Engineering Structures 66 (2014) 1023

Contents lists available at ScienceDirect

Engineering Structures
journal homepage: www.elsevier.com/locate/engstruct

Cyclic Pushover Analysis procedure to estimate seismic demands for

buildings
Phaiboon Panyakapo
Department of Civil Engineering, Sripatum University, Bangkok 10900, Thailand

a r t i c l e

i n f o

Article history:
Received 7 September 2012
Revised 31 January 2014
Accepted 3 February 2014
Available online 3 March 2014
Keywords:
Cyclic pushover
Seismic displacement demand
Loading protocol
Cumulative damage
Modal pushover

a b s t r a c t
Conventional Pushover Analysis relies on the use of monotonic lateral load distribution. The seismic displacement demands based on this procedure are considered an approximate solution that has not taken
into account the cyclic loading effects. Under earthquake loading, structural components experience stiffness degradation and strength deterioration, which are the important characteristics of reinforced concrete members under cyclic loading, causing a reduction of deformation capacity. The Cyclic Pushover
Procedure is proposed to estimate seismic demands of buildings that take into account the cumulative
damage under cyclic loading. The cyclic lateral force distribution is developed based on the mode shapes
and the prescribed displacement history. The cyclic pushover curve is converted to the equivalent SDOF
pseudo-acceleration and displacement relationship based on the rst mode response of the structure. The
seismic demands of a 9-story reinforced concrete building are evaluated by Cyclic Pushover Procedure.
Four types of loading protocol, i.e., Laboratory, ATC-24, International Organization for Standardization
(ISO), and Sequential Phased Displacement (SPD) protocols are employed to investigate the effects of displacement histories on seismic demands. The seismic demands include the peak roof displacement, the
peak oor displacement and the peak inter-story drift ratio. The results are compared with the exact
demands resulting from nonlinear time history analyses of MDOF structure subjected to 20 ground
motions, as well as the demands estimated from the Modal Pushover Analysis. The results demonstrate
that the Cyclic Pushover Analysis provides a reasonable and accurate estimate of seismic displacement
demands.
2014 Elsevier Ltd. All rights reserved.

1. Introduction
To evaluate the seismic displacement demands of a structure
under earthquake loading, a nonlinear time history analysis provides the solutions accepted as the exact demands. However, the
accuracy of each solution depends on the appropriate selection of
ground motions as well as the modeling of structural behavior.
This procedure requires computational effort. In practice, nonlinear static analysis based on the Pushover Analysis method has been
widely employed to evaluate the seismic performance of structures. During the past decade, the Pushover Analysis procedures
have been improved to estimate more accurate displacement demands. The adaptive pushover method [13] was proposed to consider the stiffness of the structure at each step of lateral
displacement. A set of lateral displacement was monotonically applied to the structure, which is displacement based rather than
force based. However, the story force could be a reversal sign
Tel.: +66 25791111x2171; fax: +66 25791111x2147.
E-mail address: phaiboon.pa@spu.ac.th
http://dx.doi.org/10.1016/j.engstruct.2014.02.001
0141-0296/ 2014 Elsevier Ltd. All rights reserved.

resulting from modal forces in the higher modes. As an alternative,

Shakeri et al. [4] introduced a force based approach in terms of a
story shear-based adaptive pushover procedure to take into account the reversal sign in the higher modes. For the multi-mode
pushover method, the Modal Pushover Analysis has been proposed
by Chopra and Goel [57] to allow for the inuence of higher
modes. This method uses invariant modal lateral force distribution
to push the structure for each mode, and the results in each mode
are combined with SRSS by assuming linear elastic behavior. The
method is widely used to estimate seismic demands for tall buildings; however, some limitations have been reported [7] regarding
the reversal in the pushover curve under higher mode lateral force
distribution, and the location of plastic hinges may not be accurately predicted. Moreover, an over-estimate of the peak demands
in the modal combination procedure has been pointed out [8]. An
alternative method, which employed the Mass Proportional Pushover procedure, was proposed by Kim and Kurama [8] to avoid
the errors due to modal combination. In this procedure, the effects
of higher modes are lumped into a single invariant lateral force distribution that is proportional to the total seismic masses. The

P. Panyakapo / Engineering Structures 66 (2014) 1023

resulting peak displacement demands were more accurate than

those of the Modal Pushover Analysis. However, the inter-story
drifts were poor estimates due to the inability of a single mode
shape to capture the changes in the relative oor displacements.
An alternative method to take into account the higher mode effects
was proposed by Poursha et al. [9,10]. This method utilized multistage and single-stage pushover analyses in terms of a consecutive
modal pushover procedure. The nal structural response was
determined by enveloping the results of multi-stage and singlestage pushover analyses. Later on, a modied consecutive modal
pushover procedure has been proposed to improve the accuracy
of the results [11]. In this modied method, the effective modal
participating ratio was employed for the contribution of mode.
Innovative pushover methods have been developed to overcome
the higher modes pushover analysis by energy concept [1215].
However, the procedure needs to construct the energy demand
and capacity diagrams, and hence, the displacement demands
may not be computed directly.
In design practice, pushover analysis has been used in seismic
design regulations, for example, the N2 method [16] has been
implemented in European regulation [17]. This method relies on
a pushover analysis using invariant load patterns to estimate
deformation demands under seismic loads. An extension of the
N2 method has been improved to take the higher mode effects
into account for medium and high-rise buildings [18]. The capacity spectrum and the coefcient methods were introduced in US
provisions [1921]. The capacity spectrum is based on an equivalent linearization, in which the target displacement is the intersection of the SDOF capacity curve and an over-damped elastic
spectrum. The coefcient method uses empirical displacement
coefcients determined from statistical analysis to dene displacement modication factors [22,23]. The application of pushover analysis on buildings with irregular plans has also been
introduced to overcome the torsional problem in buildings with
asymmetric plan. For example, the N2 method has been extended
to the pushover analysis by applying to each direction of a 3D
model [2427]. The results are then amplied by torsional correction factors to take into account the torsional effects. The Modal
Pushover Analysis has also been extended to the case of plans
for asymmetric buildings [6,10,28,29]. The method was based
on multi-mode pushover analysis, where the load vectors are proportional to each 3D elastic mode of vibration. The load vectors
are composed of modal forces in two orthogonal directions and
torsion.
The above mentioned methods typically employ monotonic lateral load in the pushover analysis. This is based on an assumption
that the behavior of structural members under earthquake loading
in the hysteretic model may be represented by a backbone curve or
an envelope curve of cyclic hysteretic behavior. However, when the
reinforced concrete members are subjected to cyclic loading,
cumulative damage occurs under several repeated loads resulting
in stiffness degradation and strength deterioration. Experimental
and numerical investigations for the effect of loading history on
the response of reinforced concrete frames show that the monotonic loading provides greater strength than the cyclic loading
[30]. In addition, the effects of reinforcement slippage of reinforced
concrete frame structure under cyclic loading cause a decrease in
the lateral stiffness and an increase in the lateral displacement
[3134]. The experimental results [35] also showed that modeling
of structural nonlinearity needs to be considered not only exural
failure, but also joint shear failure should be taken into account.
This is to achieve the realistic prediction of displacement demands.
Therefore, the monotonic lateral load in the pushover analysis may
provide an over-estimate in the lateral stiffness and strength of the
structure, and this leads to an under-estimate of displacement
demands.

11

The Cyclic Pushover Analysis is proposed to capture these

important characteristics of reinforced concrete members under
cyclic loading. The seismic demands of a 9-story reinforced concrete building are investigated by Cyclic Pushover Procedure. To
determine the effects of displacement history on seismic demands,
four types of loading protocol, i.e., Laboratory, ATC-24, ISO, and
SPD protocol are employed as the prescribed displacement in the
Cyclic Pushover Analysis. The seismic demands include the peak
roof displacement, the peak oor displacement and the peak inter-story drift ratio. The results are compared with the exact demands resulting from nonlinear time history analyses, as well as
the demands estimated from the Modal Pushover Analysis. In this
study, the Cyclic Pushover Analysis procedure is currently limited
to the structures, the responses of which are primarily governed
by the fundamental mode of vibration.
2. Cyclic Pushover Procedure
To evaluate the seismic displacement of structures using the
Cyclic Pushover Analysis, a step-by-step procedure is presented
in this section. This procedure is basically related to the concept
of modal response analysis of nonlinear structures.
2.1. Cyclic pushover lateral force distribution
The modal response analysis which was described in the Modal
Pushover Analysis procedure in Chopra and Goel [57] is applied in
this section. The equivalent governing equation of motion for a
SDOF system, that has been uncoupled from the differential equation of a multi-degree-of-freedom (MDOF) system subjected to
earthquake ground motion, can be presented in the following
form:

n 2nn xn D_ n F sn u
g t
D
Ln

n are the modal displacement, velocity and accelerawhere Dn ; D_ n ; D

tion, respectively, nn and xn are the modal damping and frequency,
respectively, F sn f/n gT ffs Dn ; signD_ n g is the modal internal resisting force; Ln = {/n}T[m]{i}, and [m] is the mass matrix of the structure, and {/n} the corresponding mode shape, and {i} is the unit
g t is acceleration of ground motion.
vector and u
To solve Eq. (1), it is common to conduct a nonlinear SDOF
dynamic time-history analysis. In the analysis, the relationship
between Fsn and Dn is to be determined using pushover analysis.
For Modal Pushover Analysis, the lateral force distribution for the
pushover analysis in each mode is fn, which can be expressed as
follows:

fn Cn mf/n gAn

where fn is the lateral force distribution in each mode, and

An x2n Dn : Cn is the modal participation factor of the nth mode
and it can be determined from

Cn

Ln
f/n gT mfig

M n f/n gT mf/n g

For Cyclic Pushover Analysis, the lateral force distribution is proposed as follows:

fn ki Cn mf/n gAn

where fn is the lateral force distribution for cyclic pushover in each
mode. ki is a variable factor which denes the direction of force, i is
dened as the sequence numbers of peak displacement for the specied displacement history, when i is an odd number (1, 3, 5,. . .)
ki 1, and for i is an even number (2, 4, 6,. . .) ki 1.

12

P. Panyakapo / Engineering Structures 66 (2014) 1023

The structure is subjected to the force distribution in the positive direction until it reaches the rst peak displacement, and then
the force distribution is reversed to the negative direction aiming
to the second peak displacement. This process repeats in cycles
according to the specied displacement history. For example, the
displacement history pattern known as laboratory-test-like-displacement history, which is typically employed in the laboratory,
is applied to control the displacement pattern in Cyclic Pushover
Analysis. In this displacement pattern, as shown in Fig. 1, each cycle is dependent on the displacement ductility ratio, l.
For structures where signicant participation from modes of
vibration other than the fundamental mode is required, higher
mode effects may be determined by conducting higher mode cyclic
pushover analyses. That is, the lateral load distributions as shown
in Eq. (4) are applied for each mode. However, the Cyclic Pushover
Procedure in this study is presented for the structures, the responses of which are primarily governed by the rst mode. For this
purpose, Eq. (4) can be simplied to

f1 ki C1 mf/1 gA1

where C1 is the Participation factor of the rst mode; A1 the acceleration in the rst mode = x21 D1 and x1, D1 is the angular frequency
and displacement in the rst mode.
2.2. Equivalent SDOF pseudo-acceleration and displacement
relationship
It is obvious that the base shear force and roof displacement
relationship obtained from the Cyclic Pushover Analysis is a cyclic
loop reversal of force and displacement. To determine the pushover curve in the form similar to monotonic loading, an envelope
of cyclic loop is normally used to represent the characteristic of
cyclic reversal curve. In this study, the criterion to determine an
envelope curve suggested in ASCE41-06 [20] was adopted. This
takes into account the strength and stiffness deterioration commonly experienced by reinforced concrete structural components.
The backbone curve is dened by points given by the intersection
of an unloading branch and the loading curve of the next load cycle
that goes to a higher level of displacement. To obtain an idealized
cyclic pushover curve, a bilinear curve that is represented for the
base shear and roof displacement relationship can be plotted based
on the following assumptions [20,21]:
(a) The effective stiffness must be such that the rst segment
passes through the curve at a point where the base shear
is 60% of the effective yield strength.
(b) The areas of the segments above and below the curve should
be approximately equal.

Displacement Ductility

25

Peak number (i)

4
2 1 3
0
2

11 13

15

17

19

21 23

10 12 14
16

18

20

6
0

Sa ; A

V b =Wbilinear

a1

The pseudo-displacement, Sd or D is given by

Sd ; D

ur;bilinear

C1  /1;roof

where a1 is the modal mass coefcient for the rst mode, which can
be calculated from

hP

a1 hPN

i1 wi

N
i1 wi /i1 =g

i2

. iP 

N
2
g
g
i1 wi /i1

wi/g is the mass assigned to level i; /i1 is the amplitude of the rst
mode at level i; N the uppermost level of structure; (Vb/W)bilinear the
base shear force normalized with building weight W obtained from
the bilinear of the envelope curve; ur,bilinear is the roof displacement
obtained from the bilinear of the envelope curve and /1,roof is the
roof displacement for the rst mode.
This relationship is developed to represent the rst mode response of the structure based on the assumption that the fundamental mode of vibration is the predominant response of the
structure.
The procedure of Cyclic Pushover Analysis can be summarized
as follows:
(a) Compute the lateral force distribution corresponding to the
rst mode shape, Eq. (5).
(b) Dene the displacement history, which is the relationship
between displacement and the number of cycles to control
the displacement pattern.
(c) Perform nonlinear static analysis using the above specied
force distribution and displacement history. The result is
plotted for the relationship between base shear and roof displacement, which is called cyclic pushover curve.
(d) Determine the envelope and its bilinear representation of
the cyclic pushover curve.
(e) Determine the equivalent single degree of freedom (SDOF)
pseudo-acceleration and displacement relationship by using
Eqs. (6) and (7).
(f) Compute the maximum displacement Dmax for SDOF systems by solving Eq. (1). This is carried out by conducting
nonlinear time history analysis of the bilinear equivalent
SDOF systems.
(g) Convert the maximum displacement Dmax for SDOF systems to the maximum roof and oor displacement for MDOF
structure by the relation

umax Dmax C/1

3. Analysis of 9-story rc building

The bilinear of the cyclic pushover curve is converted to an

equivalent single degree of freedom (SDOF) pseudo-acceleration
and displacement (as shown in Fig. 2) by using the following
relations:

10

3.1. Modeling of structure

22

24

12

14

Cycle Numbers
Fig. 1. Displacement history for laboratory type protocol.

In this study, a 9-story reinforced concrete building was employed in the Cyclic Pushover Analysis. This is a typical apartment
building, constructed in Thailand. The details of the building can be
summarized as follows: (a) the building is relatively symmetric in
plan and vertical views, oor dimensions of 14.40  35.10 m, and
an overall height of 22.50 m, (b) the oor system is precasted

13

P. Panyakapo / Engineering Structures 66 (2014) 1023

Cyclic Pushover Curve

Vb

Envelope Curve

Sa , A

Bilinear Representation

Ay

ur

2
1

Dy

(a) Cyclic pushover and envelope curve

Sd , D

(b) Pseudo-acceleration and displacement

Fig. 2. Conversion from cyclic pushover curve to pseudo-acceleration and displacement.

concrete plank supported by reinforced concrete beams, (c) the

structure is a beamcolumn reinforced concrete system with normal material strength, i.e., compressive strength of concrete is
24 MPa, and the tensile strength of reinforcing steel is 300 MPa.
The reinforced concrete structure was designed primarily for gravity load according to EIT [36]. Since this is an old building that has
been constructed before the seismic regulation became effective, it
was not designed for seismic loading. The beamcolumn frame
was modeled as a two-dimension moment resisting frame in the
NS direction by using RUAUMOKO [37]. The plan view and cross
section details of this building are shown in Fig. 3. The inelastic
behavior of beam and column members is modeled according to
the Giberson one-component concept [38], which has a plastic
hinge possible at one or both ends of the elastic central length of
the member. For Cyclic Pushover Analysis (CPA), the hysteretic
behavior of beam and column is Modied Takeda model [39] as
shown in Fig. 4a. The stiffness degradation for unloading and
reloading branches are a 0:4 and b 0:1 according to Sezen
and Chowdhury [40]. The strength loss in each direction is proportional to the maximum ductility and the numbers of inelastic
cycles.
For Nonlinear Time History Analysis (NTHA), the hysteresis
model was calibrated with the experimental results of reinforced
concrete beams and columns [41] as shown in Fig. 4b. In this analysis, the mass of building was lumped to the nodes of beamcolumn joint for the horizontal degree of freedom. The initial
stiffness Rayleigh damping was assumed to have a damping ratio
of 5%. The Newmark average acceleration method was employed
in the dynamic time history analysis.
3.2. Loading history protocols for Cyclic Pushover Analysis
It is known that the seismic performance of structures in terms
of strength and deformation capacities of structural components
depends on the cumulative damage due to the cyclic loading. The
level of damage is the result of the previous damage of cyclic loading history. Therefore, a loading history protocol plays an important role in the damage of structural components. In this study,
four types of loading history protocol were employed for the Cyclic
Pushover Analysis. These include (a) Laboratory type protocol as
described in the preceding section and shown in Fig. 1, (b) ATC24 protocol [42], in this protocol, the yield displacement is used
as the reference for increasing the amplitude of cycles as shown
in Fig. 5. To determine the yield displacement Dy of the structure
in the loading protocol, the Pushover curve has been constructed
by the conventional Pushover Analysis, then, the yield displacement Dy can be determined based on an idealized bilinear repre-

sentation of the pushover curve. (c) ISO protocol [43], the ISO
protocol has been developed by International Organization for
Standardization (ISO). It is a standard of testing for timber structures whose joints were made with mechanical fasteners. This protocol uses the maximum displacement as the control parameter of
displacement in each cycle as shown in Fig. 6. The maximum displacement Dm, as suggested by Krawinkler [44], is the expected
acceptable performance of the specimen when subjected to this
loading protocol. It could be a design target displacement for which
the component is to be qualied, (d) SPD (Sequential Phased Displacement) Protocol [45], this protocol uses the concept of the
yield displacement as the reference in each cycle, similar to ATC24. The difference is the presence of decay amplitude after reaching its peak displacement in each cycle as shown in Fig. 7. The
SPD protocol was developed by the Technical Coordinating Committee on Masonry Research (TCCMAR). The loading protocol is
based on the First Major Event (FME) which is considered as the
yield displacement of the specimen.
3.3. Ground motion records
For the ground motions, these were collected from a moderately
strong magnitude and near-fault earthquakes with a magnitude
between 6.1 (ML) and 7.1 (Ms), and epicentral distance less than
40 km, as shown in Table 1. They represent for earthquake events
that may occur in the northern part of Thailand. These ground motions were scaled to the Maximum Considered Earthquake (MCE)
response spectrum, which corresponds to a 2% probability of
exceedance in 50 years according to SPT 1302 [46]. The ground
motions were adjusted by different scale factors so that the average of the scaled SRSS 5-percent-damped spectra over the period
range 0.2T to 1.5T is not less than the MCE response spectrum. T
is the fundamental mode period of the vibration of structure. The
resulting scaled spectrum corresponds to the normal soil prole
(Soil Type D) according to SPT 1302 [46]. From Table 1, the last column shows input energy equivalent velocity, VI, which is dened
as the equivalent velocity of the normalized input energy. Previous
researchers [4750] have studied reliable parameters to measure
the damage potential of earthquake ground motion. Among many
parameters, VI is an interesting parameter that closely relates to
the earthquake damage potential for moderate seismicity region.
Therefore, this parameter was selected as an index to correlate
with the seismic demands of the structure in this study. It is known
that reliable nonlinear time history analysis results depend on the
use of appropriate ground motion records. These include the characteristics of soil conditions on the sites, frequency contents, the
duration of the ground motions, etc. The selected ground motion

P. Panyakapo / Engineering Structures 66 (2014) 1023

B2

cast-in-place slab
precast slab

ST

stairway

8
7

9@2.5 = 22.5 m

6
5
4
3
2
1
G

C3

C3

C3

C
B4

B8

B3

C3

C3

B4

B4

B4

C3
B3

B3

C3

C3

C3

B3

B3

B3

C3
B4

C2

B4

B2

S1

B4

B4

B4

B2

B3

B3

PS

B2

B2

S1

B2

B2

ST

PS

PS

PS

B3

B2

B2

B2
S1

B3

B3

B3

B3
B4

PS

PS

PS

PS

B2
PS

B4

PS

PS

B2

PS
B2

B8

B3

B2

S1

PS

B2

B2

PS

B2
S1

PS

PS

Plan view

6.0 m

B3

PS

S1

2.4 m

C3

B2
PS

B2

6.0 m

C3

C2

S1

B2

S1

PS

S1

B2

PS

PS

2.00 m

B2

PS

B2

1.90 m
B2

B3

B3
B2

PS

PS

B2

PS

3.90 m

2.00 m
B2

PS

B2

B2

1.90 m

B2

PS

10

3.90 m

2.00 m

S1

S1
B2

PS

PS

1.90 m
B2

PS

B2

B2

PS

B3

1.40 m

B2

B2

PS

B3

4.60 m

6.00 m

PS

S1

B3

B2

PS

PS

PS

1.90 m
B2

B2

B2

PS

B4

B4

B4

2.40 m

PS

S1

S1

PS

B2

2.00 m

3.90 m

3.90 m

2.00 m
B2

B2

PS

B2

B2

PS

B3

B2

ST

PS

B3
PS

1.90 m

1.90 m
B2

B2

B3

B3

6.00 m

4.60 m

PS

S1

S1
B2

B2

2.00 m

B2

PS

3.90 m

3.90 m

2.00 m

B3

B2
PS

1.90 m

B3

1.40 m

S1

1.90 m

B3

2.00 m

B2

3.90 m

3.90 m

2.00 m

B3

3.90 m
1.90 m

B4

B4

B4

B3

14

C3

5DB25mm

14DB20 mm

C3

4-RB6 mm@0.20 m

2-RB9mm@0.20m

0.25x0.40 m

5DB25mm

C3

0.25x0.45

B3

C3
14DB25 mm

2DB12mm

4-RB6 mm@0.20 m

C3
B3

C2

B3

RB6mm@0.20m
C2

0.30x0.40 m

C2

3RB12mm

0.25x0.40
C2

B3

C2

B4

C2

B3

C2

16DB25 mm

B4

4-RB6 mm@0.20 m
C2

C1

B3

B3

C1

C2

C1

B4

B4

C1
B3

C2

C1

B3

B3

C1
B4

4DB25mm

C2

0.30x0.50 m
C1

RB6mm@0.125m

C1
Column Details

C1

4DB25mm

0.25x0.45

B8

B3

Beam Details
Cross Section
Fig. 3. Plan view and cross section details of 9-story RC building.

records have variable characteristics, which are correspondent to

moderate seismicity and near-fault earthquakes. Therefore, the results presented in this paper may be limited to the site and seismic
characteristics used in this study.

4. Evaluation of seismic displacement demands

This section presents the results of seismic displacement demands based on the Cyclic Pushover Analysis (CPA) procedure for

P. Panyakapo / Engineering Structures 66 (2014) 1023

15

Yield Displacements, D y

Fig. 4. Hysteresis models used for CPA and NTHA.

5.04
4.08

3.00
1.92

2
0.48

0.96

0.72

0
-2
-4
-6
0

10

Cycle Numbers
Fig. 5. ATC-24 protocol.

15

20

the 9-story reinforced concrete building. Four types of loading history protocol, i.e., Laboratory type, ATC-24, ISO, and SPD protocols,
were employed for the Cyclic Pushover Analysis. The cyclic pushover and their envelope curves under the four loading protocols
are presented. The envelopes of cyclic pushover curves were transformed to the equivalent bilinear SDOF models. The BISPEC [51]
program was employed to conduct the SDOF dynamic analysis
for the equivalent bilinear SDOF models. The seismic demands
were evaluated for the peak roof displacements, the peak oor displacement, and the peak inter-story drift ratio. The results are
compared with those from MDOF Nonlinear Time History Analysis
(NTHA) and the Modal Pushover Analysis (MPA) procedure proposed by Chopra and Goel [57].

P. Panyakapo / Engineering Structures 66 (2014) 1023

Maximum Displacement, Dm

16

provided a decrease in yield strength and an increase in yield displacement resulting in a decrease in the lateral stiffness and
strength of structure when it is compared to the conventional
pushover curve. This is due to the yielding effect under cyclic loading, and causing plastic hinge formation in some components of
the structure. These plastic hinge regions absorbed hysteretic energy due to each cyclic loading, resulting in the cumulative damage
of some beam and column components. These components experienced stiffness degradation and strength deterioration in each
inelastic cycle, and consequently a decrease in the lateral stiffness
and strength of the structure. The strength loss is dependent on
cumulative damage due to each loading history. From Fig. 9, the
pseudo-acceleration and displacement (Sa  Sd) relationship of
the Lab-type protocol provides slightly lower stiffness and strength
than that of the Pushover Analysis (PA). Contrary to those of ATC24, SPD, and ISO protocols, the stiffness and strength are much
lower than that of PA. Because the loading histories for ATC-24,
SPD, and ISO protocols consist of many repeated cycles, particularly in the initial loading range before it reaches the yield displacement, cumulative damages occur. It is also noticed that the
ISO protocol provides the lowest strength because the loading history consists of large displacement excursions for many inelastic
cycles. As a result, the ISO protocol produces large absorbed energy
demands, resulting in high cumulative damage.

1.5
1.26
1.02

1.0

0.78
0.60

0.5
0

0.42
0.06 0.09

0.18

-0.5
-1.0
-1.5

10

15

20

Cycle Numbers
Fig. 6. ISO protocol.

Yield Displacement, Dy

4.0
3.0
2.0
1.0

0.25 0.50

1.00 1.20

0.80

1.20

1.60 1.60

2.00

2.00 2.00 2.00

1.60

2.60
2.00

3.00
2.60

0
-1.0
-2.0
-3.0
-4.0

10

20

30

40

4.2. Peak roof displacement demands

Cycle Numbers
Fig. 7. SPD protocol.

4.1. Cyclic pushover curves and equivalent bilinear SDOF models

The cyclic pushover curves and their envelope curves are plotted in comparison with the conventional pushover curve as shown
in Fig. 8ad. The equivalent bilinear SDOF models in terms of pseudo-acceleration and displacement (Sa  Sd) relationships for the
four types of loading protocols are plotted in comparison to the
conventional pushover (mode 1) as shown in Fig. 9. The bilinear
slopes in terms of x2 and ax2, as well as the yield acceleration
Ay, and the yield displacement Dy, are presented in Table 2. These
parameters were employed to conduct nonlinear time history analyses of the idealized bilinear equivalent SDOF models by BISPEC
program.It is observed that the Cyclic Pushover Analysis (CPA)

The peak roof displacements were evaluated by using nonlinear

time history analyses of the idealized bilinear equivalent SDOF
models. The peak roof displacements computed based on the CPA
are compared with those of the MPA and the exact MDOF Nonlinear Time History Analysis (NTHA) for each ground motion as
shown in Table 3. The peak roof displacements obtained from the
CPA provide the mean displacements of 10.40, 11.51, 11.33,
11.12 cm for Laboratory Type, ATC-24, ISO, SPD, respectively. These
results are compared with the mean displacement of the exact
NTHA, which is 10.75 cm. The percentages of error are 3.23 for
the under-estimate for Laboratory Type, and 7.07, 5.35, 3.43 for
the over-estimates for ATC-24, ISO, SPD, respectively. The results
of MPA provide the mean displacements of 9.25, 9.63, 9.71 cm
for one mode, two modes and three modes, respectively. The percentages of error are 13.97, 10.46, 9.70 under-estimates for one
mode, two modes and three modes, respectively. For an overview,

Table 1
Ground motions and scale factor to SPT 1302.
No

Record

Earthquake

Magnitude

Epicentral distance (km)

PGA (g)

Scale factor

Input energy equivalent

velocity, VI (m/s)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20

IMP-1
IMP-2
PARK-1
PARK-2
IMP-3
IMP-4
MAM-1
MAM-2
NAHAN-1
NAHAN-2
SPI-1
SPI-2
LOMA-1
LOMA-2
LOMA-3
LOMA-4
LOMA-5
LOMA-6
NORTH-1
NORTH-2

Imperial Valley 1940

6.3 (ML)

Parkeld
1966
Imperial Valley 1979

6.1 (ML)

9.9

6.6 (ML)

26.5

Mammoth Lake 1980

6.1 (ML)

15.5

Nahanni, Canada 1985

6.9 (Ms)

16

Spitak, Armenia 1988

7.0 (Ms)

30

Loma Prieta 1989

7.1 (Ms)

11.2

Loma Prieta 1989

7.1 (Ms)

21.4

Loma Prieta 1989

7.1 (Ms)

28.2

Northridge 1994

6.7 (Ms)

26.8

0.348
0.214
0.357
0.272
0.169
0.157
0.430
0.271
0.148
0.139
0.199
0.175
0.411
0.473
0.244
0.240
0.247
0.215
0.165
0.217

1.27
1.81
1.42
1.95
2.64
2.00
1.72
2.05
5.87
7.15
2.31
2.68
0.89
0.72
1.72
1.77
1.74
1.62
3.16
2.08

1.056
1.176
0.427
0.494
1.055
1.695
0.843
0.820
0.802
0.713
0.940
0.522
0.423
0.395
0.665
0.883
1.310
1.478
0.632
0.530

17

P. Panyakapo / Engineering Structures 66 (2014) 1023

PA, Pushover Curve (Mode 1)

Cyclic Pushover Curve

Base Shear, kN

CPA (Lab Type), Envelope Curve

2000

PA, Pushover Curve (Mode 1)

Cyclic Pushover Curve (ISO)

Base Shear, kN

CPA (ISO), Envelope Curve

2000

PA

PA

1500

1500

CPA (Lab Type)

CPA(ISO)

1000

1000

500

500
0
-1.00 -0.80 -0.60 -0.40 -0.20 0.00

0.20

-500

0.40

0.60

0.80

1.00

0
-1.00 -0.80 -0.60 -0.40 -0.20 0.00

0.20

0.40

-500

Roof Displacement, m

-1000

-1000

-1500

-1500

-2000

-2000

1.00

Roof Displacement, m

PA, Pushover Curve (Mode 1)

Cyclic Pushover Curve (ATC-24)
CPA (ATC-24), Envelope Curve

2000

0.80

(c) Cyclic pushover curve for ISO type

(a) Cyclic pushover curve for Laboratory type

Base Shear, kN

0.60

PA, Pushover Curve (Mode 1)

Cyclic Pushover Curve (SPD)

Base Shear, kN

CPA (SPD), Envelope Curve

2000
PA

1500

CPA (ATC-24)

PA

1000

1000

500

500

0
-1.00 -0.80 -0.60 -0.40 -0.20 0.00

0.20

0.40

-500

0.60

0.80

1.00

Roof Displacement, m

CPA(SPD)

1500

0
-1.00 -0.80 -0.60 -0.40 -0.20 0.00
-500

-1000

-1000

-1500

-1500

-2000

-2000

0.20

0.40

0.60

0.80

1.00

Roof Displacement, m

(d) Cyclic pushover curve for SPD type

(b) Cyclic pushover curve for ATC-24 type

Fig. 8. Cyclic pushover curves for the four types of loading protocols.

PA (Mode 1)

Sa,A(m / sec2)

12.00
10.00

CPA (SPD)
CPA (ATC-24)

8.00

CPA (LAB)

6.00

CPA (ISO)

4.00
2.00
0.00
0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0.80

Sd,D(m)
Fig. 9. Pseudo-acceleration and displacement for the Cyclic Pushover and Pushover
(mode 1).

the results of the CPA are slightly over-estimated; however, those

of MPA are under-estimated.
To gain more details, the estimated peak roof displacements of
each ground motion were evaluated for error from the exact NTHA.
The results of the CPA provide the mean errors of 11.59%, 12.55%,
15.30%, 11.68% over-estimates, and 12.31%, 6.50%, 6.96%,
9.39% under-estimates for Laboratory Type, ATC-24, ISO, SPD,
respectively. The mean absolute errors are 11.95%, 11.04%,
12.38%, and 10.99% for Laboratory Type, ATC-24, ISO, SPD,
respectively.
The results of the MPA provide the mean errors of 17.51%,
19.99%, 22.09% over-estimates, and 24.08%, 23.96%, 23.48%
under-estimates for one mode, two modes, three modes, respectively. The mean absolute errors are 20.79%, 21.58%, 22.65% for
one mode, two modes, and three modes, respectively. It can be observed that the combined modal response demands resulting from
higher mode effects do not signicantly reduce the errors. The errors tend to increase when the higher modes were considered because the estimated peak roof displacements of the MPA were
over-estimate for many ground motions. Therefore, the inclusion

18

P. Panyakapo / Engineering Structures 66 (2014) 1023

Table 2
Parameters for the Pushover Analysis (PA) and the Cyclic Pushover Analysis (CPA).
Parameters

Ay (m/s2)
Dy (m)
x2 (radian/s2)
Post-elastic stiffness, a

PA

CPA

Mode 1

Mode 2

Mode 3

(LAB)

(ATC)

(ISO)

(SPD)

8.80
0.11
77.19
0.11

54.0
0.30
180.0
0.17

140.0
0.53
264.15
0.19

8.07
0.13
62.56
0.17

7.33
0.13
56.38
0.12

6.23
0.11
54.65
0.09

6.60
0.11
57.89
0.20

Table 3
Peak roof displacements for MPA, CPA, and NTHA.
Earthquake ground motions

Cyclic Pushover Analysis

Modal Pushover Analysis

Nonlinear time history

analysis (cm)

LAB (cm)

ATC-24 (cm)

ISO (cm)

SPD (cm)

1 Mode (cm)

2 Modes (cm)

3 Modes (cm)

IMP-1
IMP-2
IMP-3
IMP-4
LOMA-1
LOMA-2
LOMA-3
LOMA-4
LOMA-5
LOMA-6
MAM-1
MAM-2
NAHAN-1
NAHAN-2
NORTH-1
NORTH-2
PARK-1
PARK-2
SPI-1
SPI-2

12.93
16.25
8.70
17.23
4.16
4.10
10.38
11.82
25.86
21.83
12.12
9.57
4.37
4.53
3.56
2.32
4.40
3.59
21.09
9.27

16.86
15.21
10.49
18.64
4.06
4.97
11.48
9.57
29.28
25.12
13.73
10.68
4.16
4.97
3.26
2.42
4.77
4.30
24.08
12.16

17.19
14.37
10.58
17.86
4.03
5.24
11.69
9.44
27.70
25.72
13.77
10.98
4.23
5.04
2.99
2.38
4.97
4.57
20.89
12.96

16.12
16.02
10.21
18.57
4.10
4.73
11.25
9.91
26.66
24.65
13.53
10.21
4.16
4.90
3.39
2.35
4.63
4.06
21.62
11.38

10.78
13.36
9.33
13.10
4.30
3.46
10.68
16.59
15.08
19.91
11.54
11.99
4.53
4.63
3.86
2.35
4.10
4.23
13.57
7.66

11.15
13.57
9.97
13.23
4.58
4.50
11.19
16.72
15.42
20.05
12.09
12.42
4.66
4.76
3.96
2.56
4.73
4.90
13.78
8.34

11.18
13.61
10.10
13.26
4.67
4.76
11.23
16.75
15.44
20.07
12.14
12.45
4.69
4.79
3.98
2.62
5.03
5.10
13.82
8.46

14.46
14.73
8.62
12.72
3.76
4.33
10.17
8.91
29.54
26.38
12.3
10.12
3.98
4.32
3.3
2.85
4.33
4.16
22.55
13.56

Mean displacements
Error of mean displacements
Mean error over-estimate
Mean error under-estimate
Mean absolute error

10.40
3.23%
11.59%
12.31%
11.95%

11.51
7.07%
12.55%
6.50%
11.04%

11.33
5.35%
15.30%
6.96%
12.38%

11.12
3.43%
11.68%
9.39%
10.99%

9.25
13.97%
17.51%
24.08%
20.79%

9.63
10.46%
19.99%
23.96%
21.58%

9.71
9.70%
22.09%
23.48%
22.65%

10.75

of higher modes increased the peak roof displacements over the

exact demands. From the above comparisons, it appears that the
results of the CPA provide lower percentages of error than those
of the MPA.
The reason that the CPA procedure provides a lower percentage
of error than those of the MPA method is due to the stiffness degradation and strength deterioration under cyclic loading. The prescribed loading protocols require many inelastic cycles, and
hence, some columns undergo inelastic behavior resulting in a
reduction of stiffness and strength. This is consistent with the
behavior of reinforced concrete structures under earthquake loading. The estimated roof displacement demand is therefore close to
the exact demand.
To determine the accuracy of the estimation in a statistical aspect, the peak roof displacements of each ground motion were veried for their accuracy by computing the Peak Roof Displacement
Ratio (PRDR).

PRDR

Destimate
Dexact

10

where Destimate is the estimated peak roof displacement and Dexact is

the exact peak roof displacement.
The PRDR of each ground motion was plotted against the percentage of frequency interval of data. The results of the MPA for
one mode, three modes, and the CPA for LAB type, ATC-24, ISO,
and SPD are shown in Fig. 10af, respectively.

From Fig. 10cf, the median of PRDR estimated from the CPA are
0.997, 1.071, 1.078, and 1.079 with Standard Deviation (SD) of
0.161, 0.129, 0.135, and 0.137 for Laboratory Type, ATC-24, ISO,
SPD, respectively. When these are compared with the MPA, the
median of PRDR estimated from the MPA are 0.982 and 1.101 with
Standard Deviation (SD) of 0.294 and 0.300 for one mode and three
modes, respectively. It was found that the CPA results are close to
the exact value (PDRD = 1.00) with a small Standard Deviation.
To determine the scatter of the estimated peak roof displacement for each ground motion, the error of the estimated peak roof
displacement of the CPA for Lab-Type, ATC-24, ISO, SPD and those
of the MPA were plotted against the input energy equivalent velocity, VI, as shown in Figs. 11(a)(d), respectively. It should be noted
that the errors of the CPA scatter in a narrow range when they are
compared with those of the MPA. The percentage errors of the CPA
for Lab-Type, ATC-24, ISO, and SPD protocols scatter mostly within
the range of 35%, 20%, 20%, and 20%, respectively. This is unlike
those of the MPA, which scatter within the range of 40%. These
are considered based on the majority result of errors.
When the results of the CPA are compared among each other,
the errors of the peak roof displacement for ATC-24, ISO, and SPD
protocols are less scattered than those of the Lab-Type protocol.
These are consistent with the characteristics of loading protocols.
ATC-24, ISO, and SPD protocols consist of many repeated cycles
resulting in cumulative damage and consequently stiffness and
strength degradation. This behavior may be consistent with the
characteristics of earthquake ground motions.

19

P. Panyakapo / Engineering Structures 66 (2014) 1023

Median = 0.982

30.0

SD = 0.294

20.0
10.0

40.0
30.0
20.0

0.5

1.0

1.5

0.0
0.0

2.0

Peak Roof Displacement Ratio

Median = 0.997

(a) PA - 1 mode

40.0
30.0

50.0

SD = 0.161

20.0
10.0

1.5

2.0

40.0

SD = 0.129

30.0
20.0
10.0

0.5

1.0

1.5

0.0
0.0

2.0

Peak Roof Displacement Ratio

30.0

1.5

2.0

(d) CPA - ATC-24 Type

50.0

SD = 0.135

20.0
10.0

Frequency (%)

Median = 1.078

40.0

1.0

Peak Roof Displacement Ratio

(c) CPA- LAB Type

50.0

0.5

Median = 1.079

0.0
0.0

0.0
0.0

1.0

(b) MPA - 3 modes

Frequency (%)

50.0

0.5

Peak Roof Displacement Ratio

Median = 1.071

0.0

Frequency (%)

SD = 0.300

10.0

0.0

Frequency (%)

Median = 1.101

40.0

50.0

Frequency (%)

Frequency (%)

50.0

40.0

SD = 0.137

30.0
20.0
10.0

0.5

1.0

1.5

2.0

Peak Roof Displacement Ratio

(e) CPA - ISO Type

0.0
0.0

0.5

1.0

1.5

2.0

Peak Roof Displacement Ratio

(f) CPA - SPD Type

Fig. 10. Peak roof displacement ratios and frequency.

4.3. Peak oor displacement demand

To determine the peak oor displacement of each story of the 9story building, the oor displacement demands for a MDOF system
can be obtained by multiplying the peak roof displacement with
the normalized mode shape, as shown in Eq. (9). The results obtained from the 20 ground motions are presented in terms of a single value as the mean of the maximum oor displacement of each
story. The mean values for the CPA procedure are plotted with the
oor levels of building and compared with those of the exact Nonlinear Time History Analysis (NTHA) and those of the MPA method,
as shown in Fig. 12. It was observed that the mean peak oor dis-

placements for the CPA, especially for LAB Type, are close to the exact displacement demands. While the displacement demands of
the other CPA types, i.e., ATC-24, ISO, SPD, are slightly greater than
the exact values, particularly for the upper oor levels. This is due
to the fact that the characteristics of the loading protocols for
ATC-24, ISO, and SPD consist of many repeated cycles at each peak
displacement. This leads to the tendency of more degradation of
stiffness and strength, and hence, they tend to produce large seismic displacement demands. Contrary to those of the MPA method,
the mean peak oor displacements are slightly under-estimated.
To determine the accuracy of the proposed procedure, the error
of the mean peak oor displacements for each story between the

20

P. Panyakapo / Engineering Structures 66 (2014) 1023

MPA (3 Modes)

MPA (3 Modes)

CPA (LAB - Type)

CPA (ISO - Type)

Input Energy Equivalent Velecity (m/sec.)

Input Energy Equivalent Velecity (m/sec.)

2.00

2.00

1.50

1.50

1.00

1.00

0.50

-100 -80

-60

-40

0.00
-20
0

0.50

20

40

60

80

100

-100 -80

Error of Peak Roof Displacement (%)

-60

-40

0.00
-20
0

20

40

60

80

100

Error of Peak Roof Displacement (%)

Fig. 11a. Error of peak roof displacement for CPA (LAB Type).

Fig. 11c. Error of peak roof displacement for CPA (ISO Type).

MPA ( 3 Modes )

MPA ( 3 Modes )

CPA (ATC-24)

CPA ( SPD Type )

Input Energy Equivalent Velocity (m/sec.)

2.00

Input Energy Equivalent Velocity (m/sec.)

2.00

1.50

1.50
1.00

1.00
0.50

0.50

-100 -80

-60

-40

0.00
-20
0

20

40

60

80

100

Error of Peak Roof Displacement (%)

Fig. 11b. Error of peak roof displacement for CPA (ATC-24 Type).

-100 -80

-60

-40

0.00
-20
0

20

40

60

80

100

Error of Peak Roof Displacement (%)

Fig. 11d. Error of peak roof displacement for CPA (SPD Type).

CPA procedure and the exact solutions were evaluated. The results
are also compared with those of the MPA method for one mode,
two modes, and three modes, as shown in Fig. 13. For an overview
consideration, the percentage errors of the mean peak oor displacement for each story were computed for the average of all nine
stories. The results are also presented in three groups as described
in the preceding section, as shown in Table 4.
From Fig. 13 and Table 4, it was found that the errors of peak
oor displacements for the CPA procedure are deviated in a narrow
boundary. They have a similar trend to that of the MPA (1 mode)
because the proposed CPA procedure employs the normalized rst
mode shape in the determination of the oor displacement. For
those of the MPA method, the combined modal response demands,
resulting from higher mode effects, tend to reduce the under-estimated errors, especially for the lower oors (third oor and fourth
oor). Nevertheless, when the whole building was considered as
shown in Table 4, the mean absolute errors for the CPA procedure

are 6.65%, 4.43%, 4.14%, and 3.88% for LAB, ATC-24, ISO, SPD,
respectively. These are less than those of the MPA, which are
16.99%, 8.84%, 11.22% for 1 mode, 2 modes, and 3 modes, respectively. Therefore, the CPA procedure provides better estimates than
those of the MPA for the MDOF peak oor displacement demands.
4.4. Peak inter-story drift ratio demand
The inter-story drift is an important parameter to determine the
seismic performance of buildings. This section presents the peak
inter-story drift ratio which is the maximum relative oor displacement between two adjacent oor levels divided by the story
height. The results obtained from the 20 ground motions are presented in terms of the mean of the peak inter-story drift ratio of
each story. The mean values for the CPA procedure are plotted with
the oor levels of building and compared with those of the exact

21

P. Panyakapo / Engineering Structures 66 (2014) 1023

7
NTHA

Floor Level

Floor Level

6
NTHA

CPA ( LAB Type )

CPA (ATC-24)
CPA (ISO)

MPA ( 1 Mode)

MPA (3 modes)

Inter Story Drift Ratio (%)

Fig. 12. Mean peak oor displacements under 20 ground motions.

T yp
e
LAB

I SO
S PD

Fig. 14. Mean peak inter-story drift ratio under 20 ground motions.

-24

Floor Level

AT C

MPA (3 Modes )

Displacement (m)

MPA (1 mode)
MPA (2 modes)

0
0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 0.20

Floor Level

CPA (ISO)
CPA (SPD)

CPA (ATC-24)

CPA (SPD)

CPA ( LAB Type )

7
7

6
6

CPA (LAB)

CPA (ATC-24)

5
mode)

PA

A(

(2
mo

-15

-10

-5

0
0

3m
o de
s)

10

CPA (SPD)

MPA (1 mode)

MPA (2 modes)
MPA (3 modes)

s)
de

-20

MP

MPA (1

1
-25

CPA (ISO)

15

20

25

Error of peak floor displacement (%)

Error of Peak Inter-Story Drift Ratio (%)

Fig. 13. Error of mean peak oor displacements under 20 ground motions.

Fig. 15. Errors of mean peak inter-story drift ratio under 20 ground motions.

Nonlinear Time History Analysis (NTHA) and those of the MPA

method, as shown in Fig. 14.
The errors of the mean peak inter-story drift ratio for each story
between the CPA procedure and the exact solutions were evaluated. The results are also compared with those of the MPA method
for one mode, two modes, and three modes, as shown in Fig. 15. For
an overview consideration, the errors of the mean peak inter-story
drift ratio for each story were computed for the averages of all nine
stories. The results are also presented in three groups similar to the
preceding section, as shown in Table 5.
For a summary, the mean peak inter-story drift ratios for both of
the CPA and the MPA procedures are consistent with the exact demands. It is observed that the peak inter-story drift ratios for the
CPA procedure are close to the exact demands, particularly for

the 1st oor to 4th oor. The results of the CPA procedure are
slightly over-estimated for the 5th oor to roof level. When the
whole building was considered as shown in Table 5, the CPA procedure provided 14.77%, 22.97%, 21.04%, 22.70% over-estimates
and 9.86%, 1.33%, 2.87%, 3.63% under-estimates for LAB,
ATC-24, ISO, SPD, respectively. The results of the CPA procedure
are better than those of the MPA method, which provide 13.73%,
28.91%, 36.75% over-estimates and 13.78%, 19.21%, 21.07%
under-estimates for one mode, two modes and three modes,
respectively. The mean absolute errors for the CPA procedure are
12.59%, 15.76%, 14.99%, and 14.23% for LAB, ATC-24, ISO, SPD,
respectively. These are comparable to that of the MPA for one
mode, which is 13.77%. However, the effect of modal combination
of the higher modes increased the errors of the MPA procedure to
23.52% and 28.04% for two modes and three modes, respectively.

Table 4
Mean percentage of errors of the mean peak oor displacement for the whole building.
Error

Mean error over-estimate

Mean error under-estimate
Mean absolute error from NTHA

Cyclic pushover

Modal pushover

LAB (%)

ATC (%)

ISO (%)

SPD (%)

1 Mode (%)

2 Modes (%)

3 Modes (%)

6.65
6.65

5.78
1.73
4.43

5.22
2.80
4.14

4.13
3.68
3.88

16.99
16.99

6.57
9.49
8.84

12.79
10.44
11.22

22

P. Panyakapo / Engineering Structures 66 (2014) 1023

Table 5
Errors of the mean peak inter-story drift ratio for the whole building.
Error

Mean error over-estimate

Mean error under-estimate
Mean absolute error from NTHA

Cyclic pushover

Modal pushover

LAB (%)

ATC (%)

ISO (%)

SPD (%)

1 Mode (%)

2 Modes (%)

3 Modes (%)

14.77
9.86
12.59

22.97
1.33
15.76

21.04
2.87
14.99

22.70
3.63
14.23

13.73
13.78
13.77

28.91
19.21
23.52

36.75
21.07
28.04

Therefore, the proposed CPA procedure provides, on average, a

more accurate estimate for the mean peak inter-story drift ratios
than the MPA method.
5. Conclusions
The Cyclic Pushover Analysis is proposed to capture the characteristics of stiffness and strength degradation of reinforced concrete structures under cyclic loading. The seismic displacement
demands for a 9-story reinforced concrete building are evaluated,
i.e., the peak roof displacement, the peak oor displacement, and
the peak inter-story drift ratio. The results are compared with
the exact MDOF nonlinear time history analysis and the Modal
Pushover Analysis. The effects of displacement history on seismic
demands are investigated. It should be remarked that the results
of this study are limited to the following assumptions: (a) one sample building in which the response is primarily controlled by the
rst mode, (b) a set of ground motion records with moderate seismicity and near-fault earthquakes. The conclusions are summarized below.
(a) The Cyclic Pushover Analysis procedure based on lateral cyclic loading provides better estimates of the displacement
demands than those of monotonic loads. This is due to the
effect of the cumulative damage resulting from cyclic load
reversal, which cannot be accounted for by monotonic pushover analysis; and so it leads to a reduction in the stiffness
and strength of the structure. This is consistent with the
behavior of reinforced concrete structures under earthquake
loading. The estimated seismic displacement demands are
therefore close to the exact demands.
(b) The displacement demands are sensitive to the characteristics of loading protocols that dene the cyclic loading history. The loading protocols that consist of many repeated
cycles at each peak displacement in the initial loading range
and beyond the yield displacement tend to produce large
cumulative damage. This effect leads to more severe degradation of stiffness and strength, and consequently structures
undergo large seismic displacement demands.
(c) The peak oor displacement and inter-story drift ratio for
the Cyclic Pushover Analysis (CPA) are close to the exact
demands. These may be explained by two reasons. Firstly,
the shape of the oor displacement for the CPA is determined based on the fundamental mode shape. This is consistent with the response of this sample building which is
dominated by the rst mode. Secondly, the estimated roof
displacements for the CPA procedure are close to the exact
demands, and the peak oor displacement and inter-story
drift ratio are in proportion to the peak roof displacements,
and as a consequence, the results have a similar trend to
the exact demands.
(d) It is worth pointing out that the seismic demands in this
study are limited to the displacement demands. In view of
the seismic damage assessment, damage of a structure is
not only dependent on the displacement demand, but also

the cumulative damage caused by cyclic loading has to be

taken into account. For this purpose, the Cyclic Pushover
Analysis is useful to estimate the seismic damage due to
the cumulative damage. Particularly for the old reinforced
concrete structures that had not been designed for seismic
resistance, the contribution of cumulative damage may play
an important role. This requires an investigation for further
study.

Acknowledgments
This research was supported by the Research Grant of Sripatum
University. The author wishes to thank the staffs of Sripatum University who contributed materials and provided helpful input for
this research project.
References
[1] Antoniou S, Pinho R. Development and verication of a displacement-based
adaptive pushover procedure. J Earthquake Eng 2004;8(5):64361.
[2] Papanikolaou VK, Elnashai AS, Pareja JF. Evaluation of conventional and
adaptive pushover analysis II: Comparative results. J Earthquake Eng
2006;10(1):12751.
[3] Abbasnia R, Davoudi AT, Maddah MM. An adaptive pushover procedure based
on effective modal mass combination rule. Eng Struct 2013;52:65466.
[4] Shakeri K, Shayanfar MA, Kabeyasawa T. A story shear-based adaptive
pushover procedure for estimating seismic demands of buildings. Eng Struct
2010;32:17483.
[5] Chopra AK, Goel RK. A modal pushover analysis procedure for estimating
seismic demand of buildings. Earthquake Eng Struct Dynam 2002;31:56182.
[6] Chopra AK, Goel RK. A modal pushover analysis procedure to estimate seismic
demands for unsymmetrical-plan buildings. Earthquake Eng Struct Dynam
2004;33(8):90327.
[7] Chopra AK, Goel RK. Role of higher mode pushover analyses in seismic analysis
of buildings. Earthquake Spectra 2005;21(4):102741.
[8] Kim S-P, Kurama YC. An alternative pushover analysis procedure to estimate
seismic displacement demands. Eng Struct 2008;30:3793807.
[9] Poursha M, Khoshnoudian F, Moghadam AS. A consecutive modal pushover
procedure for estimating the seismic demands of tall buildings. Eng Struct
2009;31:5919.
[10] Poursha M, Khoshnoudian F, Moghadam AS. A consecutive modal pushover
procedure for nonlinear static analysis of one-way unsymmetric-plan tall
building structures. Eng Struct 2011;33:241734.
[11] Khoshnoudian F, Mehdi M, Kashani B. Assessment of modied consecutive
modal pushover analysis for estimating the seismic demands of tall buildings
with dual system considering steel concentrically braced frames. J Constr Steel
Res 2012;72:15567.
[12] Hernandez-Montes E, Kwon OS, Aschheim MA. An energy based formulation
for rst and multiple-mode nonlinear static (Pushover) Analyses. J Earthquake
Eng 2004;8(1):6988.
[13] Leelataviwat S, Saewon W, Goel SC. Application of energy balance concept in
seismic evaluation of structures. J Struct Eng 2009;135(2):11321.
[14] Jiang Y, Li G, Yang D. A modied approach of energy balance concept based
multimode pushover analysis to estimate seismic demands for buildings. Eng
Struct 2010;32:127283.
[15] Manoukas G, Athanatopoulou A, Avramidis I. Static pushover analysis based on
an energy-equivalent SDOF system. Earthquake Spectra 2011;27(1):89105.
[16] Fajfar P. A nonlinear analysis method for performance based seismic design.
Earthquake Spectra 2000;16(3):57392.
[17] Eurocode 8. Design of structures for earthquake resistance, Part 1: General
rules, seismic actions and rules for buildings. European standard EN 19981.
European Committee for Standardization (CEN), Brusselsl; 2004.
[18] Kreslin M, Fajfar P. The extended N2 method taking into account higher mode
effects in elevation. Earthquake Eng Struct Dynam 2011. http://dx.doi.org/
10.1002/eqe.1104.

P. Panyakapo / Engineering Structures 66 (2014) 1023

[19] Applied Technology Council (ATC-40). Seismic evaluation and retrot of
concrete buildings, Washington, DC; 1996.
[20] American Society of Civil Engineers (ASCE). Seismic rehabilitation of existing
building: ASCE Standard No. ASCE/SEI 4106; 2007.
[21] FEMA. Improvement of nonlinear static seismic analysis procedures. Federal
Emergence Management Agency. FEMA 440, Washington D.C.; 2005.
[22] Miranda E. Inelastic displacement ratios for structures on rm sites. J Struct
Eng 2000;126(10):11509.
[23] Chopra AK. Estimating seismic demands for performance-based engineering of
buildings. Key-note lecture at the 13th world conference on earthquake
engineering. Paper No. 5007. Vancouver; 2004.
[24] Fajfar P, Marusic D, Perus I. Torsional effects in the pushover-based seismic
analysis of buildings. J Earthquake Eng 2005;9(6):83154.
[25] Fajfar P, Maruic D, Perus I. The extension of the N2 method to asymmetric
buildings. In: Proceedings of the 4th European workshop on the seismic
behavior of irregular and complex structures, CD ROM. Thessaloniki;
2005.
[26] Koren D, Kilar V. The applicability of the N2 method to the estimation of
torsional. Earthquake Eng Struct Dynam 2011;40:86786.
[27] DAmbrisi A, Stefano M, Tanganelli M. Use of pushover analysis for predicting
seismic response of irregular buildings: a case study. J Earthquake Eng
2009;13:1089100.
[28] Reyes JC, Chopra AK. Three-dimensional modal pushover analysis of buildings
subjected to two components of ground motion, including its evaluation for
tall buildings. Earthquake Eng Struct Dynam 2011;40:789806.
[29] Shakeri K, Tarbali K, Mohebbi M. An adaptive modal pushover procedure for
asymmetric-plan buildings. Eng Struct 2012;36:16072.
[30] Koutromanos I, Stavridis A, Shing PB, Willam K. Numerical modeling of
masonry-inlled RC frames subjected to seismic loads. Comput Struct
2011;89(1112):102637.
[31] Limkatanyu S, Spacone E. Effects of reinforcement slippage on the nonlinear
response under cyclic loadings of RC frame structures. Earthquake Eng Struct
Dynam 2003;32:240724.
[32] DAmbrisi A, Filippou FC. Correlation studies on an RC frame shaking-table
specimen. Earthquake Eng Struct Dynam 1997;26:102140.
[33] Di Sarno L, Manfredi G. Experimental tests on full-scale RC unretrotted frame
and retrotted with buckling restrained braces. Earthquake Eng Struct Dynam
2012;41(2):31533.
[34] Di Sarno L, Yenidogan C, Erdik M. Field evidence and numerical investigation of
the Mw = 7.1 October 23 Van, Tabanli and the Mw>5.7 November Earthquakes
of 2011. Bull Earthq Eng 2013. http://dx.doi.org/10.1007/s10518-012-9417-0.
[35] Sharma A, Reddy GR, Vaze KK, Eligehausen R. Pushover experiment and
analysis of a full scale non-seismically detailed RC structure. Eng Struct
2013;46:21833.

23

[36] Engineering Institute of Thailand. Standard for reinforced concrete building

(strength design method). EIT Standard 100838, Bangkok, Thailand; 2000.
[37] Carr
AJ.
RUAUMOKO
computer
program.
Christchurch,
New
Zealand: University of Canterbury; 2006.
[38] Sharpe RD. The seismic response of inelastic structures. Ph.D. Thesis.
Department of Civil Engineering, University of Canterbury, New Zealand;
1974.
[39] Otani S. SAKE, A computer program for inelastic response of RC frames to
earthquakes. Report UILU-Eng-74-2029: Civil Engineering Studies, University
of Illinois at Urbana-Champaign; 1974.
[40] Sezen H, Chowdhury T. Hysteretic model for reinforced concrete columns
including the effect of shear and axial load failure. J Struct Eng ASCE
2009;135(2):13946.
[41] Warnitchai P, Chintanapakdi C, Panyakapo P, Bunyapinyo V, Leelatavivat S,
Pimanmas A. Seismic evaluation and retrotting of existing buildings in
Thailand: Thailand Research Fund Report, Bangkok, Thailand; 2011.
[42] ATC-24. Guidelines for cyclic testing of components of steel structures.
Applied Technology Council, Redwood City, CA, 11; 1992.
[43] ISO. Timber structures-joints made with mechanical fasteners Quasi-static
reversed-cyclic test method. Draft standard, ISO-TC165 timber structures
working group 7; February 10, 1998.
[44] Krawinkler H. Loading histories for cyclic tests in support of performance
assessment of structural components. In: 3rd International conference on
advances in experimental structural engineering, San Francisco; October 15
16, 2009.
[45] Porter ML. Sequential phased displacement (SPD) procedure for TCCMAR
testing. In: Proceedings the 3rd meeting of the joint technical coordinating
committee on masonry research, US-Japan Coordinated Research, Program;
1987.
[46] SPT 1302. Standard of earthquake resistant design of building. Department of
Public Works and Town & Country Planning, Bangkok, Thailand; 2009.
[47] Uang CM, Bertero VV. Implication of recorded earthquake ground motions on
seismic design of building structures. UCB/EERC Report No. 88/13, University
of California, Berkeley, California; 1988.
[48] Kurama YC, Farrow KT. Ground motion scaling methods for different site
conditions and structure characteristics. Earthquake Eng Struct Dynam
2003;32:242550.
[49] Spyracos CC, Maniatakis CA, Taambas J. Evaluation of near-source seismic
records based on damage potential parameters Case study: Greece. Soil Dynam
Earthquake Eng 2008;28:73853.
[50] Climent AB, Almansa FL, Gonza DAB. Design energy input spectra for
moderate-to-high seismicity regions based on Colombian earthquakes. Soil
Dynam Earthquake Eng 2010;30:112948.
[51] Hachem M. BISPEC user manual. Berkeley (CA): University of California; 1999.