Vol. 2, No.

1 EARTHQUAKE ENGINEERINGAND ENGINEERING VIBRATION June, 2003

Article ID: 1671-3664 (2003) 01 - 0099 - 09

Lateral load pattern in pushover analysis

Sun Jingjiang (~:fJ~..~) lt, Tetsuro Ono 2,, Zhao Yangang ( i ~ J l ] ) 2. and Wang Wei( ~j~)l w
1. Institute of Engineering Mechanics, China Seismological Bureau, Harbin 150080, China
2. Department. of Architecture, Nagoya Institute of Technology, Nagoya 466-8555, Japan

Abstract: The seismic capacity curves of three types of buildings including frame, frame-shear wall and shear wall ob-
tained by pushover analysis under different lateral load patterns are compared with those from nonlinear time history analy-
sis. Based on the numerical results obtained a two-phase load pattern: an inverted triangle(first mode)load pattern until the
base shear force reaches/3 times its maximum value, Vm~, followed by a (x/H)" form, here/3 and ct being some coeffi-
cients depending on the type of the structures considered, is proposed in the paper, which can provide excellent approxima-
tion of the seismic capacity curve for low-to-mid-riseshear type buildings. Furthermore, it is shown both the two-phase load
pattern proposed and the invariant uniform pattern can be used for low-to-mid-rise shear-bending type and low-rise bending
type of buildings. No suitable load patterns have been found for high-rise buildings.

Keywords: pushover analysis; performance-basedseismic design; lateral load pattern; nonlinear time history analysis

1 Introduction of regional ground motion, in all codes. It is difficult

to provide suitable earthquake time histories as earth-
In the last ten years, much attention has been
quake motion for general design use in code. Push-
paid to performance-based seismic design in earth-
over analysis is not as complicated as nonlinear time
quake engineering research. This new method re-
history analysis, and can use response spectrum as
quires designing a building for several expected per-
demand diagram to estimate the seismic response of
formance levels associated with different earthquake
structures (Chopra and Goel, 1999). Therefore, it is
hazard levels. To meet this objective, a more rational
generally recommended in performance-based design.
design procedure based on inelastic displacement
In pushover analysis, the first step is to suppose a
rather than elastic force is needed. At present, the
certain lateral load pattern, then perform a static
method has been suggested in some recommended or
analysis of the structural model under this pattern.
guidance codes and documents (SEAOC Vision
The load pattern is applied step by step until a pre-
2000, 1995; ATC-34, 1994; ATC-40, 1996; FEMA
determined target displacement is reached. Thus, the
273 and FEMA 274, 1996; Eurocode 8; Building
relationship between base shear and roof displacement
standard law of Japan etc. ).
is obtained, which is referred to as the capacity curve
An important step in performance-based design is
of a building. It is clear that different load patterns
to estimate the nonlinear seismic response of build-
will result in different capacity curves. If the curve
ings. There are two procedures: nonlinear time histo- over-or-underestimates the seismic capacity of the
ry analysis and simplified nonlinear analysis, herein
building, then the steps used to estimate the displace-
referred to as pushover analysis. The nonlinear time ment response based on this curve and design demand
history analysis can provide more realistic results for a diagram would not be realistic. Therefore, the selec-
given earthquake ground motion. However, such ana-
tion of a reasonable lateral load pattern is particularly
lytical results tend to be highly sensitive to the earth- important in pushover analysis.
quake input. This makes it inappropriate as a code re- Several lateral load patterns have been suggested.
quirement since earthquake motion is expressed by re- They are : ( 1 ) inverted triangle distribution ( modal
sponse spectrum, which reflects an average characteristic
pattern) ; ( 2 ) uniform distribution ; ( 3 ) load distri-
bution based on linear elastic dynamic analysis or re-
Correspondence to: Sun Jingjiang, Institute of Engineering
Mechanics, CSB, Harbin 150080, China sponse spectrum analysis of the building (Tso and
E - mail : jingjiangsun@sina. com Moghadam, 1997 ) ; ( 4 ) the adaptive distribution,
t Professor; *Associate Professor; ~Assistant Professor which is varied as the inter story resistance changes in
Received date: 2002.10-05; Accepted date: 2003-05-02 each load step (Bracci et al. , 1997) ; (5) distribution
 100 EARTHQUAKE ENGINEERINGAND ENGINEERINGVIBRATION Vol. 2

proportional to the product of the mass and fundamen- roof displacement related to the earthquake can be es-
tal mode shape, which is used initially until the first tablished by connecting all the points together. This
yielding takes place, then the lateral forces are deter- relationship could be thought as the "real seismic ca-
mined based on the product of the current floor dis- pacity curve" of the building, but is only valid for the
placement and mass at each step (Fajfar and Fisch- earthquake considered. Therefore, we will take the
inger, 1988); ( 6 ) a distribution based on mode statistical average seismic capacity behavior of a
shapes derived from secant stiffness at each load step building under a number of typical earthquake excita-
(Eberhard and Sozen, 1993 ). The last three distri- tions as its "real capacity curve". Then, pushover
butions are adaptive patterns, which try to establish analyses are performed under different lateral load
equivalent lateral load distribution based on a certain patterns and the resulting curves are compared with
theoretical basis. However, their superiority over the those from time history analysis for each building.
simple fixed load patterns has not been demonstrated. All the analyses are performed using the inelastic
It was also noted that the first two patterns might analysis program IDARC-2D( Reinhorn et al. , 1996 ) ,
result in the lower and upper bound of pushover in which the proposed lateral load pattern was added.
curves, respectively (Tsopelas et a l . , 1997 ).
In this paper, many time history analyses are car- 3 Selected earthquake records and sample
ried out for eight simplified reinforced concrete (RC) buildings
buildings, which were selected to represent a variety Four earthquake records are selected, the Awa-
of structures, to obtain the capacity curves of these jishima (Jan. 17, 1995, firm soil site), E1 Centro
buildings under earthquake excitations. Then, push- (May 18, 1940, medium site), Taft (July 21,1952,
over analyses are conducted under different load pat- medium site) and Tianjin (Nov. 15, 1976, soft soil
terns including the one proposed in the paper, the ob- site). Their elastic acceleration response spectra with
tained capacity curves are compared with those from 5% damping are shown in Fig. 1. Eight RC buildings
time history analysis, the effectiveness of different are selected to cover three widely used seismically re-
load patterns is examined and suitable load patterns sistant systems, i. e. , frame, frame-shear wall and
are suggested for different types of structures. Final- shear wall, and a range of building periods. These
ly, this comparison is further made for four more three structural systems may also represent shear
practical RC frame buildings designed according to type, shear bending type and bending type of build-
seismic code under another group of earthquake exci- ings, respectively.
tations. Each of the sample buildings was designed as one
unit of typical building referred to the geometry and
2 Obtaining seismic capacity curves of buil- reinforcement of its members and so on.
dings The story heights of all buildings are: 4m for the
The seismic capacity curve of a building depends first story and 3m for the other stories.
not only on the properties of the system but also on The eight buildings are numbered from B1 to B8,
the loading process. Under different earthquake exci- their main characteristics are listed in Table 1, and
tations ( implying different loading processes) a build- the plan view of B4 to B6 is shown in Fig. 2. The first
ing will exhibit different behavior. Because a future modes of the eight example buildings are shown in
earthquake ground motion can not be predicted, it is Fig. 3.
impossible to obtain a definite seismic capacity curve
for a building, and such a curve can be achieved only
4 Analytical results
in the sense of statistic average just like the seismic The restoring force model of all the structural
response spectrum curves used in codes. members follows a tri-linear nonlinear relationship
A seismic capacity curve of a building can be ob- which is obtained from the analytical program based
tained by repeatedly performing nonlinear time history on member size, reinforcement, property of the mate-
analyses for a given earthquake excitation with differ- rials, axial force and so on.
ent intensities. A maximum base shear and a maxi- The pushover analysis is performed under the load
mum roof displacement can be obtained after each patterns of inverted triangle ( first mode ) , uniform,
time history analysis, which form one point on the ca- adaptive ( Reinhorn, 1996) and a new pattern ( pro-
pacity diagram. The relationship of base shear versus posed in this paper) , respectively.
No. 1 Sun jingjiang et al. :Lateral load pattern in pushover analysis 101

6-
.<

, 9 t , . , , , , 9 , , , 9 , , . . . . 0 , , i i ~ , 9 *

0 0.5 1 1.5 2 2.5 0 0.5 1 1.5 2 2.5

Period (s) Period (s)
(a) E1 Certro (b) Taft

1
'4

6. "4

<

3
i i i I I l i I

0 0.5 1 1.5 2 2.5 0 0.5 1.5 2.5

Period (s) Period (s)
(c) Tianjin (d) Awajishima

Fig. 1 Normalized response spectra of four earthquakes (5 % damping)

Table 1 Main characteristics of eight buildings

Number of span Total Total Funda-

No. Type Story height weight mental
Transversal Longitudinal (m) (kN) period (s)

B1 Frame 4 4 0 13 1975 0.40

B2 Frame 6 4 0 19 4025 0.58

B3 Frame 8 4 0 25 6750 0.68

B4 Frame-shear wall 9 5 2 28 33490 0.66

B5 Frame-shear wall 20 3 2 61 54945 1.55

B6 Shear wall 6 1 0 19 1320 0.36

B7 Shear wall 10 1 0 31 1600 0.68

B8 Shear wall 16 1 0 61 2400 1.10

Fig. 2 Plan view of B4 ( a ) , B5 (b) and B6 (c)

 102 EARTHQUAKEENGINEERINGAND ENGINEERINGVIBRATION Vol. 2

/ / / /
/
Fig. 3 First modes of eight buildings

The proposed pattern is a variant form transformed the seismic capacity of the buildings in the inelastic
from an inverted triangle into an ( x / H ) " pattern, in stage. The curves predicted by the inverted triangle
which x is the distance from ground to the floor, H is pattern lie near the upper bound for building B1 and
the height of the building and ot is a parameter that about in the middle for buildings B2 and B3. The
changes value for different types of buildings. The curves predicted by the proposed load pattern, name-
pushover analysis needs to be performed twice to use ly, beginning with inverted triangle load pattern till
this load pattern. First, the maximum base shear Vm~, 0.5Vma and then changed into ( x / H ) 4 load pattern
x

(which is different under the other load patterns; are a little lower than the average values of the time
here we just look for a reference value to determine history analysis. The proposed and inverted triangle
when the load pattern changes) is obtained by push- patterns are considered as the better load patterns for
ing the structure under the inverted triangle pattern these types of buildings.
from beginning to end until the drift angle (roof dis-
4.2 Frame-shear wall buildings ( shear-bending
placement/ building height) reaches the predefined
type )
value (which may be taken as 1% to 2 % ) . Next,
the inverted triangle load" pattern is used until the As shown in Fig. 5 (a) and (b) for buildings B4
base shear reaches some percentage of Vmax, then the and B5, respectively, the results predicted by the
load distribution is changed into the (x/H)" pattern. uniform load pattern no longer provide the upper
The analytical results are shown in the figures that bound of the capacity curves and overestimate the
follow. In each figure, seven capacity curves are seismic capacity of building B4 around its yield stage.
presented. Four curves are obtained by time history The inverted triangle load pattern significantly under-
analysis and the other three by pushover analysis un- estimates the capacity of building B5 and the situation
der the inverted triangle, uniform and beginning with is the same for B4 by comparing the results with those
inverted triangle then turning into ( x / H ) " load pat- under Taft, Tianjin and Awajishima earthquake exci-
tern, respectively. The capacity curves predicted by tations. The proposed load pattern is: an inverted tri-
the adaptive load pattern are not provided here, be- angle pattern until the base shear force reaches 0.8
Vmax, followed by a (x/H)-05 form for building B4
cause they show sudden up or down curves in the ine-
lastic stage, which makes it difficult to compare them and (x/H) -l.o for building B5.
with the other curves. Comparatively, the proposed and uniform load
patterns are acceptable for building B4. However, for
4.1 Frame buildings ( shear type) the 20-story building B5, all load patterns, including
The results for buildings B1, B2 and B3 are the proposed one, are unsatisfactory. Many of the
shown in Fig. 4 ( a ) , (b) and ( c ) , respectively. By other load patterns have been tried for this building
comparing the pushover analytical results with time and no suitable lateral load pattern can be found to
history analytical results it is seen that the curves pre- monotonously load the building to obtain a capacity
dicted by the uniform load pattern provide an upper curve with a better approximation to those of "real ca-
bound of the capacity curves and obviously overestimate pacity curves" in the post-yield stage.
No. 1 Sun jingjiang et al. :Lateral load pattern in pushover analysis 103

11000
1400 '' ' I ' ' ' I '' ' I ' ' ' I ' '' I ' ' ' I ' ''

10000

1200 9000

~ 800o
~. 1000
~ 7000
, Taft ~ 6000
800 I ~ Tianjin
9 Awajishima 5000 " 1 ..... Inverted tri. "
' ..... Inverted tri. .... Uniform
600 .... Uniform 4000
~ Present
3000
Ii i I i , , I i . * I l i i I | i | I * t . I I I I I 100 200 300 400 500 600
400
40 80 120 160 200 240 280 Top displacement (ram)
(a)
Top displacement (mm)
(a)
750
1800 ' I ' I ' I ' I I I ' I ' I '
=

1600 i~ elli- i~ Ie'iJeeee~'ee 600

1400
~" 450 -.--.-.-..-..-..
~ 1200
9 E1 eentro 300
lceotro E & Taft
i 10000

80O
!
=
it
~-
9
.....
Taft
Tianjin
Awajishima
Inverted tri.
t~

150 t nWLm: 9
.. :..
....
Tianjin

Uniform
600 .I . I .... Uniform i i Present
. ~ I Present 0
0 100 200 300 400 500 600
400 i
Top displacement (mm)
0 50 100 150 200 250 300 350 400
(b)
Top displacement (mm)
(b)
Fig. 5 Capacity curves of frame-shear wall buildings for
B4(a) and B5(b)
2600

2400 4.3 Shear wall buildings (bending type)

2200 The results for buildings B6 to B8 are shown in
~" 2000 Fig. 6 (a) to ( c ) , respectively. It can be seen that
the inverted triangle load pattern significantly under-
1800 "~
9 Taft estimates the seismic capacities of the buildings in the
1600[" I 9 Tianjin inelastic stage. Therefore, it is not suitable to use it
~- ~ 9 Awajishima
for shear wall (bending type) buildings in pushover
1400 ~ ~ ..... Inverted tri.
analysis, whether they are high-rise, mid-rise or low-
1200 ~" ~ .... Uniform
rise. The proposed load pattern similar to that for
1000 ~ frame-shear wall buildings is: an inverted triangle
0 100 200 300 400 500 600 pattern until the base shear force reaches 0.8 Vmax for
Top displacement (mm) building B6, and 0.7 Vmaxfor buildings B7 and BS,
followed by a (x/H)-l~ for all of them.
(e)
To this type of structures, both the proposed pat-
tern and uniform pattern are acceptable for low-rise
Fig. 4 Capacity curves of frame buildings for B1 ( a ) , B 2
(b) and B3(c) buildings. As for mid-and high-rise buildings, the
modeling results are not satisfactory, as was the case
 104 EARTHQUAKE ENGINEERINGAND ENGINEERINGVIBRATION Vol. 2

for building B5. For these buildings, no suitable lat- 4.4 High-rise buildings
eral load pattern can be found.
High-rise buildings exhibit very complex behavior
under different earthquake excitations ( Fig. 5 ( b )
800 and Fig. 6 ( c ) ) , and the real seismic capacity
"

curves" show larger variations. This makes it difficult

700
to define a rational average seismic capacity curve in
600 pushover analysis.
From the results obtained, it is shown that the
~-~ 500 i i I . ~ i Q i I I I I i I I I " I displacement response errors in pushover analysis for
high-rise building occur from ignoring the influence of
~ 4oO A Taft high modes and also from the different seismic capaci-
~- Tianjin
300 ties predicted by time history and pushover analyses.
= Awajishima
. . . . . Invertedtri. The seismic capacity levels of high-rise buildings
200 . . . . Uniform significantly depend on the soil site condition. On the
-- -- -- Present harder soils, the seismic capacity of the building ap-
100 i , I , I i , i , i i i

0 50 100 150 200 250 300 350 400 pears to be higher. This indicates that an average
Top displacement(ram) seismic capacity curve for a high-rise building could
(a) be established depending on the classification of soil
11000 "', j' t'' V ...... site, and thus avoid large errors.
The second problem is that the slope and shape of
the capacity curves of high-rise buildings obtained by
9000
i
pushover analysis are much different from those ob-
tained by time history analysis in the strong nonlinear
7000 stage. It can be concluded that none of all the load
patterns can provide satisfactory capacity curves in
5000 pushover analysis for this type of building.

3000
5 Further verification of lateral load pat-
tern for frame buildings
1000 From the above results, the pushover analysis with
0 200 400 600 800 1000 inverted triangle or proposed load patterns provide a
Top displacement(mm)
better prediction of the seismic capacity of frame
(b)
buildings. Therefore, the lateral load patterns are ex-
11000 I ' I ' I = l ' I ' I ' amined further for more practical buildings and addi-
tional earthquake records.
Eight earthquake records are selected and their
9000
characteristics are listsed in Table 2.
Four RC frame buildings designed according to
z 7000 seismic code are selected and numbered as FR1-FR4.
...D ..iD mp ~
The number of stories and plan sizes are listed in Ta-
"7 5000 ble 3. The cross sections of the beams are 650ram x
Jl. "~"-'~ ...................
t~
250mm for edge spans and 450ram x 250ram for the
~176 1 centro . . . . . Invertedtr middle span. The column sizes are 550ram x 550ram
3000
V & T a f t .. ~ for FR1, 700mm x 700ram for FR2, 500mm •
f -- T'mnjin -- m --Present
500ram for first three stories and 450ram x 450ram for
9 I Awaiishima I
1000 | i , l , , I I
the other stories of FR3, and 600ram x 600ram for
100 200 300 400 500 600 700 800
Top displacement(ram) FR4. The plan view is shown in Fig. 7 and elevation
(e) configurations and vertical node loads are shown in
Fig. 8.
Fig. 6 Capacity curves of shear wall buildings for B6 Time history analysis is performed until the PGA
(a), B7(b) and B8(c) becomes 0. 8g or the drift angle reaches 2 % . The
pushover analyses are conducted under load patterns
of inverted triangle, uniform, adaptive, generalized
No. 1 Sun jingjiang et al. :Lateral load pattern in pushover analysis 105

power (Reinhoru et al. , 1996) and the one proposed curves are presented, among which five curves are
in this paper, respectively. The results are shown in from pushover analysis and the other eight are from
Fig. 9 ( a ) to ( d ) . In each figure, thirteen capacity time history analysis.

Table 2 Characteristics of eight earthquake records

Predominant
Name of records Soil site period (s) PGA (g)

Qi' an, Aug. 20, 1976 Firm 0.11 0. 150

Parkfield, Dee. 20, 1994 Firm 0.09 0. 289
E1 Centro, May 18, 1940 Medium O. 55 O. 342
Northridge, Jan. 17, 1994 Medium 0.28 0.073
Shandon, June 27, 1996 Medium 0.62 0. 480
Taft, July 21,1952 Medium 0.36 0. 153
Tianjin (SN), Nov. 15, 1976 Soft 0.95 0. 146
Tianjin (EW) , Nov. 15, 1976 Soft 1.04 0. 104

The results exhibit the same feature as those de-

scribed in the last section for frame buildings, name-
ly, the capacity curves predicted under the inverted
load pattern lie near the middle of those obtained from
time history analysis. The curves obtained under the
proposed load pattern is a little lower than the average
values. Therefore, these two can be considered better
load patterns for frame buildings than the others.

6 Concluding remarks

The seismic capacity curves of three types of

buildings including frame, frame-shear wall and shear
wall obtained by pushover analysis under different lat-
eral load patterns are compared with those from non-
linear time history analysis performed. From the nu-
merical comparisons following findings can be
reached :
( 1 ) Based on the numerical results obtained a
two-phase load pattern: an inverted triangle (first
mode) load pattern until the base shear force reaches
/3 times its maximum value, V,.... followed by a
Fig. 7 Plan view ( x / H )" form, here/3 and o~ being some coefficients
depending on the type of the structure considered, is
Table 3 Number of stories and plan sizes of proposed in the paper.
four buildings ( 2 ) Both the invariant inverted triangle pattern
and the proposed two-phase pattern with/3 = 0.5 and
No. Story Sizes of plan (mm) a = 4 . 0 can provide an excellent approximation of
number a* 5" the seismic capacity curve of low-to-mid-rise shear
type buildings. The curves obtained under these two
FR1 4 5700 2700
load patterns lie near the middle or lower than the
FR2 5 5700 curves obtained from time history analysis, respectively.
( 3 ) Both the invariant uniform pattern and the
FR3 6 5700 2100
proposed two-phase pattern with /3 = 0 . 7 and 0 . 8
FR4 7 6000 6000 and a = - 0 . 5 and - 1.0 can be used respectively
9Note : the meaning of a and b can be seen in Fig. 7 for low-to-mid-rise shear-bending and low-rise
 106 EARTHQUAKE ENGINEERING AND ENGINEERING VIBRATION Vol. 2

bending buildings. The invariant inverted triangle

load pattern significantly underestimates the seismic
capacities of these two types of buildings in the inelas- 1~I
900
800
tic stage and therefore, it should be avoided for these
two types of buildings in pushover analysis.
50o
10 16 16 400 [ 9 @ i e history analysis

'~ l 9 ~ Uniform ~ Power

200 ~ " Presem

Im
100
18 18 12
0 ~
0 50 100 150 200 250 300 350 400 450
!~2 18 t8 Top displacement (mm)
(a) FRI

1300
[, 5,7m L2.7m l, 5.7m 1200

(a) FR1 1000

900
800
10 10 10
700
600
12 12 ~21 H 5OO
400
'1
• r~ 30O
12 12 12 200
E
100
14 0
50 100 150 200 250 300 350 400 450 500
Top displacement (ram)
(b) FR2
5.7m L 5.7m L . . . . . . , . . ., . . . . , . . . . , . . . .
700
(b) FR2
600
13 18 18 13 i 500
v
20 28 400
28 20
A 300
m
e~
20 28
200
28 20 I
21 28 28 21 100

0
50 100 150 200 250 300 350
Top displacement (ram)
L 5.7m [2.1ml, 5.7m L
rl ~ si (c) FR3
(e) FR3
1000 . . . . . , . . . . . . . . . . . . . . ..

36 36 36 36 900
800
33 qq q'~ q~ 700
A
57 600
33 331 33 33 500
r 400
24 24 24 24 300
/ 200
r
100
0
,t 6.0 [ 6.0 ; 6.0 50 100 150 200 250 300 350 400
Top displacement (ram)
(d) FR4
(d) FR4

Fig. 8 Elevation and load value ( k N ) shown near the Fig. 9 Capacity curves of four RC frame buildings
nodes
No. 1 Sun jingjiang et al. :Lateral load pattern in pushoveranalysis 107

( 4 ) Displacement response errors in pushover Deformation of Inelastic Structures: SDF systems,"

analysis for high-rise buildings occur not only from ig- Report PEER-1999/02, Pacific Engineering Research
noring the influence of higher modes but also from the Center, College of Engineering, University of Califor-
difference of the seismic capacity curve predicted by nia, Berkeley.
time history analysis and pushover analysis.
Tso WK and Moghadam AS ( 1997 ), "Seismic Re-
(5) The seismic capacity curve of high-rise build-
sponse of Asymmetric Buildings Using Pushover Anal-
ings significantly depends on the soil site conditions.
ysis," Proceedings Workshop on Seismic Design Meth-
The harder the soil, the higher the seismic capacity of
odologies for the Next Generation of Codes, Bled, Slo-
the building. In view of this, further investigation of
venia, Rotterdam, Balkema.
seismic capacity curves for high-rise buildings with
pushover analysis needs to take the classification of Bracci JM, Kunnath SK and Reinhorn AM ( 1997 ),
the soil site into account. "Seismic performance and Retrofit Evaluation of Re-
inforced Concrete Structures," Journal of Structural
References Engineering, ASCE, 123(1): 3-10.
SEAOC (1995), PeoCormanceBased Seismic Engineer- Fajfar P and Fischinger M (1988), "N2-A Method
ing of Buildings, Version 2000, Sacramento, Cali- for Nonlinear Seismic Analysis of Regular Building,"
fornia: Structural Engineers Association of California. Proceedings 9th World Conference on Earthquake En-
gineering, Tokyo, Vol. 5, pp. 111-116.
Applied Technology Council ( 1994 ) , "A critical re-
view of Current Approaches to Earthquake-resistant Eberhard MO and Sozen MA ( 1 9 9 3 ) , " Behavior-
Design , "ATC-34 Rewood City, California : Applied Based Method to Determine Design Shear in Earth-
Technology Council. quake- resistant Walls," Journal of the Structural Di-
vision, New York, New York: American Society of
Applied Technology Council (1996), "Seismic Eval- Civil Engineers, 119(2) : 619-640
uation and Retrofit of Concrete Buildings," Report
ATC-40, Rewood City, California: Applied Technolo- Tsopelas P, Constantinou MC, Kircher CA and Whit-
gy Council. taker AS ( 1997 ), "Evaluation of Simplified Methods
of Analysis for Yielding Structures ," Technical Report
Federal Emergency Management Agency ( 1996 ), NCEER-97-O012, National Center for Earthquake En-
"1994 NEHRP Guidelines for the Seismic Rehabilita- gineering Research, Buffalo, N.Y.
tion of Buildings," Reports FEMA 273 (Guidelines)
Reinhorn AM, Kunnath SK and Valles-Mattox R
and 274 (Commentary), Ballot version, Washing-
(1996), "IDARC 2D Version 4 . 0 : A Program for
ton, D.C.
the Inelastic Damage Analysis of Buildings," Techni-
Chopra AK and Goel RK ( 1 9 9 9 ) , " Capacity-De- cal Report NCEER-96-O010, National Center for
mand-Diagram Methods for Estimating Seismic Earthquake Engineering Research, Buffalo N. Y.