Chopra: Prentice-Hall

PAGES JUL. 19, 2000 14:15

ICC Oregon (503) 221-9911
Page 197

6
Earthquake Response of
Linear Systems

PREVIEW

One of the most important applications of the theory of structural dynamics is in analyzing
the response of structures to ground shaking caused by an earthquake. In this chapter we
study the earthquake response of linear SDF systems to earthquake motions. By definition,
linear systems are elastic systems, and we shall also refer to them as linearly elastic sys-
tems to emphasize both properties. Because earthquakes can cause damage to many struc-
tures, we are also interested in the response of yielding or inelastic systems, the subject of
Chapter 7.
The first part of this chapter is concerned with the earthquake responsedeforma-
tion, internal element forces, stresses, and so onof simple structures as a function of time
and how this response depends on the system parameters. Then we introduce the response
spectrum concept, which is central to earthquake engineering, together with procedures to
determine the peak response of systems directly from the response spectrum. This is fol-
lowed by a study of the characteristics of earthquake response spectra, which leads into the
design spectrum for the design of new structures and safety evaluation of existing struc-
tures against future earthquakes. The important distinctions between design and response
spectra are identified and the chapter closes with a discussion of two types of response
spectra that are not used commonly.

6.1 EARTHQUAKE EXCITATION

For engineering purposes the time variation of ground acceleration is the most useful way
of defining the shaking of the ground during an earthquake. The ground acceleration u g (t)
appears on the right side of the differential equation (1.7.4) governing the response of

197
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 198

198 Earthquake Response of Linear Systems Chap. 6

structures to earthquake excitation. Thus, for given ground acceleration the problem to be
solved is defined completely for an SDF system with known mass, stiffness, and damping
properties.
The basic instrument to record three components of ground shaking during earth-
quakes is the strong-motion accelerograph (Fig. 6.1.1), which does not record continu-
ously but is triggered into motion by the first waves of the earthquake to arrive. This is

(a)

(b)

Figure 6.1.1 Strong motion accelerographs: (a) SMA-1, an analog-recording instrument with un-
damped natural frequency of 25 Hz and damping 60% of critical; (b) SSA-2, a digital recording
instrument with undamped natural frequency of 50 Hz and damping 70% of critical. (Courtesy of
Kinemetrics, Inc.)
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 199

Sec. 6.1 Earthquake Excitation 199

because, even in earthquake-prone regions, such as California and Japan, there may not be
any strong ground motion from earthquakes to record for months, or even years, at a time.
Consequently, continual recordings of hundreds of such instruments would be a wasteful
exercise. After triggering, the recording continues for some minutes or until the ground
shaking falls again to imperceptible levels. Clearly, the instruments must be regularly
maintained and serviced so that they produce a record when shaking occurs.
The basic element of an accelerograph is a transducer element, which in its sim-
plest form is an SDF massspringdamper system (Section 3.7). Therefore, the transducer
element is characterized by its natural frequency f n and viscous damping ratio ; typi-
cally, f n = 25 Hz and = 60% for modern analog accelerographs; and f n = 50 Hz
and = 70% in modern digital accelerographs. These transducer parameters enable the
digital instrument to record, without excessive distortion, accelerationtime functions con-
taining frequencies from very low up to, say, 30 Hz; the analog instrument is accurate over
a narrower frequency range, say, up to 15 Hz.
Unfortunately, instrumental records of strong ground shaking were scarce for many
years, and even today none or very few records may be obtained from a destructive earth-
quake in some parts of the world. For example, no strong-motion records were obtained
from two earthquakes during 1993 that caused much destruction: Killari, Maharashtra,
India, September 30, 1993; and Guam, a U.S. territory, August 8, 1993. Ideally, when
a strong earthquake occurs it would be desirable to have many stations instrumented to
record the ground motions. However, not knowing when and exactly where earthquakes
will occur and having limited budgets for installation and maintenance of instruments, it
is only occasionally possible to obtain such recordings in the region of strongest shaking.
Many more records have been obtained in regions where moderate ground shaking has
occurred.
The first strong-motion accelerogram was recorded during the Long Beach earth-
quake of 1933, and since that time several hundred records have been obtained. As might
be expected, most of these records are of small motion and only a small fraction of them
have acceleration of 20% g or more. The geographical distribution of these ground motion
records is very uneven. Over half of them are from California, most of which are from
three earthquakes: the San Fernando earthquake of February 9, 1971, the Loma Prieta
earthquake of October 17, 1989, and the Northridge earthquake of January 17, 1994. The
peak values of accelerations recorded at many different locations during the Loma Prieta
earthquake are shown in Fig. 6.1.2. These acceleration values are largest near the epicenter
of the earthquake and tend to decrease with distance from the fault causing the earthquake.
However, the accelerations recorded at similar distances may vary significantly because of
several factors, especially local soil conditions.
Figure 6.1.3 shows a collection of representative accelerationtime records of earth-
quake ground motions in the region of strong shaking. One horizontal component is given
for each location and earthquake. All have been plotted to the same acceleration and time

It should be noted that most if not all of the digital accelerographs use a force-balance type of transducer,

for which two parameters will not completely define the instrument response, which is that of a higher-order (than
a massspringdamper) system.
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 200

Figure 6.1.2 Peak horizontal ground accelerations recorded during the Loma Prieta earthquake of
October 17, 1989. (Courtesy of R. B. Seed.)
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 201

Sec. 6.1 Earthquake Excitation 201

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

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
Stone Canyon 1972
Helena 1935 - S90W
Melendy Ranch - N29W
Koyna 1967 - Long -0.5

-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, sec

Figure 6.1.3 Ground motions recorded during several earthquakes. [Based in part on Hudson
(1979).]
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 202

202 Earthquake Response of Linear Systems Chap. 6

0.4

ug, g
0

-0.4 ugo = 0.319g

15
ug, in./sec

-15 ugo = 13.04 in./sec

0
ug, in.

ugo = 8.40 in.

-10
0 5 10 15 20 25 30
Time, sec

Figure 6.1.4 Northsouth component of horizontal ground acceleration recorded at the

Imperial Valley Irrigation District substation, El Centro, California, during the Imperial
Valley earthquake of May 18, 1940. The ground velocity and ground displacement were
computed by integrating the ground acceleration.

scale. The wide and very real variability of amplitude, duration, and general appearance of
different records can be clearly seen. One of these records is enlarged in Fig. 6.1.4. This is
the northsouth component of the ground motion recorded at a site in El Centro, Califor-
nia, during the Imperial Valley, California, earthquake of May 18, 1940. At this scale it
becomes apparent that ground acceleration varies with time in a highly irregular manner.
No matter how irregular, the ground motion is presumed to be known and independent of
the structural response. This is equivalent to saying that the foundation soil is rigid, im-
plying no soilstructure interaction. If the structure were founded on very flexible soil, the
motion of the structure and the resulting forces imposed on the underlying soil can modify
the base motion.
The ground acceleration is defined by numerical values at discrete time instants.
These time instants should be closely spaced to describe accurately the highly irregular
1 1
variation of acceleration with time. Typically, the time interval is chosen to be 100 to 50 of
a second, requiring 1500 to 3000 ordinates to describe the ground motion of Fig. 6.1.4.

This ground acceleration is used extensively in this book and, for brevity, will be called El Centro ground

motion, although three components of motion have been recorded at the same site during several earthquakes
after 1940.
 Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 203

Sec. 6.2 Equation of Motion 203

The top curve in Fig. 6.1.4 shows the variation of El Centro ground acceleration
with time. The peak ground acceleration u go is 0.319g. The second curve is the ground
velocity, obtained by integrating the accelerationtime function. The peak ground veloc-
ity u go is 13.04 in./sec. Integration of ground velocity provides the ground displace-
ment, presented as the lowest trace. The peak ground displacement u go is 8.40 in. It
is difficult to determine accurately the ground velocity and displacement because ana-
log accelerographs do not record the initial partuntil the accelerograph is triggeredof
the accelerationtime function and thus the base (zero acceleration) line is unknown.
Digital accelerographs overcome this problem by providing a short memory so that the
onset of ground motion is measured.
In existence are several different versions of the El Centro ground motion. The varia-
tions among them arise from differences in (1) how the original analog trace of acceleration
versus time was digitized into numerical data, and (2) the procedure chosen to introduce
the missing baseline in the record. The version shown in Fig. 6.1.4 is used throughout this
book and is tabulated in Appendix 6.

6.2 EQUATION OF MOTION

Equation (1.7.4) governs the motion of a linear SDF system (Fig. 6.2.1) subjected to ground
acceleration u g (t). Dividing this equation by m gives

u + 2 n u + n2 u = u g (t) (6.2.1)

It is clear that for a given u g (t), the deformation response u(t) of the system depends only
on the natural frequency n or natural period Tn of the system and its damping ratio, ;
writing formally, u u(t, Tn , ). Thus any two systems having the same values of Tn
and will have the same deformation response u(t) even though one system may be more
massive than the other or one may be stiffer than the other.
Ground acceleration during earthquakes varies irregularly to such an extent (see
Fig. 6.1.4) that analytical solution of the equation of motion must be ruled out. Therefore,
numerical methods are necessary to determine the structural response, and any of the meth-
ods presented in Chapter 5 could be used. The response results presented in this chapter

ut
u
m ug ut
c
c (b)
(a) k m
k
ug

Figure 6.2.1 Single-degree-of-freedom systems.

 Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 204

204 Earthquake Response of Linear Systems Chap. 6

were obtained by exact solution of the equation of motion for the ground motion varying
linearly over every time step, t = 0.02 sec (Section 5.2).

6.3 RESPONSE QUANTITIES

Of greatest interest in structural engineering is the deformation of the system, or displace-

ment u(t) of the mass relative to the moving ground, to which the internal forces are
linearly related. These are the bending moments and shears in the beams and columns of
the one-story frame of Fig. 6.2.1a or the spring force in the system of Fig. 6.2.1b. Knowing
the total displacement u t (t) of the mass would be useful in providing enough separation be-
tween adjacent buildings to prevent their pounding against each other during an earthquake.
Pounding is the cause of damage to several buildings during almost every earthquake (see
Fig. 6.3.1). Similarly, the total acceleration u t (t) of the mass would be needed if the struc-
ture is supporting sensitive equipment and the motion imparted to the equipment is to be
determined.
The numerical solution of Eq. (6.2.1) can be implemented to provide results for rel-
ative quantities u(t), u(t), and u(t) as well as total quantities u t (t), u t (t), and u t (t).

Figure 6.3.1 Pounding damage, Hotel de Carlo, Mexico City, 1985. (From National
Geophysical Data Center, photograph by C. Arnold.)
 Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 205

Sec. 6.4 Response History 205

6.4 RESPONSE HISTORY

For a given ground motion u g (t), the deformation response u(t) of an SDF system depends
only on the natural vibration period of the system and its damping ratio. Figure 6.4.1a
shows the deformation response of three different systems due to El Centro ground accel-
eration. The damping ratio, = 2%, is the same for the three systems, so that only the dif-
ferences in their natural periods are responsible for the large differences in the deformation
responses. It is seen that the time required for an SDF system to complete a cycle of vibra-
tion when subjected to this earthquake ground motion is very close to the natural period
of the system. (This interesting result, valid for typical ground motions containing a wide
range of frequencies, can be proven using random vibration theory, not included in this
book.) The peak deformation [Eq. (1.11.1)] is also noted in each case. Observe that among
these three systems, the longer the vibration period, the greater the peak deformation. As
will be seen later, this trend is neither perfect nor valid over the entire range of periods.
Figure 6.4.1b shows the deformation response of three systems to the same ground
motion. The vibration period Tn is the same for the three systems, so that the differences

(a) (b)

Tn = 0.5 sec, = 0.02 Tn = 2 sec, = 0

9.91 in.
10

0
2.67 in.
-10
Tn = 1 sec, = 0.02 Tn = 2 sec, = 0.02
Deformation u, in.

10

5.97 in. 7.47 in.

-10
Tn = 2 sec, = 0.02 Tn = 2 sec, = 0.05
10
5.37 in.

-10 7.47 in.

0 10 20 30 0 10 20 30
Time, sec Time, sec

Figure 6.4.1 Deformation response of SDF systems to El Centro ground motion.

Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 206

206 Earthquake Response of Linear Systems Chap. 6

in their deformation responses are associated with their damping. We observe the expected
trend that systems with more damping respond less than lightly damped systems. Because
the natural period of the three systems is the same, their responses display a similarity in
the time required to complete a vibration cycle and in the times the maxima and minima
occur.
Once the deformation response history u(t) has been evaluated by dynamic analysis
of the structure, the internal forces can be determined by static analysis of the structure
at each time instant. Two methods to implement such analysis were mentioned in Sec-
tion 1.8. Between them, the preferred approach in earthquake engineering is based on the
concept of the equivalent static force f S (Fig. 6.4.2) because it can be related to earthquake
forces specified in building codes; f S was defined in Eq. (1.8.1), which is repeated here for
convenience:
f S (t) = ku(t) (6.4.1)

fS(t)

Vb(t)
Mb(t) Figure 6.4.2 Equivalent static force.

where k is the lateral stiffness of the frame (Fig. 6.2.1a). Expressing k in terms of the mass
m gives
f S (t) = m n2 u(t) = m A(t) (6.4.2)
where
A(t) = n2 u(t) (6.4.3)
Observe that the equivalent static force is m times A(t), the pseudo-acceleration, not m
times the total acceleration u t (t). This distinction is discussed in Section 6.6.3.
The pseudo-acceleration response A(t) of the system can readily be computed from
the deformation response u(t). For the three systems with Tn = 0.5, 1, and 2 sec, all having
= 0.02, u(t) is available in Fig. 6.4.1. Multiplying each u(t) by the corresponding n2 =
(2/Tn )2 gives the pseudo-acceleration responses for these systems; they are presented in
Fig. 6.4.3, where the peak value is noted for each system.
For the one-story frame the internal forces (e.g., the shears and moments in the
columns and beam, or stress at any location) can be determined at a selected instant of
time by static analysis of the structure subjected to the equivalent static lateral force f S (t)
at the same time instant (Fig. 6.4.2). Thus a static analysis of the structure would be neces-
sary at each time instant when the responses are desired. In particular, the base shear Vb (t)
and the base overturning moment Mb (t) are
Vb (t) = f S (t) Mb (t) = h f S (t) (6.4.4a)
 Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 207

Sec. 6.5 Response Spectrum Concept 207

1.2 Tn = 0.5 sec, = 0.02

-1.2 1.09g
Pseudo-acceleration A, g

1.2 Tn = 1 sec, = 0.02

0
0.610g
-1.2

1.2 Tn = 2 sec, = 0.02

0
0.191g
-1.2
0 10 20 30 Figure 6.4.3 Pseudo-acceleration response
Time, sec of SDF systems to El Centro ground motion.

where h is the height of the mass above the base. We put Eq. (6.4.2) into these equations
to obtain
Vb (t) = m A(t) Mb (t) = hVb (t) (6.4.4b)
If the SDF system is viewed as a massspringdamper system (Fig. 6.2.1b), the
notion of equivalent static force is not necessary. One can readily visualize that the spring
force is given by Eq. (6.4.1).

6.5 RESPONSE SPECTRUM CONCEPT

G. W. Housner was instrumental in the widespread acceptance of the concept of the earth-
quake response spectrumintroduced by M. A. Biot in 1932as a practical means of
characterizing ground motions and their effects on structures. Now a central concept in
earthquake engineering, the response spectrum provides a convenient means to summarize
the peak response of all possible linear SDF systems to a particular component of ground
motion. It also provides a practical approach to apply the knowledge of structural dynamics
to the design of structures and development of lateral force requirements in building codes.
A plot of the peak value of a response quantity as a function of the natural vibration
period Tn of the system, or a related parameter such as circular frequency n or cyclic
frequency f n , is called the response spectrum for that quantity. Each such plot is for SDF
systems having a fixed damping ratio , and several such plots for different values of are
included to cover the range of damping values encountered in actual structures. Whether
 Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 208

208 Earthquake Response of Linear Systems Chap. 6

the peak response is plotted against f n or Tn is a matter of personal preference. We have

chosen the latter because engineers prefer to use natural period rather than natural fre-
quency because the period of vibration is a more familiar concept and one that is intuitively
appealing.
A variety of response spectra can be defined depending on the response quantity that
is plotted. Consider the following peak responses:
u o (Tn , ) max |u(t, Tn , )|
t

u o (Tn , ) max |u(t, Tn , )|

t

u to (Tn , ) max |u t (t, Tn , )|

t

The deformation response spectrum is a plot of u o against Tn for fixed . A similar plot
for u o is the relative velocity response spectrum, and for u to is the acceleration response
spectrum.

6.6 DEFORMATION, PSEUDO-VELOCITY, AND

PSEUDO-ACCELERATION RESPONSE SPECTRA

In this section the deformation response spectrum and two related spectra, the pseudo-
velocity and pseudo-acceleration response spectra, are discussed. As shown in Section 6.4,
only the deformation u(t) is needed to compute internal forces. Obviously, then, the de-
formation spectrum provides all the information necessary to compute the peak values of
deformation D u o and internal forces. The pseudo-velocity and pseudo-acceleration
response spectra are included, however, because they are useful in studying characteristics
of response spectra, constructing design spectra, and relating structural dynamics results
to building codes.

6.6.1 Deformation Response Spectrum

Figure 6.6.1 shows the procedure to determine the deformation response spectrum. The
spectrum is developed for El Centro ground motion, shown in part (a) of this figure. The
time variation of the deformation induced by this ground motion in three SDF systems is
presented in part (b). For each system the peak value of deformation D u o is determined
from the deformation history. (Usually, the peak occurs during ground shaking; however,
for lightly damped systems with very long periods the peak response may occur during
the free vibration phase after the ground shaking has stopped.) The peak deformations are
D = 2.67 in. for a system with natural period Tn = 0.5 sec and damping ratio = 2%;
D = 5.97 in. for a system with Tn = 1 sec and = 2%; and D = 7.47 in. for a system with
Tn = 2 sec and = 2%. The D value so determined for each system provides one point
on the deformation response spectrum; these three values of D are identified in Fig. 6.6.1c.
Repeating such computations for a range of values of Tn while keeping constant at
2% provides the deformation response spectrum shown in Fig. 6.6.1c. As we shall show
 Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 209

Sec. 6.6 Deformation, Pseudo-velocity, and Pseudo-acceleration Response Spectra 209

(a)
0.4

ug(t) / g
0

-0.4
0 10 20 30
Time, sec
(b)
(c)
10
20
Tn = 0.5 sec 0
= 2% 2.67 in.
Deformation u, in.

-10
10 15

D = uo, in.
Tn = 1 sec 0
= 2%
10

7.47
-10 5.97 in.

5.97
10

2.67
5
Tn = 2 sec 0
= 2%
-10 7.47 in.
0
0 10 20 30 0 1 2 3
Time, sec Tn, sec
Figure 6.6.1 (a) Ground acceleration; (b) deformation response of three SDF systems with = 2%
and Tn = 0.5, 1, and 2 sec; (c) deformation response spectrum for = 2%.

later, the complete response spectrum includes such spectrum curves for several values of
damping.

6.6.2 Pseudo-velocity Response Spectrum

Consider a quantity V for an SDF system with natural frequency n related to its peak
deformation D u o due to earthquake ground motion:

2
V = n D = D (6.6.1)
Tn
The quantity V has units of velocity. It is related to the peak value of strain energy E So
stored in the system during the earthquake by the equation

mV 2
E So = (6.6.2)
2
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 210

210 Earthquake Response of Linear Systems Chap. 6

This relationship can be derived from the definition of strain energy and using Eq. (6.6.1)
as follows:
ku 2o k D2 k(V /n )2 mV 2
E So = = = =
2 2 2 2
The right side of Eq. (6.6.2) is the kinetic energy of the structural mass m with velocity
V , called the peak relative pseudo-velocity, or simply peak pseudo-velocity. The prefix
pseudo is used because V is not equal to the peak velocity u o , although it has the correct
units. We return to this matter in Section 6.12.
The pseudo-velocity response spectrum is a plot of V as a function of the natural
vibration period Tn , or natural vibration frequency f n , of the system. For the ground mo-
tion of Fig. 6.6.1a the peak pseudo-velocity V for a system with natural period Tn can
be determined from Eq. (6.6.1) and the peak deformation D of the same system available
from the response spectrum of Fig. 6.6.1c, which has been reproduced in Fig. 6.6.2a. As
an example, for a system with Tn = 0.5 sec and = 2%, D = 2.67 in.; from Eq. (6.6.1),
V = (2/0.5)2.67 = 33.7 in./sec. Similarly, for a system with Tn = 1 sec and the same
, V = (2/1)5.97 = 37.5 in./sec; and for a system with Tn = 2 sec and the same ,
V = (2/2)7.47 = 23.5 in./sec. These three values of peak pseudo-velocity V are identi-
fied in Fig. 6.6.2b. Repeating such computations for a range of values of Tn while keeping
constant at 2% provides the pseudo-velocity spectrum shown in Fig. 6.6.2b.

6.6.3 Pseudo-acceleration Response Spectrum

Consider a quantity A for an SDF system with natural frequency n related to its peak
deformation D u o due to earthquake ground motion:
 2
2
A = n D =
2
D (6.6.3)
Tn
The quantity A has units of acceleration and is related to the peak value of base shear Vbo
[or the peak value of the equivalent static force f So , Eq. (6.4.4a)]:
Vbo = f So = m A (6.6.4)
This relationship is simply Eq. (6.4.4b) specialized for the time of peak response with the
peak value of A(t) denoted by A. The peak base shear can be written in the form
A
Vbo = w (6.6.5)
g
where w is the weight of the structure and g the gravitational acceleration. When written in
this form, A/g may be interpreted as the base shear coefficient or lateral force coefficient.
It is used in building codes to represent the coefficient by which the structural weight is
multiplied to obtain the base shear.
Observe that the base shear is equal to the inertia force associated with the mass
m undergoing acceleration A. This quantity defined by Eq. (6.6.3) is generally different
from the peak acceleration u to of the system. It is for this reason that we call A the peak
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 211

Sec. 6.6 Deformation, Pseudo-velocity, and Pseudo-acceleration Response Spectra 211

20

15

D, in.

7.47
10

5.97
(a)

2.67
5

50

37.5
33.7
40
V, in./sec

23.5
30
(b)
20
10
0

1.5
1.09

1
0.610
A, g

(c)
0.191

0.5

0
0 1 2 3
Tn, sec

Figure 6.6.2 Response spectra ( = 0.02) for El Centro ground motion: (a) deforma-
tion response spectrum; (b) pseudo-velocity response spectum; (c) pseudo-acceleration
response spectrum.

pseudo-acceleration; the prefix pseudo is used to avoid possible confusion with the true
peak acceleration u to . We return to this matter in Section 6.12.
The pseudo-acceleration response spectrum is a plot of A as a function of the natural
vibration period Tn , or natural vibration frequency f n , of the system. For the ground motion
of Fig. 6.6.1a the peak pseudo-acceleration A for a system with natural period Tn and
damping ratio can be determined from Eq. (6.6.3), and the peak deformation D of the
system from the spectrum of Fig. 6.6.2a. As an example, for a system with Tn = 0.5 sec
and = 2%, D = 2.67 in.; from Eq. (6.6.3), A = (2/0.5)2 2.67 = 1.09g, where g = 386
in./sec2 . Similarly, for a system with Tn = 1 sec and the same , A = (2/1)2 5.97 =
0.610g; and for a system with Tn = 2 sec and the same , A = (2/2)2 7.47 = 0.191g.
Note that the same values for A are also available as the peak values of A(t) presented
in Fig. 6.4.3. These three values of peak pseudo-acceleration are identified in Fig. 6.6.2c.
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 212

212 Earthquake Response of Linear Systems Chap. 6

Repeating such computations for a range of values of Tn , while keeping constant at 2%,
provides the pseudo-acceleration spectrum shown in Fig. 6.6.2c.

6.6.4 Combined DVA Spectrum

Each of the deformation, pseudo-velocity, and pseudo-acceleration response spectra for a

given ground motion contains the same information, no more and no less. The three spec-
tra are simply different ways of presenting the same information on structural response.
Knowing one of the spectra, the other two can be obtained by algebraic operations using
Eqs. (6.6.1) and (6.6.3).
Why do we need three spectra when each of them contains the same information?
One of the reasons is that each spectrum directly provides a physically meaningful quan-
tity. The deformation spectrum provides the peak deformation of a system. The pseudo-
velocity spectrum is related directly to the peak strain energy stored in the system during
the earthquake; see Eq. (6.6.2). The pseudo-acceleration spectrum is related directly to the
peak value of the equivalent static force and base shear; see Eq. (6.6.4). The second reason
lies in the fact that the shape of the spectrum can be approximated more readily for design
purposes with the aid of all three spectral quantities rather than any one of them alone; see
Sections 6.8 and 6.9. For this purpose a combined plot showing all three of the spectral
quantities is especially useful. This type of plot was developed for earthquake response
spectra, apparently for the first time, by A. S. Veletsos and N. M. Newmark in 1960.

100

50
= 0.02
10
10

23.5
20

10
10
V, in./sec

1g

7.
19

47

5
0.

n.
A,

,i
1
g

0.

D
1

2
01
0.

1
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
Tn, sec

Figure 6.6.3 Combined DV A response spectrum for El Centro ground motion; = 2%.
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 213

Sec. 6.6 Deformation, Pseudo-velocity, and Pseudo-acceleration Response Spectra 213

This integrated presentation is possible because the three spectral quantities are in-
terrelated by Eqs. (6.6.1) and (6.6.3), rewritten as
A Tn 2
= V = n D or A=V = D (6.6.6)
n 2 Tn
Observe the similarity between these equations relating D, V , and A and Eq. (3.2.21) for
the dynamic response factors Rd , Rv , and Ra for an SDF system subjected to harmonic
excitation. Equation (3.2.21) permitted presentation of Rd , Rv , and Ra , all together, on
four-way logarithmic paper (Fig. 3.2.8), constructed by the procedure described in Ap-
pendix 3 (Chapter 3). Similarly, the graph paper shown in Fig. A6.1 (Appendix 6) with
four-way logarithmic scales can be constructed to display D, V , and A, all together. The
vertical and horizontal scales for V and Tn are standard logarithmic scales. The two scales
for D and A sloping at +45 and 45 , respectively, to the Tn -axis are also logarithmic
scales but not identical to the vertical scale; see Appendix 3.
Once this graph paper has been constructed, the three response spectradeforma-
tion, pseudo-velocity, and pseudo-accelerationof Fig. 6.6.2 can readily be combined into
a single plot. The pairs of numerical data for V and Tn that were plotted in Fig. 6.6.2b on

100

50

10
10

0
50
5

20
Pseudo-velocity V, in./sec

10
10
1

5
n.
,i
Ps

D
1
eu

0.

n
do

io
at
-a

0.

m
cc

2
r
el

fo
er

01 De
a
tio .1
n
0

1
A,

0.
g

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, sec

Figure 6.6.4 Combined DV A response spectrum for El Centro ground motion; = 0,

2, 5, 10, and 20%.
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 214

214 Earthquake Response of Linear Systems Chap. 6

linear scales are replotted in Fig. 6.6.3 on logarithmic scales. For a given natural period Tn ,
the D and A values can be read from the diagonal scales. As an example, for Tn = 2 sec,
Fig. 6.6.3 gives D = 7.47 in. and A = 0.191g. (Actually, these numbers cannot be read so
accurately from the graph; in this case they were available from Fig. 6.6.2.) The four-way
plot is a compact presentation of the threedeformation, pseudo-velocity, and pseudo-
accelerationresponse spectra, for a single plot of this form replaces the three plots of
Fig. 6.6.2.
A response spectrum should cover a wide range of natural vibration periods and
several damping values so that it provides the peak response of all possible structures. The
period range in Fig. 6.6.3 should be extended because tall buildings and long-span bridges,
among other structures, may have longer vibration periods (Fig. 2.1.2) and several damping
values should be included to cover the practical range of = 0 to 20%. Figure 6.6.4 shows
spectrum curves for = 0, 2, 5, 10, and 20% over the period range 0.02 to 50 sec.
This, then, is the response spectrum for the northsouth component of ground motion
recorded at one location during the Imperial Valley earthquake of May 18, 1940. Because
the lateral force or base shear for an SDF system is related through Eq. (6.6.5) to A/g, we
also plot this normalized pseudo-acceleration spectrum in Fig. 6.6.5. Similarly, because
the peak deformation is given by D, we also plot this deformation response spectrum
in Fig. 6.6.6.
The response spectrum has proven so useful in earthquake engineering that spec-
tra for virtually all ground motions strong enough to be of engineering interest are now

3
fSo / w = A / g

0
0 0.5 1 1.5 2 2.5 3
Tn, sec

Figure 6.6.5 Normalized pseudo-acceleration, or base shear coefficient, response spec-

trum for El Centro ground motion; = 0, 2, 5, 10, and 20%.
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 215

Sec. 6.6 Deformation, Pseudo-velocity, and Pseudo-acceleration Response Spectra 215

computed and published soon after they are recorded. Enough of them have been obtained
to give us a reasonable idea of the kind of motion that is likely to occur in future earth-
quakes, and how response spectra are affected by distance to the causative fault, local soil
conditions, and regional geology.

6.6.5 Construction of Response Spectrum

The response spectrum for a given ground motion component u g (t) can be developed by
implementation of the following steps:

1. Numerically define the ground acceleration u g (t); typically, the ground motion ordi-
nates are defined every 0.02 sec.
2. Select the natural vibration period Tn and damping ratio of an SDF system.
3. Compute the deformation response u(t) of this SDF system due to the ground mo-
tion u g (t) by any of the numerical methods described in Chapter 5. [In obtaining the

50

20

10

1
D, in.

0.5

0.2

0.1

0.05

0.02

0.01

0.005
0.02 0.05 0.1 0.2 0.5 1 2 5 10 20 50
Tn, sec

Figure 6.6.6 Deformation response spectrum for El Centro ground motion; = 0, 2, 5, 10, and
20%.
 Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 216

216 Earthquake Response of Linear Systems Chap. 6

responses shown in Fig. 6.6.1, the exact solution of Eq. (6.2.1) for ground motion
assumed to be piecewise linear over every t = 0.02 sec was used; see
Section 5.2.]
4. Determine u o , the peak value of u(t).
5. The spectral ordinates are D = u o , V = (2/Tn )D, and A = (2/Tn )2 D.
6. Repeat steps 2 to 5 for a range of Tn and values covering all possible systems of
engineering interest.
7. Present the results of steps 2 to 6 graphically to produce three separate spectra like
those in Fig. 6.6.2 or a combined spectrum like the one in Fig. 6.6.4.

Considerable computational effort is required to generate an earthquake response

spectrum. A complete dynamic analysis to determine the time variation (or history) of the
deformation of an SDF system provides the data for one point on the spectrum correspond-
ing to the Tn and of the system. Each curve in the response spectrum of Fig. 6.6.4 was
produced from such data for 112 values of Tn unevenly spaced over the range Tn = 0.02
to 50 sec.

Example 6.1
Derive equations for and plot deformation, pseudo-velocity, and pseudo-acceleration response
spectra for ground acceleration u g (t) = u go (t), where (t) is the Dirac delta function and
u go is the increment in velocity, or the magnitude of the acceleration impulse. Only consider
systems without damping.
Solution
1. Determine the response history. The response of an SDF system to p(t) = (t )
is available in Eq. (4.1.6). Adapting that solution to peff (t) = m u g (t) = m u go (t)

2 10
A / ugo, sec-1
D / ugo, sec

1 5

1
0 0
0 2 10 0 2 10
Tn Tn

Figure E6.1
 Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 217

Sec. 6.7 Peak Structural Response from the Response Spectrum 217

gives
u go
u(t) = sin n t (a)
n
The peak value of u(t) is
u go
uo = (b)
n
2. Determine the spectral values.
u go u go
D uo = = Tn (c)
n 2
2 u go
V = n D = u go A = n2 D = (d)
Tn
Two of these response spectra are plotted in Fig. E6.1.

6.7 PEAK STRUCTURAL RESPONSE FROM THE RESPONSE

SPECTRUM

If the response spectrum for a given ground motion component is available, the peak value
of deformation or of an internal force in any linear SDF system can be determined readily.
This is the case because the computationally intensive dynamic analyses summarized in
Section 6.6.5 have already been completed in generating the response spectrum. Corre-
sponding to the natural vibration period Tn and damping ratio of the system, the values
of D, V , or A are read from the spectrum, such as Fig. 6.6.6, 6.6.4, or 6.6.5. Now all
response quantities of interest can be expressed in terms of D, V , or A and the mass or
stiffness properties of the system. In particular, the peak deformation of the system is
 2
Tn Tn
uo = D = V = A (6.7.1)
2 2
and the peak value of the equivalent static force f So is [from Eqs. (6.6.4) and (6.6.3)]
f So = k D = m A (6.7.2)
Static analysis of the one-story frame subjected to lateral force f So (Fig. 6.7.1) provides
the internal forces (e.g., shears and moments in columns and beams). This involves ap-
plication of well-known procedures of static structural analysis, as will be illustrated later

fSo

Vbo
Figure 6.7.1 Peak value of equivalent
Mbo static force.
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 218

218 Earthquake Response of Linear Systems Chap. 6

by examples. We emphasize again that no further dynamic analysis is required beyond

that necessary to determine u(t). In particular, the peak values of shear and overturning
moment at the base of the one-story structure are
Vbo = k D = m A Mbo = hVbo (6.7.3)
We note that only one of these response spectradeformation, pseudo-velocity, or
pseudo-accelerationis sufficient for computing the peak deformations and forces re-
quired in structural design. For such applications the velocity or acceleration spectra
(defined in Section 6.5) are not required, but for completeness we discuss these spectra
briefly at the end of this chapter.
Example 6.2
A 12-ft-long vertical cantilever, a 4-in.-nominal-diameter standard steel pipe, supports a
5200-lb weight attached at the tip as shown in Fig. E6.2. The properties of the pipe are:
outside diameter, do = 4.500 in., inside diameter di = 4.026 in., thickness t = 0.237 in.,
and second moment of cross-sectional area, I = 7.23 in4 , elastic modulus E = 29,000 ksi,
and weight = 10.79 lb/foot length. Determine the peak deformation and bending stress in the
cantilever due to the El Centro ground motion. Assume that = 2%.

Solution The lateral stiffness of this SDF system is

3E I 3(29 103 )7.23

k= = = 0.211 kip/in.
L3 (12 12)3
The total weight of the pipe is 10.79 12 = 129.5 lb, which may be neglected relative to the
lumped weight of 5200 lb. Thus
w 5.20
m= = = 0.01347 kip-sec2 /in.
g 386
The natural vibration frequency and period of the system are
 
k 0.211
n = = = 3.958 rad/sec Tn = 1.59 sec
m 0.01347

From the response spectrum curve for = 2% (Fig. E6.2b), for Tn = 1.59 sec, D = 5.0 in.
and A = 0.20g. The peak deformation is

u o = D = 5.0 in.

The peak value of the equivalent static force is

A
f So = w = 0.20 5.2 = 1.04 kips
g
The bending moment diagram is shown in Fig. E6.2d with the maximum moment at the base
= 12.48 kip-ft. Points A and B shown in Fig. E6.2e are the locations of maximum bending
stress:
Mc (12.48 12)(4.5/2)
max = = = 46.5 ksi
I 7.23
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 219

Sec. 6.7 Peak Structural Response from the Response Spectrum 219

fSo = 1.04 kips

5200 lb

4-in. pipe
12 (
A B
a a
12.48 kip-ft

(a) (c) (d) Moment (e) Section a-a

1g
1.
D

=
= .7 i

A
D
100

5.

g
=

0 n.

20
2

in

0.
.

=
50

10
10

0
n.
,i
20

D
A,
g
V, in./sec

10

10
1

5
(b)
1
0.

2
01
0.

1
0.
1

0.5
0.502

1.59

1
0.

00
01

0.

0.2
0.02 0.05 0.2 0.5 1 2 5 10 20 50
Tn, sec

Figure E6.2

As shown, = +46.5 ksi at A and = 46.5 ksi at B, where + denotes tension. The
algebraic signs of these stresses are irrelevant because the direction of the peak force is not
known, as the pseudo-acceleration spectrum is, by definition, positive.

Example 6.3
The stress computed in Example 6.2 exceeded the allowable stress and the designer decided
to increase the size of the pipe to an 8-in.-nominal standard steel pipe. Its properties are
do = 8.625 in., di = 7.981 in., t = 0.322 in., and I = 72.5 in4 . Comment on the advantages
and disadvantages of using the bigger pipe.
 Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 220

220 Earthquake Response of Linear Systems Chap. 6

Solution
3(29 103 )72.5
k= = 2.112 kips/in.
(12 12)3

2.112
n = = 12.52 rad/sec Tn = 0.502 sec
0.01347
From the response spectrum (Fig. E6.2b): D = 2.7 in. and A = 1.1g. Therefore,

u o = D = 2.7 in.

f So = 1.1 5.2 = 5.72 kips

Mbase = 5.72 12 = 68.64 kip-ft

(68.64 12)(8.625/2)
max = = 49.0 ksi
72.5
Using the 8-in.-diameter pipe decreases the deformation from 5.0 in. to 2.7 in. However,
contrary to the designers objective, the bending stress increases slightly.
This example points out an important difference between the response of structures to
earthquake excitation and to a fixed value of static force. In the latter case, the stress would
decrease, obviously, by increasing the member size. In the case of earthquake excitation,
the increase in pipe diameter shortens the natural vibration period from 1.59 sec to 0.50 sec,
which for this response spectrum has the effect of increasing the equivalent static force f So .
Whether the bending stress decreases or increases by increasing the pipe diameter depends on
the increase in section modulus, I /c, and the increase or decrease in f So , depending on the
response spectrum.

Example 6.4
A small one-story reinforced concrete building is idealized for purposes of structural analysis
as a massless frame supporting a total dead load of 10 kips at the beam level (Fig. E6.4a).
The frame is 24 ft wide and 12 ft high. Each column and the beam has a 10-in.-square cross
section. Assume that the Youngs modulus of concrete is 3 103 ksi and the damping ratio for
the building is estimated as 5%. Determine the peak response of this frame to the El Centro
ground motion. In particular, determine the peak lateral deformation at the beam level and
plot the diagram of bending moments at the instant of peak response.

17.1 kip-ft
10 kips fSo = 7.6 kips
(

17.1 17.1

10 sq. ) )
17.1
12
10 sq. ( (
28.4 28.4

24

(a) (b) (c)

Figure E6.4 (a) Frame; (b) equivalent static force; (c) bending moment diagram.
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 221

Sec. 6.7 Peak Structural Response from the Response Spectrum 221

Solution The lateral stiffness of such a frame was calculated in Chapter 1: k = 96E I /7h 3 ,
where E I is the flexural rigidity of the beam and columns and h is the height of the frame.
For this particular frame,
96(3 103 )(104 /12)
k= = 11.48 kips/in.
7(12 12)3
The natural vibration period is 
2 10/386
Tn = = 2 = 0.30 sec
k/m 11.48
For Tn = 0.3 and = 0.05, we read from the response spectrum of Fig. 6.6.4: D = 0.67 in.
and A = 0.76g. Peak deformation: u o = D = 0.67 in. Equivalent static force: f So =
(A/g)w = 0.76 10 = 7.6 kips. Static analysis of the frame for this lateral force, shown in
Fig. E6.4b, gives the bending moments that are plotted in Fig. E6.4c.

Example 6.5
The frame of Example 6.4 is modified for use in a building to be located on sloping ground
(Fig. E6.5). The beam is now made much stiffer than the columns and can be assumed to be
rigid. The cross sections of the two columns are 10 in. square, as before, but their lengths are
12 ft and 24 ft, respectively. Determine the base shears in the two columns at the instant of
peak response due to the El Centro ground motion. Assume the damping ratio to be 5%.
Solution
1. Compute the natural vibration period.
12(3 103 )(104 /12) 12(3 103 )(104 /12)
k= +
(12 12)3 (24 12)3
= 10.05 + 1.26 = 11.31 kips/in.

10/386
Tn = 2 = 0.30 sec
11.31
2. Compute the shear force at the base of the short and long columns.
u o = D = 0.67 in., A = 0.76g
Vshort = kshort u o = (10.05)0.67 = 6.73 kips

Vlong = klong u o = (1.26)0.67 = 0.84 kip

12
24

24
Figure E6.5
 Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 222

222 Earthquake Response of Linear Systems Chap. 6

Observe that both columns go through equal deformation. Undergoing equal deformations, the
stiffer column carries a greater force than the flexible column; the lateral force is distributed
to the elements in proportion to their relative stiffnesses. Sometimes this basic principle has,
inadvertently, not been recognized in building design, leading to unanticipated damage of the
stiffer elements.
Example 6.6
For the three-span box-girder bridge of Example 1.3, determine the base shear in each of
the six columns of the two bents due to El Centro ground motion applied in the longitudinal
direction. Assume the damping ratio to be 5%.
Solution The weight of the bridge deck was computed in Example 1.3: w = 6919 kips. The
natural period of longitudinal vibration of the bridge was computed in Example 2.2: Tn =
0.573 sec. For Tn = 0.573 sec and = 0.05, we read from the response spectrum of Fig. 6.6.4:
D = 2.591 in. and A = 0.807g.
All the columns have the same stiffness and they go through equal deformation u o =
D = 2.591 in. Thus, the base shear will be the same in all columns, which can be computed
in one of two ways: The total equivalent static force on the bridge is [from Eq. (6.6.5)]

f so = 0.807 6919 = 5584 kips

Base shear for one column, Vb = 5584 6 = 931 kips. Alternatively, the base shear in each
column is
2.591
Vb = kcol u o = 4313 = 931 kips
12

6.8 RESPONSE SPECTRUM CHARACTERISTICS

We now study the important properties of earthquake response spectra. Figure 6.8.1 shows
the response spectrum for El Centro ground motion together with u go , u go , and u go , the
peak values of ground acceleration, ground velocity, and ground displacement, respec-
tively, identified in Fig. 6.1.4. To show more directly the relationship between the response
spectrum and the ground motion parameters, the data of Fig. 6.8.1 have been presented
again in Fig. 6.8.2 using normalized scales: D/u go , V /u go , and A/u go . Figure 6.8.3 shows
one of the spectrum curves of Fig. 6.8.2, the one for 5% damping, together with an ideal-
ized version shown in dashed lines; the latter will provide a basis for constructing smooth
design spectra directly from the peak ground motion parameters (see Section 6.9). Based
on Figs. 6.8.1 to 6.8.3, we first study the properties of the response spectrum over various
ranges of the natural vibration period of the system separated by the period values at a, b,
c, d, e, and f : Ta = 0.035 sec, Tb = 0.125, Tc = 0.5, Td = 3.0, Te = 10, and T f = 15 sec.
Subsequently, we identify the effects of damping on spectrum ordinates.
For systems with very short period, say Tn < Ta = 0.035 sec, the peak pseudo-
acceleration A approaches u go and D is very small. This trend can be understood based on
physical reasoning. For a fixed mass, a very short period system is extremely stiff or es-
sentially rigid. Such a system would be expected to undergo very little deformation and its
mass would move rigidly with the ground; its peak acceleration should be approximately
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 223

Sec. 6.8 Response Spectrum Characteristics 223

100

50

10
10

0
20

ugo = 13.04
V, in./sec

10

10
1

ug
5

o
=
8.
1

n.

40
0.

,i
D
9g

A,
2
31

g
0.
=

01
o

1 0.
ug

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
Tn, sec

Figure 6.8.1 Response spectrum ( = 0, 2, 5, and 10%) and peak values of ground
acceleration, ground velocity, and ground displacement for El Centro ground motion.

equal to u go (Fig. 6.8.4d). This expectation is confirmed by Fig. 6.8.4, where the ground
acceleration is presented in part (a), the total acceleration u t (t) of a system with Tn = 0.02
sec and = 2% in part (b), and the pseudo-acceleration A(t) for the same system in part
(c). Observe that u t (t) and u g (t) are almost identical functions and u to u go . Further-
more, for lightly damped systems u t (t) A(t) and u to A (Section 6.12.2); therefore,
A u go .
For systems with a very long period, say Tn > T f = 15 sec, D for all damping
values approaches u go and A is very small; thus the forces in the structure, which are
related to m A, would be very small. This trend can again be explained by relying on
physical reasoning. For a fixed mass, a very-long-period system is extremely flexible. The
mass would be expected to remain essentially stationary while the ground below moves
(Fig. 6.8.5c). Thus u t (t) 0, implying that A(t) 0 (see Section 6.12.2); and u(t)
u g (t), implying that D u go . This expectation is confirmed by Fig. 6.8.5, where the
deformation response u(t) of a system with Tn = 30 sec and = 2% to the El Centro
ground motion is compared with the ground displacement u g (t). Observe that the peak
values for u o and u go are close and the time variation of u(t) is similar to that of u g (t),
but for rotation of the baseline. The discrepancy between the two arises, in part, from the
loss of the initial portion of the recorded ground motion prior to triggering of the recording
accelerograph.
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 224

224 Earthquake Response of Linear Systems Chap. 6

10
V/u go = 7

0
5

10

10
2

10
1 10 V/u go = 1

D
1
=

/u g
o
V / ugo

ug

o
0.5
A/

=
D
1

3
/u g
o
=
0.

1
1

go
0.2

A/

/u
1

u g

D
1
=

0.
o
o
ug
A/

0.1 0.
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, sec

Figure 6.8.2 Response spectrum for El Centro ground motion plotted with normalized
scales A/u go , V /u go , and D/u go ; = 0, 2, 5, and 10%.

For short-period systems with Tn between Ta = 0.035 sec and Tc = 0.50 sec, A
exceeds u go , with the amplification depending on Tn and . Over a portion of this period
range, Tb = 0.125 sec to Tc = 0.5 sec, A may be idealized as constant at a value equal to
u go amplified by a factor depending on .
For long-period systems with Tn between Td = 3 sec and T f = 15 sec, D generally
exceeds u go , with the amplification depending on Tn and . Over a portion of this period
range, Td = 3.0 sec to Te = 10 sec, D may be idealized as constant at a value equal to u go
amplified by a factor depending on .
For intermediate-period systems with Tn between Tc = 0.5 sec and Td = 3.0 sec, V
exceeds u go . Over this period range, V may be idealized as constant at a value equal to
u go , amplified by a factor depending on .
Based on these observations, it is logical to divide the spectrum into three period
ranges (Fig. 6.8.3). The long-period region to the right of point d, Tn > Td , is called
the displacement-sensitive region because structural response is related most directly to
ground displacement. The short-period region to the left of point c, Tn < Tc , is called the
acceleration-sensitive region because structural response is most directly related to ground
acceleration. The intermediate period region between points c and d, Tc < Tn < Td , is
called the velocity-sensitive region because structural response appears to be better related
to ground velocity than to other ground motion parameters. For a particular ground motion,
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 225

Sec. 6.8 Response Spectrum Characteristics 225

S p e c t r a l R e g i o n s
Acceleration Velocity Displacement
sensitive sensitive sensitive
10

0
5
10

10
2 c d

10

Td = 3.0 sec
Tc = 0.5 sec
1

1
e
b
V / ugo

0.5
Ta = 0.035 sec

1 f

0.
A/

1
Tb = 0.125 sec

go
ugo
0.2

/u
D

Te = 10 sec
1
0.1 0.

Tf = 15 sec
0.
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, sec

Figure 6.8.3 Response spectrum for El Centro ground motion shown by a solid line
together with an idealized version shown by a dashed line; = 5%.

the periods Ta , Tb , Te , and T f on the idealized spectrum are independent of damping, but
Tc and Td vary with damping.
The preceding observations and discussion have brought out the usefulness of the
four-way logarithmic plot of the combined deformation, pseudo-velocity, and pseudo-
acceleration response spectra. These observations would be difficult to glean from the
three individual spectra.
Idealizing the spectrum by a series of straight lines abcde f in the four-way
logarithmic plot is obviously not a precise process. For a given ground motion, the period
values associated with the points a, b, c, d, e, and f and the amplification factors for the
segments bc, cd, and de are somewhat judgmental in the way we have approached
them. However, formal curve-fitting techniques can be used to replace the actual spec-
trum by an idealized spectrum of a selected shape. In any case, the idealized spectrum in
Fig. 6.8.3 is not a close approximation to the actual spectrum. This may not be visually
apparent but becomes obvious when we note that the scales are logarithmic. As we shall
see in the next section, the greatest benefit of the idealized spectrum is in constructing a
design spectrum representative of many ground motions.
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 226

226 Earthquake Response of Linear Systems Chap. 6

0.4

u g, g
0 (a)

-0.4 ugo = 0.319g

Tn = 0.02 sec, = 0.02 ut

0.4
ut, g

0 (b)

ug
-0.4 uto = 0.321g
(d)
-0.4
A, g

0 (c)

0.4 A = 0.321g
0 1 2 3 4 5 6 7
Time, sec

Figure 6.8.4 (a) El Centro ground acceleration; (b) total acceleration response of an SDF
system with Tn = 0.02 sec and = 2%; (c) pseudo-acceleration response of the same
system; (d) rigid system.

The period values associated with the points Ta , Tb , Tc , Td , Te , and T f and the am-
plification factors for the segments bc, cd, and de are not unique in the sense that they
vary from one ground motion to the next. Some of the variation in these parameters re-
flects the inherent probabilistic differences that exist among ground motions even if they
are recorded under similar conditions: magnitude of the earthquake, the distance of the site
from the earthquake source, the source mechanism for the earthquake, and the local soil
conditions at the site. Greater variation exists in the response spectrum parameters among
ground motions recorded under dissimilar conditions. However, researchers have demon-
strated that response trends identified earlier from the three regions of one response spec-
trum are generally valid for the corresponding spectral regions of other ground motions.
We now turn to damping, which has significant influence on the earthquake response
spectrum (Figs. 6.6.46.6.6). The zero damping curve is marked by abrupt jaggedness,
which indicates that the response is very sensitive to small differences in the natural vibra-
tion period. The introduction of damping makes the response much less sensitive to the
period.
Damping reduces the response of a structure, as expected, and the reduction achieved
with a given amount of damping is different in the three spectral regions. In the limit as
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 227

Sec. 6.8 Response Spectrum Characteristics 227

10

ug, in.
0 (a)

ut = 0
-10 ugo = 8.4 in.

Tn = 30 sec, = 0.02
-10
ug
u, in.

0 (b) (c)

10 uo = 8.23 in.
0 10 20 30
Time, sec

Figure 6.8.5 (a) El Centro ground displacement; (b) deformation response of SDF sys-
tem with Tn = 30 sec and = 2%; (c) very flexible system.

Tn 0 damping does not affect the response because the structure moves rigidly with the
ground. In the other limit as Tn , damping again does not affect the response be-
cause the structural mass stays still while the ground underneath moves. Among the three
period regions defined earlier, the effect of damping tends to be greatest in the velocity-
sensitive region of the spectrum. In this spectral region the effect of damping depends on
the ground motion characteristics. If the ground motion is nearly harmonic over many cy-
cles (e.g., the record from Mexico City shown in Fig. 6.1.3), the effect of damping would be
especially large for systems near resonance (Chapter 3). If the ground motion is short in
duration with only a few major cycles (e.g., the record from Parkfield, California, shown
in Fig. 6.1.3), the influence of damping would be small, as in the case of pulse excitations
(Chapter 4).
Figure 6.8.6 shows the peak pseudo-acceleration A( ), normalized relative to
A( = 0), plotted as a function of for several Tn values. There are some of the data
from the response spectrum of Figs. 6.6.4 and 6.6.5 replotted in a different format. Ob-
serve that the effect of damping is stronger for smaller damping values. This means that
if the damping ratio is increased from 0 to 2%, the reduction in response is greater than
the response reduction, due to an increase in damping from 10% to 12%. The effect of
damping in reducing the response depends on the period Tn of the system, but there is no
clear trend from Fig. 6.8.6. This is yet another indication of the complexity of structural
response to earthquakes.
The motion of a structure and the associated forces could be reduced by increasing
the effective damping of the structure. The addition of dampers achieves this goal without
significantly changing the natural vibration periods of the structure. Viscoelastic dampers
have been introduced in many structures; for example, 10,000 dampers have been installed
throughout the height of each tower of the World Trade Center in New York City to reduce
 Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 228

228 Earthquake Response of Linear Systems Chap. 6

1.0


0.8



A() / A( = 0)

0.6 Tn = 5 sec
1

3 0.5
0.4

0.2
0.2

Figure 6.8.6 Variation of peak

0.0
pseudo-acceleration with damping for
0 2 5 10 20
systems with Tn = 0.2, 0.5, 1, 3, and 5 sec;
Damping ratio , % El Centro ground motion.

wind-induced motion to within a comfortable range for the occupants. In recent years
there is a growing interest in developing dampers suitable for structures in earthquake-
prone regions. Because the inherent damping in most structures is relatively small, their
earthquake response can be reduced significantly by the addition of dampers. These can be
especially useful in improving the seismic safety of an existing structure. We will return to
this topic in Chapter 7.

6.9 ELASTIC DESIGN SPECTRUM

In this section we introduce the concept of earthquake design spectrum for elastic systems
and present a procedure to construct it from estimated peak values for ground acceleration,
ground velocity, and ground displacement.
The design spectrum should satisfy certain requirements because it is intended for
the design of new structures, or the seismic safety evaluation of existing structures, to resist
future earthquakes. For this purpose the response spectrum for a ground motion recorded
during a past earthquake is inappropriate. The jaggedness in the response spectrum, as
seen in Fig. 6.6.4, is characteristic of that one excitation. The response spectrum for another
ground motion recorded at the same site during a different earthquake is also jagged, but the
peaks and valleys are not necessarily at the same periods. This is apparent from Fig. 6.9.1,
where the response spectra for ground motions recorded at the same site during three past
earthquakes are plotted. Similarly, it is not possible to predict the jagged response spectrum
in all its detail for a ground motion that may occur in the future. Thus the design spectrum
should consist of a set of smooth curves or a series of straight lines with one curve for each
level of damping.
 Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 229

Sec. 6.9 Elastic Design Spectrum 229

4
1956
A / ugo

3
1968
2

Figure 6.9.1 Response spectra for the

1 1940 northsouth component of ground motions
recorded at the Imperial Valley Irrigation
0 District substation, El Centro, California,
0 0.5 1.0 1.5 2.0 2.5 3.0 during earthquakes of May 18, 1940;
February 9, 1956; and April 8, 1968.
Tn, sec = 2%.

10

5
0
10

Mean + 1
10

g o
2

/u
D
c d
10
A/
u go

1 Mean
Tc = 0.349 sec

Tf = 33 sec
V/ugo

b
0.5
1

e
Ta = 1/33 sec

Td = 3.135 sec
0.
1

0.2
Tb = 1/8 sec

Te = 10 sec
0.

0.1 f
0.
01

a
0.05
01
0.
0.
00
1

0.02
0.02 0.05 0.1 0.2 0.5 1 2 5 10 20 50
Tn, sec

Figure 6.9.2 Mean and mean +1 spectra with probability distributions for V at
Tn = 0.25, 1, and 4 sec; = 5%. Dashed lines show an idealized design spectrum.
(Based on numerical data from R. Riddell and N. M. Newmark, 1979.)
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 230

230 Earthquake Response of Linear Systems Chap. 6

The design spectrum should, in a general sense, be representative of ground motions

recorded at the site during past earthquakes. If none have been recorded at the site, the
design spectrum should be based on ground motions recorded at other sites under similar
conditions. The factors that one tries to match in the selection include the magnitude of
the earthquake, the distance of the site from the causative fault, the fault mechanism, the
geology of the travel path of seismic waves from the source to the site, and the local soil
conditions at the site. While this approach is feasible for some parts of the world, such as
California and Japan, where numerous ground motion records are available, in many other
regions it is hampered by the lack of a sufficient number of such records. In such situa-
tions compromises in the approach are necessary by considering ground motion records
that were recorded for conditions different from those at the site. Detailed discussion of
these issues is beyond the scope of this book. The presentation here is focused on the nar-
row question of how to develop the design spectrum that is representative of an available
ensemble (or set) of recorded ground motions.
The design spectrum is based on statistical analysis of the response spectra for the
ensemble of ground motions. Suppose that I is the number of ground motions in the en-
semble, the ith ground motion is denoted by u ig (t), and u igo , u igo , and u igo are its peak
displacement, velocity, and acceleration, respectively. Each ground motion is normalized
(scaled up or down) so that all ground motions have the same peak ground acceleration,
say u go ; other bases for normalization can be chosen. The response spectrum for each
normalized ground motion is computed by the procedures described in Section 6.6. At
each period Tn there are as many spectral values as the number I of ground motion records
in the ensemble: D i , V i , and Ai (i = 1, 2, . . . , I ), the deformation, pseudo-velocity, and
pseudo-acceleration spectral ordinates. Such data were generated for an ensemble of 10
earthquake records, and selected aspects of the results are presented in Fig. 6.9.2. The
quantities u go , u go , and u go in the normalized scales of Fig. 6.9.2 are the average values
of the peak ground displacement, velocity, and accelerationaveraged over the I ground
motions. Statistical analysis of these data provide the probability distribution for the spec-
tral ordinate, its mean value, and its standard deviation at each period Tn . The probability
distributions are shown schematically at three selected Tn values, indicating that the coef-
ficient of variation (= standard deviation mean value) varies with Tn . Connecting all
the mean values gives the mean response spectrum. Similarly connecting all the mean-
plus-one-standard-deviation values gives the mean-plus-one-standard-deviation response
spectrum. Observe that these two response spectra are much smoother than the response
spectrum for an individual ground motion (Fig. 6.6.4). As shown in Fig. 6.9.2, such a
smooth spectrum curve lends itself to idealization by a series of straight lines much better
than the spectrum for an individual ground motion (Fig. 6.8.3).
Researchers have developed procedures to construct such design spectra from ground
motion parameters. One such procedure, which is illustrated in Fig. 6.9.3, will be summa-
rized later. The recommended period values Ta = 33 1
sec, Tb = 18 sec, Te = 10 sec,
and T f = 33 sec, and the amplification factors A , V , and D for the three spectral
regions were developed by the preceding analysis of a larger ensemble of ground mo-
tions recorded on firm ground (rock, soft rock, and competent sediments). The amplifi-
cation factors for two different nonexceedance probabilities, 50% and 84.1%, are given
in Table 6.9.1 for several values of damping and in Table 6.9.2 as a function of damping
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 231

Sec. 6.9 Elastic Design Spectrum 231

Elastic design c Vugo d

spectrum

D
u go
e

go
A u
Pseudo-velocity (log scale)
ugo

u go
b

o
ug
f
Peak ground acceleration,
velocity, and displacement
a

1/33 sec 1/8 sec 10 sec 33 sec

33 Hz 8 Hz 1/10 Hz 1/33 Hz
Natural vibration period (log scale)

Figure 6.9.3 Construction of elastic design spectrum.

TABLE 6.9.1 AMPLIFICATION FACTORS: ELASTIC DESIGN SPECTRA

Median (50th percentile) One Sigma (84.1th percentile)

Damping,
(%) A V D A V D

1 3.21 2.31 1.82 4.38 3.38 2.73

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
20 1.17 1.08 1.01 1.26 1.37 1.38
Source: N. M. Newmark and W. J. Hall, Earthquake Spectra and Design, Earth-
quake Engineering Research Institute, Berkeley, Calif., 1982, pp. 35 and 36.

TABLE 6.9.2 AMPLIFICATION FACTORS: ELASTIC DESIGN SPECTRAa

Median (50th percentile) One Sigma (84.1th percentile)

A 3.21 0.68 ln 4.38 1.04 ln

V 2.31 0.41 ln 3.38 0.67 ln
D 1.82 0.27 ln 2.73 0.45 ln

Source: N. M. Newmark and W. J. Hall, Earthquake Spectra and

Design, Earthquake Engineering Research Institute, Berkeley, Calif.,
1982, pp. 35 and 36.
a Damping ratio in percent.
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 232

232 Earthquake Response of Linear Systems Chap. 6

ratio. The 50% nonexceedance probability represents the median value of the spectral or-
dinates and the 84.1% approximates the mean-plus-one-standard-deviation value assuming
lognormal probability distribution for the spectral ordinates.
Observe that the period values Ta , Tb , Te , and T f are fixed; the values in Fig. 6.9.3
are for firm ground. Period values Tc and Td are determined by the intersections of the
constant-A (= A u go ), constant-V (= V u go ), and constant-D (= D u go ) branches of the
spectrum. Because A , V , and D are functions of (Tables 6.9.1 and 6.9.2), Tc and Td
depend on the damping ratio.

Summary. A procedure to construct a design spectrum is now summarized with

reference to Fig. 6.9.3:

1. Plot the three dashed lines corresponding to the peak values of ground acceleration
u go , velocity u go , and displacement u go for the design ground motion.
2. Obtain from Table 6.9.1 or 6.9.2 the values for A , V , and D for the selected.
3. Multiply u go by the amplification factor A to obtain the straight line bc represent-
ing a constant value of pseudo-acceleration A.

200
48 2.3 = 110.4 36
100
g

2.
71

01
2.

=
=

50 72
71

ugo = 48 in./sec .4
2.

10

10

ug

0
1g

1g

o
=
=

36

20
o
ug

in
.
V, in./sec

10
10
1

n.

5
,i
D
1
A,

0.

1
g

2
01
0.

1
0.
1

0.5
0.02 0.05 0.1 0.2 0.5 1 2 5 10 20 50
1
0.

00
01

0.

Tn, sec

Figure 6.9.4 Construction of elastic design spectrum (84.1th percentile) for ground
motions with u go = 1g, u go = 48 in./sec, and u go = 36 in.; = 5%.
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 233

Sec. 6.9 Elastic Design Spectrum 233

4. Multiply u go by the amplification factor V to obtain the straight line cd represent-

ing a constant value of pseudo-velocity V .
5. Multiply u go by the amplification factor D to obtain the straight line de represent-
ing a constant value of deformation D.
6. Draw the line A = u go for periods shorter than Ta and the line D = u go for periods
longer than T f .
7. The transition lines ab and e f complete the spectrum.

We now illustrate use of this procedure by constructing the 84.1th percentile design
spectrum for systems with 5% damping. For convenience, a peak ground acceleration
u go = 1g is selected; the resulting spectrum can be scaled by to obtain the design spec-
trum corresponding to u go = g. Consider also that no specific estimates for peak ground
velocity u go and displacement u go are provided; thus typical values u go /u go = 48 in./sec/g
and u go u go /u 2go = 6, recommended for firm ground, are used. For u go = 1g, these ratios
give u go = 48 in./sec and u go = 36 in.

04 2.71
0.7
Tn
.70
11
1.8
0T

1
-1
n
Pseudo-acceleration A, g

7.40
Tn
-2

0.1
28.4
4T n
-2.58
5
1/33 sec

0.66 sec

4.12 sec
1/8 sec

10 sec

0.01
0.02 0.1 1 10
Natural vibration period Tn, sec

Figure 6.9.5 Elastic pseudo-acceleration design spectrum (84.1th percentile) for ground
motions with u go = 1g, u go = 48 in./sec, and u go = 36 in.; = 5%.
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 234

234 Earthquake Response of Linear Systems Chap. 6

The design spectrum shown in Fig. 6.9.4 is determined by the following steps:

1. The peak parameters for the ground motion: u go = 1g, u go = 48 in./sec, and u go =
36 in. are plotted.
2. From Table 6.9.1, the amplification factors for the 84.1th percentile spectrum and
5% damping are obtained: A = 2.71, V = 2.30, and D = 2.01.
35. The ordinate for the constant-A branch is A = 1g 2.71 = 2.71g, for the con-
stant-V branch: V = 48 2.30 = 110.4, and for the constant-D branch: D =
36 2.01 = 72.4. The three branches are drawn as shown.
6. The line A = 1g is plotted for Tn < 33
1
sec and D = 36 in. for Tn > 33 sec.
7. The transition line ba is drawn to connect the point A = 2.71g at Tn = 18 sec to
u go = 1g at Tn = 33
1
sec. Similarly, the transition line e f is drawn to connect the
point D = 72.4 at Tn = 10 sec to u go = 36 in. at Tn = 33 sec.

200
100
50

20
10
5

2
D, in.

1
1/33 sec

0.5

0.2
0.1
0.66 sec

4.12 sec
1/8 sec

0.05
10 sec

33 sec

0.02
0.01
0.005
0.02 0.05 0.1 0.2 0.5 1 2 5 10 20 50
Tn, sec

Figure 6.9.6 Deformation design spectrum (84.1th percentile) for ground motions with u go = 1g,
u go = 48 in./sec, and u go = 36 in.; = 5%.
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 235

Sec. 6.9 Elastic Design Spectrum 235

With the pseudo-velocity design spectrum known (Fig. 6.9.4), the pseudo-
acceleration design spectrum and the deformation design spectrum are determined using
Eq. (6.6.6) and plotted in Figs. 6.9.5 and 6.9.6, respectively. Observe that A approaches
u go = 1g at Tn = 0 and D tends to u go = 48 in./sec at Tn = 50 sec. The design spectrum
can be defined completely by numerical values for Ta , Tb , Tc , Td , Te , and T f , and equations
for A(Tn ), V (Tn ), or D(Tn ) for each branch of the spectrum. As mentioned before, some of
these periodsTa , Tb , Te , and T f are fixed, but othersTc and Td depend on damping.
The intersections of A = 2.71g, V = 110.4 in./sec, and D = 72.4 in. are determined from
Eq. (6.6.6): Tc = 0.66 sec and Td = 4.12 sec for = 5%. Equations describing various
brances of the pseudo-acceleration design spectrum are given in Fig. 6.9.5.
Repeating the preceding construction of the design spectrum for additional values of
the damping ratio leads to Figs. 6.9.76.9.10. This then is the design spectrum for ground
motions on firm ground with u go = 1g, u go = 48 in./sec, and u go = 36 in. in three diff-
erent forms: pseudo-velocity, pseudo-acceleration, and deformation. Observe that the
pseudo-acceleration design spectrum has been plotted in two formats: logarithmic scales
(Fig. 6.9.8) and linear scales (Fig. 6.9.9).

200

10
0

100 1
= 1%
0.9
50
0.8
= 20%

10
10

0
0.7
20
0.6
V, in./sec

10

10
1

0.5

5 0.4
n.
i
A,

D,
g

0.3
0.

2 0.2

0.1
01
0.

1
0.
1

0
0.5
0 0.2 0.4 0.6 0.8 1
0.02 0.05 0.1 0.2 0.5 1 2 5 10 20 50
T , sec
n

Figure 6.9.7 Pseudovelocity design spectrum for ground motions with u go = 1g, u go = 48 in./sec,
and u go = 36 in.; = 1, 2, 5, 10, and 20%.
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 236

236 Earthquake Response of Linear Systems Chap. 6

The elastic design spectrum provides a basis for calculating the design force and
deformation for SDF systems to be designed to remain elastic. For this purpose the design
spectrum is used in the same way as the response spectrum was used to compute peak
response; see Examples 6.2 to 6.6. The errors in reading spectral ordinates from a four-way
logarithmic plot can be avoided, however, because simple functions of Tn define various
branches of the spectrum in Figs. 6.9.46.9.6.
Parameters that enter into construction of the elastic design spectrum should be
selected considering the factors that influence ground motion mentioned previously. Thus
the selection of design ground motion parameter u go , u go , and u go should be based on
earthquake magnitude, distance to the earthquake fault, fault mechanism, wave-travel-
path geology, and local soil conditions. Results of research on these factors and related
issues are available; they are used to determine site-dependent design spectra for im-
portant projects. Similarly, numerical values for the amplification factors A , V , and
D should be chosen consistent with the expected frequency content of the ground
motion.

= 1%
1

= 20%
A, g

0.1

0.01
0.02 0.1 1 10 20
Tn, sec

Figure 6.9.8 Pseudoacceleration design spectrum (84.1th percentile) for ground mo-
tions with u go = 1g, u go = 48 in./sec, and u go = 36 in.; = 1, 2, 5, 10, and 20%.
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 237

Sec. 6.9 Elastic Design Spectrum 237

5
= 1%

4
= 2%

3
= 5%
A, g

= 10%
2

= 20%
1

0
0 1 Tn, sec 2 3

Figure 6.9.9 Pseudoacceleration design spectrum (84.1th percentile) for ground mo-
tions with u go = 1g, u go = 48 in./sec, and u go = 36 in.; = 1, 2, 5, 10, and 20%.

200
100
50

20
10
5 = 1%

2
D, in.

1 = 20%
0.5

0.2
0.1
0.05

0.02
0.01
0.005
0.02 0.05 0.1 0.2 0.5 1 2 5 10 20 50
Tn, sec

Figure 6.9.10 Deformation design spectrum (84.1th percentile) for ground motions with
u go = 1g, u go = 48 in./sec, and u go = 36 in.; = 1, 2, 5, 10, and 20%.
 Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 238

238 Earthquake Response of Linear Systems Chap. 6

The selected values of u go , u go , and u go are consistent with u go /u go = 48 in./sec/g

and u go u go /u 2go = 6. These ratios are considered representative of ground motions
on firm ground. The resulting spectrum shape is widely used in engineering practice, by
scaling the spectrum to conform to the peak ground acceleration estimated for the site.
Thus if this estimate is 0.4g, the spectrum of Figs. 6.9.76.9.9 multiplied by 0.4 gives the
design spectrum for the site.

6.10 COMPARISON OF DESIGN AND RESPONSE SPECTRA

It is instructive to compare the standard design spectrum developed in Section 6.9 for
firm ground with an actual response spectrum for similar soil conditions. Figure 6.10.1
shows a standard design spectrum for u go = 0.319g, the peak acceleration for the El
Centro ground motion; the implied values for u go and u go are 15.3 in./sec and 11.5 in.,
respectively, based on the standard ratios mentioned in the preceding paragraph. Also
shown in Fig. 6.10.1 is the response spectrum for the El Centro ground motion; recall that
the actual peak values for this motion are u go = 13.04 in./sec and u go = 8.40 in. The
El Centro response spectrum agrees well with the design spectrum in the acceleration-
sensitive region, largely because the peak accelerations for the two are matched. However,
the two spectra are considerably different in the velocity-sensitive region because of the dif-
ferences (15.3 in./sec versus 13.04 in./sec) in the peak ground velocity. Simil-
arly, they are even more different in the displacement-sensitive region because of
the larger differences (11.5 in. versus 8.4 in.) in the peak ground displacement.
The response spectrum for an individual ground motion differs from the design
spectrum even if the peak values u go , u go , and u go for the two spectra are matched. In
Fig. 6.10.2 the response spectrum for the El Centro ground motion is compared with the
design spectrum for ground motion parameters u go = 0.319g, u go = 13.04 in./sec, and
u go = 8.40 in.the same as for the El Centro ground motion. Two design spectra are
included: the 50th percentile spectrum and the 84.1th percentile spectrum. The agree-
ment between the response and design spectra is now better because the ground motion
parameters are matched. However, significant differences remain: over the acceleration-
sensitive region the response spectrum is close to the 84.1th percentile design spectrum;
over the velocity- and displacement-sensitive regions the response spectrum is between
the two design spectra for some periods and below the median design spectrum for other
periods.
Such differences are to be expected because the design spectrum is not intended
to match the response spectrum for any particular ground motion but is constructed to
represent the average characteristics of many ground motions. These differences are due
to the inherent variability in ground motions as reflected in the probability distributions of
the amplification factors and responses; see Fig. 6.9.2.
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 239

Sec. 6.10 Comparison of Design and Response Spectra 239

100
Standard design spectrum, ugo = 0.319g
50

10
10

0
n.
A,

,i
20

D
10 Response spectrum

10
1
V, in./sec

1
0.

1
2

01
0.
1

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
Tn, sec

Figure 6.10.1 Comparison of standard design spectrum (u go = 0.319g) with elastic

response spectrum for El Centro ground motion; = 5%.

100
Design spectra
84.1th percentile
50 50th percentile
10
10

0
n.
A,

,i

20
g

10
10

Response spectrum
1
V, in./sec

5
1
0.

2
01
0.

1
0.
1

0.5
1
0.

00
0

0.
1

0.2
0.02 0.05 0.1 0.2 0.5 1 2 5 10 20 50
Tn, sec

Figure 6.10.2 Comparison of design spectra (u go = 0.319g, u go = 13.04 in./sec,

u go = 8.40 in.) with elastic response spectrum for El Centro ground motion; = 5%.
 Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 240

240 Earthquake Response of Linear Systems Chap. 6

6.11 DISTINCTION BETWEEN DESIGN AND RESPONSE SPECTRA

A design spectrum differs conceptually from a response spectrum in two important ways.
First, the jagged response spectrum is a plot of the peak response of all possible SDF
systems and hence is a description of a particular ground motion. The smooth design
spectrum, however, is a specification of the level of seismic design force, or deformation,
as a function of natural vibration period and damping ratio. This conceptual difference
between the two spectra should be recognized, although in some situations, their shapes
may be similar. Such is the case when the design spectrum is determined by statistical
analysis of several comparable response spectra.
Second, for some sites a design spectrum is the envelope of two different elas-
tic design spectra. Consider a site in southern California that could be affected by two
different types of earthquakes: a Magnitude 6.5 earthquake originating on a nearby fault
and a Magnitude 8.5 earthquake on the distant San Andreas fault. The design spectrum
for each earthquake could be determined by the procedure developed in Section 6.9. The
ordinates and shapes of the two design spectra would differ, as shown schematically in
Fig. 6.11.1, because of the differences in earthquake magnitude and distance of the site
from the earthquake fault. The design spectrum for this site is defined as the envelope
of the design spectra for the two different types of earthquakes. Note that the short-period
portion of the design spectrum is governed by the nearby earthquake, while the long-period
portion of the design spectrum is controlled by the distant earthquake.

Moderate-sized earthquake at small

distance from site
Pseudo-acceleration A, g

Design spectrum for site

Large earthquake at large

distance from site

Natural vibration period Tn, sec

Figure 6.11.1 Design spectrum defined as the envelope of design spectra for earthquakes
originating on two different faults.
 Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 241

Sec. 6.12 Velocity and Acceleration Response Spectra 241

6.12 VELOCITY AND ACCELERATION RESPONSE SPECTRA

We now return to the relative velocity response spectrum and the acceleration response
spectrum that were introduced in Section 6.5. In one sense there is little motivation to
study these true spectra because they are not needed to determine the peak deformations
and forces in a system; for this purpose the pseudo-acceleration (or pseudo-velocity or
deformation) response spectrum is sufficient. A brief discussion of these true spectra
is included, however, because the distinction between them and pseudo spectra has not
always been made in the early publications, and the two have sometimes been used inter-
changeably.

3
= 0.1 (a)
Relative velocity uo
V / ugo or uo / ugo

1
Pseudo-velocity V

4
(b)

= 0.2
= 0.1
V / uo

1
=0

0
0.02 0.1 1 10 50
Tn, sec

Figure 6.12.1 (a) Comparison between pseudo-velocity and relative-velocity response

spectra; = 10%; (b) ratio V /u o for = 0, 10, and 20%.
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 242

242 Earthquake Response of Linear Systems Chap. 6

To study the relationship between these spectra, we write them in mathematical

form. The deformation response of a linear SDF system to an arbitrary ground motion
with zero initial conditions is given by Duhamels integral, Eq. (4.2.3), with p(t) replaced
by peff (t) = m u g (t):
 t
1
u(t) = u g ( )e n (t ) sin[ D (t )] d (6.12.1)
D 0
Using theorems from calculus to differentiate under the integral sign leads to
 t
u(t) = n u(t) u g ( )e n (t ) cos[ D (t )] d (6.12.2)
0

An equation for the acceleration u t (t) of the mass can be obtained by differentiating
Eq. (6.12.2) and adding the ground acceleration u g (t). However, the equation of motion
for the system [Eq. (6.2.1)] provides a more convenient alternative:

u t (t) = n2 u(t) 2 n u(t) (6.12.3)

As defined earlier, the relative-velocity spectrum and acceleration spectrum are plots of u o
and u to , the peak values of u(t) and u t (t), respectively, as functions of Tn .

6.12.1 Pseudo-velocity and Relative-Velocity Spectra

In Fig. 6.12.1a the relative-velocity response spectrum is compared with the pseudo-velocity
response spectrum, both for El Centro ground motion and systems with = 10%. The lat-
ter spectrum is simply one of the curves of Fig. 6.6.4 presented in a different form. Each
point on the relative-velocity response spectrum represents the peak velocity of an SDF
system obtained from u(t) determined by the numerical methods of Chapter 5. The differ-
ences between the two spectra depend on the natural period of the system. For long-period
systems, V is less than u o and the differences between the two are large. This can be under-
stood by recognizing that as Tn becomes very long, the mass of the system stays still while
the ground underneath moves. Thus, as Tn , D u go (see Section 6.8 and Fig. 6.8.5)
and u o u go . Now D u go implies that V 0 because of Eq. (6.6.1). These trends
are confirmed by the results presented in Fig. 6.12.1a. For short-period systems V exceeds
u o , with the differences increasing as Tn becomes shorter. For medium-period systems, the
differences between V and u o are small over a wide range of Tn .
In Fig. 6.12.1b the ratio V /u o is plotted for three damping values, = 0, 10, and
20%. The differences between the two spectra, as indicated by how much the ratio V /u o
differs from unity, are smallest for undamped systems and increase with damping. This can
be explained from Eqs. (6.12.1) and (6.12.2) by observing that for = 0, u(t) and n u(t)
are the same except for the sine and cosine terms in the integrand. With damping, the first
term in Eq. (6.12.2) contributes to u(t), suggesting that u(t) would differ from n u(t) to a
greater degree. Over the medium-period range V can be taken as an approximation to u o
for the practical range of damping.
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 243

Sec. 6.12 Velocity and Acceleration Response Spectra 243

3
= 0.1 (a)
Acceleration uot

A / ugo or uot / ugo 2

Pseudo-acceleration A

2
(b)

=0
A / uot

1
= 0.1

= 0.2

0
0.02 0.1 1 10 50
Tn, sec

Figure 6.12.2 (a) Comparison between pseudo-acceleration and acceleration response

spectra; = 10%; (b) ratio A/u to for = 0, 10, and 20%.

6.12.2 Pseudo-acceleration and Acceleration Spectra

The pseudo-acceleration and acceleration response spectra are identical for systems with-
out damping. This is apparent from Eq. (6.12.3), which for undamped systems specializes
to
u t (t) = n2 u(t) (6.12.4)
The peak values of the two sides are therefore equal, that is,
u to = n2 u o = n2 D = A (6.12.5)
With damping, Eq. (6.12.4) is not valid at all times, but only at the time instants when
u(t) = 0, in particular when u(t) attains its peak u o . At this instant, n2 u o represents
the true acceleration of the mass. The peak value u to of u t (t) does not occur at the same
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 244

244 Earthquake Response of Linear Systems Chap. 6

instant, however, unless = 0. The peak values u to and A occur at the same time and are
equal only for = 0.
Equation (6.12.3) suggests that the differences between A and u to are expected to
increase as the damping increases. This expectation is confirmed by the data presented in
Fig. 6.12.2, where the pseudo-acceleration and the acceleration spectra for the El Centro
ground motion are plotted for = 10%, and the ratio A/u to is presented for three damping
values. The difference between the two spectra is small for short-period systems and is of
some significance only for long-period systems with large values of damping. Thus for a
wide range of conditions the pseudo-acceleration may be treated as an approximation to
the true acceleration.
As the natural vibration period Tn of a system approaches infinity, the mass of the
system stays still while the ground underneath moves. Thus, as Tn , u to 0 and
D u go ; the latter implies that A 0 because of Eq. (6.6.3). Both A and (u t )o 0 as
Tn , but at different rates, as evident from the ratio A/u to plotted as a function of Tn ;
A 0 at a much faster rate because of Tn2 in the denominator of Eq. (6.6.3).
Another way of looking at the differences between the two spectra is by recalling
that m A is equal to the peak value of the elastic-resisting force. In contrast, m u to is equal
to the peak value of the sum of elastic and damping forces. As seen in Fig. 6.12.2b, the
pseudo-acceleration is smaller than the true acceleration, because it is that part of the true
acceleration which gives the elastic force.
Parenthetically, we note that the widespread adoption of the prefix pseudo is in one
sense misleading. The literal meaning of pseudo (false) is not really appropriate since we
are dealing with approximation rather than with concepts that are in any sense false or in-
appropriate. In fact, there is rarely the need to use the pseudo-spectra as approximations
to the true spectra because the latter can be computed by the same numerical procedures
as those used for the former. Furthermore, as emphasized earlier, the pseudo quantities
provide the exact values of the desired deformation and forces.

FURTHER READING

Benioff, H., The Physical Evaluation of Seismic Destructiveness, Bulletin of the Seismological
Society of America, 24, 1934, pp. 398403.
Biot, M. A., Theory of Elastic Systems under Transient Loading with an Application to Earthquake
Proof Buildings, Proceedings, National Academy of Sciences, 19, 1933, pp. 262268.
Biot, M. A., A Mechanical Analyzer for the Prediction of Earthquake Stresses, Bulletin of the
Seismological Society of America, 31, 1941, pp. 151171.
Bolt, B. A., Earthquakes, W.H. Freeman, New York, 1993, Chapters 17.
Clough, R. W., and Penzien, J., Dynamics of Structures, McGraw-Hill, New York, 1993, pp.
586597.
Housner, G. W., Calculating the Response of an Oscillator to Arbitrary Ground Motion, Bulletin
of the Seismological Society of America, 31, 1941, pp. 143149.
 Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 245

Appendix 6: El Centro, 1940 Ground Motion 245

Housner, G. W., and Jennings, P. C., Earthquake Design Criteria, Earthquake Engineering Research
Institute, Berkeley, Calif., 1982, pp. 1941 and 5888.
Hudson, D. E., Response Spectrum Techniques in Engineering Seismology, Proceedings of the
First World Conference in Earthquake Engineering, Berkeley, Calif., 1956, pp. 41 to 412.
Hudson, D. E., Reading and Interpreting Strong Motion Accelerograms, Earthquake Engineering
Research Institute, Berkeley, Calif., 1979, pp. 2270 and 9597.
Hudson, D. E., A History of Earthquake Engineering, Proceedings of the IDNDR International
Symposium on Earthquake Disaster Reduction Technology30th Anniversary of IISEE, Tsukuba,
Japan, 1992, pp. 313.
Mohraz, B., and Elghadamsi, F. E., Earthquake Ground Motion and Response Spectra, Chapter 2
in The Seismic Design Handbook (ed. F. Naeim), Van Nostrand Reinhold, New York, 1989.
Newmark, N. M., and Hall, W. J., Earthquake Spectra and Design, Earthquake Engineering Research
Institute, Berkeley, Calif., 1982, pp. 2937.
Newmark, N. M., and Rosenblueth, E., Fundamentals of Earthquake Engineering, Prentice Hall,
Englewood Cliffs, N.J., 1971, Chapter 7.
Riddell, R., and Newmark, N. M., Statistical Analysis of the Response of Nonlinear Systems Sub-
jected to Earthquakes, Structural Research Series No. 468, University of Illinois at Urbana-Champaign,
Urbana, Ill., August 1979.
Rosenblueth, E., Characteristics of Earthquakes, Chapter 1 in Design of Earthquake Resistant
Structures (ed. E. Rosenblueth), Pentech Press, London, 1980.
Seed, H. B., and Idriss, I. M., Ground Motions and Soil Liquefaction during Earthquakes, Earthquake
Engineering Research Institute, Berkeley, Calif., 1982, pp. 2156.
Veletsos, A. S., Maximum Deformation of Certain Nonlinear Systems, Proceedings of the 4th
World Conference on Earthquake Engineering, Santiago, Chile, Vol. 1, 1969, pp. 155170.
Veletsos, A. S., and Newmark, N. M., Response Spectra for Single-Degree-of-Freedom Elastic
and Inelastic Systems, Report No. RTD-TDR-63-3096, Vol. III, Air Force Weapons Laboratory,
Albuquerque, N.M., June 1964.
Veletsos, A. S., Newmark, N. M., and Chelapati, C. V., Deformation Spectra for Elastic and Elasto-
plastic Systems Subjected to Ground Shock and Earthquake Motion, Proceedings of the 3rd World
Conference on Earthquake Engineering, New Zealand, Vol. II, 1965, pp. 663682.

APPENDIX 6: EL CENTRO, 1940 GROUND MOTION

The northsouth component of the ground motion recorded at a site in El Centro, Cal-
ifornia during the Imperial Valley, California earthquake of May 18, 1940 is shown in
Fig. 6.1.4. This particular version of this record is used throughout this book, and is re-
quired in solving some of the end-of-chapter problems. Numerical values for the ground
acceleration in units of g, the acceleration due to gravity, are presented in Table A6.1.
This includes 1559 data points at equal time spacings of 0.02 sec, to be read row by row;
the first value is at t = 0.02 sec; acceleration at t = 0 is zero. These data are also
available electronically from the National Information Service for Earthquake Engineering
(NISEE), University of California at Berkeley, on the World Wide Web at the following
URL: <http://www.eerc.berkeley.edu/>.
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 246

246 Earthquake Response of Linear Systems Chap. 6

TABLE A6.1 GROUND ACCELERATION DATA

0.00630 0.00364 0.00099 0.00428 0.00758 0.01087 0.00682 0.00277

0.00128 0.00368 0.00864 0.01360 0.00727 0.00094 0.00420 0.00221
0.00021 0.00444 0.00867 0.01290 0.01713 0.00343 0.02400 0.00992
0.00416 0.00528 0.01653 0.02779 0.03904 0.02449 0.00995 0.00961
0.00926 0.00892 0.00486 0.01864 0.03242 0.03365 0.05723 0.04534
0.03346 0.03201 0.03056 0.02911 0.02766 0.04116 0.05466 0.06816
0.08166 0.06846 0.05527 0.04208 0.04259 0.04311 0.02428 0.00545
0.01338 0.03221 0.05104 0.06987 0.08870 0.04524 0.00179 0.04167
0.08513 0.12858 0.17204 0.12908 0.08613 0.08902 0.09192 0.09482
0.09324 0.09166 0.09478 0.09789 0.12902 0.07652 0.02401 0.02849
0.08099 0.13350 0.18600 0.23850 0.21993 0.20135 0.18277 0.16420
0.14562 0.16143 0.17725 0.13215 0.08705 0.04196 0.00314 0.04824
0.09334 0.13843 0.18353 0.22863 0.27372 0.31882 0.25024 0.18166
0.11309 0.04451 0.02407 0.09265 0.16123 0.22981 0.29839 0.23197
0.16554 0.09912 0.03270 0.03372 0.10014 0.16656 0.23299 0.29941
0.00421 0.29099 0.22380 0.15662 0.08943 0.02224 0.04495 0.01834
0.08163 0.14491 0.20820 0.18973 0.17125 0.13759 0.10393 0.07027
0.03661 0.00295 0.03071 0.00561 0.01948 0.04458 0.06468 0.08478
0.10487 0.05895 0.01303 0.03289 0.07882 0.03556 0.00771 0.05097
0.01013 0.03071 0.07156 0.11240 0.15324 0.11314 0.07304 0.03294
0.00715 0.06350 0.13415 0.20480 0.12482 0.04485 0.03513 0.11510
0.19508 0.12301 0.05094 0.02113 0.09320 0.02663 0.03995 0.10653
0.17311 0.11283 0.05255 0.00772 0.01064 0.02900 0.04737 0.06573
0.02021 0.02530 0.07081 0.04107 0.01133 0.00288 0.01709 0.03131
0.02278 0.07686 0.13095 0.18504 0.14347 0.10190 0.06034 0.01877
0.02280 0.00996 0.04272 0.02147 0.00021 0.02104 0.01459 0.05022
0.08585 0.12148 0.15711 0.19274 0.22837 0.18145 0.13453 0.08761
0.04069 0.00623 0.05316 0.10008 0.14700 0.09754 0.04808 0.00138
0.05141 0.10420 0.15699 0.20979 0.26258 0.16996 0.07734 0.01527
0.10789 0.20051 0.06786 0.06479 0.01671 0.03137 0.07945 0.12753
0.17561 0.22369 0.27177 0.15851 0.04525 0.06802 0.18128 0.14464
0.10800 0.07137 0.03473 0.09666 0.15860 0.22053 0.18296 0.14538
0.10780 0.07023 0.03265 0.06649 0.10033 0.13417 0.10337 0.07257
0.04177 0.01097 0.01983 0.04438 0.10860 0.17281 0.10416 0.03551
0.03315 0.10180 0.07262 0.04344 0.01426 0.01492 0.02025 0.05543
0.09060 0.12578 0.16095 0.19613 0.14784 0.09955 0.05127 0.00298
0.01952 0.03605 0.05259 0.04182 0.03106 0.02903 0.02699 0.02515
0.01770 0.02213 0.02656 0.00419 0.01819 0.04057 0.06294 0.02417
0.01460 0.05337 0.02428 0.00480 0.03389 0.00557 0.02274 0.00679
0.00915 0.02509 0.04103 0.05698 0.01826 0.02046 0.00454 0.01138
0.00215 0.00708 0.00496 0.00285 0.00074 0.00534 0.01141 0.00361
0.01863 0.03365 0.04867 0.03040 0.01213 0.00614 0.02441 0.01375
0.01099 0.00823 0.00547 0.00812 0.01077 0.00692 0.02461 0.04230
0.05999 0.07768 0.09538 0.06209 0.02880 0.00448 0.03777 0.01773
0.00231 0.02235 0.01791 0.05816 0.03738 0.01660 0.00418 0.02496
0.04574 0.02071 0.00432 0.02935 0.01526 0.01806 0.02086 0.00793
0.00501 0.01795 0.03089 0.01841 0.00593 0.00655 0.02519 0.05693
0.04045 0.02398 0.00750 0.00897 0.00384 0.00129 0.00642 0.01156
0.02619 0.04082 0.05545 0.04366 0.03188 0.06964 0.05634 0.04303
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 247

Appendix 6: El Centro, 1940 Ground Motion 247

TABLE A6.1 GROUND ACCELERATION DATA (Continued )

0.02972 0.01642 0.00311 0.01020 0.02350 0.03681 0.05011 0.02436

0.00139 0.02714 0.00309 0.02096 0.04501 0.06906 0.05773 0.04640
0.03507 0.03357 0.03207 0.03057 0.03250 0.03444 0.03637 0.01348
0.00942 0.03231 0.02997 0.03095 0.03192 0.02588 0.01984 0.01379
0.00775 0.01449 0.02123 0.01523 0.05170 0.08816 0.12463 0.16109
0.12987 0.09864 0.06741 0.03618 0.00495 0.00420 0.00345 0.00269
0.05922 0.12112 0.18303 0.12043 0.05782 0.00479 0.06740 0.13001
0.08373 0.03745 0.06979 0.10213 0.03517 0.17247 0.13763 0.10278
0.06794 0.03310 0.03647 0.03984 0.00517 0.02950 0.06417 0.09883
0.13350 0.05924 0.01503 0.08929 0.16355 0.06096 0.04164 0.01551
0.01061 0.03674 0.06287 0.08899 0.05430 0.01961 0.01508 0.04977
0.08446 0.05023 0.01600 0.01823 0.05246 0.08669 0.06769 0.04870
0.02970 0.01071 0.00829 0.00314 0.02966 0.06246 0.00234 0.06714
0.04051 0.01388 0.01274 0.00805 0.03024 0.05243 0.02351 0.00541
0.03432 0.06324 0.09215 0.12107 0.08450 0.04794 0.01137 0.02520
0.06177 0.04028 0.01880 0.04456 0.07032 0.09608 0.12184 0.06350
0.00517 0.05317 0.03124 0.00930 0.01263 0.03457 0.03283 0.03109
0.02935 0.04511 0.06087 0.07663 0.09239 0.05742 0.02245 0.01252
0.00680 0.02611 0.04543 0.01571 0.01402 0.04374 0.07347 0.03990
0.00633 0.02724 0.06080 0.03669 0.01258 0.01153 0.03564 0.00677
0.02210 0.05098 0.07985 0.06915 0.05845 0.04775 0.03706 0.02636
0.05822 0.09009 0.12196 0.10069 0.07943 0.05816 0.03689 0.01563
0.00564 0.02690 0.04817 0.06944 0.09070 0.11197 0.11521 0.11846
0.12170 0.12494 0.16500 0.20505 0.15713 0.10921 0.06129 0.01337
0.03455 0.08247 0.07576 0.06906 0.06236 0.08735 0.11235 0.13734
0.12175 0.10616 0.09057 0.07498 0.08011 0.08524 0.09037 0.06208
0.03378 0.00549 0.02281 0.05444 0.04030 0.02615 0.01201 0.02028
0.02855 0.06243 0.03524 0.00805 0.04948 0.03643 0.02337 0.03368
0.01879 0.00389 0.01100 0.02589 0.01446 0.00303 0.00840 0.00463
0.01766 0.03069 0.04372 0.02165 0.00042 0.02249 0.04456 0.03638
0.02819 0.02001 0.01182 0.02445 0.03707 0.04969 0.05882 0.06795
0.07707 0.08620 0.09533 0.06276 0.03018 0.00239 0.03496 0.04399
0.05301 0.03176 0.01051 0.01073 0.03198 0.05323 0.00186 0.05696
0.01985 0.01726 0.05438 0.01204 0.03031 0.07265 0.11499 0.07237
0.02975 0.01288 0.01212 0.03711 0.03517 0.03323 0.01853 0.00383
0.00342 0.02181 0.04704 0.07227 0.09750 0.12273 0.08317 0.04362
0.00407 0.03549 0.07504 0.11460 0.07769 0.04078 0.00387 0.00284
0.00182 0.05513 0.04732 0.05223 0.05715 0.06206 0.06698 0.07189
0.02705 0.01779 0.06263 0.10747 0.15232 0.12591 0.09950 0.07309
0.04668 0.02027 0.00614 0.03255 0.00859 0.01537 0.03932 0.06328
0.03322 0.00315 0.02691 0.01196 0.00300 0.00335 0.00970 0.01605
0.02239 0.04215 0.06191 0.08167 0.03477 0.01212 0.01309 0.01407
0.05274 0.02544 0.00186 0.02916 0.05646 0.08376 0.01754 0.04869
0.02074 0.00722 0.03517 0.00528 0.04572 0.08617 0.06960 0.05303
0.03646 0.01989 0.00332 0.01325 0.02982 0.01101 0.00781 0.02662
0.00563 0.01536 0.03635 0.05734 0.03159 0.00584 0.01992 0.00201
0.01589 0.01024 0.03636 0.06249 0.04780 0.03311 0.04941 0.06570
0.08200 0.04980 0.01760 0.01460 0.04680 0.07900 0.04750 0.01600
0.01550 0.00102 0.01347 0.02795 0.04244 0.05692 0.03781 0.01870
0.00041 0.01952 0.00427 0.01098 0.02623 0.04148 0.01821 0.00506
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 248

248 Earthquake Response of Linear Systems Chap. 6

TABLE A6.1 GROUND ACCELERATION DATA (Continued )

0.00874 0.03726 0.06579 0.02600 0.01380 0.05359 0.09338 0.05883

0.02429 0.01026 0.04480 0.01083 0.01869 0.02655 0.03441 0.02503
0.01564 0.00626 0.01009 0.01392 0.01490 0.04372 0.03463 0.02098
0.00733 0.00632 0.01997 0.00767 0.03532 0.03409 0.03287 0.03164
0.02403 0.01642 0.00982 0.00322 0.00339 0.02202 0.01941 0.06085
0.10228 0.07847 0.05466 0.03084 0.00703 0.01678 0.01946 0.02214
0.02483 0.01809 0.00202 0.02213 0.00278 0.01656 0.03590 0.05525
0.07459 0.06203 0.04948 0.03692 0.00145 0.04599 0.04079 0.03558
0.03037 0.03626 0.04215 0.04803 0.05392 0.04947 0.04502 0.04056
0.03611 0.03166 0.00614 0.01937 0.04489 0.07040 0.09592 0.07745
0.05899 0.04052 0.02206 0.00359 0.01487 0.01005 0.00523 0.00041
0.00441 0.00923 0.01189 0.01523 0.01856 0.02190 0.00983 0.00224
0.01431 0.00335 0.00760 0.01856 0.00737 0.00383 0.01502 0.02622
0.01016 0.00590 0.02196 0.00121 0.01953 0.04027 0.02826 0.01625
0.00424 0.00196 0.00031 0.00258 0.00486 0.00713 0.00941 0.01168
0.01396 0.01750 0.02104 0.02458 0.02813 0.03167 0.03521 0.04205
0.04889 0.03559 0.02229 0.00899 0.00431 0.01762 0.00714 0.00334
0.01383 0.01314 0.04011 0.06708 0.04820 0.02932 0.01043 0.00845
0.02733 0.04621 0.03155 0.01688 0.00222 0.01244 0.02683 0.04121
0.05559 0.03253 0.00946 0.01360 0.01432 0.01504 0.01576 0.04209
0.02685 0.01161 0.00363 0.01887 0.03411 0.03115 0.02819 0.02917
0.03015 0.03113 0.00388 0.02337 0.05062 0.03820 0.02579 0.01337
0.00095 0.01146 0.02388 0.03629 0.01047 0.01535 0.04117 0.06699
0.05207 0.03715 0.02222 0.00730 0.00762 0.02254 0.03747 0.04001
0.04256 0.04507 0.04759 0.05010 0.04545 0.04080 0.02876 0.01671
0.00467 0.00738 0.00116 0.00506 0.01128 0.01750 0.00211 0.02173
0.04135 0.06096 0.08058 0.06995 0.05931 0.04868 0.03805 0.02557
0.01310 0.00063 0.01185 0.02432 0.03680 0.04927 0.02974 0.01021
0.00932 0.02884 0.04837 0.06790 0.04862 0.02934 0.01006 0.00922
0.02851 0.04779 0.02456 0.00133 0.02190 0.04513 0.06836 0.04978
0.03120 0.01262 0.00596 0.02453 0.04311 0.06169 0.08027 0.09885
0.06452 0.03019 0.00414 0.03848 0.07281 0.05999 0.04717 0.03435
0.03231 0.03028 0.02824 0.00396 0.02032 0.00313 0.01406 0.03124
0.04843 0.06562 0.05132 0.03702 0.02272 0.00843 0.00587 0.02017
0.02698 0.03379 0.04061 0.04742 0.05423 0.03535 0.01647 0.01622
0.01598 0.01574 0.00747 0.00080 0.00907 0.00072 0.01051 0.02030
0.03009 0.03989 0.03478 0.02967 0.02457 0.03075 0.03694 0.04313
0.04931 0.05550 0.06168 0.00526 0.07220 0.06336 0.05451 0.04566
0.03681 0.03678 0.03675 0.03672 0.01765 0.00143 0.02051 0.03958
0.05866 0.03556 0.01245 0.01066 0.03376 0.05687 0.04502 0.03317
0.02131 0.00946 0.00239 0.00208 0.00654 0.01101 0.01548 0.01200
0.00851 0.00503 0.00154 0.00195 0.00051 0.00092 0.01135 0.02363
0.03590 0.04818 0.06045 0.07273 0.02847 0.01579 0.06004 0.05069
0.04134 0.03199 0.03135 0.03071 0.03007 0.01863 0.00719 0.00425
0.01570 0.02714 0.03858 0.02975 0.02092 0.02334 0.02576 0.02819
0.03061 0.03304 0.01371 0.00561 0.02494 0.02208 0.01923 0.01638
0.01353 0.01261 0.01170 0.00169 0.00833 0.01834 0.02835 0.03836
0.04838 0.03749 0.02660 0.01571 0.00482 0.00607 0.01696 0.00780
0.00136 0.01052 0.01968 0.02884 0.00504 0.03893 0.02342 0.00791
0.00759 0.02310 0.00707 0.00895 0.02498 0.04100 0.05703 0.02920
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 249

Appendix 6: El Centro, 1940 Ground Motion 249

TABLE A6.1 GROUND ACCELERATION DATA (Continued )

0.00137 0.02645 0.05428 0.03587 0.01746 0.00096 0.01937 0.03778

0.02281 0.00784 0.00713 0.02210 0.03707 0.05204 0.06701 0.08198
0.03085 0.02027 0.07140 0.12253 0.08644 0.05035 0.01426 0.02183
0.05792 0.09400 0.13009 0.03611 0.05787 0.04802 0.03817 0.02832
0.02970 0.03993 0.05017 0.06041 0.07065 0.08089 0.00192 0.08473
0.01846 0.00861 0.03652 0.06444 0.06169 0.05894 0.05618 0.06073
0.06528 0.04628 0.02728 0.00829 0.01071 0.02970 0.03138 0.03306
0.03474 0.03642 0.04574 0.05506 0.06439 0.07371 0.08303 0.03605
0.01092 0.05790 0.04696 0.03602 0.02508 0.01414 0.03561 0.05708
0.07855 0.06304 0.04753 0.03203 0.01652 0.00102 0.00922 0.01946
0.07032 0.05590 0.04148 0.05296 0.06443 0.07590 0.08738 0.09885
0.06798 0.03710 0.00623 0.02465 0.05553 0.08640 0.11728 0.14815
0.08715 0.02615 0.03485 0.09584 0.07100 0.04616 0.02132 0.00353
0.02837 0.05321 0.00469 0.06258 0.12048 0.09960 0.07872 0.05784
0.03696 0.01608 0.00480 0.02568 0.04656 0.06744 0.08832 0.10920
0.13008 0.10995 0.08982 0.06969 0.04955 0.04006 0.03056 0.02107
0.01158 0.00780 0.00402 0.00024 0.00354 0.00732 0.01110 0.00780
0.00450 0.00120 0.00210 0.00540 0.00831 0.02203 0.03575 0.04947
0.06319 0.05046 0.03773 0.02500 0.01227 0.00046 0.00482 0.00919
0.01355 0.01791 0.02228 0.00883 0.00462 0.01807 0.03152 0.02276
0.01401 0.00526 0.00350 0.01225 0.02101 0.01437 0.00773 0.00110
0.00823 0.01537 0.02251 0.01713 0.01175 0.00637 0.01376 0.02114
0.02852 0.03591 0.04329 0.03458 0.02587 0.01715 0.00844 0.00027
0.00898 0.00126 0.00645 0.01417 0.02039 0.02661 0.03283 0.03905
0.04527 0.03639 0.02750 0.01862 0.00974 0.00086 0.01333 0.02752
0.04171 0.02812 0.01453 0.00094 0.01264 0.02623 0.01690 0.00756
0.00177 0.01111 0.02044 0.02977 0.03911 0.02442 0.00973 0.00496
0.01965 0.03434 0.02054 0.00674 0.00706 0.02086 0.03466 0.02663
0.01860 0.01057 0.00254 0.00063 0.00128 0.00319 0.00510 0.00999
0.01488 0.00791 0.00093 0.00605 0.00342 0.01288 0.02235 0.03181
0.04128 0.02707 0.01287 0.00134 0.01554 0.02975 0.04395 0.03612
0.02828 0.02044 0.01260 0.00476 0.00307 0.01091 0.00984 0.00876
0.00768 0.00661 0.01234 0.01807 0.02380 0.02953 0.03526 0.02784
0.02042 0.01300 0.03415 0.00628 0.00621 0.00615 0.00609 0.00602
0.00596 0.00590 0.00583 0.00577 0.00571 0.00564 0.00558 0.00552
0.00545 0.00539 0.00532 0.00526 0.00520 0.00513 0.00507 0.00501
0.00494 0.00488 0.00482 0.00475 0.00469 0.00463 0.00456 0.00450
0.00444 0.00437 0.00431 0.00425 0.00418 0.00412 0.00406 0.00399
0.00393 0.00387 0.00380 0.00374 0.00368 0.00361 0.00355 0.00349
0.00342 0.00336 0.00330 0.00323 0.00317 0.00311 0.00304 0.00298
0.00292 0.00285 0.00279 0.00273 0.00266 0.00260 0.00254 0.00247
0.00241 0.00235 0.00228 0.00222 0.00216 0.00209 0.00203 0.00197
0.00190 0.00184 0.00178 0.00171 0.00165 0.00158 0.00152 0.00146
0.00139 0.00133 0.00127 0.00120 0.00114 0.00108 0.00101 0.00095
0.00089 0.00082 0.00076 0.00070 0.00063 0.00057 0.00051 0.00044
0.00038 0.00032 0.00025 0.00019 0.00013 0.00006 0.00000
 100
250
ICC Oregon

50
10
Chopra: Prentice-Hall
PAGES JUL. 19, 2000

10 0
50
5
20
(503) 221-9911
14:15

10
10
1 5

5
Ps
eu 1 n.
do 0. 1 ,i
-a D

Pseudo-velocity V, in./sec
cc n
el 0.
5 io
er at
Page 250

2 at rm
io o
n ef
A, D
g
0. 01
1 1 0.

0.5
0. 1
01 00
0.
Earthquake Response of Linear Systems

0.2
0.02 0.05 0.1 0.2 0.5 1 2 5 10 20 50
Natural vibration period Tn, sec
Chap. 6

Figure A6.1 Graph paper with four-way logarithmic scales.

Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 251

Chap. 6 Problems 251

PROBLEMS

6.1 Determine the deformation response u(t) for 0 t 15 sec for an SDF system with natural
period Tn = 2 sec and damping ratio = 0 to El Centro 1940 ground motion. The ground
acceleration values are available at every t = 0.02 sec in Appendix 6. Implement the
numerical time-stepping algorithm of Section 5.2. Plot u(t) and compare it with Fig. 6.4.1.
6.2 Solve Problem 6.1 for = 5%.
6.3 Solve Problem 6.2 by the central difference method.
6.4 Derive equations for the deformation, pseudo-velocity, and pseudo-acceleration response spec-
tra for ground acceleration u g (t) = u go (t), where (t) is the Dirac delta function and u go
is the increment in velocity or the magnitude of the acceleration impulse. Plot the spectra for
= 0 and 10%.
6.5 An SDF undamped system is subjected to ground motion u g (t) consisting of a sequence of
two acceleration impulses, each with a velocity increment u go , as shown in Fig. P6.5.
(a) Plot the displacement response of the system td /Tn = 18 , 14 , 12 , and 1. For each case show
the response to individual impulses and the combined response.
(b) Determine the deformation response spectrum for this excitation by plotting u o /(u go /n )
as a function of td /Tn . Indicate separately the maximum occurring during t td and during
t td .
(c) Determine the pseudo-velocity response spectrum for this excitation with td = 0.5 sec by
plotting V /u go as a function of f n = 1/Tn .

.. ..
ug ug
. . .
ugo ugo ugo

t t
td . td
ugo

Figure P6.5 Figure P6.6

6.6 Repeat Problem 6.5 with the two velocity pulses acting in the same direction (Fig. P6.6).
6.7 Consider harmonic ground motion u g (t) = u go sin(2 t/T ).
(a) Derive equations for A and for u to in terms of the natural vibration period Tn and the
damping ratio of the SDF system. A is the peak value for the pseudo-acceleration, and u to
is the peak value of the true acceleration. Consider only the steady-state response.
(b) Show that A and u to are identical for undamped systems but different for damped systems.
(c) Graphically display the two response spectra by plotting the normalized values A/u go and
u to /u go against Tn /T , the ratio of the natural vibration period of the system and the period of
the excitation.

Denotes that a computer is necessary to solve this problem.

Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 252

252 Earthquake Response of Linear Systems Chap. 6

6.8 Certain types of near-fault ground motion can be represented by a full sinusoidal-cycle of
6.9 ground acceleration (Fig. P6.8) or a full cosine cycle of ground acceleration (Fig. P6.9).
Assuming the ground velocity and displacement is zero at time zero, plot the ground ve-
locity and ground displacement as a function of time. Determine the pseudo-acceleration
response spectrum for undamped systems. Plot this spectrum against td /Tn . How will the
true-acceleration response spectrum differ?
.. ..
ug
u g
.. ..
ugo ugo

t t
td td

.. ..
ugo ugo

Figure P6.8 Figure P6.9

6.10 A 10-ft-long vertical cantilever made of a 6-in.-nominal-diameter standard steel pipe supports
a 3000-lb weight attached at the tip, as shown in Fig. P6.10. The properties of the pipe are:
outside diameter = 6.625 in., inside diameter = 6.065 in., thickness = 0.280 in., second moment
of cross-sectional area I = 28.1 in4 , Youngs modulus E = 29,000 ksi, and weight = 18.97
lb/ft length. Determine the peak deformation and the bending stress in the cantilever due to
the El Centro ground motion; assume that = 5%.

3000 lb

steel pipe
6 std

10

Figure P6.10

6.11 (a) A full water tank is supported on an 80-ft-high cantilever tower. It is idealized as an SDF
system with weight w = 100 kips, lateral stiffness k = 4 kips/in., and damping ratio = 5%.
The tower supporting the tank is to be designed for ground motion characterized by the design
spectrum of Fig. 6.9.5 scaled to 0.5g peak ground acceleration. Determine the design values
of lateral deformation and base shear.
(b) The deformation computed for the system in Part (a) seemed excessive to the structural
designer, who decided to stiffen the tower by increasing its size. Determine the design val-
ues of deformation and base shear for the modified system if its lateral stiffness is 8 kips/in.;
assume that the damping ratio is still 5%. Comment on how stiffening the system has affected
the design requirements. What is the disadvantage of stiffening the system?
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 253

Chap. 6 Problems 253

(c) If the stiffened tower were to support a tank weighing 200 kips, determine the design re-
quirements; assume for purposes of this example that the damping ratio is still 5%. Comment
on how the increased weight has affected the design requirements.
6.12 Solve Problem 6.11 modified as follows: w = 16 kips in part (a) and w = 32 kips in part (c).
6.13 Solve Problem 6.11 modified as follows: w = 1600 kips in part (a) and w = 3200 kips in
part (c).
6.14 A one-story reinforced-concrete building is idealized for structural analysis as a massless
frame supporting a dead load of 10 kips at the beam level. The frame is 24 ft wide and 12 ft
high. Each column, clamped at the base, has a 10-in.-square cross section. The Youngs
modulus of concrete is 3 103 ksi, and the damping ratio of the building is estimated as 5%.
If the building is to be designed for the design spectrum of Fig. 6.9.5 scaled to a peak ground
acceleration of 0.5g, determine the design values of lateral deformation and bending moments
in the columns for two conditions:
(a) The cross section of the beam is much larger than that of the columns, so the beam may
be assumed as rigid in flexure.
(b) The beam cross section is much smaller than the columns, so the beam stiffness can be
ignored. Comment on the influence of beam stiffness on the design quantities.
6.15 The columns of the frame of Problem 6.14 with condition (a) (i.e., rigid beam) are hinged at
the base. For the same design earthquake, determine the design values of lateral deformation
and bending moments on the columns. Comment on the influence of base fixity on the design
deformation and bending moments.
6.16 Determine the peak response of the one-story industrial building of Example 1.2 to ground
motion characterized by the design spectrum of Fig. 6.9.5 scaled to a peak ground motion
acceleration of 0.25g.
(a) For north-south excitation determine the lateral displacement of the roof and the bending
moments in the columns.
(b) For east-west excitation determine the lateral displacement of the roof and the axial force
in each brace.
6.17 A small one-story reinforced-concrete building shown in Fig. P6.17 is idealized as a massless
frame supporting a total dead load of 10 kips at the beam level. Each 10-in.-square column
is hinged at the base; the beam may be assumed to be rigid in flexure; and E = 3 103 ksi.
Determine the peak response of this structure to ground motion characterized by the design
spectrum of Fig. 6.9.5 scaled to 0.25g peak ground acceleration. The response quantities of
interest are the displacement at the top of the frame and the bending moments in the two
columns. Draw the bending moment diagram.

10

20

Figure P6.17
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 254

254 Earthquake Response of Linear Systems Chap. 6

6.18 A one-story steel frame of 24-ft span and 12-ft height has the following properties: The second
moments of cross-sectional area for beam and columns are Ib = 160 in4 and Ic = 320 in4 ,
respectively; the elastic modulus for steel is 30 103 ksi. For purposes of dynamic analysis
the frame is considered massless with a weight of 100 kips lumped at the beam level; the
columns are clamped at the base; the damping ratio is estimated at 5%. Determine the peak
values of lateral displacement at the beam level and bending moments throughout the frame
due to the design spectrum of Fig. 6.9.5 scaled to a peak ground acceleration of 0.5g.
6.19 Solve Problem 6.18 assuming that the columns are hinged at the base. Comment on the
influence of base fixity on the design deformation and bending moments.
6.20 The ash hopper in Fig. 6.20 consists of a bin mounted on a rigid platform supported by four
columns 24 ft long. The weight of the platform, bin, and contents is 150 kips and may be
taken as a point mass located 6 ft above the bottom of the platform. The columns are braced
in the longitudinal direction, that is, normal to the plane of the paper, but are unbraced in the
transverse direction. The column properties are: A = 20 in2 , E = 29,500 ksi, I = 2000
in4 , and S = 170 in3 . Taking damping ratio to be 5%, find the peak lateral displacement and
the peak stress in the columns due to gravity and the earthquake characterized by the design
spectrum of Fig. 6.9.5 scaled to 13 g acting in the transverse direction. Take the columns to be
clamped at the base and at the rigid platform. Neglect axial deformation of the column and
gravity effects on the lateral stiffness.

24

20
Figure P6.20

6.21 The structure of Example 1.7 subjected to rotational acceleration u g = (t) of the foundation.
Derive an equation for the rotation u (t) of the roof slab in terms of I O , k x , k y , b, and d.
Neglect damping.
6.22 The peak response of the system described in Examples 1.7 and 2.4 due to rotational ground
acceleration u g (see Fig. E1.7) is to be determined; = 5%. The design spectrum for
translational ground acceleration (b/2)u g is given by Fig. 6.9.5 scaled to a peak ground
acceleration of 0.05g. Determine the displacement at each corner of the roof slab, the base
torque, and the bending moments about the x and y axes at the base of each column.
6.23 For the design earthquake at a site, the peak values of ground acceleration, velocity, and dis-
placement have been estimated: u go = 0.5g, u go = 24 in./sec, and u go = 18 in. For systems
with 2% damping ratio, construct the 50th and 84.1th percentile design spectra.
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 255

Chap. 6 Problems 255

(a) Plot both spectra, together, on four-way log paper.

(b) Plot the 84.1th percentile spectrum for pseudo-acceleration on log-log paper, and deter-
mine the equations for A(Tn ) for each branch of the spectrum and the period values at the
intersections of the branches.
(c) Plot the spectrum of part (b) on a linear-linear graph (the Tn scale should cover the range
of 0 to 5 sec).
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:15
ICC Oregon (503) 221-9911
Page 256