EARTHQUAKE ENGINEERING AND STRUCTURAL DYNAMICS, VOL. 19.

77-90 (1990)

EVALUATION OF SEISMIC ENERGY IN STRUCTURES

CHIA-MING UANG*
Department of Civil Engineering, Northeastern University, Boston. MA 02115. U.S.A.
AND
VITELMO V. BERTERO'
University of Cali&ornia, Berkeley, C A 94720, U.S.A.

SUMMARY
Research engineers use two types of energy equations to study single-degree-of-freedom (SDOF) systems subject to
earthquake induced ground motions. The first method uses an absolute energy formulation; the second method uses a
relative energy formulation. While the relative energy formulation has been used in the majority of previous investigations,
this study shows that the absolute energy equation is physically more meaningful. For a given ductility ratio, the input
energy demands calculated by both methods are significantly differentfor both the short and long period ranges although
the results are similar in the intermediate period range. A comparison between the analytically predicted absolute input
energy of a SDOF system with the experimentally measured input energy of a six-storey braced steel frame shows good
correlation.

INTRODUCTION
Traditionally displacement ductility has been used as a criterion to establish inelastic design response spectra
l 9 The maximum required strength (or capacity for
(IDRS) for the earthquake-resistant design of b~ildings.'~.
lateral force) of the building structure is then based on selected IDRS. As an alternative, an energy-based
design method was proposed by G . W. Housner.' Since then this approach has gained extensive
attention.l.2.4,5.9-12. 1 5 . 2 0 This design method is based on the premise that the energy demand during an
earthquake (or an ensemble of earthquakes) can be predicted and that the energy supply of a structural
element (or a structural system) can be established. A satisfactory design implies that the energy supply should
be larger than the energy demand.
The first objective of this paper is to analyse the physical meaning of two energy equations that are derived
and used in the literature. The second objective is to use these two definitions to construct inelastic input
energy spectra for a single-degree-of-freedom(SDOF) system, and then to compare the spectra, as well as to
evaluate the reliability of using them for SDOF systems to predict the input energy to multi-storey buildings.

ENERGY EQUATIONS FOR A SDOF SYSTEM

Evaluation of the energy equations is limited to a linearly elastic-perfectly plastic SDOF system. Three
earthquake ground motions (see Table I and Figure 1) including two recently recorded destructive
earthquakes are used in this study.

*Assistant Professor of Civil Engineering.

'Professor of Civil Engineering.

0098-8847/90/010077-14$07~00
0 1990 by John Wiley & Sons, Ltd.

Receiued 13 December 1988

Revised 30 May 1989

C.-M. UANG AND V. V. BERTERO

78

Table I. Earthquake record data

No.

'
2

Earthquake

Record

Comp.

Focal
depth (km)

M,

160

6*3

4.2-5.0

8.1

IX

8.0

5-4

VIII-IX

Imperia1
El Centro NWE
May 18, 1940
Mexico City
N90E
SCT
September 19, 1985
San Salvador
CIG
N90E
October 10, 1986

Acceleration
0.31
0 2
0.1
0.0

MMI

Epicentral t ,
dist. (km) (see)

Geology

30 m stiff clay
vllLvllrvolcanic rock

9.3

24.4

Soft lacustrine clay

360

38.8

Fluviatile

9.0

4.3

(g)

El C e n t r o

-0.1
-0.2
-0.31

' -

20

10

30

40

50

60

Time ( s e c )
Acceleration

(g)

0.20

Mexico C i t y

0.10
0.0
-0.10

-0.20

100

50

150

200

Time ( s e c )
Acceleration

-0.8

(g)

9 1 0 1 1 1 2

Time ( s e c )

Figure 1. Earthquake ground motion acceleration time histories

Given a viscous damped SDOF system subjected to a horizontal earthquake ground motion [Figure 2(a)],
the equation of motion can be written as
mut

+ cir +f, = 0

(1)

where rn = mass, c = viscous damping coefficient, f,= restoring force, u , = u + uq = absolute (or total)
displacement of the mass, v=relative displacement of the mass with respect to the ground and, v g
=earthquake ground displacement.
Note thatf, may be expressed as kv for a linear elastic system (k = stiffness). By letting iit = u+ u,, equation
may be rewritten as
mu+cd+f,=

-mu,

(2)

Therefore the structural system in Figure 2(a) can be treated conveniently as the equivalent system in Figure
2(b) with a fixed base and subjected to an effective horizontal dynamic force of magnitude -mu,. Although

SEISMIC ENERGY IN STRUCTURES

(a) Moving base system

79

(b) Equivalent fixed-base system

Figure 2. Mathematical model of a SDOF system subjected to an earthquake ground motion

both systems give the same relative displacement, this convenience does cause some confusion in the
definition of input energy and kinetic energy. Depending upon whether equation (1) or (2) is used to derive the
energy equation, different definitions of input and kinetic energies may result.
Method I-Derivation of absolute energy equation
Integrate equation (1) with respect to v from the time that the ground motion excitation starts:

s s s

mu,dv + cddo + f,dv=O

(3)

Replacing v by (v, - v g ) in the first term of equation (3), then

s s

mv,dv, = m(d,)2 - smfi,dvg

2

mv,do = mv,(do, - dug)=

(4)

Substituting equation (4) into equation (3) yields

+
+s
4
s
cddo

m(d1)2

f,dv = mu,dv,

The first term of the above equation is the absolute kinetic energy Ek,
m(dd2
Ek = 2

because absolute velocity (d,) is used to calculate the kinetic energy. The second term in equation (5) is the
damping energy (Er), which is always non-negative because
E,=

cddv=

Cljdt

(7)

The third term in equation (5) is the absorbed energy ( E J , which is composed of recoverable elastic strain
energy ( E , ) and irrecoverable hysteretic energy ( Eh):

E , = f,dv = E, + E ,

(8)

where E , = (f,)/2k.
The right-hand-side term in equation (5) is, by definition, the input energy ( E J :

Ei = (mu,)du,

(9)

80

C.-M. UANG AND V. V. BERTERO

In this study Ei is defined as the absolute input energy. This definition is physically meaningful in that the
term mut represents the inertia force applied to the structure. This force, which from equation (1) is equal to
restoring force plus damping force, is the same as the total force applied to the structure foundation. Therefore
Ei represents the work done by the total base shear at the foundation on the foundation displacement. The
absolute energy equation [equation ( 5 ) ] then can be written as follows:
Ei = E;

+ E , + E , = E; + E , + E , + E ,

(10)

Method 2-Derivation of relative energy equation

Integrate equation (2) with respect to u:

I J I I
mvdv+

cddv+ &do = - mvgdv

Notice that the second term ( = Ec) and the third term ( = E,) on the left side of the equation remain unchanged.
The first term in the previous equation can be rewritten as

which is the relative kinetic energy (E;)calculated from the relative velocity:
m(d)2
E; = __
2

The right-hand-side term of equation ( 1 1 ) is conventionally defined as the input energy (Ef):

Ef = - mvgdv

(13)

In this study Ef is defined as the relativeinput energy. This definition of input energy physically represents the
work done by the static equivalent lateral force (-mug) on the equivalent fixed-base system, that is, it neglects
the effect of the rigid body translation of the structure. The relative energy equation is then expressed as

Ef = EL+ E , + E , = E k + E t + E , + E h

(14)

Comparison of energy time histories

Input energy as defined by either equation (9) or equation (13) is a function of time. Figure 3 shows the
energy time histories for a short period (T=0.2 sec) and a long period ( T = 5.0 sec) elastic-perfectly plastic
SDOF structure subjected to the 1986 San Salvador earthquake ground motion.
Damping energy (Ec),strain energy (E,) and hysteretic energy (Eh)terms are uniquely defined, irrespective
of what method is used. But the input energy and kinetic energy are different, depending upon which method
is used. Note the significant difference in the magnitudes of Ei and EI for the long period structure. When the
period of the structure is significantly larger than the predominant excitation period of the ground motion, the
mass of the structure practically does not move. Therefore the absolute input energy for the relatively long
period structure should be low, as is reflected in the E , time history.
To construct input energy spectra, the time at which the input energy is evaluated should be specified. Most
of the previous researchers evaluated the input energy at (i) the end of the ground motion duration, or (ii) this
duration plus a time equal to one half the period of free vibration of the structure,20or (iii) this duration plus a
time at which the velocity of the structure changes sign. If the relative energy equation is used, the time at
which the input energy is evaluated by the methods just described is suitable for short period structures (see
Figure 3); for long period structures these methods may significantly underestimate the maximum input
energy that may occur early in the ground motion shaking [see Figure 3(b)]. For this reason, in this study the
maximum input energy evaluated throughout the whole duration is used to construct the input energy
spectra. It should be noted that if Ei and Ei are evaluated a t the end of the ground motion duration, which

81

SEISMIC ENERGY IN STRUCTURES

Energy (kN-m)

0.2 sec

Energy (kN-m)

0.6

0.06

0.4

0.04

0.2

0.02

5.0 sec

0.0

0 0

6
8
1
Time (sec)

6
8
1
Time ( s e c )

(a) Energy Terms Defined by Method 1

Energy (kN-m)

0.2

Energy (kN-rn)

SEC

5.0 sec

0.4,

Time (sec)

Time ( s e c )

(b) Energy Terms Defined by Method 2

Figure 3. Energy time histories for an elastic-perfectly plastic system subjected to the 1986 San Salvador earthquake
(ductility ratio 5, 5 per cent damping)

corresponds to the time at which 6,= 0, the rigid body kinetic energy is zero and hence the values of Ei and E;
are identical.
To solve the problem of the non-zero initial condition of each ground motion, the method proposed by
Pecknold and RiddleI4 to prefix a two second acceleration pulse was adopted in the analysis.
Estimation of the difference between input energies from different dejnitions
Energy equation (9) was used by Berg and T h o m a i d e ~ Goel
, ~ and Berg,5 Mahin and Lin, Uang and
Bertero, among others. Equation (13) was used by most of the other researchers. The difference between the
input energies of Methods 1 and 2 is derived below.
r

E i = (mu,)du, = (mu,)(do,-du) = (mu,)do,m

= -(I#

m
2

- -(d)

m(u + ii,)du

m
+ EI = -(d,)
+ mddg + E;
2

i.e.
Ei-E;

m
2

= -(fig)

+ mdd,

It can be proved easily that the difference between the kinetic energies due to the different definitions is

Because the last term in the above equation contains the term d, the difference cannot be estimated easily.
However, the values of Ei and EI for very long and very short period structures can be calculated as follows.

82

C.-M.
UANG AND V. V. BERTERO

For a structure with uery long period (T+co), the input energy tends to converge to a constant value,
depending upon what definition of input energy is used. For a structure with infinitely long period,

u= - u B
u,=u+u,=o;

ut=o

therefore,
Method 1:

3
m = [Cldug = [(O)du,

=0

Method 2:

i.e. the difference between the input energies Ei and Ei for a structure with T+co is equal to m(d,)/2. If the
input energy EI is evaluated at the end of duration, its value will be very small because d, tends to be
vanishingly small. If Ei is evaluated as the maximum throughout the duration, then E i / m will then converge to
d.$,,,,,/2for long period structures.
For a structure with uery short period (T+O), the input energy will also converge to a constant value,
depending upon the definition used. For a structure with zero period, i.e. a rigid structure.
i; -ij
1-

u,=ug,

or u = O

therefore,
Method 1:
Method 2:

i.e. the difference between the input energy spectra for a structure with zero period is equal to m d ~ ~ , , , , / 2 .
Comparison of input energy spectra
By specifying a constant displacement ductility ratio of 5, the input energy spectra for three earthquake
records (see Table I) are generated. The input energy is converted to an equivalent velocity by the following
relationship:

The input energy equivalent velocity spectra are shown in Figure 4.

Figure 4 shows that Viand Vi are very close in the intermediate period ranges; to be more specific, the input
energies calculated by Method 1 or 2 are very close in the vicinity of the predominant excitation periods of the
earthquake ground motions. The difference between Vi and Vi increases for structures with longer and shorter
periods. The level of maximum ground velocity Zjg(,,)
is also shown in Figure 4 for each earthquake record.
The trend that Vi converges to
as the period of the structure tends to zero and that VI converges to
dgtrnax)as the period of the structure tends to infinity [as stated in equations (16b) and (17a)l is clearly shown in
Figure 4. The tendency for V fin the short period range and for Vi in the long period range to decrease to zero
can also be observed [see equations (16a) and (17b)l.
InJuence of displacement ductility ratios on input energy spectra
It has been concluded that Ef (or Vi in the form of equivalent velocity) spectral values evaluated at the end of
the ground motion duration are relatively insensitive to the displacement ductility level.20The variation of

SEISMIC ENERGY IN STRUCTURES

v, Or v; (m/SCC)

Period (se c )

Mexico C i t y

Period (sec)

83
vi Or

vl

(m/sec)

San Salvador

Period (sec)

Figure 4. Comparison of input energy equivalent velocity spectra for ductility ratio 5 (5 per cent damping):

-v---

.-v,

Pel-iod ( s e c )

- _ _ _vs(msl)

Period ( s e c )

Period ( s e c )

Figure 5. Comparison of absolute input energy equivalent velocity spectra for ductility ratios 2, 5 and 8 (5 per cent damping):
__-_- p=2
-p = 5 - . -. - p = 8

absolute input energy equivalent velocity spectra for displacement ductility ratios of 2,5 and 8 are shown in
Figure 5. It can be observed that the input energy spectra are generally insensitive to the level of ductility ratio.
The only exceptions to this observation are the spectra of the 1985 Mexico City earthquake. For this highly
harmonic, long duration earthquake record the input energy is significantly affected by the ductility level
(especially from p = 2 to p = 5 ) in the period range that is to the left side of the predominant excitation period.
The peak of the spectrum shape, which corresponds to the predominant period of the ground motion, tends
to shift slightly towards a smaller period value as the displacement ductility ratio is increased. Therefore as the
value of the displacement ductility ratio increases, the values of Vi in the period range that is immediately to
the left of the peak increase and the values in the period range that is to the right of the peak decrease.
Verijkation of Housners assumption
For a linear elastic system the maximum input energy that is stored in a SDOF system is
1

E --k(S,)
D-2

1
2

= -m(S,,)

where S , is the linear elastic spectral displacement and S,, is the linear elastic pseudo-velocity, both being a
function of period and damping ratio. It should be noted that ED is the maximum elastic energy that is stored
in the structure; the damping energy is not included. Housner assumed that ED (or S,, in the form of
equivalent velocity) can be used as the energy demand for an inelastic system in his proposed limit design

84

C.-M. UANG AND V. V. BERTERO

method. If S,, spectra with 5 per cent damping are compared with the Vi spectra with 5 per cent damping and
a ductility ratio of 5, it is seen from Figure 6 that S,, may significantly understimate Vi.
Approximate inelastic input energy spectra
Inelastic behaviour has the effect of (i) increasing the effective natural period, and (ii) increasing the effective
damping ratio of a structure. On the basis of a study of a class of hysteretic structures subjected to a total of 12
earthquake ground motions, Iwan* found that an inelastic response spectrum can be approximated by an
elastic spectrum corresponding to an equivalent viscous damping (C,) and an equivalent natural period ( Te):
{,=<+0.0587 ( p - 1)0'371

(204

Te ' 1 +0*121 ( p -

-=

where 5 is the nominal viscous damping ratio, T is the natural period in the elastic range and p is the ductility
ratio.
For a given ductility ratio, the elastic input energy equivalent velocity spectra, constructed by using an
equivalent damping ratio of [equation 20(a)] and then performing a period shift using equation (20b) are
compared with the inelastic spectra shown in Figure 5. Figure 7 shows such a comparison for p= 5. It can be

re

Period (sec)

Period (sec)

Period (sec)

Figure 6. Comparison of absolute input energy equivalent velocity spectra and linear elastic pseudo-velocity response spectra for
ductility ratio 5 (5 per cent damping for and Spv):-Vi- - - - SPV
"g(max)

vi

(nl/scc)

~l Centre

Vi(m/scc)

Mexico City

"i

(m/scc)
50

San Salvador

1
0.5

0.1

0.1

0.05

(1.01
0.01

0.1 0.5
5
['el-iod ( s e c )

!L

0.01

0.5
5
Period ( s e c )

0.1

50

0.01.
0.01

0.1 0.5
5
Period (sec)

50

Figure 7. Comparisonof absolute input energy and Iwan's elastic input energy equivalent velocity spectra for ductility ratio 5 (5 per cent
damping): - -~
- - - - approximate V, based on Iwan's procedure
~

SEISMIC ENERGY IN STRUCTURES

85

observed that, although inelastic input energy equivalent veIocity spectra appear to be predicted very well by
elastic spectra constructed using Iwans procedure, there are some significant differences. For example, for a
period of about 2 sec Iwans elastic Vi spectral value for the Mexico City earthquake is twice the inelastic Vi
spectral value; and therefore the elastic E, value will be four times the value of the inelastic Ei. It is believed that
this can be attributed to the particular highly harmonic nature of the ground shaking in the case of the Mexico
City earthquake and that this type of motion was not taken into account in Iwans derivation of equation (20).
In his study of the relationship of 5, and T, for both harmonic and typical earthquake excitations, Hadjian
has shown that the equivalent damping ratios due to harmonic excitation are about five times those due to
earthquake excitation, and the period changes due to harmonic excitation are about twice those due to
earthquake excitation. It is believed that equation (20)significantly underestimates the values of 5, and T, for
the 1985 Mexico City earthquake. An increase in the value of 5, will lower the magnitude of Iwans elastic
input energy spectra, making them more comparable to the actual inelastic input energy spectra. Deriving
appropriate values of 5, and T, for the 1985 Mexico City earthquake is outside the scope of this study.
Amplification factor of Vi and strong motion duration
It is well known that elastic spectral values like elastic pseudo-acceleration cannot reflect the effect of strong
motion duration. This shortcoming carries through to any inelastic design spectra derived from them. Since
input energy reflects the effect of the duration directly through integration, it is worthwhile to investigate the
relationship between the maximum equivalent velocity of input energy and the strong motion duration. Two
quantities-amplification factor and the strong motion duration used in this study-are described first.
The amplification factor ( Y )of an input energy equivalent velocity spectrum for a given ductility ratio (p)
and a viscous damping ratio ( 5 ) is defined by the following:

where V Y x @ , 5 ) is the maximum value of Vi evaluated throughout the whole period range. In general
V r ( p ,5 ) occurs in the immediate vicinity of the predominant period of the earthquake ground motion.
One commonly used definition of strong motion duration is that due to Trifunac and Brady:I6
(22)

D = t0,95-t0,05

where ro.05 and to.,, define the times at which 5 and 95 per cent, respectively, of the value of the Arias intensity
( I , ) is achieved. Arias intensity is defined as f01lows.~
I

1.

A-29

td

Ui(t)dt

where t , is the total duration of the earthquake record. The calculated values oft for each earthquake record
are listed in Table I. A plot of Y (p= 5 , 5 = 5 per cent) versus tD for the three earthquake ground motions
together with the 1971 Pacoima Dam record (S 16E) and the 1952 Taft record (N21E) is shown in Figure 8. It
is observed that Y and t, are linearly dependent; by letting the intercept of the line of best fit, shown in Figure
8, be 1.0, the following equation is obtained by the method of least-squares:
Y ( p= 5 , 5 = 5 per cent) = 1.0

+ 0.12 ,

(24)

Therefore, if the strong motion duration at a given site is known, it is possible to predict the maximum energy
input to a structure with a specified ductility ratio [ S for the case used in developing equation (24)]. The
period of the structure at which this maximum input energy occurs is close to the predominant excitation
period of the expected earthquakes at the site under consideration.
For example, if it is expected from previous earthquake records at a certain site that the maximum ground
velocity is 75 cmjsec (30 in/sec) and that the strong motion duration t, is 20 sec, the maximum input energy
per unit mass for a structure having a damping ratio of 5 per cent and a ductility ratio of 5 can be estimated by

86

C.-M. UANG AND V. V. BERTERO

'V

= 1.0

+ 0.12fD

10

15

20

25

30

35

40

ID (see)

Figure 8. Amplification factor of V,and strong motion duration relationship

the following procedure:

Y = 1.0
max

Vi

+ 0.12tD = 1.0 + 0.12(20) = 3.4

- Y dg(max)= 3.4(75) = 255 cm/sec

-

Era' 1
-=-(Vyx)2

= -(255)2 = 32,513cm2/sec2= 3.25m2/sec2

INPUT ENERGY TO MULTI-STOREY BUILDINGS

The 'absolute' energy equation for an N-storey building subjected to an earthquake ground motion has been
derived as follows.l 7

where m, c, and v are the diagonal mass matrix, viscous damping matrix and relative displacement vector,
respectively; mi is the lumped mass associated with the ith floor, ifti
is the absolute acceleration a t the ith floor.
The kinetic energy and input energy are calculated as follows:

I(1
N

Ei=

i= 1

mi&) dug

Nhere E , is the summation of the kinetic energy at each floor level, calculated using an absolute velocity (dti) at
the ith floor, and Eiis the summation of the work due to an inertia force (rniu,i) at each floor for ground
displacement.
Akiyama' has shown that the relative input energy Efbased on a SDOF system can provide a very good
estimate of the input energy for multi-storey buildings. Although no parametric study is attempted here to

SEISMIC ENERGY IN STRUCTURES

87

verify the same conclusion for the absolute input energy Ei, shaking table test results for a six-storey
concentrically braced steel structure will be used to support this conclusion.
Figure 9 shows the 0.3-scale test model during the shaking table test; the 1978 Miyagi-Ken-Oki (MO)
earthquake recorded in Japan was used as the input ground motion. The test structure, which weighs 476 kN
(107 kips), is classified by the UBC18 as a dual system with two exterior ductile moment-resisting frames and
one interior concentrically K-braced frame in the excitation direction. The magnitude of the earthquake
record was scaled to different levels to represent different limit states of the structure responses. Details of the
test results are reported in Reference 17. During the collapse level test (MO-65 Test, which had a measured
peak base horizontal acceleration of 0.65g), the model experienced severe brace buckling in the bottom five
storeys and the braces in the fifth storey even ruptured. Figure 10 shows the envelope of base shear versus
critical inter-storey drift index obtained from different limit state tests. As a result of brace buckling and
rupture, the envelope exhibits strength deterioration. Figure 1 1 shows the energy time histories of the MO-65
Test. Note that the viscous damped energy curve was calculated indirectly by the following expression:

E , = Ei - E , - E , - E ,

(27)

In order to compare the experimental input energy of this frame with an elastic-perfectly plastic SDOF
system, an estimate of the displacement ductility ratio for this frame from the test envelope in Figure 10 is
needed. By approximating this non-linear envelope by two straight segments, where the yield level was
calculated from simple plastic analysis, the corresponding ductility ratio is 2.6. The calculated input energy
spectrum of a SDOF system with a ductility ratio is 2.6. The calculated input energy spectrum of a SDOF
system with a ductility ratio of 2.6 and a viscous damping ratio of 2 per cent,which was the measured first
mode equivalent viscous damping ratio, is shown in Figure 12. The quantities presented in Figure 12 have
been scaled to the prototype level by similitude laws. The correlation between the experimentally measured Vi
for the multi-storey structure and the calculated Vi for a SDOF system is very good. It is concluded from this

Figure 9. Overall view of the O.3-scale model with reference frame

Base Shear Ratio

1.2

r 7 ,

0.6

lcst Envclopc

3
MO-65

0.4

0.2

MO-6.3
I

0.0
0.0

0.5

1.0

1.5

2.0

2.5

Critical Inter-storyDrift Index (%)

Figure 10. Envelope of critical inter-storey drift index versus base. shear ratio

Energy
(k-in)

(kN-111)

500
50
400

Input Energy

40

300
30

200
20

100

10

10

limc (sec)
Figure 1 1 . Model collapse level test (MO-65 Test) energy time histories

12

SEISMIC ENERGY IN STRUCTURES

89

0.62 (Measured Fundamental Period

of t.he Six-Story Steel Frame)

case study that the input energy spectra for a SDOF system can be used to predict the input energy demand
for this type of multi-storey building structure reliably.
CONCLUSIONS
From the results obtained in this study, the following observations can be made.
1. The use of an absolute energy equation rather than a relativeenergy equation has the advantage that
the physical energy input is reflected.
2. The profiles of the energy time histories calculated by the absolute energy equation differ significantly
from those calculated by the conventional relative energy equation (see Figure 3).
3. The absolute and the relative input energies for a constant displacement ductility are very close in the
period range of practical interest, namely 0 3 to 5.0 sec (see Figure 4). The difference between these two
input energies increases as the structure period differs more and more from the previous range. As the
period decreases, the absolute input energy approaches rnd~(,,,,/2, where
is the maximum ground
velocity, and the relative input energy approaches zero. The situation is reversed for long period
structures.
4. For certain types of earthquake ground motion, the absolute input energy spectra are sensitive to the
variation of ductility ratio.
5. Except for the highly harmonic earthquakes (1985 Mexcio City earthquake, for example), the absolute
input energy spectra for a constant ductility ratio can be predicted reliably by the elastic input energy
spectra using Iwans procedure which takes into consideration the effect of increasing damping ratio and
natural period.
6. The maximum energy input to a structure whose fundamental period is close to the predominant
excitation period of an expected earthquake can be predicted reliably with the expected maximum
ground velocity and the amplification factor Y [one such expression for ductility ratio 5 and damping
ratio 5 per cent is presented in equation (24)]. The amplification factor Y is approximately linearly
related to the strong motion duration t , defined in equation (22).

90

C.-M. UANG AND V. V. BERTERO

7. For steel dual systems of medium rise buildings it is possible to estimate with sufficient accuracy the
input energy for a multi-storey building structure from the absolute input energy spectra for a SDOF
system and the fundamental period of the multi-storey structure.
ACKNOWLEDGEMENTS

This research is partially sponsored by the National Science Foundation Grant No. CES-8810563 to the first
author and Grant Nos. CES-8804305 and ECE-8610870 to the second author. Any opinions, discussions,
findings, conclusions and recommendations are those of the authors and do not necessarily reflect the views of
the sponsor.
REFERENCES

I . H. Akiyama, Earthquake Resistant Limit-State Design for Buildings, University of Tokyo Press, 1985.
2. J. C. Anderson and V. V. Bertero, Seismic behavior of multistorey frames by different philosophies, Report No. UCB/EERC-69/11,
Earthquake Engineering Research Center, University of California, Berkeley, CA, 1969.
3. A. Arias, A measure of earthquake intensity, in Seismic Designfor Nuclear Power Plants (Ed. R. J. Hansen), Massachusetts Institute
of Technology Press, Cambridge, MA, 1970, pp. 438469.
4. G. V. Berg and S. S. Thomaides, Energy consumption by structures in strong-motion earthquakes, Proc. 2nd world con5 earthquake
eng. Tokyo, Japan 681496, (1960).
5. S. C. Goel and G. V. Berg, Inelastic earthquake response of tall steel frames, J. struct. diu. ASCE 94, 1907-1934 (1968).
6. A. H. Hadjian, A re-evaluation of equivalent linear models for simple yielding systems, Earthquake eng. struct. dyn. 10, 759-767
(1982).
7. G. W. Housner, Limit design of structures to resist earthquake, Proc. 1st world con5 earthquake Eng. Berkeley, California, 5-1 to 5-13
(1956).
8. W. D. Iwan, Estimating inelastic response spectra from elastic spectra, Earthquake eng. struct. dyn. 8, 375-388 (1980).
9. P. C. Jennings, Earthquake response of a yielding structure, J . eng. mech. din ASCE 90, 41-68 (1965).
10. B. Kato and H. Akiyama, Seismic design of steel buildings, J . struct. diu. ASCE, 108, 1709-1721, (1982).
11. S. A. Mahin and J. Lin, Construction of inelastic response spectrum for single degree of freedom system, Report N o . UCBIEERC83/17, Earthquake Engineering Research Center, University of California, Berkeley, CA, 1983.
12. W. E. McKevitt, D. L. Anderson, N. D. Nathan and S. Cherry, Towards a simple energy method for seismic design ofstructure, Proc.
2nd U.S. nat. conf. earthquake eng. 383-392 (1979).
13. N. M. Newmark and W. J. Hall, Procedures and criteria for earthquake resistant design, Building Science Series N o . 46, pp. 209-236,
Building Practices for Disaster Mitigation, National Bureau of Standards, 1973.
14. D. A. Pecknold and R. Riddle, Effect of initial base motion on response spectra, J . eng. mech. diu. ASCE 104,485-491 (1978).
15. J. M. Tembulkar and J. M. Nau, Inelastic modeling and seismic energy dissipation, J. struct. diu. ASCE, 113, 1373-1377 (1987).
16. M. D. Trifunac and A. G. Brady, A study on the duration of strong earthquake ground motion, Bull. seism. SOC. Am. 65,581-626
(1975).
17. C. -M. Uang and V. V. Bertero, Earthquake simulation tests and associated studies of a 0.3-scale model of a 6-story concentrically
braced steel structure, Report N o . UCB/EERC-86/10,Earthquake Engineering Research Center, University of California, Berkeley,
CA 1986.
18. Uniform Building Code, International Conference of Building Officials, Whittier, California, 1985.
19. A. S. Veletsos, N. M. Newmark and C. V. Chelapati, Deformation spectra for elastic and elastoplastic systems subjected to ground
shock and earthquake motions, Proc. 3rd world conj earthquake eng. Wellington, New Zealand 11-663 to 11-678 (1965).
20. T. F. Zahrah and W. J. Hall, Seismic energy absorption in simple structures, Structural Research Series No. 501, University of
Illinois, Urbana, Illinois, 1982.