moment: Determine the base shear force using
Equation 6.40,

Vmax b = ∑f j =1
jo = 32.1kips + 64.2 kips + 48.1kips = 144.4 kips

Determine the overturning moment using Equation 6.41,

Mmax b = ∑h f
j=1
j oj = 32.1kips(12 ft ) + 64.2 kips( 24 ft ) + 48.1kips(36 ft ) = 3658 kip ft

We can also obtain the internal story shear forces and moments at each level using
Equations 6.38 and 6.39, respectively. Alternatively, we can draw shear force and
­bending moment diagrams by treating the building as a cantilever beam in order to
obtain the internal story shear forces and moments shown in Figure E6.2. This diagram
also summarizes the results of the analysis.

48.1 kips 0.93 in

64.2 kips 12 ft
0.62 in
48 –578
12 ft
32.1 kips 0.31 in
112 –1925
12 ft

144.4 kips 144 –3658

3658 kips ft V (kips) M (kips ft)

FIGURE E6.2  Lateral forces, internal shear, and internal moment diagram.
162 Introduction to Earthquake Engineering

6.2  GENERALIZED SDOF CONTINUOUS SYSTEM

The equation of motion for a cantilever tower (or beam) with distributed mass, m(x), and distributed
flexural stiffness, EI(x), can be formulated using the tower shown in Figure 6.4. The deformed shape
of the system along with the virtual deformation used in the formulation of the equation of motion
is depicted in the figure. In this case, the deformation of the tower along its length is u(x,t), which
can be decoupled into two functions, one in terms of position, ψ(x), and the other in terms of time,
z(t), representing the dynamic displacement at the top of the tower. Also shown in the figure is an
equivalent generalized SDOF.

6.2.1 Equation of Motion
The equation of motion for a tower subjected to a time-dependent input force or ground excita-
tion can be formulated in terms of z(t) only, provided we can find an appropriate shape of the
deformed system as it vibrates, ψ(x). The generalized equation of motion in terms of z(t) can again
be derived using the principle of virtual work as described in the last section; the resulting second-
order ordinary differential equation is of the same form as Equation 6.29 and is rewritten below for
convenience,

  + cz
mz  = −Lu
 + kz  g (t )

In the continuous system, integration is used to determine properties of the generalized
SDOF as follows:
L
m =
∫ 0
m( x )[ψ( x )]2 dx is the generalized mass
L
k =
∫ 0
EI ( x )[ψ ′′( x )]2 dx is the generalized stiffness
L
L =
∫ 0
m( x )ψ( x ) dx is the generalized force

Generalized damping, c , is not explicitly determined; it’s estimated using the damping ratio,
L is the tower height,

δz z(t)

ut

z(t)
EI(x), m(x)

δu u(x,t) = ψ(x)⋅z(t)
~
m
L

x
~
k

ug(t) ug(t)

FIGURE 6.4  Tower with distributed properties and equivalent generalized SDOF system.
Generalized SDOF System Analysis 163

m(x) is the mass per unit height at an arbitrary height x,

EI(x) is the flexural stiffness per unit height at an arbitrary height x, and
ψ ′′( x ) = d 2ψ dx 2 , the second derivative of ψ(x) with respect to x.

Again, this generalized equation of motion is the same as the one that characterizes the vibration
of the SDOF oscillator introduced in Chapter 3; thus, all analysis procedures described so far can
be used to determine the response z(t). With z(t) and the shape function ψ(x), we can determine a
function of the shape into which the tower vibrates, which is the relative displacement at an arbitrary
height x,

u( x, t ) = ψ ( x )z(t ) (6.42)

The accuracy of the results in this case depends on how close the assumed shape function
tracks the actual shape of the deformed tower. A satisfactory approximation for the shape func-
tion, ψ(x), must satisfy the geometric boundary conditions of the tower, which for this case
include zero ground lateral relative displacement and rotation; that is, u(0) = 0 and u′(0) = 0,
where the prime notation indicates differentiation with respect to x, u′(x) = du/dx. Also, since z
is only a function of t, these boundary conditions reduce to ψ(0) = 0 and ψ′(0) = 0. Furthermore,
the function ψ(x) must be ­normalized at the top of the tower such that ψ(L) = 1.0, where L is the
height of the tower.
This analysis can be used to determine internal stress resultants in the tower, which can then
be used to determine the stresses. The analysis procedure can be used with a variety of cantilever
­systems subjected to different time-dependent loads (Chopra 2012). Like the shear building, here
we assume a tower subjected to a time-dependent ground lateral displacement of ug(t) as shown in
Figure 6.4, which gives a total lateral displacement at a height x of

ut ( x, t ) = u( x, t ) + ug (t ) (6.43)

6.2.2  Generalized Participation Factor, Frequency, and Natural Period

Rearranging Equation 6.43 into the form of Equation 3.1, we obtain the same results as Equation
6.30 in Section 6.1.2, except that the generalized properties are now computed using the integral
relationships. That is,

z + 2ζωn z + ωn2 z = −Γ ug (t )



where:
k
ωn = is the generalized natural frequency
m
L
Γ = is the generalized participation factor
m
ζ is an estimate of the damping ratio
m
generalized natural period is given as before, Tn = 2π
k

6.2.3 Deflections, Base Shear, and Moments Using

Participation Factors and Response Spectra
As discussed in Section 6.1.3, the maximum response of a generalized SDOF system due to a
dynamic excitation (or force) can be determined using the generalized period and an appropriate
164 Introduction to Earthquake Engineering

design response spectrum. The maximum displacement at the top of the tower based on response
spectrum analysis is

z = Γ ⋅ D (6.44)

where:
D is the spectral displacement

The maximum dynamic displacement relative to the ground along the entire tower can be
obtained using this displacement

uo ( x ) = z ⋅ ψ( x ) = Γ ⋅ D ⋅ ψ( x ) (6.45)

This can also be written in terms of the spectral acceleration, A, as

A
uo ( x ) = Γ ⋅ 2 ⋅ ψ( x ) (6.46)
ωn

The acceleration is approximately proportional to the displacement; that is, A ≅ ωn2 D.

The equivalent static forces associated with these floor displacements can be obtained using the
product of the maximum displacement and associated flexural stiffness, or the product of the mass
and maximum total acceleration; that is,

fo ( x, t ) = m( x ) ⋅ ψ( x ) ⋅ Γ ⋅ A (6.47)

This can also be written in terms of the spectral displacement, D, as

fo ( x, t ) = ωn2 ⋅ m( x ) ⋅ ψ( x ) ⋅ Γ ⋅ D (6.48)

The internal shear force, Vmax, and internal story moment, Mmax, at an arbitrary height, x, can be
obtained by applying static equilibrium,

L L

Vmax ( x ) =
∫ fo (ξ )d ξ = Γ A
∫ m(ξ)ψ(ξ)dξ (6.49)
x x
L L

M max ( x ) =
∫ (ξ − x ) fo (ξ )d ξ = Γ A
∫ (ξ − x)m(ξ)ψ(ξ)dξ (6.50)
x x

where:
ξ ranges from x to L

Setting x equal to 0 in Equations 6.49 and 6.50 gives the base shear force, Vmax b, and overturning
moment, Mmax b,

Vmax b = Γ AL (6.51)

Generalized SDOF System Analysis 165

M max b = Γ A
∫ xm( x)ψ( x)dx (6.52)
0

EXAMPLE 6.2
The reinforced concrete chimney shown in Figure E6.3 is subjected to a ground acceleration due
to an earthquake characterized by an 84.1% design spectrum, scaled to 0.25g PGA. Determine
(i) period, (ii) peak displacement, (iii) maximum base shear, and (iv) maximum overturning moment
of the chimney. Assume diameter, d = 3.0 ft; thickness = 4 in; modulus of elasticity, Ec = 3600 ksi;
­concrete weight, γc = 150 pcf; damping, ζ = 5%; and the shape function shown in Figure E6.3.

Solution
1. Determine the properties of the chimney: The properties per unit length at an arbitrary
height, x, are as follows:
Cross-sectional area at x,

Area(x ) = π(Ro )2 − π(Ri )2 = π[(1.5 ft )2 − (1.167 ft )2 ] = 2.79 ft 2

Mass at x per unit foot of tower,

m(x ) = Area(x ) ⋅ γ c /g = 2.79 ft 2(150 pcf ) / 32.2 ft /s2 = 13.0(lbs ⋅ s2 /ft)) /ft

Moment of inertia at x,

I(x ) = π / 4 ⋅ (Ro )4 − π / 4 ⋅ (Ri )4 = π / 4[(1.5 ft )4 − (1.167 ft )4 ] = 2.52 ft 4

Flexural stiffness at x

EcI(x ) = 3600 ksi(1000 lb /kip)(12 in /ft )2( 2.52 ft 4 ) = 1.307 ×109 lbs ⋅ ft 2

2. Determine the generalized properties: The generalized mass, stiffness, and force of the
generalized SDOF system are calculated as follows:
Generalized mass,

L 30 ft
 x2 2
 =
m
∫ 2
m( x )[ψ( x )] dx = 13 lb s /ft 2 2
∫   2
 (30 ft )2  dx = 78.1lb s /ft
 
0 0

Generalized stiffness,

d 2ψ( x ) 2
ψ ′′( x ) = =
dx 2 (30 ft )2

L 30 ft 2
 2 
k =
∫ EI( x )[ψ ′′( x )] dx = 1.307 ×10 lbs ⋅ ft
2 9 2
∫ 



5
 (30 ft )2  dx = 1.94 ×10 lb /ft
0 0
166 Introduction to Earthquake Engineering

The generalized force is obtained using

L 30 ft
13lb s2 /ft 2
L =
∫ m( x )ψ( x )dx =
(30 ft )2 ∫ x dx = 130 lb s /ft
2 2

0 0

3. Determine the natural period, frequency, and participation factor of the generalized
SDOF system.
Natural period,

m 78.1lb − s2 /ft
Tn = 2π = 2π = 0.126 s
k 1.94 ×105 lb /ft

Natural frequency,

k 1.94 ×105 lb /ft

ωn = = = 49.8 rad /s
m 78.1lb − s2 /ft

Participation factor,

L 130 lb − s2 /ft
Γ = = = 1.667
m 78.1lb − s2 /ft

4. Determine maximum displacement at the top of the tower: This displacement is z(t) or
D due to 84.1% design spectrum, scaled to 0.25g PGA; first, enter the response spec-
trum in Figure 5.11 with Tn = 0.126 s, which is greater than 1/8 s, so the acceleration is
constant,

A = η ⋅ 2.71g = 0.25( 2.71)(386.4 in/s2 ) = 261.8 in/s2 = 21.8 ft /s2

where:
η is the PGA scale for a given earthquake; 0.25 in this case

x2
ψ(x) =
L2
30 ft

FIGURE E6.3  Schematic of a reinforced concrete chimney.

Generalized SDOF System Analysis 167

The maximum spectral displacement is

A 261.8 in/ s2
D= 2
= = 0.106 in
ωn ( 49.8 rad /s)2

We can now determine the maximum displacement at the top of the tower,

uo = Γ ⋅ D = (1.667)(0.106 in) = 0.176 in

5. Determine base shear and overturning moment: Determine the base shear force using
Equation 6.51,

Vmax b = Γ AL = 1.667( 261.8 in /s2 )(130 lb s2 /ft )(ft /12 in)(kip /1000 lb) = 4.73 kips

Determine the overturning moment using Equation 6.52

L 30 ft
13 lb s2 /ft 2
Mmax b
∫
= Γ A xm( x )ψ( x )dx = 1.667(21.8 ft /s )2
(30 ft )2 ∫ x dx = 106 kip ⋅ ft
3

0 0

PROBLEMS
6.1 Solve Example 6.1 with a shape function given for a low-rise building.
6.2 Solve Example 6.1 with a shape function of ψ(x) = (x/H)2, where H is the building
height.
6.3 Use the generalized SDOF analysis for the following building frame to determine the
story displacements, story forces, base shear, and overturning moment due to an earth-
quake characterized by an 84.1% design spectrum, scaled to 0.25g PGA. Let us assume
damping of 5% and a shape function of ψ(x) = sin (πx/2H), where H is the building
height.
w2 = 150 kips

k2 = 20 kips/in w1 = 150 kips 15 ft

k1 = 40 kips/in 15 ft

6.4 Use the generalized SDOF analysis for the following building frame to determine the
story displacements, story forces, base shear, and overturning moment due to an earth-
quake characterized by an 84.1% design spectrum, scaled to 0.25g PGA. Let us assume
damping of 5% and a shape function given for a low-rise building.

w2 = 20 kips

k2 = 10 kips/in 12 ft
w1 = 50 kips

k1 = 20 kips/in 14 ft

20 ft 20 ft
168 Introduction to Earthquake Engineering

6.5 Given a five-story building frame subjected to a ground acceleration due to a 0.25g PGA
earthquake characterized by the 84.1% design spectrum, use generalized SDOF analy-
sis to determine (i) peak displacements, (ii) maximum base shear, and (iii) m­ aximum
floor overturning moments. k = 326.3 kips/in, w5 = 50 kips, and the weights on the
other four floors equal 100 kips. Let us assume rigid beams, damping of 5%, and a shape
­function given for a mid-rise building.
6.6 Solve Problem 6.5 with a shape function given for a high-rise building.
6.7 Solve Example 6.2 with a shape function of ψ(x) = 1.5 (x/L)2 – 0.5 (x/L)3.
6.8 Solve Example 6.2 with a shape function of ψ(x) = 1 – cos(πx/2L).

REFERENCES
Chopra, A. K., Dynamics of Structures: Theory and Applications to Earthquake Engineering, 4th edition,
Prentice-Hall, Upper Saddle River, NJ, 2012.
Naeim, F., The Seismic Design Handbook, 2nd edition, Springer, Dordrecht, The Netherlands, 2001.
 7 Multi-Degree-of-Freedom
System Analysis
After reading this chapter, you will be able to:

1. Setup the equation of motion for an multi-degree of freedom (MDOF) system

2. Determine frequencies and periods for an MDOF system
3. Determine participation factors for an MDOF system
4. Use response spectra and participation factors to calculate the maximum shear and moment
response of an MDOF system

In Chapter 6, we modeled multi-degree-of-freedom (MDOF) systems as generalized single-

degree-of-freedom (SDOF) systems to obtain the dynamic response caused by a ground excita-
tion. This analysis procedure is convenient for structures that can assume a unique shape during
vibration. However, practical structures can assume several different shapes (mode shapes) during
their vibration (see Figure 7.1), with each mode contributing to the total response of the structure.
Also, each mode shape shown in Figure 7.1 has a unique associated frequency, with the first (or
fundamental) mode of vibration defined as having the lowest frequency, or the longest period. The
other modes have shorter periods and are known as higher modes, or harmonics. Depending on the
frequency content of the time-dependent excitation, different mode shapes may contribute differ-
ently to the total response of the system; typically, the first or fundamental mode has the greatest
influence on the response, but because of resonance other modes can dominate the response at
higher harmonics.
Since each individual mode of vibration has its own period, they each can be represented by a
generalized SDOF system of the same period (following the same process as in Chapter 6). Also,
each mode shape remains constant regardless of the amplitude of the displacement. Thus, each
mode shape can be assigned a reference unit amplitude value at a specified location, say the top of
the structure as in Chapter 6. The actual amplitude can then be obtained from the dynamic analysis
of the generalized SDOF. Therefore, once we obtain equations or vectors representing the mode
shapes and associated periods, the maximum response for the separate modes can be obtained
by analyzing each mode as a distinct generalized SDOF system. However, it is not likely that the
maximum response values will all occur simultaneously at the same level; therefore, the maximum
responses from all generalized SDOF cases are combined statistically in order to obtain the total
system response. A number of procedures have been developed to combine the maximum results;
two methods will be discussed later in the chapter, square root of the sum of the squares (SRSS)
or the complete quadratic combination (CQC) methods, both of which are mentioned in the code-
based dynamic seismic analysis procedures.
All modes for shear buildings such as the one shown in Figure 7.2 can be modeled simultane-
ously using matrix methods; this is known as modal analysis and will be the subject of the remain-
der of the chapter. For example, the frame shown in Figure 7.2 has four degrees of freedom and thus
four mode shapes. Each mode shape has a differential equation associated with it. For this discrete
case, a system of four dependent differential equations is formulated. The process entails decou-
pling the system of differential equations into four independent equations, which can then be used to
find maximum response values using a response spectrum. The key concept to obtaining the mode
shapes and the associated periods is a solution to an eigenvalue problem; the eigenvalues are related
to the periods and the corresponding eigenvectors give the mode shapes.

169
170 Introduction to Earthquake Engineering

P(t)

And so on
h

FIGURE 7.1  Structural mode shapes.

The eigen matrix operation can be viewed as a linear transformation that maps vectors into mul-
tiples of themselves. That is,

Ax = λx or [ A]{ x} = λ{ x} (7.1)

where λ is the scalar proportionality constant.

A must be a square matrix, with as many values of λ as there are elements in each row. This
equation can also be rewritten as

( A − λI)x = 0 or ([ A] − λ[ I ]){ x} = {0} (7.2)

where:
I is the identify matrix
The nonzero solutions to this equation require that λ be chosen so that

det( A − λI) = 0 or det([ A] − λ[ I ]) = 0 (7.3)

The values of λ that satisfy this relationship correspond to eigenvalues of matrix A; each λ yields
an eigenvector {x} solution of Equation 7.1 to within a multiplicative constant. That is, scaling the
eigenvector by any constant factor will still satisfy Equation 7.1. Consequently, each eigenvector is
normalized with respect to one of its elements, usually the first element.

m4
u4
4
k4
m3
3 u3
k3
m2 u2
2
k2
m1
1 u1

k1

FIGURE 7.2  Idealized four-story shear building system and mode shapes.
Multi-Degree-of-Freedom System Analysis 171

These solutions are practically intractable via hand calculations for systems with more than two
degrees of freedom, which is why solutions are regularly found via computer using one of the widely
available numerical packages, such as MATLAB®. With this program, we can use the eig operator
to determine the eigenvectors denoted as phi and the eigenvalues denoted as lambda as follows,

[phi, lambda]= eig(A) (7.4)

In the sections that follow, we present the application of the eigenvalue problem to the solution of
structural dynamics problems. This solution will then be extended to encompass the calculation of
the maximum shear and moment response of MDOF systems.

7.1  EQUATIONS OF MOTION FOR AN MDOF SYSTEM

To formulate the equation of motion for an MDOF system, we first use the free vibration response of
the two-degree-of-freedom shear building shown in Figure 7.3, and then expand it to an N-degree-
of-freedom system. The solution to the resulting equation of motion yields the natural periods (or
frequencies) and the corresponding vibration mode shapes. Recall from Section 3.1.1 that free vibra-
tion is caused by initial conditions (displacement and velocity), not a time-dependent applied force.
To characterize the free vibration motion of an MDOF system, we need to establish a system of
equations of motion for the model in terms of relative displacements (for stiffness forces) and total
displacements (for inertial forces). These equations can be determined by applying equilibrium to
the free-body diagrams of the masses in Figure 7.3 using D’Alembert’s principle. Notice that we
have not included vertical forces or internal moments in the column stems because we only need to
apply horizontal equilibrium to obtain the equations of motion.
Applying horizontal equilibrium to the free-body diagrams of the two masses shown in Figure
7.3 yields two equations of motion.
Equilibrium of mass 1:

+ → ∑ Fx = 0; − m1u1 − k1u1 + k2 (u2 − u1 ) = 0 ⇒ m1u1 + (k1 + k2 )u1 − k2u2 = 0 (7.5)

u2

m2 u 2 – u1 m2u2

k2 (u2 – u1)
k2 (u2 – u1)
k2
k2
u1
k2 (u2 – u1)
m1 m1u1
k2 (u2 – u1)
k1u1
k1u1

k1
k1

k1u1
u1

FIGURE 7.3  Two-degree-of-freedom system and FBD of the two masses.

172 Introduction to Earthquake Engineering

where m1 is the mass of the first floor, which can be determined following the procedure discussed
in Section 3.1.2, and k1 and k2 are the story stiffnesses of the first and second levels, respectively,
which can be determined by adding the lateral stiffnesses of all columns in the story. Each story can
be treated as a portal frame as discussed in Chapter 3; thus, for a story of height h, column flexural
stiffness EI, and assuming rigid floor diaphragms (rigid frame beams), the lateral stiffness of a col-
umn is 12EI/h3 and the stiffness of story j is kj = ∑columns 12EI/h3.
Equilibrium of mass 2:

+ → ∑ Fx = 0; − m2u2 − k2 (u2 − u1 ) = 0 ⇒ m2u2 − k2u1 + k2u2 = 0 (7.6)

where m2 is the mass of the second floor.

These two dependent equations of motion (Equations 7.5 and 7.6) are coupled second-order,
linear, and homogeneous differential equations with constant coefficients, and can be rewritten in
matrix form as

 m1 0  u1   k1 + k2 −k2  u1  0

  +    =   (7.7)
0 m2  u2   −k2 k2  u2  0


In general,

[ m]{u} + [ k ]{u} = {0} (7.8)

where [m] is the mass matrix, which is a diagonal matrix (nonzero elements only along the forward
diagonal), and [k] is the stiffness matrix; this can alternatively be obtained by directly finding each
coefficient kij, which represents a force needed at level i to hold a unit displacement at level j, holding
zero displacements at all other levels.
Notice that an Nth-degree-of-freedom system results in N-dependent equations, or N × N mass
and stiffness matrices.

7.1.1  Periods and Mode Shapes for an MDOF System

As discussed in the last section, and SDOF analysis in Chapter 3, the free-vibration response
yields the natural periods (or frequencies). Since this is similar to the SDOF case, we can assume
that the resulting motion is described by a simple harmonic equation with the following form for
displacement:

{u} = {a} sin(ω n t − α) (7.9)

where:
ωn = k m is the natural circular frequency with units of radians per second (rad/s), which can
be related to the period as Tn = 2π /ωn

Taking the first and second derivatives of Equation 7.9 with respect to time,

{u} = {a}ωn cos(ωn t − α) (7.10)

{u} = −{a}ωn2 sin(ωn t − α) = −ωn2 {u} (7.11)

Multi-Degree-of-Freedom System Analysis 173

Substituting these equations into Equation 7.8 yields

[m](−ωn2 {u}) + [k ]{u} = {0} (7.12)

Factoring out the displacement vector {u}, we get

[k ] − ωn2 [m] {u} = {0} (7.13)

 

This equation is similar to Equation 7.2 and has a nontrivial solution (one that gives nonzero
values for {u}) if Equation 7.3 is satisfied. That is,

det ([k ] − ωn2 [m]) = 0 (7.14)

This is the characteristic (or eigen) equation of the system, also known as the frequency equa-
tion because expanding the determinate results in an Nth degree polynomial in terms of ωn2 . That
is, Equation 7.14, in general, should be satisfied for N values of ωn2 . We can then solve the homoge-
neous system of equations (Equation 7.12) for a1, a2, … , aN for each value of ωn2 , totaling N modes
of vibration.
Solving this eigenvalue problem is convenient using a computer program such as MATLAB; for
this case, the eig operator includes the stiffness, [k] and the mass, [m] matrices to determine eigen-
vectors, denoted by phi and eigenvalues denoted by lambda, as

[phi,lambda]= eig(k,m) (7.15)

For illustration purposes, let us conduct the analysis of the two-degree-of-freedom case in Figure
7.3 using hand calculations. Recall that the assumed displacement response vector is given as

u1   a1 

  =   sin(ω n t − α) (7.16)
u2  a2 

Taking two derivatives of Equation 7.16 with respect to time and substituting into Equation 7.7
gives the following equation:

 m1 0   a1   k1 + k2 −k2   a1  0

−ωn2   +    =   (7.17)
0 m2  a2   −k2 k2  a2  0


which can be rewritten as

 k1 + k2 − m1ωn2 −k2   a1  0

  =  (7.18)
 −k2 k2 − m2ωn2  a2  0

As discussed above, the trivial solution a1 = a2 = 0 does not provide any information, and the
nontrivial solution requires that the determinate of the matrix be zero; that is,

 k + k2 − m1ωn2 −k2 
det  1 =0 (7.19)
 −k2 k2 − m2ωn2 

174 Introduction to Earthquake Engineering

Expanding this determinate results in a second-order polynomial in ωn2 , the frequency equation,

m1m2ωn4 − ((k1 + k2 )m2 + m1k2 )ω 2n + k1k2 = 0 (7.20)

We simplify the analysis further by assuming k1 = k2 = k and m1 = m2 = m, which leads to the

following polynomial:

m 2ωn4 − 3 kmωn2 + k 2 = 0 (7.21)

We use the quadratic formula, x = (−b ± b2 − 4ac ) / 2a , to find the roots of Equation 7.21,
which are the eigen (characteristic) values:

3km ± (3km)2 − 4m 2 k 2 3± 5 k (7.22)

ωn2 = =
2m 2 2 m

or

3− 5 k 3+ 5 k
ωn21 = and ωn22 = (7.23)
2 m 2 m

The corresponding frequencies are

k k
ωn1 = 0.618 and ωn 2 = 1.618 (7.24)
m m

Frequencies are usually listed in ascending order; that is, from the smallest to largest.
Also, each of these eigenvalues is associated with a distinct eigenvector (or mode shape). The
actual values of the eigenvector elements in Equation 7.18, a1 and a2 for this case, are indeterminate.
However, each eigenvector can be obtained by setting the values of the top element equal to one,
2
and solving for the values of the other eigenvector elements relative to top value. Substituting ωn1
into Equation 7.18 yields

   
 2k − m  3 − 5 k  −k 
  2 m  
     a1  0
   =   (7.25)
  3 − 5 k  a2  0
 
 −k k − m  

  2 m 


Simplifying the elements of this matrix and setting a2 = 1, we get

1.618 −1   a1  0 1.618 −1  a1 0

k    =   or    =   (7.26)
 −1 0.618 a2  0  −1 0.618  1  0
 

Solving for a1, the resulting eigenvector (mode shape) is

φ11   a1  0.618

 =  =  (7.27)
φ21 a2 1  1 

Multi-Degree-of-Freedom System Analysis 175

1 1

m2

k2
a1 0.618 a1 –1.618
φ1 = = φ2 = =
0.618 a2 1
1 a2 2
1
–1.618

m1
k k
ω 1 = 0.618 m ω 2 = 1.618 m
k1

Mode 1, or fundamental mode Mode 2,

associated with the lowest frequency harmonic or higher harmonic

FIGURE 7.4  Mode shapes and associated frequencies.

Similarly, we can obtain the second eigenvector (mode shape) as

φ12   a1  −1.618

 =  =  (7.28)
φ22  a2 2  1 

These eigenvectors are referred to as normal modes because they are normalized so that the
displacements of the mass at the top of the structure are equal to one. The resulting eigenvectors
describe the mode shapes and are defined as {φ}j (this was the process followed in Chapter 6 to
obtain a shape vector that describes the deflected shape of a vibrating shear building). A summary
of these results, including the normal modes and frequencies, is shown in Figure 7.4.
Regardless of how the eigenvectors are normalized, the resulting mode shapes are orthogonal;
that is, for two distinct mode shapes of the structure, r ≠ s,

{φr }T {φs} = 0 (7.29)

which leads to the following relationships:

{φr }T  k  {φs} = 0 and {φr }T [ m]{φs} = 0 (7.30)

where the superscript T denotes the transpose matrix operation.

We can now obtain the total solution of the system by combining the two modes as

u1 = φ11q1 + φ12q2
(7.31)
u2 = φ21q1 + φ22q2

where we now assume a new set of harmonic displacement functions with constants Aj and Bj,

q1 = A1 sin ω1t + B1 cos ω1t

(7.32)
q2 = A2 sin ω2t + B2 cos ω2t
176 Introduction to Earthquake Engineering

Rewriting Equations 7.31 in the matrix form

u1  φ11 φ12   q1 

  =     orr {u} = [φ]{q} (7.33)
u2  φ21 φ22  q2 

The matrix [φ] is a coordinate transformation matrix (also known as the modal matrix) that maps
the original coordinates {u} to the principal coordinates {q}. The inverse transformation is

{q} = [φ]−1{u} (7.34)

The orthogonality condition given by Equation 7.29 implies that [φ]−1 = [φ]T. Thus, we can then
rewrite Equation 7.34 as

{q} = [φ]T {u} (7.35)

Substituting the transformation relationship, Equation 7.33, into the equations of motion,
Equation 7.8, and premultiplying by [φ]T yields

[φ]T [ m][φ]{q} + [φ]T [ k ][φ]{q} = {0} or [ M ]{q} + [ K ]{q} = {0} (7.36)

The matrices [M] and [K] are diagonal because of the orthogonality condition given by
Equation 7.30 (i.e., all off diagonal elements are zero). These are known as modal mass and modal
stiffness matrices, respectively. The diagonal elements of these matrices are given by

N N

Mi = {φi}T [ m]{φi} = ∑ j=1

m jφ 2ji and K i = {φi}T [ k ]{φi} = ∑k φ
j=1
j
2
ji (7.37)

Notice that Equations 7.36 are now uncoupled, or independent. That is,

Mi qi + K i qi = 0 (7.38)

If each equation is divided by the associated modal mass, the equations can be rewritten as

qi + ωni2 qi = 0 (7.39)

So, each equation has the same form as the equation of motion for the SDOF system we described
in detail in Chapter 3. This implies that we can use the solutions to the different loadings or forcing
functions examined previously, including response spectrum analysis.
To illustrate the process of decoupling the equations of motion, let us revisit the two-degrees-of-
freedom case introduced earlier in this chapter. Using Equation 7.36, we get the mass transforma-
tion coefficients M1 and M2:

 M1 0   0.618 1  m 0  0.618 −1.618  1.382m 0 

0 = = (7.40)
 M 2   −1.618 1  0 m   1 1   0 3.618m 

Multi-Degree-of-Freedom System Analysis 177

and the stiffness coefficients, K1 and K2,

 K1 0   0.618 1  2k −k  0.618 −1.618  0.528k 0 

 =   =  (7.41)
0 K 2  −1.618 1 −k k   1 1   0 9.472k 


This also proves the orthogonality condition ({φ1}T [m]{φ2} = {φ2}T [m]{φ1} = {0} and {φ1}T [k]
{φ2} = {φ2}T [k]{φ1} = {0}) since the off-diagonal elements of the two matrices are zero. That is,

m 0  −1.618 −1.618
{φ1}T [m]{φ 2 } = [0.618 1]    = [0.618m m]  =0 (7.42)
0 m   1   1 
  

The equations of motion are

1.382 ⋅ m ⋅ q1 + 0.528 ⋅ k ⋅ q1 = 0

(7.43)
3.618 ⋅ m ⋅ q2 + 9.472 ⋅ k ⋅ q2 = 0

EXAMPLE 7.1
For the two-story building frame shown in Figure E7.1, determine (1) mass and stiffness matrices,
(2) periods and modal matrix (normalized mode shapes), and (3) the stiffness and modal matrices.
Assume beams are rigid. k = 20 kips/in and m = 0.4  kips · s2/in.

Solution
1. Determine the mass and stiffness matrices. These can be obtained following the process
shown in Figure 7.3 by setting m1 = m2 = m, k1 = 2k, and k 2 = k, and applying equilib-
rium to the free-body diagrams of each mass, resulting in equations similar to Equations
7.5 and 7.6. The two equations can be combined into a matrix equation as shown in
Equation 7.7:

m 0  u1   3k −k  u1  0

  +    =  
0 m u2  −k k  u2  0

The mass and stiffness matrices are

m 0   0 .4 0 
[ m] =  =  kip ⋅ s2 /in
0 m  0 0.4

k m 15 ft

2k 15 ft

FIGURE E7.1  Schematic of two-story building frame.

178 Introduction to Earthquake Engineering

 3k −k   60 −20
[ k ] =  =  kips/in
−k k  −20 20 

2. Determine the periods and modal matrix. The natural frequencies and mode shapes
can be determined by solving for the eigenvalues and eigenvectors using a standard
eigen solution; natural frequencies can be obtained using the eigenvalues and the mode
shapes with the eigenvectors. Also, the periods are obtained using the natural frequen-
cies. The frequency equation (Equation 7.20) resulting from the expansion of the eigen
equation is

m2ωn4 − 4 kmωn 2 + 2k 2 = 0

Using the quadratic formula, we can find the roots of this equation, which are the
eigenvalues:

w n2 =
4km ± ( 4km)2 − 8m2k 2 4 ± 2 2 k
2m2
=
2 m
= 2± 2
20 kips /in
(
0.4 kip ⋅ s2 / in
)

The frequencies are (from smallest to largest)

ωn1 = 5.41 rad/s and ωn 2 = 13.1 rad/s

The natural periods, obtained using Tnj = 2π /ωnj , are

Tn1 = 1.16 s and Tn 2 = 0.48 s

2
To obtain mode shapes, we substitute each eigenvalue (ωn1 first) into Equation 7.18
and set a2 = 1,



(
3k − 2 − 2 k
) −k 
  a1  0
  =  
 2.414

−1  a1 0
  =  
or
  −1 0.414  1 0
 −k k − 2 − 2 k  a2  0
( ) 


Solving for a1, the resulting eigenvector (mode shape) is

φ11  a1  0.414

 =  = 
φ21 a2   1 
1

Similarly, we can obtain the second eigenvector (mode shape) as

φ12   a1  −2.414

  =   =  
φ22  a2 2  1 

The modal matrix is

0.414 −2.414
[φ] =  
 1 1 

Multi-Degree-of-Freedom System Analysis 179

3. Determine the modal mass and stiffness matrices. Use Equations 7.37 to determine the
elements of the modal mass and stiffness matrices. Or, we can use the matrix operations
in Equation 7.36:
  Modal mass matrix:

 0.414 1 0.4 0  0.414 −2.414

[M] = [φ]T [m][φ] =    
−2.414
 1  0 0.4  1 1 
0.469 0 
=  kip ⋅ s2 /in
 0 2.73


Modal stiffness matrix:

 0.414 1  60 −20 0.414 4 −2.414

[K ] = [φ]T [k ][φ] =    
−2.414 1 −20 20   1 1 
 
13.7 0 
=  kips/in
 0 466.3


4. Alternatively, we can write a MATLAB script to perform all the operations; the most
important of the operations being the eig operator in Equation 7.15:

clear all % clears any previously defined variables

clc % clears the screen
m = [0.4 0; 0 0.4]; % mass matrix
k = [60 −20; −20 20]; % stiffness matrix
[phi,lam]=eig(k,m); % compute eigenvalues and eigenvectors
omegas=sqrt(lam) % determine and show frequencies from eigenvalues
periods = 2*pi*diag(inv(omegas)) % determine the periods from omegas
% mode shapes by normalizing eigenvectors to get top displ equal to 1
[N_rows, N_cols] = size(phi); % finds N, the size of the matrices
for i = 1:N_cols; % loops over the N modes to normalize them
norm_phi(:,i) = phi(:,i)./phi(N_rows,i);
end
norm_phi % display normalized modal matrix
% determine the modal mass and stiffness matrices
M = norm_phi’*m*norm_phi
K = norm_phi’*k*norm_phi

The results of this script are the same as those of parts 1–3:

omegas =
5.4120 0
0 13.0656
periods =
1.1610
0.4809
norm_phi =
0.4142 −2.4142
1.0000 1.0000
M =
0.4686 0.0000
0.0000 2.7314
K =
13.7258 0.0000
0 466.2742
180 Introduction to Earthquake Engineering

7.1.2  Participation Factors for an MDOF System

As in Chapter 6, we can account for ground excitation loading in the formulation of the equations of
motion for MDOF systems by determining the total displacement response of the system at level j,
which is a combination of the relative displacement, uj(t) and the ground level lateral displacement,
ug(t), caused by seismic ground motion:

utj (t ) = u j (t ) + ug (t ) (7.44)

For MDOF systems, we can substitute the second derivative of Equation 7.44 into the first term
of Equation 7.8 to obtain the equations of motion including ground excitation,

[m]{u} + [k ]{u} = −[m]{1}ug (7.45)

where {1} is a column vector of ones.

We can use Equation 7.34, the coordinate transformation matrix [φ], to transform the original
coordinates {u} into the principal coordinates {q}. If we substitute the transformation relationship
into the equations of motion, and premultiply by [φ]T, we get

[φ]T [m][φ]{q} + [φ]T [k ][φ]{q} = −[φ]T [m]{1}ug (7.46)

Substituting modal mass and stiffness relations, Equations 7.37, into Equation 7.46, we get

[ M ]{q} + [ K ]{q} = −[φ]T [m]{1}ug (7.47)

Alternatively, these equations of motion can be written as N uncoupled equations as

qj + ωnj2 q j = −Γ jug (7.48)

which after including the effects of damping can be written as

qj + 2ζωnj q j + ωn2q j = −Γ jug (7.49)

where the damping is assumed to be constant for all mode shapes. For elastic analysis of MDOF
systems, where damping effects are significant, there are a number of approaches used to establish
the damping matrix, including a relationship based on the stiffness and mass matrices. This can be
found in conventional structural dynamics texts such as Chopra (2012) and Villaverde (2009). Also,
we can now define the participation factor for each mode shape (the modal participation factor) as

Lj ∑N m φ [φ]T [ m]{1} (7.50)

Γj = = iN=1 i ij2 or {Γ} =
Mj ∑ i=1 miφij [φ]T [ m][φ]

where Lj is the generalized force associated with the jth mode shape.
Multi-Degree-of-Freedom System Analysis 181

EXAMPLE 7.2
Consider the two-story building frame shown in Figure E7.1 and determine the participation factors.

Solution
1. Determine the participation factors. First, apply the summation form of Equation 7.50:

∑
2

L miφi1 m1φ11 + m2φ21 0.4(0.414) + 0.4(1) 0.566

Γ1 = 1 = i =1
= = = = 1.207
∑
2
M1 mφ 2
2
m1φ11 2
+ m2φ21 0.4(0.414)2 + 0.4(1)2 0.469
i i1
i =1

∑
2

L miφi1 m1φ11 + m2φ21 0.4(−2.414) + 0.4(1) −0.566

Γ2 = 2 = i =1
= = = = −0.207
∑
2
M2 mφ 2
2
m1φ11 2
+ m2φ21 0.4(−2.414)2 + 0.4(1)2 2.731
i i1
i =1

2. Determine the participation factors by expanding the MATLAB script presented in

Example 7.1 (we present the entire script for completeness):
clear all % clears any previously defined variables
clc % clears the screen
m = [0.4 0; 0 0.4]; % mass matrix
k = [60 -20; -20 20]; % stiffness matrix
[phi,lam]=eig(k,m); % compute eigenvalues and eigenvectors
omegas=sqrt(lam); % determine and show frequencies from eigenvalues
periods = 2*pi*diag(inv(omegas)); % determine the periods from omegas
% mode shapes by normalizing eigenvectors to get top displ equal to 1
[N_rows, N_cols] = size(phi); % finds N, the size of the matrices
for i = 1:N_cols; % loops over the N modes to normalize them
norm_phi(:,i) = phi(:,i)./phi(N_rows,i);
end
% determine the participation factors
LT = sum(norm_phi’*m,2) % sum(A,2) returns a vector of sums of each row
MT= sum(norm_phi’*m*norm_phi,2)
par_fac = LT./MT % ./ operation divides each element of two vectors

The results of this script are the same as those of part (i):
LT =
0.5657
-0.5657
MT =
0.4686
2.7314
par_fac =
1.2071
-0.2071

Also, in MDOF systems, a modal participation factor can be viewed as providing a measure
of the degree to which the jth mode contributes (or participates) to the total dynamic response.
Equation 7.50 is essentially the same as Equation 6.32 for the generalized SDOF case, except the
use of N eigenvectors rather than a single assumed mode shape. Therefore, we can use the formula-
tions presented in Chapter 6 to obtain the response of a system due to each eigenvector. That is, the
original system of N simultaneous differential equations of motion is now transformed into a system
182 Introduction to Earthquake Engineering

of N-independent differential equations, each of which can be solved using the SDOF procedures
described in Chapters 3–5. Thus, for linear elastic systems the total displacement can be obtained
using superposition of the modal contributions:

∑
N
u j (t ) = φ ji ⋅ qi (t ) or {u(t )} = {φ}1 q1(t ) + {φ}2 q2 (t ) +  + {φ} N qN (t ) (7.51)
i=1

where i denotes the level, from 1 to N, and j the range of eigenvectors, from 1 to N. With these
displacements, we can compute equivalent lateral forces, which can then be used to conduct a
static structural analysis to determine the internal element forces and stresses. This superposition
approach produces the entire time history of the structural response (displacements, forces, etc.).
However, for design of various elements of the structure, we only need the absolute maximum val-
ues of the response. This is the basis for the code-based seismic response history modal analysis
procedure introduced in Section 8.7. A more efficient approach is to use the response spectrum
analysis described in the next section.

7.1.3 Deflections, Base Shear, and Moments Using

Participation Factors and Response Spectra
The approach covered in Section 6.1.3 for a generalized SDOF (a single mode of vibration) can be
used to solve MDOF systems by treating each mode shape (eigenvector) as an independent general-
ized SDOF system. After computing the periods for each mode of vibration, Tnj, using the eigen-
values, we can determine the maximum dynamic response caused by any dynamic excitation (or
force) using the SDOF system analyses discussed in Chapters 3 and 4. Also, we can use the response
spectrum analysis method discussed in Chapter 5 to determine the peak response to an earthquake
characterized by a design response spectrum with spectral displacement, velocity, and accelera-
tion (Dj, Vj, and Aj, respectively). The values of Dj, Vj, and Aj are obtained as the ordinates of the
response spectrum corresponding to the periods Tnj associated with the jth mode of vibration, and
an appropriate damping ratio. Thus, the solution to the equation of motion associated with the jth
mode of vibration based on response spectrum analysis using the spectral displacement, Dj, gives
the maximum displacements at the top of the building (maximum principle coordinates) as

q jo = Γ j ⋅ D j (7.52)

These maximum displacements can be distributed to the other floors using the normalized eigen-
vectors, or mode shapes. That is, the maximum displacement value of the ith floor, induced, when
only the jth mode shape is excited, is

uijo = φij ⋅ q jo = φij ⋅ Γ j ⋅ D j (7.53)

The displacement subscript ij can be interpreted as the contribution to the total response at level
i due to mode of vibration j. In matrix form, each matrix column represents the floor displacements
associated with each mode shape, that is,

[uo ] = [φ][qo ] = [φ][Γ][ D] (7.54)

where:
[qo] is a diagonal matrix of maximum principle coordinates
[Γ] is a diagonal matrix of participation factors
[D] is a diagonal matrix of spectral displacements
Multi-Degree-of-Freedom System Analysis 183

The displacements can also be written in terms of the spectral accelerations Aj as

Aj
uijo = φij ⋅ q jo = φij ⋅ Γ j ⋅ (7.55)
ωnj2

The maximum acceleration value of the ith floor, induced when only the jth mode shape is
excited, is

uijo = φij Γ j A j = φij Γij ωnj2 D j (7.56)

Similar to displacement, this can be represented in matrix form as

[uo ] = [φ][Γ][ A] = [φ][Γ][ωn2 ][ D] (7.57)

where:
[ωn2 ] is a diagonal matrix of the square modal frequencies (also known as the spectral matrix)
[A] is a diagonal matrix of spectral accelerations

The equivalent static forces associated with these floor accelerations can be obtained using
Newton’s second law; the product of the associated story mass and maximum story acceleration;
that is, the inertial force at level i due to vibration mode j (using Equations 7.56) is

fij = miφij Γ j ωnj2 D j or fij = miφij Γ j A j (7.58)

Similar to accelerations, this can be represented in matrix form as

[ f ] = [m][φ][Γ]  ωn2  [ D] = [m][φ][Γ][ A] (7.59)

 

From static analysis, we can determine the force at the base of the building, which is known as
the base shear, Vjb, for each mode j,

N
V jb = f1 j + f2 j + f3 j + = Γ j ωnj2 D j ∑ miφij = Γ j ωnj 2 D j L j or V jb = Γ j A j L j (7.60)
i =1

The base shear in matrix form is

{Vb } = [ f ]T {1} = [Γ][ A][φ]T [m]{1} = [Γ][ D]  ωn2  [φ]T [m]{1} (7.61)
 

Figure 7.5 shows the equivalent static forces along with the base shear and overturning moment
for the jth mode shape. To obtain the overturning moment, we can apply static equilibrium to the
free-body diagram.
184 Introduction to Earthquake Engineering

mN qj
fNjo uNjo
. kN
.
.

fijo mi uijo

. ki
.
.
.
. hi
.
m1 u1jo
f1jo
k1

Mjb
Vjb

FIGURE 7.5  Maximum dynamic displacements and associated equivalent static forces.

The inertial forces (or lateral story forces) at level i due to vibration mode j can also be written
in terms of the base shear by substituting the second of Equations 7.60 as Γ j A j = V jb /L j into the
second of Equations 7.58:

V jb m jφijV jb (7.62)
fij = miφij Γ j A j = miφij or fij =
Lj ∑ Nj=1 m jφ ji

Also, writing the second of Equations 7.60 in the form of Newton’s second law (F = mA), we get

V jb = M ej A j (7.63)

where M ej is the effective mass given as

2
e
M = Γ jLj or M =eL2j
=
( ∑ iN=1 miφij ) (7.64)
j j
Mj ∑ iN=1 miφij2

Substituting into Equation 7.63 the definition of the mass in terms of the floor weights, mi = wi/g
(weight associated with the ith floor divided by the acceleration due to gravity),

V jb =
( ∑ iN=1 wiφij ) Aj A
= W je j
(7.65)
∑ iN=1 wiφij2 g g

where W je is the effective weight given as

W e
=
( ∑ iN=1 wiφij ) (7.66)
j N 2
∑ i=1 wφ
i ij

Multi-Degree-of-Freedom System Analysis 185

Combining the contribution of all the modes of vibration, we get the total weight of the
building as
N

Wtotal = ∑W
j =1
j
e
(7.67)

This implies that the effective weight of the jth mode shape is the fraction of the total weight that
participates in the jth mode of vibration. The code (ASCE-7) requires including enough modes of
vibration in the analysis to obtain an effective weight of at least 90% of the actual building weight,
as will be discussed in Section 8.6.
Displacement, acceleration, and lateral story forces resulting from different mode shapes must be
combined to acquire the total response, though a simple summation is not used since modal maxima
generally occur at different times during the response history. Furthermore, response spectra only
provide the values of the modal maxima and not the time at which each value occurs. Therefore,
appropriate combination rules have been developed to obtain the total response. These rules are
based on random vibration theory to estimate the average maximum response. The most popular
rules include the SRSS and CQC methods, the latter being more complicated, but yielding more
accurate results for a wider range of mode shapes. The following relationships for the SRSS and CQC
are given in terms of generic maximum modal values, Rj, which are intended to represent maximum
values due to vibration mode j for any response parameter, such as displacement, acceleration, etc.
The total response for the SRSS rule

(7.68)
Rmax ≈ ∑ Nj =1 R 2j

The total response for the CQC rule

(7.69)
Rmax ≈ ∑ iN=1 ∑ Nj =1 R jρij R j

where:

8ς 2 (1 + βij )βij1.5
ρij = (7.70)
(1 + βij2 )2 + 4βijς 2 (1 + βij )2

ρij  is the cross-modal coefficient that varies from 0 to 1 (1 for the case of i = j), and is generally
expressed in terms of the modal frequencies of two distinct mode shapes, βij = ωj/ωi and d­ amping
characteristics, which reduces to the damping ratio given by Equation 3.32 (repeated here for con-
venience, ζ = c/ccr) when the modal damping is constant for the entire modal spectrum. Also, when
the modal frequencies are well separated, this matrix tends to the identity matrix and the CQC rule
approaches the SRSS rule results.
To determine an upper bound to both of these rules, we can sum the maximum absolute values as
N

Rmax ≤ ∑R
j=1
j (7.71)

Also, as discussed in Section 3.3, with the maximum dynamic displacement or acceleration, we can
conduct a static structural analysis to determine element forces (bending moment, shear force, and axial
force), and stresses needed for design; no additional dynamic analysis is necessary. These maximum
element force or stress results from each mode shape can be combined using either SRSS or CQC.
186 Introduction to Earthquake Engineering

EXAMPLE 7.3
Consider the two-story building frame of Figure E7.1 and determine (1) effective masses, (2) peak
displacements, (3) maximum equivalent static floor forces, (4) maximum base shear, and (5) maxi-
mum overturning moment. Assume beams are rigid, damping of 5%, and that the frame is sub-
jected to a ground acceleration due to an earthquake characterized by the 84.1% design spectrum
scaled to 0.25g PGA, the solution is given as follows.

Solution
1. Determine the effective masses. Use Equation 7.64:

M1e =
L12
=
(∑i2=1 miφi1) = (m1φ11 + m2φ21)2 = (0.4(0.414) + 0.4(1))2 = 0.320 = 0.683
M1 ∑ i2=1 miφi21 2
m1φ11 2
+ m2φ21 0.4(0.414)2 + 0.4(1)2 0.469

2
L2
M = 2 =
e (∑i2=1 miφi1) = (m1φ11 + m2φ21)2 = (0.4(−2.414) + 0.4(1))2 = 0.320 = 0.117
2
M2 ∑ i2=1 miφi21 2
m1φ11 2
+ m2φ21 0.4(−2.414)2 + 0.4(1)2 2.731

The operation can be performed using MATLAB after determining participation fac-
tors, using the same factors, MT and LT; that is,
Eff_mass = LT.∧2./MT % .∧ operation squares each element of LT vector

2. Determine peak (maximum) floor displacements. Determine the maximum displacement

at the top of the building, q jo = Γ j ⋅ D j , where Dj is due to an 84.1% design spectrum,
scaled to 0.25g PGA; first, enter the response spectrum in Figure 5.11 with the following
two values for Tn:

Tn1 = 1.16 s, which is greater than 0.66 s, so the acceleration is given as,
A1 = η · 1.8g/Tn1 = 0.25(1.8 s)(386.4 in/s2)/1.16 s = 149.9 in/s2, and
Tn2 = 0.48 s, which is greater than 1/8, but less than 0.66 s, so the acceleration is
constant,
A 2 = η · 2.71g = 0.25(2.71)(386.4 in/s2) = 261.8 in/s2,

where:
η is the PGA scale factor for a given earthquake, in this case 0.25
The maximum spectral displacements are

A1 149.9 in/s2
D1 = = = 5.12 in
ωn21 (5.41rad/s)2

A1 261.8 in/s2
D2 = = = 1.52 in
ωn 2 (13.1rad/s)2
2

Now, determine the maximum floor displacements using Equation 7.53, uijo = φij ⋅ Γ j ⋅ D j

0.414 2.56
ui1o = φi1Γ1 ⋅ D1 =   (1.207)(5.12 in) =   in
 1  6.18

−2.414  0.76 

ui 2o = φi 2Γ 2 ⋅ D2 =   (−0.207)(1.52 in) =   in
 1  −0.32

Multi-Degree-of-Freedom System Analysis 187

Finally, combine these displacements using the SRSS rule (Equation 7.68) to obtain the
peak displacements at each level:

u1max = ( 2.56 in)2 + (0.76 in)2 = 2.67 in

u2max = (6.18 in)2 + (−0.32in)2 = 6.19 in

3. Determine maximum equivalent static floor forces. Determine the equivalent static floor
forces using Equation 7.58 or 7.59; let us use the matrix form:

[f ] = [m][φ][Γ][ A]
 0 .4 0  0.414 −2.414 1.2077 0  149.9 0 
=    
 0 0 . 4 1 1  0 − 0 . 207  0 261 .8
   
0.4(0.414)(1.207)(149.9) 0.4(−2.414)(−0.207)( 261.8)
= 
 0.4(1.207)(149.9) 0.4(−0.207)( 261.8) 
 
30.0 52.3 
=  kips
72.4 −21.7

Finally, combine these forces using the SRSS rule (Equation 7.68) to obtain the maxi-
mum floor forces at each level:

f1max = (30 kips)2 + (52.3 kips)2 = 60.3 kips

f2max = (72.4 kips)2 + (−21.7 kips)2 = 75.5 kips

4. Determine maximum base shear. Add the equivalent static floor forces for each mode
using Equation 7.60 or 7.61; let us use the matrix form again:

30.0 72.4  1 30.0 + 72.4 102.4

{Vb } = [f ]T {1} =    =  =  kips
 52.3 −21.7 1 52.3 − 21.7  30.6 

Finally, combine these using the SRSS rule (Equation 7.68) to obtain the maximum
base shear:

Vbmax = (102.4 kips)2 + (30.6 kips)2 = 106.9 kips

5. Determine maximum overturning moment. Use static equilibrium to obtain the contribu-
tion from each mode to the maximum overturning moment:

Mb1 = 30 kips(15 ft ) + 72.4 kips(30 ft ) = 2622 kip ⋅ ft

Mb2 = 52.3 kips(15 ft ) − 21.7 kips(30 ft) = 133 kip ⋅ ft

Finally, combine these using the SRSS rule (Equation 7.68) to obtain the maximum
overturning moment:

Mbmax = ( 2622 kip ⋅ ft)2 + (133 kip ⋅ ft )2 = 2625 kip ⋅ ft

188 Introduction to Earthquake Engineering

7.2  RESPONSE SPECTRUM ANALYSIS METHOD SUMMARY

The following is a brief step-by-step procedure to estimate the maximum response of a structure
subjected to a ground excitation due to an earthquake characterized by a response spectrum scaled
to η peak ground acceleration:

1. Determine the mass matrix, [m] from the given floor weights.
2. Determine the stiffness matrix, [k] from the column properties for a shear building.
3. With the stiffness and mass matrices, solve the associated eigenvalue problem for eigenval-
ues, ω2, which are used to determine the natural frequencies, ωn and periods, Tn.
4. After selecting an appropriate damping ratio, determine spectral accelerations (or dis-
placements) using the ordinates of the acceleration (or displacement) response spectrum of
the excitation for each of the natural periods.
5. With the eigenvalues, we can also determine the eigenvectors, which are normalized to
obtain the modal matrix, [φ].
6. Obtain the modal participation factors, {Γ} = [φ]T [m]{1}/[φ]T [m][φ]; (the contribution of
each mode shape to the total response).
7. Determine the displacements associated with each mode shape, [u] = [φ][Γ][D] = [φ][Γ]
[A][ω2]−1
where:
[Γ] = diagonal matrix of participation factors
[D] = diagonal matrix of spectral displacements
[A] = diagonal matrix of spectral accelerations
[ω2] = diagonal matrix of squared modal frequencies
8. The resultant maximum displacement at each node is obtained using the SRSS rule
(Equation 7.68) for each row vector: umaxi = (Σui2)½.
9. The matrix of lateral forces at each node is: [f] = [k][u].
10. The resultant maximum lateral force at each node is obtained using SRSS of each row vec-
tor: fmaxi = (Σfi2)½.
11. The column vector of total base shear forces is: {Vb} = [f]T{1}
12. The maximum base shear force is obtained using SRSS as: Vbmax = (ΣVi2)½.
13. The maximum overturning moment is obtained from SRSS using static equilibrium.

EXAMPLE 7.4
Given the three-story building frame shown in Figure E7.2 with damping of 5% and subjected to a
ground acceleration due to an earthquake characterized by the 84.1% design spectrum scaled to
0.25g PGA, determine (1) periods, (2) mode shapes, (3) peak displacements, (4) maximum equiva-
lent static floor forces, (5) maximum base shear, and (6) maximum overturning moment. Assume
beams are rigid. Each story has a stiffness k = 326.3 kips/in (Figure E7.2).

u3
50 kips
m3

k 12 ft u2 k3
100 kips
m2
k 12 ft
u1
100 kips k2
m1
12 ft
k
k1

FIGURE E7.2  Building frame schematic (left) and idealized MDOF structural model (right).
Multi-Degree-of-Freedom System Analysis 189

Solution
1. Determine the equations of motion, frequencies, periods, and shape vectors. The equa-
tions of motion for this case can be determined using D’Alembert’s principle by apply-
ing horizontal equilibrium to the free-body diagrams of the three masses shown in
Figure E7.3. The three equations of motion are

+ → ∑ Fx = 0; − m1u1 − k1u1 + k2(u2 − u1) = 0 ⇒ m1u1 + (k1 + k2 )u1 − k2u2 = 0

+ → ∑ Fx = 0; − m2u2 − k2(u2 − u1) + k3(u3 − u2 ) = 0 ⇒ m2u2 − k2u1 + (k2 + k3 )u2 − k3u3 = 0

+ → ∑ Fx = 0; − m3u3 − k3(u3 − u2 ) = 0 ⇒ m3u3 − k3u2 + k3u3 = 0

These three dependent equations of motion are second-order, linear, and homogeneous
differential equations with constant coefficients, and can be rewritten in the matrix form as

m1 0 0  u1  k1 + k2 −k2 0  u1  0

       
0
 m2 0  u2  +  −k2 k2 + k3 −k3  u2  = 0
     
0
 0 m3  u3   0 −k3 k3  u3  0

Since k1 = k 2 = k3 = k = 326.3 kips/in and m1 = m2 = m = 100 kips/g and k3 =
50 kips/g,
100 kips 0 0 
 
[m] =  0 100 kips 0  / 386.4 in/ s2

 0 0 50 kips

2 −1 0
 
[k ] = −1 2 −1 ⋅ 326.3 kips/in
0 −1 1 


We use a MATLAB script to solve the eigenvalue problem:

u3 – u2
m2u3
k3(u3 – u2) k3(u3 – u2)

k3
u2 – u1 k3(u3 – u2)
k3(u3 – u2)
k2(u2 – u1) m2u2
k2(u2 – u1)
k2
k2(u2 – u1)
k2(u2 – u1)
m1u1
k1u1 k1u1

k1

k1u1

u1

FIGURE E7.3  FBDs for each mass used to determine equations of motion.
190 Introduction to Earthquake Engineering

clear all % clears any previously defined variables

clc % clears the screen
 m = [100 0 0; 0 100 0; 0 0 50]/386.4% mass matrix in kips-sec2/in
k = [2 -1 0; -1 2 -1; 0 -1 1]*326.3% stiffness matrix in kips/in
[phi,lam]=eig(k,m); % compute eigenvalues and eigenvectors
omegas=sqrt(lam) % determine and show frequencies from eigenvalues
periods = 2*pi*diag(inv(omegas)) % determine the periods from omegas
% mode shapes by normalizing eigenvectors to get top displ equal to 1
[N_rows, N_cols] = size(phi); % finds N, the size of the matrices
for i = 1:N_cols; % loops over the N modes to normalize them
norm_phi(:,i) = phi(:,i)./phi(N_rows,i);
end
norm_phi % display normalized modal matrix

The results of the script are:

omegas =
18.3803 0 0
0 50.2160 0
0 0 68.5963
periods =
0.3418
0.1251
0.0916
norm_phi =
0.5000 −1.0000 0.5000
0.8660 0.0000 −0.8660
1.0000 1.0000 1.0000

2. With the eigenvalues and eigenvectors, first we follow the same procedure as for the
generalized SDOF case using the first mode (following the same process as in Example
6.1). Then, we use matrix operations to complete the entire solution.
a. Use the first mode, φi1, as the shape vector:

 0.5 
 
φi1 = 0.866
 
 1 

b. Determine the generalized properties. The building frame can be modeled as a gen-
eralized SDOF system assuming that the beams are rigid and only lateral deforma-
tions of columns occur. The stiffness of each floor is then obtained by summing the
column lateral stiffnesses at each level. The generalized mass, stiffness, and force of
the generalized SDOF system are calculated as follows:

Level, i mi ki φi1 Δφi1 mi ⋅ φi1 mi ⋅ φ 2i1 ki · (Δφi1)2

3 m/2 k 1 0.134 m/2 · (1) m/2 · (1)2 k · (0.134)2
2 m k 0.866 0.366 m · (0.866) m · (0.866)2 k · (0.366)2
1 m k 0.5 0.5 m · (0.5) m · (0.5)2 k · (0.5)2
∑ 3i=1 = 1.866m 1.5m 0.402k

Substituting the given values for mass and stiffness, the following properties of the
“generalized SDOF,” which are the first elements of the modal mass and stiffness
matrices, are calculated as follows.
Multi-Degree-of-Freedom System Analysis 191

Generalized mass:

3 W 100 kips(1000 lb/kip)

M1 = ∑ mi ⋅ φi21 = 1.5m = 1.5 = 1 .5 = 388.2 lb-s2 /in
i =1 g 386.4 in/s 2

Generalized stiffness:

3
K1 = ∑ ki ⋅ (Δφi1)2 = 0.402 k = 0.402(326.3 kips/in)(1000 lb/kip) = 131144
, lb/in
i =1

Generalized force:

3 W 100 kips(1000 lb/kip)

L1 = ∑ mi ⋅ φi1 = 1.866 ⋅ m = 1.866 = 1.866 = 482.9 lb-s 2 /in
i =1 g 386.4 in/s 2

c. Determine the natural period, frequency, and participation factor for the first mode.
Natural period, which is the same as the MATLAB result in step 1:

M1 388.2lb-s2 /in
Tn1 = 2π = 2π = 0.3418 s
K1 ,
131144 lb/in

Natural frequency, which is the same as the MATLAB result in step 1:

K1 131144
, lb/in
ωn1 = = = 18.38 rad/s
M1 388.2lb-s2 /in

Participation factor:

L1 482.9 lb-s2 / in
Γ1 = = = 1.244
M1 388.2lb-s2 / in

d. Determine maximum floor displacements. Determine the maximum displacement at the

top of the building, q(t) or D1 due to an 84.1% design spectrum, scaled to 0.25g PGA;
first, enter the response spectrum in Figure 5.11 with Tn1 = 0.342 s, which is greater than
1/8 s so the acceleration is constant:

A1 = η ⋅ 2.71g = 0.25(2.71)(386.4in/s2 ) = 2618in/s

. 2

The maximum spectral displacement is

A1 261.8 in/s2
D1 = = = 0.775in
ω n21 (18.38 rad/s)2

We can now determine the maximum floor displacements using Equation 6.35:

 0.5  0.482
   
  n) = 0.835 in
uio = φi1Γ1 ⋅ D1 = 0.866 (1.244)(0.775in
   
 1  0.964

192 Introduction to Earthquake Engineering

 0.311
 
vs. Example problem 1 in Chapter 6, where u jo = 0.623 in
 
0.934
e. Determine the equivalent static story forces. Determine the equivalent static story forces
using Equation 6.37,

fio = Γ1miφi1A1

f3o = Γ1m3φ31A1 = 1.244(50 kips/g)(1)(0.25 ⋅ 2.71⋅ g ) = 42.1kips

f2o = Γ1m2φ21A1 = 1.244(100 kips/g)(0.866)(0.25 ⋅ 2.71⋅ g ) = 73.0 kipss

f1o = Γ1m1φ11A1 = 1.244(100 kips/g)(0.5)(0.25 ⋅ 2.71⋅ g) = 42.1kips

f. Determine base shear and overturning moment. Determine the base shear force using
Equation 6.40,

3
Vmax b = ∑ fio = 42.1 kips + 73.0 kips + 42.1 kips = 157.3 kips
1=1

Determine the overturning moment using Equation 6.41,

3
Mmax b = ∑ hj foj = 42.1 kips(12 ft ) + 73.0 kips(24 ft) + 42.1 kips(36 ft ) = 3775 kip ⋅ ft
j =1

As in Chapter 6, we can obtain the internal story shear force and moment at each
level by drawing shear force and bending moment diagrams considering the building as
a cantilever beam. The diagrams also compare these results to those of the analysis of
Example 6.1, which are given in parentheses (Figure E7.4).
At this point, we could follow steps 2-a to 2-f for each of the remaining mode shapes
and then apply one of the combination rules to obtain the total response. Alternatively,
we could use the matrix operation process listed in Section 7.3, beginning with step 6 to
perform the analysis of all the modes simultaneously, see next step.
3. Use the matrix operations listed in Section 7.3:

42.1 kips 0.96 in

(48.1 kips) (0.93 in)
12 ft
73.0 kips 0.83 in
(64.2 kips) (0.62 in) 42 –505
12 ft
42.1 kips 0.48 in (48) (–578)
(32.1 kips) (0.31 in) 115 –1885
12 ft (112) (–1925)

3775 kip-ft 157.3 kips 157 –3775

(144.4 kips) V( kips) M(kip-ft)
(3658 kip-ft)

FIGURE E7.4  Lateral forces, internal shear, and internal moment diagram for mode 1.
Multi-Degree-of-Freedom System Analysis 193

a. Obtain the modal participation factors, {Γ} = [φ]T[m]{1}/[φ]T[m][φ]: First, get the
modal matrix, [M] = [φ]T[m][φ],

 0 .5 0.866 1 100 0 0   0 .5 −1 0 .5 
   
[M] =  −1 0 1  0 100 0  0.866 0 −0.866 / 386.4 in/ss2
 0 .5 −0.866 1  0 0 50  1 1 1 


0.388 0 0 
 
[M] =  0 0.388 0  lb-s2 /in
 0 0 0.388


And, the modal force, [L] = [φ]T[m]{1},

 0 .5 0.866 1 100 0 0  1  0.483 

     
[L] =  −1 0 1  0 100 0  1 / 386.4 in/s2 = −0.129 lb-s2 /in
  
 0.0347 
 0 .5
 −0.866 1  0 0 50 1  

Now, get the participation factors, {Γ} = [φ]T[m]{1}/[φ]T[m][φ],

 2.576 0 0   0.483   1.244 

  2    
−1 
{Γ} = [M] {L} =  0 2.576 0  in/lb-s ⋅ −0.129 lb-s /in = −0.333
  2

 0 
 0.0347  
 0.0893
 0 2.576    

b. Determine the displacements associated with each mode shape, [u] = [φ][Γ][A]
[ω2]−1. First, assemble the participation factors and square of the frequencies into
diagonal matrices:
  Diagonal matrix of participation factors:

1.244 0 0 
 
[Γ] =  0 −0.333 0 
 0 0 0.0893


Diagonal matrix of squared modal frequencies, which is equal to the eigenvalues:

337.8 0 0 
 
2 
[ω ] =  0
n 2521.6 0  rad2 /s2
 0 0 4705.5


To obtain the diagonal matrix of accelerations, we first enter the response spec-
trum in Figure 5.11 with the following three values for Tn:
– Tn1 = 0.342 s, which is greater than 1/8 s so the acceleration is constant,
A1 = η · 2.71g = 0.25(2.71)(386.4 in/s2) = 261.8 in/s2;
– Tn2 = 0.1251 s, which is also greater than 1/8 s so the acceleration is constant,
A 2 = η · 2.71g = 0.25(2.71)(386.4 in/s2) = 261.8 in/s2; and
– Tn3 = 0.0916 s, which is less than 1/8 s so the acceleration is given as,
A3 = η · 11.7(Tn3)0.704g = 0.25(11.7) (0.0916)0.704 (386.4 in/s2) = 210.1 in/s2.
194 Introduction to Earthquake Engineering

Thus, the diagonal matrix of spectral accelerations is given as

 261.8 0 0 
 
[ A] =  0 261.8 0  in/s2
 0 0 210.1


The displacements then, [u] = [φ][Γ][A][ω2]−1

 0 .5 −1 0.5  1.244 0 0 
  
[u] = 0.866 0 −0.866  0 −0.333 0 
 1 1 1   0 0 0.0893

 261.8 0 0   3 3 7 .8 0 0  −1
  
*  0 261.8 0   0 2521.6 0 
 0 0 210.1  0 0 4705.5

 0.482 0.0346 0.0020 
 

= 0.835 0 −0.0035 in
0.964 −0.0346 0.0040 


c. The resultant maximum displacement at each node is obtained from the SRSS rule
(Equation 7.68) of each row vector: umaxi = (Σui2)½:

u1max = (0.482 in)2 + (0.0346 in)2 + (0.002 in)2 = 0.483 in

u2max = (0.835 in)2 + (−0.0035 in)2 = 0.835 in

u3max = (0.964 in)2 + (0.0346 in)2 + (0.004 in)2 = 0.965 in

d. The matrix of lateral forces at each node is, [f] = [k][u]:

2 −1 0  0.482 0.0346 0.0020 

   
[f ] = −1 2 −1 ⋅ 326.3 kip / in ⋅ 0.835 0 −0.0035 in
0 −1 1  0.964 −0.0346 0.0040 
 
 42.1 22.6 2 .4 
 
= 73.0 0 −4.2 kips
 42.1 −11.3 2.4 


These are shown in Figure E7.5.

e. The resultant maximum lateral force at each node is obtained from the SRSS rule
(Equation 7.68) of each row vector: fimax = (Σfi2)½,

f1max = ( 42.1 kips)2 + ( 22.6 kips)2 + ( 2.4 kips)2 = 47.9 kips

f2max = (73 kips)2 + (−4.2 kips)2 = 73.1 kips

Multi-Degree-of-Freedom System Analysis 195

(a) (b) (c)

44.1k 11.3k 2.4k

73.0k 4.2k

44.1k 22.6k 2.4k

Mode 1 Mode 2 Mode 3

FIGURE E7.5  Resulting lateral forces at each node for a given mode of vibration: (a) Mode 1,
(b) Mode 2, and (c) Mode 3.

f3max = ( 42.1 kips)2 + (−11.3 kips)2 + ( 2.4 kips)2 = 43.7 kips

f. The column vector of total base shear forces is: {Vb} = [f]T{1},

 42.1 73.0 42.1  1 157.6

     
[Vb ] =  22.6 0 −11.3 kips ⋅ 1 =  11.3  kips
   
 2 .4
 −4 .3 2.4  1  0.7 

Alternatively, we can obtain all the story shear forces using static equilibrium at each
level, for each mode shape:

42.1+ 73 + 42.1 157.6

   
{V1} =  42.1+ 73  kips = 115.2 kips
 
 42.1   42.1 
   

−11.3 + 22.6  11.3 

   
{V2} =  −11.3  kips = −11.3 kips
 
   
 −11.3  −11.3

2.4 − 4.3 + 2.4  0.7 

   
{V3 } =  2.4 − 4.3  kips = −1.8 kips
 2 .4   2.4 
   

g. The maximum base shear force is obtained from SRSS as: Vbmax = (ΣVi2)½,

Vbmax = (157.6 kips)2 + (11.3 kips)2 + (0.7 kips)2 = 158 kips

The maximum story shear forces can be obtained using the SRSS rule at each level for
all the mode shape story shears obtained in step 3-f:

 (157.6 kips)2 + (11.3kips)2 + (0.7 kips)2 

  158
   
{Vmax } =  (115.2kips)2 + (−11.3kips)2 + (−1.8 kips)2  = 116 kips
   
 ( 42.1kips)2 + (−11.3kips)2 + (2.4 kips)2   44 
 

196 Introduction to Earthquake Engineering

43.7 kips 0.96 in

12 ft
73.1 kips 0.83 in
44 –524
12 ft
47.9 kips 0.48 in
116 –1906
12 ft

3777 kip-ft 158 kips 158 –3777

V (kips) M (kip-ft)

FIGURE E7.6  Lateral forces, internal shears, and internal moment diagrams.

h. The maximum overturning moment is obtained from SRSS using static equilibrium. We
first obtain the story overturning moments using static equilibrium at each level, for each
mode shape:

42.136 ( )
( ) + 73( 24) + 42.112 3774
   
{M1} =  ( ) + 73(12)
42.124  kips ⋅ ft = 1887 kips ⋅ ft
   
 42.112
( )   505 

−11.3(36) + 22.6(12) −135
   
{M2} =  −11.3( 24)  kips ⋅ ft = −271 kips ⋅ ft
   
 −11.3(12)  −135

2.4(36) − 4.3( 24) + 2.4(12) 13
   
{M3} =  2.4(24) − 4.3(12)  kips ⋅ ft =  7  kips ⋅ ft.
   
 2.4(12)  29

  The maximum story overturning moments can now be obtained using the SRSS rule:
 (3774)2 + (135)2 + (13)2 
  3777
   
2 
{Mmax } =  (1887) + ( 271) + (7)  kips ⋅ ft = 1906 kips ⋅ ft
2 2
   
 (505)2 + (135)2 + ( 29)2   524 
 

  Figure E7.6 summarizes the results of the maximum response obtained:
4. We can write a MATLAB script to perform all the operation as follows:
clear all % clears any previously defined variables
clc % clears the screen
g = 386.4; % acceleration due to gravity in in/s2
pA = 0.25* g; % peak ground acceleration
m = [100 0 0; 0 100 0; 0 0 50]/386.4 % mass matrix in kip · s2/in
k = [2 −1 0; −1 2 −1; 0 −1 1]*326.3 % stiffness matrix in kips/in
[phi,lam]=eig(k,m); % compute eigenvalues, lam and eigenvectors, phi
omegas=sqrt(lam) % determine and show frequencies from eigenvalues
periods = 2*pi*diag(inv(omegas)) % determine the periods from omegas
% mode shapes by normalizing eigenvectors to get top displ equal to 1
[N_rows, N_cols] = size(phi); % finds N, the size of the matrices
for i = 1:N_cols; % loops over the N modes to normalize them
norm_phi(:,i) = phi(:,i)./phi(N_rows,i);
end
norm_phi % display normalized modal matrix
% The spectral accelerations from an applicable response spectrum
Multi-Degree-of-Freedom System Analysis 197

SA1 = pA*2.71; % THIS CHANGES WITH THE PERIOD

SA2 = pA*2.71; % THIS CHANGES WITH THE PERIOD
SA3 = pA*11.7*(periods(3))∧0.704; % THIS CHANGES WITH THE PERIOD
SA = [SA1,SA2,SA3]
% Top floor spectral displacements
SD = SA*inv(lam)
% determine the participation factors
LT = sum(norm_phi’* m,2) % sum(A,2) returns a vector of sums of each row
MT= sum(norm_phi’* m*norm_phi,2)
par_fac = LT./MT % ./ operation divides each element of the two vectors
% Maximum displacements at each level for each mode shape
for i = 1:N_rows;
for j = 1:N_cols;
ui_max(i,j) = SD(j)*norm_phi(i,j)*par_fac(j);
end
end
ui_max % display displs.; each column corresponds to a mode shape
% Use the SRSS rule to get max displacements at each level
u_maxsrss=sqrt(sum((ui_max).∧2,2)) %.∧ operation squares each element
% Floor forces at each level for each mode shape
f= k*ui_max
% Use SRSS method to get total max forces at each level
f_maxsrss=sqrt(sum((f).∧2,2))
% Base shear for each mode shape and max base shear using SRSS rule
V=(sum(f,1))
V_maxsrss=sqrt(sum((V).∧2,2))
% Overturning moment for each mode shape and max moment using SRSS
heights=[12;24;36];
OTM = f’*heights
OMT_maxsrss=sqrt(sum((OTM).∧2,1))

The results of the script are:

m=
0.2588 0 0
0 0.2588 0
0 0 0.1294

k=
652.6000 –326.3000 0
−326.3000 652.6000 −326.3000
0 –326.3000 326.3000

omegas =
18.3803 0 0
0 50.2160 0
0 0 68.5963

periods =
0.3418
0.1251
0.0916

norm_phi =
0.5000 −1.0000    0.5000
0.8660   0.0000  −0.8660
1.0000   1.0000   1.0000

SA =
261.7860 261.7860 210.0515
198 Introduction to Earthquake Engineering

SD =
0.7749 0.1038 0.0446

LT =
0.4829
−0.1294
0.0347

MT =
0.3882
0.3882
0.3882

par_fac =
  1.2440
−0.3333
  0.0893

ui_max =
0.4820 0.0346 0.0020
0.8348 –0.0000 −0.0035
0.9640 –0.0346 0.0040

u_maxsrss =
0.4832
0.8348
0.9646

f=
42.1411 22.5833 2.4277
72.9905 −0.0000 −4.2049
42.1411 −11.2917 2.4277

f_maxsrss =
47.8724
73.1115
43.6951

V=
157.2726 11.2917 0.6505

V_maxsrss =
157.6788

OTM =
1.0e+003 *
 3.7745
−0.1355
 0.0156

OMT_maxsrss =
3.7770e+003
Multi-Degree-of-Freedom System Analysis 199

PROBLEMS
7.1 Using the definition of stiffness and mass influence coefficients, formulate the equation of
motion for the frame shown below. Assume the beams are rigid.

m/2
p2(t)

h EI

m
p1(t)

h EI

2h

7.2 Formulate the equation of motion (in matrix form) using D’Alembert’s principle for the
given three-story building, which has rigid beams and flexible steel (E = 29,000 ksi) col-
umns with total moment of inertial for each floor as shown.

400 kips

I = 850 in4 500 kips 12 ft

I = 1240 in4 500 kips 12 ft

I = 1240 in4 14 ft

7.3 Formulate the equation of motion (in matrix form) using D’Alembert’s principle for the
given two-story building, which has rigid beams and flexible steel (E = 29,000 ksi) col-
umns with total moment of inertial for each floor as shown.

w2 = 20 kips

I = 1110 in4 I = 1240 in4 12 ft

w1 = 50 kips

I = 1110 in4 I = 1240 in4 I = 850 in4 14 ft

20 ft 20 ft
200 Introduction to Earthquake Engineering

7.4 Formulate the equation of motion (in matrix form) using D’Alembert’s principle for the
given three-story building, which has rigid beams and flexible steel (E = 29,000 ksi) col-
umn sections as shown.
WDL = 1 kips/ft

W14×99 W14×109 W14×99

12 ft

WDL = 1.2 kips/ft WDL = 1.2 kips/ft

W14×109 W14×120
12 ft W14×109

WDL = 1.2 kips/ft WDL = 1.2 kips/ft

12 ft W14×120 W14×145 W14×120

20 ft 20 ft

7.5 Formulate the equation of motion (in matrix form) using D’Alembert’s principle for the
given two-story building, and determine the natural frequencies, periods, normalized
modal matrix, and participation factors.

3m

7.6 The following two-story building is supported with steel frames spaced at 15 ft on center,
and has the floor load shown plus wall load of 20 psf. Formulate the equation of motion (in
matrix form) using D’Alembert’s principle and determine the natural frequencies, periods,
normalized modal matrix, and participation factors using MATLAB.
Multi-Degree-of-Freedom System Analysis 201

50 psf
u2(t)

W10×21
12 ft
I = 106.3 in4
100 psf
u1(t)

W10×45 15 ft

I = 248.6 in4

30 ft

7.7 Given the following mode shapes and frequencies, compute the participation factors for a
3-story building with floor weights of 120 kips for the first floor, and 80 kips for the second
floor and roof.

 1.68 −1.208 −0.714  8.77 

   
[φ] =  1.22 0.704 1.697  and {ω} = 25.18 rad/s
 
 0.572
 1.385 −0.984 48.13

Note that[φ]T [M][φ] = [I], so {Γ} = [φ]T [M]{1}.

7.8 Use an MDOF analysis for the following building frame to determine the story displace-
ments, story forces, base shear, and overturning moment due to an earthquake character-
ized by an 84.1% design spectrum, scaled to 0.25g PGA. Assume rigid beams and damping
of 5%.
w2 = 150 kips

k2 = 20 kips/in w1 = 150 kips 15 ft

k1 = 40 kips/in 15 ft

7.9 Given a five-story building frame subjected to a ground acceleration due to a 0.25g PGA
earthquake characterized by the 84.1% design spectrum, use an MDOF analysis to deter-
mine (i) peak displacements, (ii) maximum base shear, and (iii) maximum overturning
moments. k = 326.3 kips/in, w5 = 50 kips, and the weights on the other four floors equal
100 kips. Assume rigid beams and damping of 5%.

REFERENCES
Chopra, A. K., Dynamics of Structures: Theory and Applications to Earthquake Engineering, 4th edition,
Prentice-Hall, Upper Saddle River, NJ, 2012.
Villaverde, R., Fundamental Concepts of Earthquake Engineering, CRC Press, Boca Raton, FL, 2009.
 8 Seismic Code Provisions

After reading this chapter, you will be able to:

1. Compute seismic load combinations

2. Determine the redundancy factor
3. Explain the overstrength factor
4. Perform calculations used to determine earthquake loads using the equivalent lateral force
procedure (ELFP)
5. Perform calculations used to determine earthquake loads using modal response spectrum
(MRS) analysis
6. Describe seismic response history procedures

Design of structures must be based on a minimum standard of care to safeguard the safety,
health, and welfare of the public. Municipal, state, or federal governments concerned with the
safety of the public have established building codes used to control the construction of struc-
tures within their jurisdiction. These design codes specify among other things the design loads.
While building codes may vary from city to city, most municipalities rely on regional codes and
design standards for the specification of loads. These codes and standards include general build-
ing codes, such as the International Building Code (IBC) and the ASCE/SEI (2010) “Minimum
Design Loads for Buildings and Other Structures” standard, from here on referred to as ASCE-7.
These codes and standards are developed by various organizations and present the best opinion
of those organizations as to what constitutes good practice. In this chapter, we provide an over-
view of structural design philosophies and explain how seismic loads are incorporated as part
of structural design. The primary focus of this chapter is to describe the parameters associated
with determining earthquake loads in accordance with ASCE-7 and to demonstrate the proce-
dures for determining earthquake loads using the ELFP, seismic response history analysis, and
MRS analysis.

8.1  STRUCTURAL DESIGN PHILOSOPHIES

ASCE-7 and IBC adhere to two general design philosophies: working stress design also known as
allowable stress design (ASD) and limit states design also known as the load and resistance factor
design (LRFD); the primary objective of both design philosophies is to obtain structures that will
not fail under an extreme loading condition. Failure of a structure is defined as the state where a
loading condition exceeds a limit state. Thus, a limit state is a condition beyond which a structure,
or structural component, ceases to fulfill the function for which it was designed. There are two limit
states of interest, structural failure and serviceability:

1. A structural failure (or strength) limit state is the more critical of the two because it is
concerned with safety and it relates to the maximum load carrying capacity of a structure.
2. A serviceability limit state relates to structural performance under normal service
conditions.

Both design philosophies focus on the ultimate limit state and allow the designer some freedom
of judgment in serviceability. The main concern from a strength limit state standpoint is ensuring
that the capacity (resistance, R) of the structure exceeds the demand (effect of loads, Q). Since it is

203
204 Introduction to Earthquake Engineering

economically impracticable to design 100% safe structures, we must assess the risk of failure, or
the reliability of the structure. To assess the reliability of a structure, we must understand how the
uncertainties on material properties, geometry, and material durability relate to the load uncertain-
ties; not all the extremes of loads can occur simultaneously on the structure, or, at least, it is very
unlikely. In ASD, all uncertainties are accounted using a single quantity, the factor of safety (FS),
while the LRFD uses a rational approach to obtain a more uniform assessment of the reliability of a
structure. Thus, the main difference between ASD and LRFD is that LRFD rationally accounts for
the statistical variability in the loads (and materials), and is based on reliability analysis to obtain a
probability-based assessment of structural safety.

8.1.1 Allowable Stress Design

In ASD, the risk assessment of failure is made on the basis of an allowable FS, which was learned
from previous experiences for the system in question under its anticipated environment. The FS is
defined as the ratio of nominal values of resistance R and the sum of load effects ΣQ. That is, to
ensure safety in structural design,

R (8.1)
FS ≥
∑Q

It is important to note that both sides of the inequality must be evaluated for the same conditions.
For example, load effect that produces tensile stresses should be compared with the tensile strength
of the member in question. The strength of a member is material dependent; a properly designed
system of members must ensure vertical and lateral load-resisting system integrity by providing
ductility, energy dissipation, and avoiding progressive collapse by providing redundancy and pre-
serving the load path—all while maintaining deformations within the prescribed limits.
The stress produced by the loads, ΣQ, is determined once the structural form has been selected
using combinations of loads that can be reasonably expected to act on the structure. That is, not
all possible loads will act concurrently; thus, ASCE-7 (Section 2.4) specifies the following load
combinations:

1. D
2. D + L
3. D + (Lr or S or R)
4. D + 0.75 L + 0.75 (Lr or S or R)
5. D + (0.6 W or 0.7 E)
6a. D + 0.75 L + 0.75(0.6 W) + 0.75 (Lr or S or R)
6b. D + 0.75 L + 0.75(0.7 E) + 0.75 S
7. 0.6 D + 0.6 W
8. 0.6 D + 0.7 E

where:
D represents the dead loads, which are loads that have constant magnitude, remain in one posi-
tion, and are usually known to a high degree of certainty. They include the weight of vari-
ous structural elements and weight of objects permanently attached to the structure, such
as walls, roofs, ceilings, and equipment
L represents the live loads, which are those loads that vary in magnitude and position with time.
They are caused by the structure being occupied, used, and maintained. In general, live
loads are induced by gravity, and can be stationary or transient, which may still be treated
as a static load, unless applied rapidly, in which case, the dynamic effect creates additional
inertial load called impact
Seismic Code Provisions 205

Lr represents the roof live loads, which are loads produced by repair/maintenance workers, their
equipment and materials, or during the life of the structure by movable objects such as
planters. There is a limit to the amount of load that can be realistically placed upon roofs
because most roofs are sloped
S represents the snow load, which depends on ground snow weight, the building’s general shape, and
roof geometry (flat roofs are subjected to higher loads compared to sloped roofs)
R represents the rain load, which depends on the functionality of the drainage facilities, and
only a concern when rainwater accumulates faster than it runs off, ponding. According
to ASCE-7, elements should be designed to support all rainwater that accumulates when
primary drains are blocked and the water rises above the inlet of the secondary drainage
system
W represents the wind load, which is a lateral load caused by wind pressure resulting from the
kinetic energy of moving air as it strikes an object in its path
E represents the earthquake load, which is the main subject of this book

These loads are actions that result from the weight of all building materials, occupants and their
possessions, environmental effects, and are categorized based on their character and duration.
The history of each of these loads can be represented schematically as shown in Figure 8.1.
Dead load

Time
Sustained
live load

Several years

Time
Transient

Several
live load

hours/days

Time
Seconds
Earthquake

n years

Time
Wind

3s

Time
Load combos

Time
D D+L D + 0.75 L + 0.75(0.6 W) D + 0.75 L + 0.75(0.7 E)

FIGURE 8.1  Schematic history of loads and their superposition to obtain the load combinations.
206 Introduction to Earthquake Engineering

This figure shows that the dead load remains constant over its entire history, the live load can be
divided into a sustained load (one that may have variations at different points in its history, such
as building partitions) and a transient load (one that changes frequently, both in duration and
magnitude), the earthquake load is based on short events as was discussed in earlier chapters, and
a wind load that includes sustained intensity and wind gusts. As shown in the figure, the largest
magnitudes of each of these loads are not likely to occur simultaneously, which is why various
load combinations are needed to represent potential critical loadings at different points during the
life of a structure.
Gravity-induced loads on any floor system are assumed to act as a uniform pressure on the floor
slab. We are only concerned with the structural system; so to analyze the effects of these pressure-
like forces on each structural member, we need to rationally distribute them to the supporting mem-
bers. To accomplish this, we use the concept of the tributary area, the area of the slab that is carried
by a particular structural member.

8.1.2 Load and Resistance Factor Design

In LRFD, the risk assessment of failure uses probability theory to account for the variability of
resistance and load effects, providing a more uniform measure of reliability. The resistance side,
R, is multiplied by statistically based resistance factors φ that take into account uncertainties in
material and geometric properties, and equations that predict strength; while the load effects side is
multiplied by statistically based load factors γi that take into account uncertainties in the magnitude
and position of loads. Because of uncertainties, both Q and R are assumed to be random variables,
and their behavior is characterized by a mean, standard deviation, and frequency distribution. The
safety criterion is then stated as

φR ≥ ∑γ Q
i i (8.2)

Again, both sides of the inequality must be evaluated for the same conditions. The left-hand side
of this inequality is the available strength, while the right-hand side is the required strength, which
is computed using the following load combinations (ASCE-7, Section 2.3):

1. 1.4 D
2. 1.2 D + 1.6 L + 0.5 (Lr or S or R)
3. 1.2 D + 1.6 (Lr or S or R) + (L or 0.5 W)
4. 1.2 D + W + L + 0.5 (Lr or S or R)
5. 1.2 D + E + L + 0.2 S
6. 0.9 D + W
7. 0.9 D + E

where all quantities were previously defined. Also, in this design philosophy, unfactored loads are
used for serviceability requirements because safety is not usually of concern for serviceability.

8.1.3 Allowable Seismic Force-Resisting System

The resistance side of the inequalities for both design philosophies is related to the force-resisting
system. Allowable seismic force-resisting systems are listed in ASCE-7, Table 12.2-1, and include:

a. Bearing wall systems (in concrete, masonry, and timber)

b. Building frame systems (in steel, concrete, masonry, and timber)
c. Moment-resisting frame systems (in steel and concrete)
Seismic Code Provisions 207

d. Dual systems with special moment frames (in steel and concrete)
e. Dual systems with intermediate moment frames (in steel and concrete)
f. Ordinary shearwall frame interactive systems (in concrete)
g. Cantilever column systems (in steel, concrete, and timber)
h. Steel systems not detained for seismic resistance

Some of these systems are not permitted in seismic-resistant design; and those permitted are sub-
ject to building height and other limitations. These limitations depend on the various seismic
design categories (SDCs, discussed later in Section 8.4.2, and ASCE-7, Section 11.6), which are
classifications of building requirements based on a desired performance that depends on a build-
ing’s occupancy, the effects of probable ground-shaking intensity at the building site, and struc-
tural irregularities.

8.2  OVERVIEW OF SEISMIC DESIGN CODES

Section 1.3 of the book provides a brief history of the development of mitigation strategies for the
effects of seismic hazards, primarily shaking effects. Since 2000, a new model-building code has been
in use, the aforementioned IBC, which refers the user to the ASCE-7 standard for s­ eismic-loading
specifications. ASCE-7 incorporates a substantial portion of the National Earthquake Hazard
Reduction Program (NEHRP) provisions for seismic design. The explicit objective of these seismic
provisions is to ensure that strength and deflection (drift) requirements are met while preserving the
integrity of the load path throughout the structural system during a seismic event. And, while not
explicitly stated, the objectives of the IBC seismic provisions are as follows:

1. Resist a minor earthquake with no damage.

2. Resist a moderate earthquake with some nonstructural damage, but no structural damage.
3. Resist a major earthquake with some structural damage, but no collapse.

In all cases, damage depends on: intensity and duration of ground shaking, building configuration
(irregular buildings are more susceptible to damage), lateral force-resisting system type (e.g., shear-
walls, braced or unbraced frames), building material (e.g., concrete, steel, wood, or masonry), and
quality of construction. Note that IBC or ASCE-7 provides no guidance for protection against earth
movement, earth slides, liquefaction, or direct fault displacement. The ASCE-7 seismic loading-
related specifications are provided in 13 chapters, the titles of which are as follows:

Chapter 11: Seismic Design Criteria

Chapter 12: Seismic Design Requirements for Building Structures
Chapter 13: Seismic Design Requirements for Nonstructural Components
Chapter 14: Material-Specific Seismic Design and Detailing Requirements
Chapter 15: Seismic Design Requirements for Nonbuilding Structures
Chapter 16: Seismic Response History Procedures
Chapter 17: Seismic Design Requirements for Seismically Isolated Structures
Chapter 18: Seismic Design Requirements for Structures with Damping Systems
Chapter 19: Soil Structure Interaction for Seismic Design
Chapter 20: Site Classification Procedure for Seismic Design
Chapter 21: Site-Specific Ground Motion Procedures for Seismic Design
Chapter 22: Seismic Ground Motion and Long-Period Transition Maps
Chapter 23: Seismic Design Reference Documents

This chapter only covers some of the seismic code provisions, mainly procedures described in
Chapters 11, 12, and 16 of ASCE-7.
208 Introduction to Earthquake Engineering

8.3  SEISMIC LOAD COMBINATIONS

The earthquake load E is computed using one of the methods allowed by the code that will be
presented later in this chapter. This load characterizes vertical and lateral load effects, which are
combined following the provisions presented in ASCE-7, Section 12.4. All members of the vertical
and lateral force-resisting systems must be designed using the load combinations presented earlier;
when seismic effects govern, load combinations 5, 6b, and 8 are used in ASD, while load combina-
tions 5 and 7 are used in LRFD. The values of E used in load combinations of both design philoso-
phies are based on a combination of horizontal, Eh, and vertical, Ev, components, and are given by
the following relationships:

1. When effects of seismic and gravity loads are additive,

E = Eh + Ev = ρQE + 0.2S DS D (8.3)

2. When gravity loads counteract the seismic effects,

E = Eh – Ev = ρQE – 0.2S DS D (8.4)

For cases where structural overstrength should be included:

1. When effects of seismic and gravity loads are additive,

E = Em = Eh + Ev = ΩoQE + 0.2S DS D (8.5)

2. When gravity loads counteract the seismic effects,

E = Em = Eh – Ev = ΩoQE – 0.2S DS D (8.6)

where:
Ev is not a peak value; rather it recognizes that the peak values of the horizontal and vertical
seismic forces are unlikely to occur simultaneously. Also, Ev need not be included when
SDS ≤ 0.125g
ρ is the redundancy factor and is discussed in Section 8.3.1
QE is the effect of horizontal seismic forces (exhibited as internal forces in the axial, shear, and
flexure members), and will be discussed in Sections 8.6−8.8
D is the dead load previously defined
SDS is the design spectral acceleration for short periods, and will be discussed in Section 8.5
Ωo is the overstrength factor discussed in Section 8.3.2

Regarding the direction of the design seismic loads, ASCE-7 (Section 12.5) specifies that E be
applied in directions that produce the largest effect on structural members. The code does, however,
provide guidelines to satisfy this requirement. For example, design seismic forces are permitted to
be applied separately and independently in each of two orthogonal directions for structures assigned
to SDC B (discussed in Section 8.4.2) and regular structures in all other categories.

EXAMPLE 8.1
A column of a steel special concentrically braced frame in a single story medical office build-
ing (SDC = D) supports an axial dead load, D = 35 kips, and horizontal seismic load effect,
QE = ±15 kips. Given SDS = 1.25, an overstrength factor, Ωo = 2.5, and a redundancy factor,
Seismic Code Provisions 209

ρ = 1.3, determine the maximum and minimum axial forces in the column using LRFD load com-
binations. Repeat the problem accounting for overstrength.

Solution
1. Determine the maximum axial load. Determine the maximum earthquake load, E, using
Equation 8.3:

. (±15 kips) + 0.2(125

E = Eh + Ev = ρQ E + 0.2SDSD = 13 . )(35 kips) = ±19.5 kips + 8.8 kips

Thus, taking the positive quantity in the first term, the maximum earthquake load is

E = Eh + Ev = 19.5 kips + 8.8 kips = 23.8 kips

Now, determine the maximum axial load using load combination 5:

. D + E + L + 0.2S = 12
5) 12 . (35 kips) + 28.3 kips = 70.3 kips

2. Determine the minimum axial load. Determine the minimum earthquake load, E, using
Equation 8.4:

E = Eh – Ev = ρQE – 0.2SDSD = 13
. (±15 kips) – 0.2(125
. )(35 kips) = ±19.5 kip
ps – 8.8 kips

Thus, taking the negative quantity in the first term, the minimum earthquake load is

E = Eh – Ev = –19.5 kips – 8.8 kips = –28.3 kips

Now, determine the minimum axial load using load combination 7:

7) 0.9D + E = 0.9(35 kips) − 28.3 kips = 3.2 kips

3. Determine the maximum axial load including overstrength. Determine the maximum
earthquake load, Em, using Equation 8.5:

Em = Eh + Ev = ΩoQE + 0.2SDSD = 2.5(±15 kips) + 0.2(125

. )(35 kips) = ±37.5 kips + 8.8 kips

Thus, taking the positive quantity in the first term, the maximum earthquake load is

Em = Eh + Ev = 37.5 kips + 8.8 kips = 46.3 kips

Now, determine the maximum axial load using load combination 5:

. D + Em + L + 0.2S = 12
5) 12 . (35 kips) + 46.3 kips = 88.3 kips

4. Determine the minimum axial load including overstrength. Determine the minimum
earthquake load, Em, using Equation 8.6:

Em = Eh – Ev = ΩoQE – 0.2SDSD = 13
. (±15 kips) – 0.2(125
. )(35 kips) = ±19.5 kips – 8.8 kips

210 Introduction to Earthquake Engineering

Thus, taking the negative quantity in the first term, the minimum earthquake load is

Em = Eh – Ev = –37.5 kips – 8.8 kips = –46.3 kips

Now, determine the minimum axial load using load combination 7:

7) 0.9D + Em = 0.9(35 kips) − 46.3 kips = −14.8 kips

8.3.1 Redundancy Factor, ρ
Redundancy is a characteristic in which multiple paths of resisting lateral loads are provided. ρ is
intended to encourage system redundancy; when redundancy is lacking, ρ in Equations 8.3 and 8.4
can result in a penalty, or increase in the seismic load of up to 30%. Also, ρ is assigned to the seis-
mic force-resisting system in each of two orthogonal directions for all structures. For the following
conditions, ρ can be taken as 1.0, and thus there is no penalty:

1. All structures in SDC B or C.

2. Calculations of the drift and P-delta (P-Δ) effects.
3. For the design of nonstructural components and nonbuilding structures.
4. For the design of members, connections, collector elements, and splices, when load com-
binations with the overstrength factor are required.
5. Diaphragm design forces specified in ASCE-7, Section 12.10.1.1.
6. Design of structures with damping systems designed in accordance with ASCE-7, Chapter 18.

The redundancy factor should be taken as 1.3 (a 30% penalty) for structures assigned to SDC D, E,
and F. Unless one of the following two conditions is met, whereby ρ is permitted to be taken as 1.0:

1. Stories resisting forces larger than 35% of the base shear in any given direction that com-
plies with Table 12.3-3 in ASCE-7.
2. Structures that are regular in plan at all levels and are provided with at least two bays of
seismic force-resisting perimeter framing on each side of the structure, in each orthogonal
direction, with each story resisting more than 35% of the base shear. For masonry, concrete,
or steel plate shearwalls, the number of bays shall be calculated as the length of shearwall
divided by the story height. For light-framed construction shearwalls, the number of bays
shall be calculated as two times the length of shearwall divided by the story height.

EXAMPLE 8.2
A single-story building shown in Figure E8.1 with an average roof height of 12 ft is constructed
with a flexible roof diaphragm assigned to SDC D. The base shear was determined to be 100 kips,

1 2
100 ft

SFRS
50 ft

FIGURE E8.1  Single-story building plan view.

Seismic Code Provisions 211

which can be resolved into total forces along wall lines 1 and 2 of 50 kips each. Determine the
redundancy factor for each of three lateral force-resisting systems consisting of (1) 18-foot long
concrete shearwalls, (2) single-steel ordinary concentrically braced frames, and (3) double-steel
special moment frames.

Solution
1. 18-foot long concrete shearwalls.
Determine the number of bays using one of the following relationships: Light-frame con-
struction number of bays is given as

#bays = 2Lw /hsx

where:
Lw is the length of the shearwall
hsx is the story height

For all other cases, such as concrete,

#bays = Lw /hsx

Thus, the number of bays for the concrete shearwalls is

# bayscsw = Lw /hsx = 18 ft/12 ft = 15

. < 2 bays minimum

Therefore, ρ = 1.3 (no redundancy)

2. Single-steel ordinary concentrically braced frames.

#baysSCBF = 1< 2 bays minimum

Therefore, ρ = 1.3
3. Double-steel special moment frames.

#baysSSMF = 2 = 2 bays minimum

Therefore, ρ = 1.0

8.3.2 Overstrength Factor Ωo and other Seismic Force-Resisting System Parameters

To ensure safety of the overall system, a complete lateral load path must be provided. A proper load
path is guaranteed by providing sufficient ductility of the system, which is achieved by properly
designing and detailing the overall system. Detailing refers to special connection/component designs
to integrate the overall lateral force-resisting system in order to transfer the seismic loads from their
point of origin to the foundation of the structural system. Ductility of a system is mobilized when
seismic-resistant structures experience inelastic deformations during an earthquake. As shown in
Section 5.2, ductility greatly reduces the seismic response of the overall structure mainly because of
two reasons: (a) yielding of the structure leads to inelastic deformations that change the stiffness of
the structure, which leads to the lengthening of the effective period, reducing the required strength
demand; and (b) substantial energy dissipation (hysteretic damping) from inelastic behavior.
The effect of ductility is taken into account with the response modification coefficient, R, which
reduces the lateral seismic force to strength level design forces. As discussed in Section 5.2, we
use elastic response spectra to obtain inelastic ones, which based on ASCE-7 specifications can be
212 Introduction to Earthquake Engineering

Base shear, V
Elastic response spectrum

VE

VY = ΩoVS

VS = VE I ∆D
Cd =
R ∆S

∆S ∆D ∆E Drift, ∆
Period, T

FIGURE 8.2  Inelastic force–deformation curve.

accomplished by dividing elastic spectra by R. The response modification coefficient, R, is given

as the ratio of the lateral force developed in a linear elastic structure, VE , to the prescribed design
lateral force, VS, and represents an adjustment of elastic force to design load level for the entire struc-
tural system in order to reach the required ductility demand; see Figure 8.2. Values of R are listed in
ASCE-7, Table 12.2-1, for various seismic force-resisting systems. Table 12.2-1 in ASCE 7 provides
values for two additional parameters, the overstrength factor Ωo and the deflection amplification fac-
tor Cd. Values of R vary from 1 to 8, Ωo vary from 1.25 to 3, and Cd vary from 1.25 to 6; for example,
both steel and reinforced concrete special moment frames have R = 8, Ωo = 3, and Cd = 5.5. These
three parameters are related and collectively characterize the inelastic behavior of structures. Figure
8.2 illustrates the relationship between the parameters.
When a component does not reach the necessary level of ductility, the seismic forces can be greatly
underestimated. In the methods used to obtain seismic loads discussed later in this chapter, it is
assumed that all components have the same level of ductility, which is an extremely rare condition.
Some components (like bolted connections) have much lower levels of ductility as compared to other
components of the lateral force-resisting system, such as shearwalls and beams. The overstrength
factor, Ωo, attempts to oversize low-level-ductile (or nonyielding) portions of the structure so that first
yield failure occurs in the more ductile elements, avoiding sudden brittle failures during a seismic
event. The overstrength factor, Ωo, is given as the ratio of the actual lateral force developed in a struc-
ture, VY, to the calculated design lateral force, VS, and represents an adjustment of design force to the
actual load level for some components that cannot reach the required ductility demand; see Figure 8.2.
Deflections are typically calculated using elastic structural analyses, and since excessive
deformation can lead to instabilities of the overall system, it is important to estimate deflections
­accurately. Rather than conducting a full inelastic structural analysis, the deflection amplification
factor Cd magnifies the deflection of the structure based on an elastic analysis. These amplified lat-
eral deflections (or drifts) must be limited to prescribed values. The deflection amplification factor
Cd represents the ratio of the design drift, ΔD, to the drift under design forces, ΔS, and relates the
elastic drift to the inelastic drift; see Figure 8.2. It is important to note that the specifications require
that elastic deflections be calculated at the strength level; that is, using load combinations, which
implies a strength limit state and not a serviceability limit state.

EXAMPLE 8.3
Use ASCE-7, Table 12.2-1, to determine seismic force-resisting system parameters and limitations
for a steel special moment frame (SSMF) depicted in Figure E8.2. Also, list an advantage and a
disadvantage for SSMF systems.
Seismic Code Provisions 213

FIGURE E8.2  Special steel moment frame schematic.

Solution
1. Determine seismic force-resisting system parameters. From ASCE-7, Table 12.2-1, the
SSMF system is part of group C, C.1. The values of design coefficients and factors are as
follows:
Response modification coefficient, R = 8
The overstrength factor Ωo = 3
The deflection amplification factor Cd = 5.5
2. Identify the limitations for the seismic force-resisting system. From ASCE-7, Table 12.2-1,
group C, C.1, structural system limitations are available according to SDC as

SDC B C D E F
Restriction NL NL NL NL NL

For each SDC of the SSMF, NL = not limited is listed, which implies that the system is
permitted everywhere and has no height limits.
3. Describe an advantage and a disadvantage for SSMF systems.
Advantage: because of the high R value, the base shear is relatively low.
Disadvantage: because of the high Cd value, the drifts are relatively high.

8.4  OVERVIEW OF SEISMIC LOAD ANALYSIS PROCEDURES

As discussed in Section 8.3, earthquakes cause vertical and horizontal accelerations, which result in
vertical and horizontal forces on structures. Vertical forces are generally smaller than lateral forces
and are considered only in overturning analyses. The focus of this section is on determining the
effect of horizontal seismic forces, QE , in seismic load combinations shown in Equations 8.3−8.6.
The methods discussed in this section are used to conduct a complete structural analysis of horizon-
tal forces. After determining a site-specific response spectrum and selecting an allowable seismic
force-resisting system, the process entails, calculating the base shear, distributing this base shear
force to various levels of the structure, and using structural analysis to determine lateral deflections
as well as internal forces and moments generated in each of the structural elements.

8.4.1 Design Response Spectrum

The key step in obtaining lateral forces using any method is the development of appropriate
design response spectra. ASCE-7’s approach is similar to that followed by Newmark and Hall
as described in Chapter 5. ASCE-7 uses two spectral response accelerations (SS and S1) that are
based on a maximum considered earthquake (MCE) ground motion (as described in Section 2.4).
214 Introduction to Earthquake Engineering

San Joaquin County

SS S1

Stockton, CA

2% Probability of being
exceeded in 50 years

FIGURE 8.3  Maximum considered ground motion intensity in San Joaquin County, California (as ­percent
of acceleration due to gravity).

SS and S1 are presented in map form in ASCE-7, Figures 22-1–22-6, assuming 5% damping and a
site underlaid by rock, see Figure 8.3, for the contour maps in San Joaquin County, California. S1
represents acceleration for a period of 1.0 s, while SS represents acceleration for a period of 0.2 s
(short period). These maps are difficult to read, especially in cases where contour lines are closely
spaced. A more accurate and convenient approach is to use the USGS Seismic Hazards Mapping
Application, available at: http://earthquake.usgs.gov/designmaps/us/application.php, as shown in
Figure 8.4. This application generates detailed reports that provide a site-specific response spec-
trum; see Example 8.4.
As indicated in Chapter 5, Figure 5.13, the response spectrum is affected by local soil condi-
tions. Since SS and S1 are based on rock subsurface conditions, they must be adjusted for other

FIGURE 8.4  USGS seismic hazard mapping application input screen. (Courtesy of USGS.)
Seismic Code Provisions 215

types of soil conditions. ASCE-7, Chapter 20, classifies a site as Site Class A (hard rock), B (rock),
C (very dense soil and soft rock), D (stiff soil), E (soft soil), or F (soil) based on soil shear wave
velocity, vs; if vs is not known, ASCE-7 allows the use of standard penetration resistance or und-
rained shear strength values. This information is typically included as part of a geotechnical
engineering report. If no soil properties are known, ASCE-7 allows the use of Site Class D as a
conservative assumption.
The response spectrum associated with MCE SS and S1 can be adjusted to include site class
effects, SMS and SM1; the results are given as

S MS = Fa SS and S M 1 = Fv S1 (8.7)

where:
Fa is a site coefficient per ASCE-7, Table 11.4-1, that depends on soil type and ranges in values
from 0.8 for rock to 2.5 for soft soil
Fv is a site coefficient per ASCE-7, Table 11.4-2, that also depends on soil type and ranges in
values from 0.8 for rock to 3.5 for soft soil
SMS is the mapped risk-targeted maximum considered earthquake (MCER) spectral response
acceleration parameter at short periods
SM1 is the mapped MCER spectral response acceleration parameter at a period of 1 s

Also, the values Fa and Fv are smaller for large spectral accelerations, which is intended to cap-
ture the nonlinear behavior of soils that prevent ground motion amplifications.
The response spectrum associated with SMS and SM1 (based on MCER ground motion) is intended
to produce the most severe earthquake effects. MCE ground motion corresponds to a 2% probabil-
ity of being exceeded in a 50-year period (∼2%/50), which corresponds to an approximate return
period, or a recurrence interval of 2500 years, as obtained in Section 2.4. However, the nominal
design level is based on a maximum design earthquake (MDE) ground motion, which corresponds
to a 10% probability of being exceeded in a 50-year period, or approximately a 500-year recurrence
interval (∼10%/50). ASCE-7 accounts for two-thirds of the MCE ground motion spectrum accelera-
tions, SMS and SM1, in order to reduce the MCE ground motion from 2%/50 to 10%/50 hazard level;
the results are given as

S DS = 2 / 3 S MS and S D1 = 2 / 3 S M 1 (8.8)

where:
SDS is the design, 5% damped, spectral response acceleration parameter at short periods
SD1 is the design, 5% damped, spectral response acceleration parameter at a period of 1 s

The design response spectrum associated with SDS and SD1 can be obtained using ASCE-7,
Section 11.4.5. Draw the design spectral response acceleration, Sa, as a piecewise ­continuous func-
tion, with four segments as shown in Figure 8.5 or algebraically as follows:
For

 T
T ≤ To , Sa = S DS 0.4 + 0.6 
 To 

where the transition periods To and Ts are:

To = 0.2 Ts

216 Introduction to Earthquake Engineering

SDS

Design spectral acceleration, Sa (g)

T
Sa = SDS 0.4 + 0.6
To

SD1
Sa =
SD1 T SD1TL
Sa =
T2

SD1
Ts =
SDS

To = 0.2Ts Ts 1.0 TL
Structure period, T

FIGURE 8.5  Elastic design response spectrum.

Ts = SD1/SDS, which represents the transition period from acceleration- to velocity-controlled seg-
ments of the design spectrum (see Figure 5.8).
T is the fundamental period of the structure.
Note that for T = 0, the spectral acceleration, Sa, is approximately equal to 0.4SDS, which repre-
sents the design-level peak ground acceleration.

For To < T ≤ Ts, Sa = SDS

For Ts < T ≤ TL , Sa = (SD1/T)
and
S D1TL
For T > TL , Sa =
T2
The long-period transition period, TL , is obtained from ASCE-7, Figure 22-12, which is only appli-
cable to tall buildings, or for the sloshing of fluids in tanks. In addition, TL represents the transition
period from velocity- to displacement-controlled portions of the design spectrum (see Figure 5.8).

EXAMPLE 8.4
Determine the spectral response acceleration parameters and draw the response spectrum for a
site located on the campus of University of the Pacific in Stockton, California, which will be used
for the construction of an office building. The site (soil) is class D.

Solution
1. Determine the maximum considered spectral accelerations for short and 1 s periods, SS
and S1. From the seismic intensity maps in ASCE-7, Chapter 22, SS and S1 (as fractions of
acceleration due to gravity) are estimated from Figure 8.3,
SS = 0.85
S1 = 0.30
2. Determine the site coefficients for the given site class. For Site Class D, using ASCE-7,
Tables 11.4-1 and 11.4-2, identify the site coefficients.
Fa = 1.16 using linear interpolation of values in Tables 11.4-1
Fv = 1.80 using Tables 11.4-2

3. Determine the maximum considered spectral accelerations, adjusted for site effects, for
short and 1 s periods, SMS and SM1,
Seismic Code Provisions 217

SMS = Fa SS = 1.16(0.85) = 0.896
SM1 = Fv S1 = 1.80(0.30) = 0.54

4. Determine the design spectral accelerations for short and 1 s periods, SDS and SD1,
SDS = 2/3SMS = 2/3(0.896) = 0.65
SD1 = 2/3SM1 = 2/3(0.54) = 0.36

5. Draw the design response spectrum. The design response spectrum, Sa, is a piecewise
continuous function with four segments:

 T  T 
For T ≤ To , Sa = SDS 0.4 + 0.6  = 0.650.4 + 0.6  = 0.26 + 3.5 T
 To   0.11 s 

where the transition periods To and Ts are:

To = 0.2Ts = 0.11

SD1 0.36
Ts = = = 0.55 s
SDS 0.65

For To < T ≤ Ts, Sa = SDS = 0.65

For Ts < T ≤ TL, Sa = (SD1/T) = (0.36/T)
and
For T > TL, Sa = ( SD1TL /T 2 ) = (0.36TL /T 2 )

The long-period transition period TL obtained from ASCE-7, Figure 22-12, is equal to 8 s. Also,
notice that for T = 0, the spectral acceleration is approximately equal to 0.26, which represents the
design-level peak ground acceleration. Figure E8.3 shows the entire spectrum.
The sample report for this example from the USGS Seismic Hazards Mapping Application at http://
earthquake.usgs.gov/designmaps/us/application.php also provides the response spectrum (Figure E8.4).

0.6

0.5

0.4
Sa (g)

0.3

0.2

0.1

0
0 1 2 3 4 5 6 7 8 9 10
T (s)

FIGURE E8.3  Acceleration response spectrum at the University of the Pacific, Site Class D.
218 Introduction to Earthquake Engineering

FIGURE E8.4  USGS seismic hazard mapping application output.

Seismic Code Provisions 219

8.4.2  Permitted Lateral Analysis Procedures

SDCs are not only used to determine the allowable seismic force-resisting system and their build-
ing height limit as discussed in Section 8.1.3, but also to determine the permitted lateral analysis
procedure, as well as restrictions on buildings with irregularities, seismic detailing requirements,
and requirements for nonstructural components. SDCs classify building systems into categories A
through F based on their desired performance depending on a building’s occupancy, the effects of
probable ground-shaking intensity at the building site, and structural irregularities. The various
SDCs and hazard-level descriptions are:

A → Very low seismic risk

B → Low seismic risk
C → Moderate seismic risk
D → Moderate-to-high seismic risk
E → High seismic risk for structures in Occupancy Categories I, II, and III near an active
fault
F → High seismic risk for structures in Occupancy Category IV near an active fault

To determine the SDC for a structure using Table 8.1 (the most severe case of SDS, SD1, or S1 governs),
we must determine the soil site class using ASCE-7, Chapter 20, the MCE spectral response accel-
erations (SS and S1), the design spectral response accelerations (SDS and SD1) as described in Section
8.4.1, and finally the occupancy category using ASCE-7, Table 1.5-1 (discussed in the following
paragraphs), or use the USGS application with results provided in the detailed report.
ASCE-7 also classifies buildings based on the occupancy category and importance factor for the
building system. Occupancy categories are assigned to various buildings and other structures based
on their use and nature of occupancy (see ASCE-7, Table 1.5-1), and they are intended to increase
the safety level of essential facilities and those that pose a significant risk to human life. The nomi-
nal design level is Occupancy Category II, which corresponds to an MDE ground motion (10%/50).
The higher risk levels correspond to the probabilities listed in Table 8.2. As this table shows, essen-
tial and critical structures are designed for an MCE ground motion (2%/50), which is the level given
by the mapped spectrum parameters in ASCE-7, Chapter 22. Thus, the importance factors listed in
Table 8.2 effectively adjust the design response spectrum for various hazard levels.
The importance factors indirectly address the objectives of IBC seismic provisions discussed in
Section 8.2, which are related to the performance of systems for various levels of earthquake shak-
ing, that is,

1. Resist a minor earthquake with no damage.

2. Resist a moderate earthquake with some nonstructural damage, but no structural damage.
3. Resist a major earthquake with some structural damage, but no collapse.

TABLE 8.1
Seismic Design Categories; Summary of ASCE-7, Section 11.6
Based on SDS (Table 11.6-1) Based on SD1 (Table 11.6-2) I or II or III IV

SDS < 0.167 SD1 < 0.067 A A
0.167 ≤ SDS < 0.33 0.067 ≤ SD1 < 0.133 B C
0.33 ≤ SDS < 0.50 0.133 ≤ SD1 < 0.20 C D
SDS ≥ 0.50 SD1 ≥ 0.20 D D
S1 ≥ 0.75 E F
220 Introduction to Earthquake Engineering

TABLE 8.2
Occupancy Categories and Importance Factors
Occupancy
Category Nature of Occupancy Return Period (Years) Importance Factor, Ie
I Low hazard structures 475 1.00
(10% probability of exceedance in 50 years)
II Standard occupancy, other than I, III, or IV 475 1.00
(10% probability of exceedance in 50 years)
III Assembly structures 975 1.25
(5% probability of exceedance in 50 years)
IV Essential or critical structures 2475 1.50
(2% probability of exceedance in 50 years)

Occupancy Categories I and II provide the minimum level of protection and are intended to
address the risk of structural failure and loss of life, but not to limit structural damage, whereas
Occupancy Categories III and IV provide a higher level of protection against loss of life and prop-
erty by limiting the damage to a minimum in order to provide a continued function of facilities
during and after a seismic event.
The analysis procedures allowed by the code include ASCE-7, Section 12.8, ELFP as well as two
dynamic methods: ASCE-7, Section 12.9, MRS analysis, and ASCE-7, Chapter 16, seismic response
history procedures, all of which will be discussed in this chapter. The dynamic methods are permit-
ted for all building systems in all six SDCs. The ELFP is permitted for all building systems in SDCs
A, B, and C, and is allowed for most cases in SDCs D, E, and F, except buildings exceeding 160 ft
in height with T > 3.5 Ts and those with T < 3.5 Ts and horizontal irregularities 1a or 1b, or vertical
irregularities 1a, 1b, 2, or 3 (as noted in ASCE-7, Table 12.6-1). Even for the exempt cases, the ELFP
is used in scaling the results of the dynamic methods.
Building irregularities are discussed in ASCE-7, Section 12.3.2, and are defined as horizontal
(ASCE-7, Table 12.3-1) or vertical (ASCE-7, Table 12.3-2). Horizontal irregularities include: (1a)
torsional, (1b) extreme torsional, (2) reentrant corners, (3) diaphragm discontinuities, (4) out-of-
plane offsets, and (5) nonparallel resisting systems. Vertical irregularities include: (1a) stiffness-soft
story, (1b) stiffness-extreme soft story, (2) weight (mass), (3) vertical geometric, (4) In-plane discon-
tinuity in vertical lateral force-resisting element, (5a) discontinuity in lateral strength–weak story,
and (5b) discontinuity in lateral strength−extreme weak story.

8.5 EARTHQUAKE LOADS BASED ON ASCE-7

EQUIVALENT LATERAL FORCE PROCEDURE
As discussed in Section 8.4.2, earthquake loads can be obtained using an equivalent static method
(as presented here) or a full dynamic analysis approach presented in the following sections. The
main purpose of any seismic analysis is to ensure that strength and lateral deflection (drift) require-
ments are satisfied. ASCE-7 ELFP is similar to the generalized SDOF system analysis for shear
buildings presented in Chapter 6. The process involves determining the natural period, the base
shear, equivalent static story forces, story drifts, and overturning moment.
The ASCE-7 ELFP analysis steps are as follows (corresponding ASCE-7 sections are given in
parenthesis):

1. Compute seismic weight W using details presented in Section 3.1.2 and repeated here for
convenience (ASCE-7, Section 12.7.2).
Seismic Code Provisions 221

W = DL + 0.25 StL + larger of PL or 10 − psf + WPE + 0.2SL + La L (3.8)

where:
DL is the dead load of the structural system that is tributary to each floor
StL is the storage load, which is a live load in areas used for storage, such as a
warehouse
PL is the partition load, when applicable
WPE is the operating weight of permanent equipment
SL is the flat roof snow load when it exceeds 30-psf, regardless of roof slope
LaL is the landscape loads associated with roof and balcony gardens

2. Compute the natural period of the structure, T, following ASCE-7, Section 12.8.2. T is the
calculated fundamental (first mode) period of a structure in the direction under consider-
ation. It can be determined using the structural characteristics of the lateral force-resisting
elements as introduced in Section 3.1.1, or the generalized SDOF properties covered in
Chapter 6, or a full eigenanalysis covered in Chapter 7. When one of these methods is used
to determine T, it is subject to lower and upper limits as follows:

T = CuTa if T > CuTa

Use T if Ta ≤ T ≤ CuTa

T = Ta if T < Ta

where:
Cu is the coefficient for the upper limit on the calculated period from ASCE-7,
Table 12.8-1
Ta is the approximate fundamental period (ASCE-7, Section 12.8.2.1)

Alternatively, the code permits the use of Ta, calculated using the details presented in
Section 3.1.4, and repeated here for convenience,

Ta = Ct hnx (8.9)

where:
hn is the height in feet above the base to the highest level of the structure
Ct and x are determined from Table 3-2 (or ASCE-7, Table 12.8-2)

Ta is based on the measured response of buildings in high seismic regions; its value can
be adjusted to get a more accurate approximation of the actual period by including local
seismicity using Cu to obtain the upper bound period, CuTa. This is based on the best-fit of
the measured response.

3. Computer seismic base shear (ASCE-7, Section 12.8.1). The base shear is caused by the
inertial force (mass times acceleration) from the earthquake motion, F = ma = W(a/g),
where g is the acceleration due to gravity. ASCE-7 modifies this expression to obtain an
equivalent static lateral force acting at the base of the structure, the base shear, V:

V = C sW (8.10)
222 Introduction to Earthquake Engineering

where:
W is the seismic weight computed in step 1
Cs is the seismic response coefficient, expressed as a fraction of g (ASCE-7, Section
12.8.1.1)

S DS
Cs = (8.11)
R /I e

where:
SDS is the design spectral response acceleration in the short period range, computed
using the first of Equation 8.8 (ASCE-7, Section 11.4.4)
R is the response modification factor in ASCE-7, Table 12.2-1
Ie is the importance factor in Table 8.2 (ASCE-7, Section 11.5.1)
The value of Cs need not exceed

S D1 S D1TL
Cs = for T ≤ TL or Cs = for T > TL (8.12)
T ( R /I e ) T 2 ( R /I e )

Collectively, these three relationships for Cs represent the inelastic design response spec-
trum (Equation 8.11 is the constant acceleration portion and Equations 8.12 represents the
constant velocity and displacement portions, respectively), obtained by dividing the elastic
design response spectrum (shown in Figure 8.5) by the factor R/Ie. Figure 8.6 provides an
illustration of the resulting inelastic design response spectrum.
Also,

Cs > 0.044S DS I e or 0.01, whichever is larger (8.13)

or

0.5S1
Cs > , for S1 ≥ 0.6 g (8.14)
R /I e

SDS
(R/Ie) SD1 Cs (min) = 0.044SDSIe

T(R/Ie) or 0.01
0.5S1
or , for S1 ≥ 0.6g
R/Ie
Cs(g)

SD1 SD1TL
Ts =
SDS
T 2(R/Ie)

Ts TL
T (s)

FIGURE 8.6  Inelastic design response spectrum.

Seismic Code Provisions 223

where:
SD1 is the design spectral response acceleration at a period of 1 s, computed using the
second of Equations 8.8 (ASCE-7, Section 11.4.4)
T is the fundamental period of the structure computed in step 2
TL is the long-period transition period (ASCE-7, Section 11.4.5)
S1 is the MCE spectral response acceleration at a period of 1 s discussed in Section
8.4.1 (ASCE-7, Section 11.4.4)

The two limits in Equation 8.13 are intended to represent the minimum base shear force
levels to safeguard against the collapse of long-period structures (uncertainty related to
P-Δ effects), while Equation 8.14 accounts for the effects of near-fault directivity.

4. Compute the equivalent lateral forces (ASCE-7, Section 12.8.3). The lateral seismic force,
Fx, induced at any level x as illustrated in Figure 8.7, can be determined using the following
relationship:

Fx = CvxV (8.15)

where:
V is the base shear computed in step 3
Cvx is the vertical distribution factor

wx ⋅ hxk
Cvx = (8.16)
∑
n
wi ⋅ hik
i=1

where:
wx and wi are the weights at level x or i, depicted in Figure 8.7
h x and hi are the heights from the base to the level x or i, depicted in Figure 8.7
n is the number of stories
k depends on the structural period and accounts for higher mode effects
k = 1 for T ≤ 0.5 s

F4 w4 h4

k=2
F3 w3 h3

F2 w2 h2
k=1
F1 w1 h1

V
Mb

FIGURE 8.7  Equivalent lateral forces and overturning moment.

224 Introduction to Earthquake Engineering

k = 2 for 0.5 s < T < 2.5 s, or can be determined by linear interpolation between

1 and 2
k = 2 for T ≥ 2.5 s

These formulas are approximately the same as Equation 6.37 for a generalized SDOF
shear building, where we assumed that the lateral response is dominated by the funda-
mental mode. The shape vector, ψi, in this case is assumed to vary from linear (k = 1) to
parabolic (k = 2), as shown in Figure 8.7. The parabolic mode is intended to account for
the response of flexible structures.
5. Compute overturning moment (ASCE-7, Section 12.8.5). As discussed in Section 6.1.3,
with the equivalent static forces, we can determine the overturning moment by applying
static equilibrium to the free-body diagram depicted in Figure 8.7, that is,

Mb = ∑h F i i (8.17)
i=1

As shown in Figure 6.3, we can also obtain the internal story shear force, Vmax i, and the
internal story moment, Mmax i, at an arbitrary level i by applying static equilibrium to the
right-hand side free-body diagram as shown by Equations 6.38 and 6.39, respectively, and
are repeated as follows for convenience:

Vmax i = ∑f
j =i
jo

M max i = ∑ (h − h ) f
j =i
j i jo

Again, in the derivation of Equations 8.17 and 6.39, we assumed that the floor weights are
directly in the center of the building frame, thus making no contribution to the moment
equilibrium equation. However, for buildings with off-center weights, a full moment static
equilibrium analysis is conducted to account for the contribution of these off-center forces.
6. Determine story drifts (ASCE-7, Section 12.8.6). With the lateral forces, we can perform
a structural analysis to determine lateral displacements. The relative displacement of
each story at the center of mass is defined as the drift. For cases where centers of mass
do not align and for those cases with irregularities, restrictions are specified in ASCE-7.
Note that the loads used to calculate drifts are at strength level (not service), as indicated
in ASCE-7, Figure 12.8-2. The drift is computed using the following inelastic displacement
at level x:

Cdδxe
δx = (8.18)
Ie

where:
Cd is the deflection amplification factor discussed in Section 8.3.2 and given in ASCE-7,
Table 12.2-1
Seismic Code Provisions 225

Ie is the importance factor introduced in Section 8.4.2 and used in step 3

δ xe is the deflection at level x determined from an elastic analysis
The drift is then computed as

∆ x = δx − δx−1 (8.19)

where:
δ x is the deflection at the top of level x determined using Equation 8.18
δ x−1 is the deflection at the bottom of level x determined using Equation 8.18

To ensure the stability of the overall system (i.e., avoid P-Δ effects) and to minimize dam-
age to nonstructural elements, these drifts must be kept within allowable limits per ASCE-7,
Section 12.12.1, which are listed in Table 12.12-1 and depend on the structure type and
occupancy category.

When P-Δ effects are significant, provisions in ASCE-7, Section 12.9.6, should be followed to
ensure stability. For cases where torsional effects are significant, ASCE-7, Section 12.8.4.3, requires
the application of torsional amplification factors in the analysis.

EXAMPLE 8.5
Given the following three-story office building frame located on the campus of University of
the Pacific in Stockton, California, compute the approximate structural period Ta, the design
base shear V, the design earthquake loads acting on each floor Fx, and the overturning moment
at the base. The building in plan view is square; the structural system being proposed is a
special steel concentrically braced frame (SSCBF) as shown in Figure E8.5 with an estimated
weight on each of the first two floors of 100 and 80 kips on the roof. The geotechnical engineer
has classified the site (soil) class as D.

Solution
1. Determine the seismic weight. The weight of the roof and two floors is given; the total
weight of the system is the sum of these weights,

W = 100 kips + 100 kips + 80 kips = 280 kips

2. Determine the period. No information about structural stiffness is given, so we use the
approximate period given by Equation 8.9,

Ta = Ct hnx

12 ft

12 ft

12 ft

FIGURE E8.5  Elevation view of the special steel concentrically braced frame.
226 Introduction to Earthquake Engineering

where:
hn = 36 ft and is the building height
for an SSCBF, which corresponds to all other structural systems in Table 3.2,
Ct = 0.02
x = 0.75
Thus,

Ta = Ct hnx = 0.02(36 ft )0.75 = 0.29 s

Note that this value is very close to the result of Equation 3.19, Ta = 0.1N = 0.1(3) = 0.3 s.
This relationship can be used to check the validity of the calculated values of the period.

3. Calculate the base shear, V = CsW per ASCE-7, Section 12.8. The base shear is obtained
using Equation 8.10, with Equations 8.11 and 8.12, or the response spectrum shown in
Figure 8.6. Using the spectral values obtained in Example 8.4, draw the response spec-
trum for this case as shown in Figure E8.6.
With Ta = 0.29 s, we can enter this spectrum. Since To < Ta ≤ Ts, Sa = SDS = 0.65 and

SDS
Cs =
R /Ie

As noted in the problem statement, the building will be used for office space, which
corresponds to an occupancy category of II; thus, the importance factor obtained from
Table 8.2 is

Ie = 1.0

Also, we can determine the SDC using Table 8.1 (or the USGS application discussed in
Example 8.4) to ensure that the selected seismic load-resisting system is permitted, that

0.6

0.5

0.4
Sa (g)

0.3

0.2

0.1

0
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
Ta T (s)

FIGURE E8.6  Design acceleration response spectrum.

Seismic Code Provisions 227

is, for Occupancy Category II, and SDS = 0.65 > 0.5 and SD1 = 0.36 > 0.2, the governing
SDC is D, as shown in the following table.

Based on SDS (Table 11.6-1) Based on SD1 (Table 11.6-2) I or II or III IV

SDS < 0.167 SD1 < 0.067 A A
0.167 ≤ SDS < 0.33 0.067 ≤ SD1 < 0.133 B C
0.33 ≤ SDS < 0.50 0.133 ≤ SD1 < 0.20 C D
SDS ≥ 0.50 SD1 ≥ 0.20 D D
S1 ≥ 0.75 E F

Thus, for SSCBF (ASCE-7, Table 12.2-1, case 2 in part B), the height limit is 160 ft, which
is larger than the proposed 36 ft building height. The seismic response coefficient is also
obtained from ASCE-7, Table 12.2-1,

R = 6.0

The base shear is calculated as follows:

SDS 0.65
V = C sW = W= ( 280 kips) = 30.7 kips
R /Ie 6.0 /1.0

4. Determine lateral forces at each level of the structure, Fx = CvxV per ASCE-7, Section
12.8.3. The forces are obtained using Equations 8.15 and 8.16 for each level, x:

Fx = CvxV

where:
V = 30.7 kips, computed in step 3
Cvx is the vertical distribution factor with k = 1 since Ta = 0.29 s ≤ 0.5 s

w x ⋅ hx w x ⋅ hx w x ⋅ hx
Cvx = = =
∑
3
w i ⋅ hi 100 kips(12 ft ) + 100 kips( 24 ft ) + 80 kips(36 ft ) 6480 k ⋅ ft
i =1

So the lateral forces are:

w1 ⋅ h1 100 kips(12 ft )
F1 = Cv1V = (30.7 kips) = (30.7 kips) = 5.68 kips
6480 k ⋅ ft 6480 k ⋅ ft

w 2 ⋅ h2 100 kips( 24 ft )
F2 = Cv 2V = (30.7 kips) = (30.7 kips) = 11.36 kips
6480 k ⋅ ft 6480 k ⋅ ft

w 3 ⋅ h3 80 kips(36 ft )
F3 = Cv 3V = (30.7 kips) = (30.7 kips) = 13.63 kips
6480 k ⋅ ft 6480 k ⋅ ft

5. Determine the overturning moment using static equilibrium. With the lateral forces com-
puted in the last step, we can determine the overturning moment at the base using
Equation 8.17,
3

Mb = ∑ h F = 5.68 kips(12 ft) + 11.36 kips(24 ft) + 13.63 kips(36 ft)) = 831kip ⋅ ft
i i

i =1

228 Introduction to Earthquake Engineering

13.63 kips
12 ft
11.36 kips
13.6 –490
12 ft
5.68 kips
25 –763
12 ft

831 kip-ft 30.68 kips 30.7 –831

V (kips) M (kip-ft)

FIGURE E8.7  Summary of analysis results using the equivalent lateral force method.

We can also obtain the internal story shear forces and moments at each level using
Equations 6.38 and 6.39, respectively. Alternatively, we can draw shear force and bend-
ing moment diagrams by treating the building as a cantilever beam in order to obtain
the internal story shear forces and moments. Figure E8.7 summarizes the results of the
analysis.

EXAMPLE 8.6
Check the allowable story drifts for the three-story office building frame given in Example 8.5.
A structural analysis of the building frame was conducted with lateral forces determined in Example
8.5 in order to determine the elastic displacements at each level (δ1e = 0.65 in, δ2e = 1.3 in, and
δ3e = 1.6 in) and illustrated in Figure E8.8.

SOLUTION
1. Determine inelastic displacements per ASCE-7, Section 12.8.6. Compute inelastic dis-
placements at each level x using Equation 8.18,

Cd δxe
δx =
Ie

where:
Cd = 5.0 for steel special concentrically braced frames given in ASCE-7, Table 12.2-1
Ie = 1 for the office building
δxe is the deflection at level x given in Figure E8.8.

So the inelastic displacements are:

δ1 = Cd δ1e /Ie = 5(0.65 in) /1 = 3.25 in

δ2 = Cd δ2e /Ie = 5(1.3 in) /1 = 6.5 in

13.63 kips 1.6 in

12 ft
11.36 kips 1.3 in
12 ft
5.68 kips
0.65 in
12 ft

FIGURE E8.8  Lateral forces and elastic displacements at each level of the building frame.
Seismic Code Provisions 229

δ3 = Cd δ3e /Ie = 5(1.6 in) /1 = 8.0 in

2. Determine story drifts. Compute drifts at each level using Equation 8.19,

∆1 = δ1 − δ1−1 = δ1 − δ0 = 3.25 in − 0 = 3.25 in

∆2 = δ2 − δ2−1 = δ2 − δ1 = 6.5 in − 3.25 in = 3.25 in

∆3 = δ3 − δ3−1 = δ3 − δ2 = 8 in − 6.5 in = 1.5 in

3. Check the allowable story drifts per ASCE-7, Section 12.12.1. For buildings in Occupancy
Category II, ASCE-7, Table 12.12-1, gives an allowable drift (for four stories or less struc-
tures with interior walls, partitions, etc. designed to accommodate story drifts) of

∆a = 0.025 hsx

where:
hsx is the story height for level x. In this case, each level is 12 ft high; thus,

∆a = 0.025 hsx = 0.025(12 ft )(12 in/ft ) = 3.6 in

None of the story drifts exceed the allowable drift.

8.6  MODAL RESPONSE SPECTRUM ANALYSIS

As discussed in Section 8.4, earthquake loads can be obtained using one of two structural dynam-
ics-based methods, the MRS (discussed in this section) or response time-history procedure (briefly
covered in the next section) analyses. ASCE-7 MRS analysis is roughly based on the MDOF system
analysis for shear buildings presented in Chapter 7. After determining the modal properties for each
mode (natural period, shape, modal participation factor, and effective modal weight), the remainder
of the process is similar to the ELFP method in that it involves determining the base shear, equiva-
lent static story forces, story drifts, and overturning moments for each participating mode. The
number of participating modes in each horizontal direction should be enough to obtain the modal
weight participation of at least 90% of the actual weight.
The ASCE-7 MRS analysis steps are as follows and are similar to the summary presented in
Section 7.2 (corresponding ASCE-7 sections are given in parentheses). Note that many of the steps
required for the ELFP method are the same as the ones that follow, and many of the additional steps
are usually performed by a computer.

1. Determine the mass matrix [m] from the floor and roof seismic weight using details pre-
sented in Section 3.1.2 (ASCE-7, Section 12.7.2).
2. Determine the stiffness matrix [k] from the properties of the structural system using details
presented in Section 7.1.
3. With [m] and [k], solve the associated eigenvalue problem to determine modal properties
for each mode using details presented in Section 7.1.1:
a. Natural frequencies ωnj and periods Tnj using the eigenvalues ω 2j .
b. Mode shapes φij using normalized eigenvectors.
230 Introduction to Earthquake Engineering

c. Modal participation factors Γj (which are the contribution of each mode shape to the
total response) using Equation 7.50, repeated here for convenience,

∑
n
L miφij [Φ]T [ m]{1}
Γj = j = i=1
or {Γ} =
∑
n
Mj miφij2 [Φ]T [ m][Φ]
i=1

Where we have assumed that mii = mi since the mass matrix is diagonal.
d. Effective modal weight using Equation 7.66, repeated here for convenience,

 2
∑
n

 wiφij 

∑
i=1 n
W je = or W je = Γ j wiφij or W je = Γ j L j g
∑
n
2 i=1
wφ
i ij
i=1

where:
wi = mi ⋅ g, which is the fraction of the building seismic weight at level i
n is the number of stories

4. Determine the number of modes to be used in the analysis (ASCE-7, Section 12.9.1).
Use a sufficient number of modes φij in each direction to account for at least 90% of the
actual weight W.
5. Develop the design response spectrum and determine the spectral accelerations for partici-
pating modes.
a. Develop an elastic response spectrum using details presented in Section 8.4.1, Figure
8.5 (ASCE-7, Section 11.4.5).
b. Determine spectral accelerations for the mth mode, Sam, using the elastic response
spectrum and the corresponding periods computed in step 3 for each contributing
mode found in step 4.
c. Divide Sam by the quantity R/Ie to obtain the modal seismic response coefficient,

Sam
Csm = (8.20)
R /I e

where:
R is the response modification factor in ASCE-7, Table 12.2-1
Ie is the importance factor in Table 8.2 (ASCE-7, Section 11.5.1)

6. Determine the system base shear, Vt.

a. Compute seismic modal base shear Vm for each of the participating modes; that is, the
fraction of the total base shear contributed by the mth mode of vibration is
n

Vm = CsmWm = Csm Γm ∑wφ i im

(8.21)
i =1
where:
Csm is given by Equation 8.20 and is usually expressed as a fraction of g
Wm is the effective modal weight for the participating modes computed in step 3d
Seismic Code Provisions 231

wi is the weight located at level i

φim is the displacement at the ith level for the mth mode
b. Statistically combine the modal base shears using SRSS (Equation 7.68) or CQC
(Equations 7.69 and 7.70) to obtain the system base shear Vt (Section 12.9.3).
c. Ensure that Vt is larger than a minimum value based on Section 8.5 ELFP analysis, V
(ASCE-7, Section 12.9.4.1). If Vi is less than 85% of V (with T = CuTa when T > CuTa),
then lateral force, story shears, and story moments determined using the modal analy-
sis must be multiplied by 0.85 V/Vt. Cu and Ta are discussed in Section 8.5.
7. Determine the lateral forces at each level x (ASCE-7, Section 12.8.3). The fraction of the
total lateral force contributed by the mth mode of vibration at each level x, Fxm, is

Fxm = CvxmVm (8.22)

where:
Vm is the seismic modal base shear computed in step 6a
Cvxm is the vertical distribution factor

wxφxm
Cvxm = (8.23)
∑
n
wiφim
i=1

where:
wx and wi are the weights at levels x and i, depicted in Figure 8.7
φim and φxm = ith and xth level displacements for the mth mode

We can then statistically combine these modal lateral forces using SRSS (Equation 7.68) or
CQC (Equations 7.69 and 7.70) to obtain the design lateral forces at each level Fx (ASCE-7,
Section 12.9.3).
8. Compute overturning moment. With the modal lateral forces obtained in the last step Fxm,
we can determine the modal overturning moments by applying static equilibrium,

M bm = ∑h F i im (8.24)
i=1

We can then statistically combine these modal overturning moments using SRSS (Equation
7.68) or CQC (Equations 7.69 and 7.70) to obtain the design overturning moment at the
base Mb (ASCE-7, Section 12.9.3).

9. Determine story drifts. Story drifts are the relative displacements of each story, computed
by taking the difference between displacements of two consecutive floors. Note that the
modal lateral loads used to calculate drifts are at the strength level (not service), as indicated
in ASCE-7, Figure 12.8-2.
a. The fraction of modal inelastic displacement contributed by the mth mode of vibration
at level x is

Cdδxem
δxm = (8.25)
Ie
232 Introduction to Earthquake Engineering

where:
Cd is the deflection amplification factor discussed in Section 8.3.2 and given in
ASCE-7, Table 12.2-1
Ie is the importance factor introduced in Section 8.4.2
δ xem is the deflection at level x in the mth mode, and can be determined using the
following relationship:

 g  T 2F
δxem =  2  m xm (8.26)
 4π  wx

where:
g is the acceleration due to gravity
Tm is the period of the mth mode obtained following step 3a
wx is the weight at level x
Fxm is the modal lateral forces determined in step 7
b. Statistically combine the modal inelastic displacements using SRSS (Equation 7.68) or
CQC (Equations 7.69 and 7.70) to obtain the design displacements.
c. The drift is then computed as

Δ xm = δxm − δxm−1 (8.27)

where:
δ xm is the deflection at the top of level x determined using Equation 8.25
δ xm−1 is the deflection at the bottom of level x determined using Equation 8.25
d. Statistically combine the modal drifts using SRSS (Equation 7.68) or CQC (Equations
7.69 and 7.70) to obtain the design drifts.

EXAMPLE 8.7
A four-story office building frame, with equal story heights of 12 ft, is located at a site with site soil
class D, SD1 = 0.75 and SDS = 1.33. The structural system is a special reinforced concrete moment
frame with the following stiffness matrix [k] and story weights of w1 = w2 = w3 = 789 kips and
w4 = 645 kips (Villaverde 2009). Use the MRS analysis to compute the design earthquake loads
acting on each floor, the story drifts, and the overturning moment:

 5611 −2916 458 −45  789 0 0 0 

   
−2916 3863 −1885 269   0 789 0 0 
[k ] =  kips/in and [w ] =  kips
 458 −1885 2893 −1398  0 0 78
89 0 
 −45 269 −1398 1168   0 0 0 645
 

Solution
1. Determine the mass matrix [m]. Use the floor and roof seismic weights given in the
problem statement (the matrix can be obtained using D’Alembert’s principle as shown
in Chapter 7, Example 8.4). The first floor is listed in the first row, while the roof is listed
along the last row (some books will reverse this order).

 2.04 0 0 0 
 
 0 2.04 0 0 
[m] = [w ]/g =  kips ⋅ s2 /in
 0 0 2.04 0 
 0 0 0 1.67

Seismic Code Provisions 233

where:
g = 386.4 in/s2

2. Determine the stiffness matrix [k]. Use the stiffness matrix given in the problem
­statement (this matrix can be obtained using techniques from matrix structural analysis).
3. With [m] and [k], solve the associated eigenvalue problem to determine modal proper-
ties for each mode using MATLAB®:
a. Natural frequencies ωnj and periods Tnj using the eigenvalues ω 2j .

 8.8  0.713

 
 
26.1  2π   0.241
{ωn } =   rad/s and {Tn } =   =  s
 44.1  ωn  0.143
 
63.7 0.099
 

b. Mode shapes φij using the normalized eigenvectors [Φ].

 0.2191 −0.6832 1.0193 −6.4994

 
0.5257 −1.0288 0.3319 5.5143 
[Φ] = 
0.8369 −0.1517 −1.4521 −2.7391
 1.0 1.0 1.0 1.0 


c. Modal participation factors {Γ} using Equation 7.50, {Γ} = [Φ]T[m]{1}/[Φ]T[m][Φ].

First, determine the modal mass matrix, [M] = [Φ]T[m][Φ],

 0.2191 0.5257 0.8369 1.0  2.04 4 0 0 0 

  
−0.6832 −1.0288 −0.1517 1.0  0 2.04 0 0 

[M] =  
 1.0193 0.3319 −1.4521 
1 .0   0 0 2.04 0 
−6.4994
 5.5143 −2.7391 1.0  0 0 0 1.67
 0.2191 −0.6832 1.0193 −6.4994
 
0.5257 −1.0288 0.3319 5.5143 
×  kips ⋅ s2 /in
0.8369 −0.1517 −1.4521 −2.7391
 1.0
 1.0 1.0 1.0 
3.76 0 0 0 
 
 0 4.83 0 0 
[M] =  kips ⋅ s2 /in
 0 0 8.32 0 
 0 0 0 165.35


And the modal force, [L] = [Φ]T[m]{1},

 0.2191 0.5257 0.8369 1.0  2.04 4 0 0 0  1

   
−0.6832 −1.0288 −0.1517  
1 .0   0 2.04 0 0  1
[L] =  2
  kips ⋅ s /in
 1.0193 0.3319 −1.4521 1.0  0 0 2.04 0  1
−6.4994 
 5.5143 −2.7391 1.0  0 0 0 1.67 1

234 Introduction to Earthquake Engineering

 4.90 

 
−2.14 2
[L] =   kips ⋅ s /in
 1.46 
 
−5.94

Next determine the participation factors,

1/ 3.76 0 0 0   4.90 

   
 0   
{Γ} = [M]−1{L} = 
1/ 4.83 0 0  kips ⋅ s2 /in −2.14 kips ⋅ s2 /in
  
 0 0 1/ 8.32 0   1.46 
 0 0 0 1/165.35 −5.94
  

 1.3023 

 
−0.4422
{Γ} =  
 0.1758 
 
−0.0359

d. Calculate the effective modal weight using Equation 7.66, Wje = Γ j ∑ ni=1 w iφij .

∑
4
W1e = Γ1 w iφi1 = 1.3023(789(0.219 + 0.526 + 0.837) + 645(1)) = 2466 kips
i =1

Similarly,

W2e = 365 kips

W3e = 99 kips

W4e = 82 kips

4. Determine the number of modes to be used in the analysis. Use enough modes φij in
each direction to capture at least 90% of the actual weight W. The total weight of the
system is W = 4(789 kips) + 645 kips = 3012 kips. The percentages of the total weight
per mode are as follows:
Mode 1, W1e /W ×100 = 2466 kips/ 3012 kips = 81.8%
Mode 2, W2e /W ×100 = 365 kips/ 3012 kips = 12.1%
Mode 3, W3e /W ×100 = 99 kips/ 3012 kips = 3.3%
Mode 4, W4e /W ×100 = 82 kips/ 3012 kips = 2.7%
Therefore, we only need to include the first two modes, which capture 93.9% of W.
5. Develop the design response spectrum and determine spectral accelerations for partici-
pating modes.
a. Develop an elastic response spectrum using details presented in Section 8.4.1,
Figure 8.5.
Seismic Code Provisions 235

1.5

SDS = 1.33
1.25

T
Sa = SDS 0.4 + 0.6 = 0.53 = 7.25T
1 To
Sa (g)

0.75 SD1 = 0.75

SD1 0.75
0.5 Sa = =
T T

0.25
SD1 0.75
Ts = = = 0.56 s
SDS 1.33
0
0 0.5 1 1.5 2 2.5 3 3.5 4
To = 0.2Ts = 0.11 T (s)

FIGURE E8.9  Design acceleration response spectrum.

b. Determine spectral accelerations using Figure E8.9 and the periods computed in
step 3 for each contributing mode found in step 4. Enter the graph with T1 and T2
(since only the first two modes are required) to acquire spectral accelerations for the
two participating modes:
T1 = 0.713 s > Ts = 0.56 s; therefore, Sa1 = SD1/T1 = 0.75/0.713 = 1.05
T2 = 0.241 s < Ts = 0.56 s; therefore, Sa2 = 1.33
c. Divide Sam by R/Ie to obtain the design modal seismic response coefficients, Csm,
Ie = 1.0 from Table 8.2, since an office building is classified with an Occupancy
Category II. R = 8, for a special moment reinforced concrete frame (case C5 in
ASCE-7, Table 12.2-1).
Thus,
Sa1 1.05
C s1 = = = 0.13
R /Ie 8 /1

Sa2 1.33
C s2 = = = 0.17
R /Ie 8 /1
6. Determine the system base shear, Vb.
a. Compute the seismic modal base shear for each of the participating modes with Csm
from step 5c and Wm from step 3d:

V1 = C s1W1 = 0.13( 2465 kips) = 320 kips

V2 = C s 2W2 = 0.17(365 kips) = 62 kips

b. Combine V1 and V2 using SRSS (Equation 7.68) to obtain the base shear Vt = ΣVi 2 .

Vt = (320 kips)2 + (62 kips)2 = 326 kips

236 Introduction to Earthquake Engineering

c. Verify that Vt is larger than a minimum base shear value, V, obtained with an ELFP
analysis described in Section 8.5. If Vt is less than 85% of V (with T = CuTa when
T > CuTa), then lateral force, story shears, and story moments determined using the
modal analysis should be multiplied by 0.85 V/Vt.
First, calculate the base shear V using the ELFP method, with T = CuTa:
Cu = 1.4 since SD1 > 0.4 (ASCE-7, Table 12.8-1)
Ta = Ct hnx as discussed in Section 8.5
where:
hn = 48 ft and is the building height
for a concrete moment-resisting frame in Table 3.2 (ASCE-7, Table 12.8-2)

Ct = 0.016

x = 0.9

So,

Ta = Ct hnx = 0.016(48 ft)0.9 = 0.53 s

Thus,

CuTa = 1.4(0.53 s) = 0.73 s > Ts = 0.56 s;  therefore, Sa = SD/T = 0.75/0.73 = 0.13

and

V = C sW = 0.13(3012 kips) = 387 kips

Furthermore,

V = 0.85(387 kips) = 329 kips > Vt = 326 kips
0.85 

Therefore, we must multiply Vt by 0.85 V/Vt to obtain the actual design base

shear:

Vb = (Vt)(0.85 V/Vt) = 0.85 V = 329 kips

7. Determine the lateral forces, Fx.

a. Use Equations 8.22 through 8.24 to determine the modal lateral force Fxm: Mode 1
forces at each level x,

Fx1 = Cvx1V1

where:
V1 = 0.85 V/Vt (320 kips) = 1.01(320 kips) = 323 kips, computed in step 6a

w xφx1 w xφx1 w xφx1

Cvx1 = = =
∑
n
w iφi1 789 kips(0.219 + 0.526 + 0.837) + 645 kips 1890 k ⋅ ft
i =1

So the lateral forces are:

w1φ11 789 kips(0.219)

F11 = (323 kips) = (323 kips) = 30 kips
1890 k ⋅ ft 1890 k ⋅ ft

Seismic Code Provisions 237

w 2φ21 789 kips(0.526)

F21 = (323 kips) = (323 kips) = 71kips
1890 k ⋅ ft 1890 k ⋅ ft

w 3φ31 789 kips(0.837)

F31 = (323 kips) = (323 kips) = 113 kips
1890 k ⋅ ft 1890 k ⋅ ft

w 4φ41 641kips(1)
F41 = (323 kips) = (323 kips) = 110 kips
1890 k ⋅ ft 1890 k ⋅ ft

Mode 2 forces at each level x are similarly computed:

Fx 2 = Cvx 2V2

where:
V2 = 0.85 V/Vt (62 kips) = 1.01(62 kips) = 62.6 kips, computed in step 6a

w xφx 2 w xφx 2 w xφx 2

Cvx 2 = = =
∑
n
w iφi 2 789 kips(−0.683 − 1.029 − 0.152) + 645 kips −826k ⋅ ft
i =1

So the lateral forces are:

w1φ12 789 kips(−0.152)

F12 = (62.6 kips) = (62.6 kips) = 41kips
−826 k ⋅ ft −826 k ⋅ ft

w 2φ22 789 kips(−1.029)

F22 = (62.6 kips) = (62.6 kips) = 61kips
−826 k ⋅ ft −826 k ⋅ ft

w 3φ32 789 kips(−0.683)

F32 = (62.6 kips) = (62.6 kips) = 9.1kips
−826 k ⋅ ft −826 k ⋅ ft

w 4φ42 641kips(1)
F42 = (62.6 kips) = (62.6 kips) = −48 kips
−826k ⋅ ft −826 k ⋅ ft

b. Combine lateral forces for mode 1, Fx1, and mode 2, Fx2, using SRSS (Equation 7.68)
to obtain the lateral forces, Fx = ΣFxi2 .

F1 = (30 kips)2 + ( 41kips)2 = 51kips

F2 = (71kips)2 + (61kips)2 = 94 kips

F3 = (113 kips)2 + (9.1kips)2 = 113 kips

F4 = (110 kips)2 + (−48 kips)2 = 120 kips

238 Introduction to Earthquake Engineering

8. Determine the overturning moment.

a. With the modal lateral forces obtained in the last step, Fxm, we can determine the
modal overturning moments by applying static equilibrium for each mode: The
mode 1 overturning moment at the base is

Mb1 = ∑ h F = 30(12) + 7124

i i1 ( ) + 113(36) + 110( 48) = 11, 412 k ⋅ ft
i =1

Mode 2 overturning moment at the base:

Mb 2 = ∑hF i i2 = 4112
( ) + 6124
( ) + 9.136
( ) − 48(48) = −20 k ⋅ ft
i =1

b. Combine Mb1 and Mb2 using SRSS (Equation 7.68) to obtain the system overturning
moment at the base Mb = ΣMbi2 .

Mb = (11, 412 k ⋅ ft )2 + (−20 k ⋅ ft )2 = 11, 412 k ⋅ ft

Note the factor 0.85 V/Vt was included in the forces.

9. Determine story drifts.
a. The modal inelastic displacement, δxm, at level x is

Cd δxem
δxm =
Ie

where:
Cd = 5.5 for special moment reinforced concrete frames given in ASCE-7, Table
12.2-1
Ie = 1 for an office building

 g  T 2F
δxem =  2  m xm
 4π  w x

where:
Tm is the period of the mth mode obtained in step 3a
wx is the weight at level x from step 1
Fxm is the modal lateral forces obtained in step 7a without the factor 0.85(V/Vt)

Mode 1 inelastic displacement is calculated as follows:

Cd 5.5  386.4 in/s 2  (0.713 s)2Fx1 F

δx1 = δxe1 =   = 27.37 in x1
Ie 1  4π 2  wx wx

The inelastic displacements at each level x are then determined as follows:

F11 29.3 kips

δ11 = 27.37 in = 27.37 in = 1.02 in
w1 789 kips

Seismic Code Provisions 239

F21 70.3 kips

δ21 = 27.37 in = 27.37 in = 2.44 in
w2 789 kips

F31 111.8 kips

δ31 = 27.37 in = 27.37 in = 3.88 in
w3 789 kips

F41 108.6 kips

δ41 = 27.37 in = 27.37 in = 4.61in
w4 645 kips

Similarly, mode 2 inelastic displacement is calculated as follows:

Cd 5.5  386.4 in/s 2  (0.241 s)2Fx 2 F

δx 2 = δxe 2 =   = 3.13 in x 2
Ie 1  4π 2  wx wx

The inelastic displacements at each level x are then determined as follows:

F12 40.4 kips

δ12 = 3.13 in = 3.13 in = 0.16 in
w1 789 kips

F22 60.8 kips

δ22 = 3.13 in = 3.13 in = 0.24 in
w2 789 kips

F32 9.0 kips

δ32 = 3.13 in = 3.13 in = 0.036 in
w3 789 kips

F42 −48 kips

δ42 = 3.13 in = 3.13 in = −0.23 in
w4 645 kips

b. Combine inelastic displacements using SRSS (Equation 7.68) to obtain total inelastic
displacements δx = Σδxi2 .

δ1 = (1.02 in)2 + (0.16 in)2 = 1.03 in

δ2 = ( 2.44 in)2 + (0.24 in)2 = 2.45 in

δ3 = (3.88 in)2 + (0.036 in)2 = 3.88 in

δ4 = ( 4.61in)2 + (−0.23 in)2 = 4.61in

c. Determine drifts at each level using Equation 8.19. The story drifts associated with
mode 1 are determined below:

Δ11 = δ11 − δ1−11, = δ11 − δ01 = 1.02 in − 0 = 1.02 in

240 Introduction to Earthquake Engineering

Δ21 = δ21 − δ2−11, = δ21 − δ11 = 2.44 in − 1.02 in = 1.42 in

Δ31 = δ31 − δ3−11, = δ31 − δ21 = 3.88 in − 2.44 in = 1.44 in

Δ41 = δ41 − δ4−11, = δ41 − δ31 = 4.61in − 3.88 in = 0.73 in

Similarly, the story drifts associated with mode 2 are determined as follows:

Δ12 = δ12 − δ1−1,2 = δ12 − δ02 = 0.16 in − 0 = 0.16 in

Δ22 = δ22 − δ2−1,2 = δ22 − δ12 = 0.24 in − 0.16 in = 0.08 in

Δ32 = δ32 − δ3−1,2 = δ32 − δ22 = 0.036 in − 0.24 in = −0.21in

Δ42 = δ42 − δ4−1,2 = δ42 − δ32 = −0.23 in − 0.036 in = −0.27 in

d. Combine drifts using SRSS (Equation 7.68) to obtain total drifts Δ x = ΣΔ2xi .

Δ1 = (1.02 in)2 + (0.16 in)2 = 1.03 in

Δ2 = (1.42 in)2 + (0.08 in)2 = 1.42 in

Δ3 = (1.44 in)2 + (−0.21in)2 = 1.45 in

∆ 4 = (0.73 in)2 + ( −0.27 in)2 = 0.78 in

e. Check allowable story drifts per ASCE-7, Section 12.12.1 and Table 12.12-1. For
buildings in Occupancy Category II, this table gives an allowable drift (for four sto-
ries or less structures with interior walls, partitions, etc. designed to accommodate
story drifts) of

Δa = 0.025 hsx

where:
hsx is the story height for level x. In this case each level is 12 ft high; thus,

Δa = 0.025 hsx = 0.025(12 ft )(12 in/ft ) = 3.6 in

Therefore, none of the story drifts exceed the allowable drift!

8.7  INTRODUCTION TO SEISMIC RESPONSE HISTORY PROCEDURES

The code allows both linear and nonlinear seismic response history analyses following specifica-
tions in ASCE-7, Chapter 16. Here, we briefly introduce the linear analysis procedure; details of the
nonlinear analysis are beyond the scope of this textbook. In both cases, we first develop a mathemat-
ical model of the structural system, which typically involves a finite element model. After creating
the mathematical model (which can be two- or three-dimensional), we must select a suite of three or
more ground motion acceleration histories that are compatible with the design response spectrum
for the site. Appropriate ground motion records include those that have magnitudes, fault distance,
and source mechanisms consistent with those that control the maximum considered ground motion
at the site. Simulated ground motion records may be used when recorded ground motions are not
Seismic Code Provisions 241

available. In order to include the effects of ductility and adjust the design response for various haz-
ard levels, the response parameters should be divided by R/Ie.
The mathematical model of the structure is analyzed by directly solving the coupled equations
of motion discussed in Chapter 7, Equation 7.45, but including damping,

[ m]{u} + [c]{u} + [ k ]{u} = −[ m]{1}ug [8.28]

These equations can be solved using one of two procedures: direct analysis or modal analysis.
Direct analysis requires the numerical integration of the coupled equations of motion using one of
several time-stepping techniques, such as Newmark’s method. This requires the explicit formation of
the system damping matrix [c], which is not required in a modal analysis. The numerical integration of
MDOF equations is similar to the process for SDOF systems covered in Sections 4.4 and 4.5, but can
be much more complex depending on the number of degrees of freedom (see Villaverde 2009 for more
details). The modal analysis, on the other hand, uses mode shapes to transform the coupled MDOF
system of equations into a series of SDOF equations (with modal amplitudes as the unknowns), one
for each mode—this is similar to an MRS analysis. After the amplitudes are determined, they are
transformed back to the nodal displacements using mode shapes. The two approaches produce equal
results; however, the modal response history analysis method enjoys several advantages over the
direct integration method, including the fact that each SDOF equation can be solved using the meth-
ods covered in Sections 4.4 and 4.5, the explicit formation of the system damping matrix [c] is not
required, and accurate results may be obtained using a limited number of modes (Villaverde 2009).
Once the deformation response history is obtained, a complete analysis can be performed fol-
lowing the approach presented in Section 3.3. That is, after determining the maximum dynamic
displacement from a dynamic excitation, we can conduct a static structural analysis to determine
element forces (bending moment, shear force, and axial force) and stresses needed for design; no
additional dynamic analysis is necessary.

PROBLEMS
8.1 A column of a steel-braced frame in a single-story building carries the following external
loading: dead load, PD = 35 kips, roof live load, PLr = 15 kips, and seismic load (from
horizontal effects), QE = ±15 kips. Also, SDS = 1.25, Ωo = 2.5, and ρ = 1.3. Determine the
maximum and minimum axial forces in the column using ASD load combinations. Repeat
the problem accounting for overstrength.
8.2 Given the following frame and loading (moment and axial load) shown, determine the
maximum and minimum design loads in column C5 using LRFD load combinations.
SDS = 0.65 and redundancy factor ρ = 1.0.
–20 k

F2
N3 N4

–50 k

F1

N2 N5 N6 D=0
L=0
QE = ±38 k-ft
C5
D = 18 kips
L=0
N7
QE = ±5 kips
N1 N8
242 Introduction to Earthquake Engineering

8.3 Given the following single-story office building frame and loading, determine the maxi-
mum and minimum design loads in the brace member A, and maximum design load in
column B using ASD load combinations. SDS = 0.65 and redundancy factor ρ = 1.3.

F
D=0
L=0
QE = ±33 kips
B
A

D = 50 k
L=0
QE = ±17 kips

8.4 Use ASCE-7, Table 12.2-1, to determine seismic force-resisting system parameters and
restrictions for a steel special concentrically braced frame (SSCBF). Also, list an advan-
tage and a disadvantage for SSCBF systems.

8.5 Given Ss = 1.82 and S1 = 0.68, and assuming a stiff soil profile, determine the site coef-
ficients Fa and Fv, the maximum spectral accelerations SMS and SM1, and the design spectral
accelerations SDS and SD1.
8.6 Determine the spectral response acceleration parameters and draw the response spec-
trum for the Golden 1 Center Arena site, located at 500 David J Stern Walk, Sacramento,
California. Assume that the site (soil) class is D.
8.7 Draw the design response spectrum for a site where there is very dense soil (Site Class C).
The geotechnical engineer has provided the mapped acceleration parameters (SS = 1.25
and S1 = 0.52) and the long-period TL = 16 s.
8.8 A seismic vulnerability assessment of a two-story ordinary building located in Stockton,
California, is being performed. If the building’s first story is 14 ft high and the second
story is 12 ft high, compute the approximate structural period Ta, the static design base
shear V, the lateral forces at each level Fx, and the overturning moment at the base. The
proposed structural system is a steel ordinary moment frame with an estimated weight of
50 kips on the first level and 20 kips on the roof. The geotechnical engineer has provided
the mapped design acceleration parameters (SDS = 0.672 and SD1 = 0.362).
8.9 Given a two-story, 25-foot average height, wood frame single-family residence located in
Stockton, California, compute the design base shear. The geotechnical engineer estimates
Seismic Code Provisions 243

that the house will be on Site Class D, with mapped acceleration parameters of SS = 1.75
and S1 = 0.90. The building in plan view is regular. The structural system being proposed
is a wood structural panel shearwall. The estimated weight on level 1 is 83 kips and level
2 is 52 kips.
8.10 Given a six-story office building frame with equal story heights of 12 ft located in
Brookside (Stockton, California), compute the design earthquake loads acting on each
floor. The building in plan view is square. The structural system being proposed is a steel
special concentrically braced frame with an estimated weight on each floor of 100 kips and
80 kips on the roof. The geotechnical engineering report has classified the site soil class
as D.
8.11 First, sketch the design response spectrum using SDS = 0.82 and SD1 = 0.47. Then, use
the equivalent lateral force procedure to compute the approximate structural period Ta,
the design base shear V, the lateral forces at each level Fx, and the overturning moment
at the base. The structural system for the office building is a steel ordinary moment frame.

w2 = 20 kips

12 ft
w1 = 50 kips

14 ft
C5

20 ft 20 ft

8.12 A four-story special moment-resisting steel frame with equal story heights of 15 ft has a
period of 0.7 s, occupancy category of III, and elastic displacement of 0.7 in. What is the
maximum inelastic displacement (δ x) for the first story?
8.13 A three-story hospital building with reinforced concrete special moment frames has floor
heights of h1 = 14 ft, h2 = 12 ft, and h3 = 10 ft. The elastic displacements at each story
are δ1e = 0.8 in, δ2e = 1.45 in, and δ3e = 2.15 in. SDS = 0.82 and SD1 = 0.47, SDC of D, and
period, T = 0.75 s. Find the design story drifts and check to ensure that they are within the
allowable drifts.
8.14 Given a four-story County Jail with 12-foot story heights and steel special concentrically
braced frames, determine design story drifts and verify that they are within allowable drift
limits. The equivalent lateral force procedure analysis results are shown below.

Level wx (kips) Fx (kips) δxe (in)

4 (roof) 100 25 1.23
3 120 20 0.80
2 120 13 0.45
1 130 8 0.20

8.15 Use the equivalent lateral force procedure to compute the approximate structural period
Ta, the base shear V, the lateral forces at each level Fx, the overturning moment at the base,
and the inelastic story drifts given the elastic displacements shown. Also, determine if the
244 Introduction to Earthquake Engineering

drifts are within the allowable limits. The structural system is a special moment reinforced
concrete frame at a site with SDS = 0.82 and SD1 = 0.47.

400 kips
2.3 in
12 ft
500 kips
1.5 in
12 ft
500 kips
0.7 in
14 ft

8.16 A three-story office building with a special moment-resisting frame has floor heights
of h1 = 15 ft, h2 = 12 ft, and h3 = 12 ft. The building has an estimated weight on each
floor of 850 kips and 700 kips on the roof. SDS = 0.81, SD1 = 0.50 and seismic design cat-
egory D. A preliminary analysis yielded mode shapes and periods shown. First, use the
ELFP method to determine the seismic base shear, and then use the MRS analysis method
to determine effective weight for each mode, story forces for each mode, and the maximum
story shear forces by SRSS.

1.0 1.0

0.675 0.250

0.252 –0.345

T1 = 1.5 s T2 = 0.64 s

8.17 Given a three-story building and corresponding eigensolution (normalized eigenvectors

and eigenvalues), determine the periods, participation factors, and effective modal weights
for each mode shape. Also, use the SRSS method to determine the maximum elastic-based
shear using the fewest number of modes such that they contribute 90% of the effective
modal weight, per ASCE-7 MRS analysis.

1.00 1.00 1.00 

 
[Φ] =  0.64 −0.60 −2.57
 0.30 −0.68 2.47 


 21 0 0 
 
[λ] =  0 96.6 0  s−2
0 0 221.4


 88.8 


 
Assume {Sa} = 
187.2 in/s
2

 


272.4

Seismic Code Provisions 245

m3 = 1 kip s2/in
k3 = 60 kips/in
m2 = 1.5 kip s2/in

k2 = 120 kips/in
m1 = 2 kip s2/in
k1 = 180 kips/in

8.18 Given the following eigensolution for a three-story building with floor weights of 1800 kips
for the first floor and roof, and 1200 kips for the second floor:
 2.86 −0.657 0.387  15.1 377
     
[Φ] = 1.95 0.725 −1.61 and {ω} = 38.5 rad/s, which yield {Sa} = 346 in/s2
 
   
1.00
 1.00 1.00  61.7 308

Determine the participation factors for each mode, effective weight for each mode, base
shears for each mode using the given design spectral accelerations, and the maximum base
shear using the SRSS method.
8.19 Given the following six-story office building frame, ground motion parameters, and mass
and stiffness matrices, use MATLAB to compute the displacements, design earthquake
loads acting at each floor, base shear, and overturning moment using the MRS analysis. Also,
compute the effective modal masses for the six modes and accumulated mass percentage.
Ground motion parameters:
Site Class = D
SDS = 0.323
SD1 = 0.186
Building frame parameters (steel intermediate moment frame):
Risk category = II (office building)
SDC = C
R = 4.5
Cd = 4
12 ft

12 ft

12 ft

12 ft

12 ft

12 ft

3.95 0 0 0 0 0 
 
 0 3.88 0 0 0 0 

 0 0 2.65 0 0 0 
[ m] =  kips ⋅ s2 /in
 0 0 0 2.59 0 0 
 
 0 0 0 0 1.36 0 
 
 0 0 0 0 0 1.29

246 Introduction to Earthquake Engineering

 2260.0 −1374.7 198.5 −28.4 2.0 −0.2 

 
−1374.7 1822.9 −896.6 191.6 −12.9 1.44 

 198.5 −896.6 1352.6 −714.8 91.9 −10.5 
[ k ] =  kips/in
 −28.4 191.6 −714.8 885.8 −410.2 73.0 
 
 2.0 −12.9 91.9 −410.2 521.9 −192.5
 
 −0.2 1.44 −10.5 73.0 −192.5 128.7 

REFERENCES
American Society of Civil Engineers, ASCE/SEI, Minimum Design Loads for Buildings and Other Structures,
Reston, VA, 2010.
Villaverde, R., Fundamental Concepts of Earthquake Engineering, CRC Press, Boca Raton, FL, 2009.
Index
A Critical damping coefficient, 68
Critically damped system, 67
Accelerations, 34, 37, 80 CSMIP, see California Strong Motion Instrumentation
Aftershocks, 20 Program
Allowable seismic force-resisting system, 206–207
Allowable stress design (ASD), 203–206
American Society of Civil Engineers (ASCE), 15 D
Ancash earthquake, 5 D’Alembert’s principle, 61, 65–66, 152, 171
ASCE-7, 42, 203, 215, 219 Damped single-degree-of-freedom system, 64; see also
earthquake loads based on ASCE-7 equivalent lateral Undamped single-degree-of-freedom system
force procedure, 220–229 free vibration response of damped systems, 65–71
ELFP analysis steps, 220–225 structural damping, 71–73
MRS analysis steps, 229 time-dependent forced damped vibration response,
ASCE, see American Society of Civil Engineers 73–78
ASD, see Allowable stress design time-dependent support accelerations, 78–82
Ashlar masonry, 15 Damped systems, free vibration response of, 65–71
Attenuation, 29 Damping, 32, 91
relations, 138 coefficient, 66
ratio, 71
B Dead loads (DL), 42, 204, 206, 221
Decoupling process, 176–177
Base shear Deflections, 157–159, 163–165, 182–187, 212
using participation factors and response spectra, Deformation response factor, 63, 74
157–159, 163–165, 182–187 Degree-of-freedom (DOF), 37–38, 171
by time-dependent forces and support excitations, Design respsponse spectrum, 213–218
83–85 Detailing, 211
Bending moment, 53–54 Deterministic approach, 34
Body waves, 21, 22 Diagonal matrix, 172
Braced frames, 51, 53–56 Dip-slip, 20
Bracketed duration, 30–31 Dip angle, 19
Direct integration methods, 114
C central difference method, 117–120
Newmark’s Beta method for linear systems, 120–121
California Strong Motion Instrumentation Program Nigam−Jennings algorithm, 114–117, 118
(CSMIP), 33 Discrete system, 152; see also Continuous system
Cantilever shear wall, 57 deflections, base shear, and moments, 157–159
Central difference method, 117–120 equation of motion, 152–156
Circular frequency, 42 example, 159–161
City of Los Angeles building code (1943), 16 generalized participation factor, frequency, and
Collapse earthquakes, 9 natural period, 157
comparison of building collapse and natural Displacements, 37
earthquakes, 10 DL, see Dead loads
Complementary solutions, 62–63, 74 DOF, see Degree-of-freedom
Complete quadratic combination methods Dry-stone walls, 15
(CQC methods), 169, 185 Ductility, 33, 71
Compliant systems, 33 Duhamel’s integral, 91, 95
Contemporary design spectral analysis, 138 Dynamic analysis, 37, 83
Continental lithosphere, 11 Dynamic magnification factor, see Deformation response
Continuous system, 165–167; see also Discrete system factor
deflections, base shear, and moments, 163–165
equation of motion, 162–163 E
generalized participation factor, frequency, and
natural period, 163 Earthquake, 1, 37
generalized SDOF, 162 accelerogram, 123
conv function, 102 Ancash earthquake, 5
Convolution integral, 91, 114 charactereristics, 29–31
CQC methods, see Complete quadratic combination development of mitigation strategies, 14–17
methods earthquake-resistant design, 14

247
248 Index

Earthquake (Continued) generalized SDOF continuous system, 162–163

effects on structures, 29 for MDOF system, 171
epicenters of 358,214 earthquakes, 14 participation factors for MDOF system, 180–182
extent of San Francisco fire, 6 periods and mode shapes for MDOF system, 172–179
ground failures, 2–5 Equivalent static method of analysis, 60
ground shaking, 7–8 Euler method, 102
hazard assessment, 34–35 Euler’s identities, 40–41
indirect effects, 5–6 Explicit methods, 114
intensity, 23–25 central difference method, 117–120
intensity and extent of 2014 South Napa Earthquake, 8 Nigam−Jennings algorithm, 114–117, 118
intensity of 1985 Mexico City earthquake, 9
Loma Prieta earthquake (1989), 32 F
magnitude, 25–29
man-made, 9–11 Factor of safety (FS), 204
measurement, 23 Fault(s), 13
Mexico City earthquake (1985), 31 creep, 2
Nobi earthquake, 15 plane, 19
Parkfield earthquake, 29–30 slip, 19, 20
prediction, 36 FBD, see Free-body diagram
site charactereristics, 31–32 Fires, 5–6
structural charactereristics, 32–34 Fluid viscous dampers, 65
structural effects, 2 Focal depth, 19
surface rupture, 3 Focus, 13, 19
surface subsidence, 4 Foothill Zone, 31
tectonic, 11–14 Footwall, 20
terminology, 20 Force, 47
types, 8 forced damped vibration motion, 73
volcanic, 11 Foreshocks, 20
Earthquake load (E), 203, 205, 206, 208 Fourier method, 61
based on ASCE-7 equivalent lateral force procedure, Fracking process, 10
220–229 Free-body diagram (FBD), 38, 171
Effective mass, 184 Free vibration response
Effective weight, 184 of damped systems, 65–71
Eigenvalues, 171 of undamped systems, 38–42
Eigenvector, 175 Frequency equation, 173
Elastic design response spectrum, 216 FS, see Factor of safety
Elastic rebound theory, 21
Elastic response spectrum, 123; see also Inelastic G
response spectrum
design response spectrum, 133–134 General forcing function, 91, 94
displacement response, 124 Generalized forcing function, 99–100
elastic design response spectrum, 132, 134 discretization of general loading function and
El Centro earthquake, 125–126, 132 response, 94
estimating PPGA, 132–133 Duhamel’s integral, 95
examples, 129–132, 135–140 examples, 95–99
Mexico City earthquake, 126–127 MATLAB, 99–102
Newmark–Hall elastic design spectrum, 134–135 Generalized SDOF continuous system, 162–167
pseudo-acceleration, 127–128 deflections, base shear, and moments, 163–165
pseudo-velocity, 128–129 equation of motion, 162–163
seismic time-history, 123–125 example, 165–167
El Centro earthquake, 29–30, 125–126 generalized participation factor, frequency, and
Energy production processes, earthquakes caused by, 10 natural period, 163
Engineering seismology, 19 Generalized SDOF system analysis, 151
earthquake effects on structures, 29–34 discrete system, 152–161
earthquake hazard assessment, 34–35 generalized SDOF continuous system, 162–167
earthquake prediction, 36 General loading
engineered seismic-resistant design, 15 direct integration methods, 114–121
measuring earthquakes, 23–29 function, 94
Epicenter, 13, 19 general forcing function, 94–102
Epicentral distance, 19 response of SDOF system to impulse, 91–93
Equations of motion response to ground motion, 111–114
deflections, base shear, and moments, 182–187 SDOF system, 91
discrete system, 152–156 shock spectra, 102–111
equilibrium of mass, 171–172 Gravity-induced loads, 206
Index 249

Great Peruvian earthquake, see Ancash earthquake script to solving eigenvalue, 189–190
Ground cracking, 3 Matrix methods, 169
Ground failures, 2–5 Maximum considered earthquake (MCE), 34, 213–214
Ground motion, response to, 111–114 Maximum design earthquake (MDE), 34, 215
Ground shaking, 7–8 Maximum probable earthquake (MPE), 35
MDOF system, see Multi-degree of freedom system
H Messina-Reggio earthquake, 15
Mexico City earthquake (1985), 31
Hanging wall, 20 Mitigation strategies development for seismic hazard
Harmonic excitation force, 61 effects, 14–17
Harmonic motion, 40 MM scale, see Modified Mercalli scale
Heat, 65 Modal analysis, 169
Horizontal equilibrium of FBD, 66 Modal mass matrices, 176
Hypocenter, 13, 19 Modal participation factor, 180
Modal response spectrum analysis, 229–240
I Modal stiffness matrices, 176
Mode shapes for MDOF system, 172, 177–179
IBC, see International Building Code decoupling process, 176–177
Implicit methods, 114 eigenvector, 174–175
Newmark’s Beta method for linear systems, 120–121 simple harmonic equation, 172–173
Impulse, 91 Modified Mercalli scale (MM scale), 24
impulsive force, 91 MM-I, 24
loading function, 91, 92 MM-XII, 24
response function, 94–95 Moment(s), 27, 28, 47, 57, 157–159, 163–165
response of SDOF system to, 91–93 using participation factors and response spectra, 182–187
Inelastic response spectrum, 140; see also Elastic MPE, see Maximum probable earthquake
response spectrum Multi-degree of freedom system (MDOF system), 1, 151,
design response spectrum, 222 169
elastoplastic response spectra for El Centro quake, 142 degrees of freedom, 171
example, 142–143 equations of motion, 171–187
inelastic design response spectrum, 143–146 response spectrum analysis method, 188–198
nonlinear equation of motion, 140–141 spectrum analysis, 2
Intensity of earthquake, 23–25 structural mode shapes, 170, 172–179
Internal shear force, 53–54 Multistory building systems, 152
International Building Code (IBC), 203, 207 Multistory shear buildings, 1

K N
Kanamori wave energy, 28 National Earthquake Hazards Reduction Program
(NEHRP), 32, 207
L Natural circular frequency, 39, 71
Natural cyclic frequency, 41
Lake Zone, 31 Natural frequency, see Natural circular frequency
Lateral force-resisting systems, 50 NEHRP, see National Earthquake Hazards Reduction
Left-lateral faults, 20 Program
Linear impulse and momentum, 92 Newmark’s Beta method for linear systems, 120–121
Liquefaction, 3 Newmark–Hall method, 1, 133, 134
Lithosphere, 11 Newton’s second law, 34, 91, 184
Live loads (L), 16, 204 Nigam−Jennings algorithm, 114–117, 118
Load and resistance factor design (LRFD), 203, 206 Nobi earthquake, 15
Local magnitude scale, see Richter scale Nonhomogeneous differential equations, 62, 80
Logarithmic decrement, 69 Nonlinear equation of motion, 140–141
Loma Prieta earthquake (1989), 3, 32 Nonperiodic forcing function, 91
Love waves, 22, 23 Normal faults, 20
LRFD, see Load and resistance factor design Normal modes, see eigenvector

M O
Magnitude of earthquake, see Earthquake—magnitude Oblique-slip, 20
Magnitude saturation, 27 Observation station, 19
Man-made earthquakes, 9–11 Occupancy categories, 219
Mass, 184 and importance factors, 220
MATrix LABoratory (MATLAB), 99–102, 171, 173 Occupancy Category II, 219
script to performing operation, 196–198 Oceanic lithosphere, 11
250 Index

Oscillator model, 66 Roof live loads (Lr), 205

Overdamped system, 67, 68 Rossi–Forel scale (RF scale), 24
Overstrength factor, 211 RP, see Return period
Overturning moment, 53–54
S
P
San Francisco Bay Mud, 32
Panel shear, 53–54 SDCs, see Seismic design categories
Parkfield earthquake, 29–30 SDOF structural dynamic analysis; see also Single-
Participation factors, 182, 186–187 degree of-freedom systems (SDOF systems)
CQC rule, 185 base shears and stresses, 83–85
deflections, base shear, and moments using, 157–159, damped single-degree-of-freedom system, 64–82
163–165 undamped single-degree-of-freedom system, 37–64,
displacement, 182–183 65
mass, 184–185 SDOF systems, see Single-degree of-freedom systems
for MDOF system, 180–182 SEAOC, see Structural Engineers Association of
Particular solutions, 62–63 California
PE, see Professional Engineering Secondary waves (S-waves), 22
Peak ground acceleration (PGA), 23, 24, 123, 159 Seiches, 5
Portal frame model, 61 Seismic code provisions
PPGA, see Probable PGAs earthquake loads based on ASCE-7 equivalent lateral
Pressure waves, see Primary waves (P-waves) force procedure, 220–229
Primary waves (P-waves), 22 modal response spectrum analysis, 229–240
Principal coordinates, 176 seismic design codes, 207
Probabilistic approach, 34 seismic load analysis procedures, 213–220
Probabilistic seismic hazard map, 35 seismic load combinations, 208–213
Probable PGAs (PPGA), 132 seismic response history procedures, 240–241
Professional Engineering (PE), 1 structural design philosophies, 203–207
Pseudo-acceleration, 127–128 Seismic design categories (SDCs), 207
Pseudo-velocity, 128–129 Seismic design codes, 207
P-waves, see Primary waves Seismic force-resisting system parameters, 211–213
Seismic gaps, 21
R Seismic hazard effects, mitigation strategies development
for, 14–17
Radians per second (rad/s), 39 Seismic load analysis procedures, 213
Rain load (R), 205 design respsponse spectrum, 213–218
Rayleigh waves, 22, 23 permitted lateral analysis procedures, 219–220
Redundadancy factor, 210–211 Seismic load combinations, 208
Reinforced concrete bridge structure, 130 maximum axial load, 209
Response of SDOF system to impulse, 91–93 overstrength factor, 211
Response spectrum analysis, 123, 188 redundadancy factor, 210–211
building frame, 188–189 seismic force-resisting system parameters, 211–213
elastic response spectrum, 123–140 Seismic response history procedures, 240–241
generalized SDOF case, 190–192 Seismic time-history, 123–125
inelastic response spectrum, 140–146 Seismic waves, 7, 21–23
MATLAB script to performing operation, 196–198 excitations, 78–79
MATLAB script to solving eigenvalue, 189–190 Seismology, 13, 19; see also Engineering seismology
matrix operations, 192–196 elastic rebound theory, 21
of SDOF system, 123 horizontal movement, strike-slip fault, 21
Response spectrum/spectra, 37, 169, 182, 186–187 seismic waves, 21–23
CQC rule, 185 seismograph, 22
deflections, base shear, and moments using, 157, vertical movement, dip-slip faults, 20
163–165 Shallow earthquakes, 19
displacement, 182–183 Shear, 57
mass, 184–185 shear-wave velocity, 31
Return period (RP), 35 shear buildings, 152, 169
Reverse faults, 20 shear walls, 51, 57–60, 61
RF-I, see Roman numerals from 1 Shear waves, see Secondary waves (S-waves)
RF scale, see Rossi–Forel scale Shock spectra, 102–111, 123
Richter scale, 25, 26 Simpson’s method, 102
Ridge push, 13 Single-degree of-freedom systems (SDOF systems), 1,
Right-lateral faults, 20 38, 91, 169; see also Multi-degree of freedom
Roman numerals from 1 (RF-I), 24 system (MDOF system)
Roman numerals from 10 (RF-X), 24 elastic response spectrum, 123–140
Index 251

inelastic response spectrum, 140–146 Time-dependent forced undamped vibration response,

response spectrum analysis of, 123 61–64
response to impulse, 91–93 forcing function and resulting response, 65
Slabs, 13 Time-dependent forces and support excitations, base
Snow load (SL), 42, 205, 221 shears and stresses caused by, 83–85
Soil liquefaction, 3 Time-dependent support accelerations, 78–82
Square root of sum of squares (SRSS), 169 Time-stepping methods, 114
Static analysis, 37 Total effective weight (W), 42
Static displacement, 80 Transmissibility, 77
Steady-state solution, 80 Trapezoidal method, 102
Stiffness, 47, 91 Tripartite graph paper, 135
Stresses caused by time-dependent forces and support Tsunamis, 5, 14, 19
excitations, 83–85 Two-degree-of-freedom system, 171
Strike angle, 19
Strike-slip, 20 U
Structural damping, 71–73
Structural design philosophies, 203 UBC, see Uniform Building Code
allowable seismic force-resisting system, 206–207 Unbraced frames, 51–53
ASD, 204–206 Undamped single-degree-of-freedom system, 37; see also
LRFD, 206 Damped single-degree-of-freedom system
Structural dynamics, 37 approximate structural period, 60–61
Structural Engineers Association of California free vibration response of undamped systems,
(SEAOC), 16 38–42
Blue Book, 16 idealized single-degree-of-freedom system for portal
Structural stiffffness, 46 frame, 38
braced frames, 53–56 structural stiffffness, 46–60
equivalent stiffness constants, k, 48 structural weight, 42–46
force–displacement relationship, 49 time-dependent forced undamped vibration response,
force or moment, 47 61–64, 65
lateral force-resisting systems, 50 Underdamped system, 67, 68, 74, 111
shear walls, 57–60 Uniform Building Code (UBC), 16
unbraced frames, 51–53 Unit impulse response, 93–94
Structural weight, 42–46 USGS Seismic Hazard Mapping Application, 214, 217,
Surface faulting, 2–3 218
Surface rupture, 20 input screen, 214
Surface waves, 21 output, 218
scale, 27
S-waves, see Secondary waves V
Vibration theory, see Structural dynamics
T
Virtual forces, 54
Taylor devices, 65 Virtual work equation, 57
Tectonic earthquakes, 11–14 “Viscous damping”, 65
Tectonic plates, 11; see also Earthquake Volcanic earthquakes, 8, 11
movement, 21
Thrust faults, 20 W
Time-dependent excitation force, 61
Time-dependent forced damped vibration response, Wagner’s theory, 11
73–78 Wind load (W), 51, 205, 206