Journal of Constructional Steel Research 72 (2012) 155167

Contents lists available at SciVerse ScienceDirect

Journal of Constructional Steel Research

Assessment of modied consecutive modal pushover analysis for estimating the

seismic demands of tall buildings with dual system considering steel concentrically
braced frames
Faramarz Khoshnoudian , M. Mehdi B. Kashani
Faculty of Civil Engineering, AmirKabir University of Technology, Hafez St., Tehran, Iran

a r t i c l e

i n f o

Article history:
Received 22 February 2011
Accepted 3 December 2011
Available online 11 January 2012
Keywords:
Modied consecutive modal pushover
(MCMP) procedure
CMP procedure
Tall buildings
Higher-mode effects
Seismic demands

a b s t r a c t
According to the previous researches, conventional nonlinear static procedure (NSP), which is limited to single mode response, cannot predict the seismic demands of tall buildings with reliable accuracy. To estimate
the seismic demands in upper stories for tall buildings the effects of higher modes should be included. In the
recent years, developing traditional pushover analysis to consider the effects of higher modes conducted researchers to propose several methods, such as N2, MPA and MMPA procedures, that have a specic approach
to estimate seismic demands of structures but the accuracy of them is doubtable for estimating of hinge plastic rotations. Recently consecutive modal pushover (CMP) procedure was proposed to consider the effects of
higher modes with acceptable accuracy especially in prediction of hinge plastic rotations. The CMP procedure
was limited to include two or three modes, and use of higher modes might cause some inaccuracy at results
of upper stories. In CMP procedure, estimation of modal participating factors is important and choosing inadequate modes may cause large errors. In this paper some changes have been applied to the CMP procedure to
improve accuracy of the results and the modied method is proposed and named modied consecutive
modal pushover (MCMP) procedure. In this modied method the contribution of mode is used of effective
modal participating mass ratio. The comparison of MCMP procedure to exact values derived by nonlinear response history analysis (NL-RHA) demonstrated the reliable predictions and it can overcome the limitations
of traditional pushover analysis.
2011 Elsevier Ltd. All rights reserved.

1. Introduction
Nonlinear static analysis, or pushover analysis, has been developed
over the past decades and has become the common procedure for building evaluation and design verication; however the procedure involves
certain approximations and simplications. According to literature, the
accuracy of pushover analysis in predicting seismic demands has been
the controversial discussion, so the proposed approaches are leading researchers to achieve more accurate and reliable method; however nonlinear static procedures suffer a lot of limitations especially for high-rise
buildings. The invariant load pattern is one of the most signicant limitations of traditional methods, because the actual inertia force distribution changes continuously during seismic events due to higher mode
contribution and structural degradation, which modies the stiffness
of individual structural elements and, consequently of the structure as
a whole [1]. Therefore the effects of higher modes should be considered
for estimating seismic demands of tall buildings.

Corresponding author.
E-mail address: khoshnud@aut.ac.ir (F. Khoshnoudian).
0143-974X/$ see front matter 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jcsr.2011.12.002

Generally, using modal properties of the structure in nonlinear static

analysis is most accessible approach to take into account the dynamic
characteristics of the system. MMP (Multi Mode Pushover procedure)
[2] was proposed to involve higher modes effects in pushover analysis.
MMP provides better estimation of seismic demands comparing to traditional pushover methods based on load pattern using rst mode.
Although higher modes have being participated in MMP analysis, responses estimation and distribution of them over height of the structure
were inadequate. More recently, PCR (Pushover Results Combination)
[3] was proposed to consider the effects of higher modes. In this method
several load patterns using mode shapes should be applied and then
nal responses would be determined as a weighted (using modal participation factors) summation of the results from each pushover analysis. MPA (Modal Pushover Analysis) [4] was developed to consider
higher modes effects by analyze each mode as an equivalent singledegree-of-freedom (SDOF) system including nonlinear properties related
to that mode. The MPA procedure was able to predict seismic demands
with reliable errors at the level of displacements, but in order to calculate hinge plastic rotations MPA procedure had underestimate predictions. Then a modied version of the MPA (MMPA) [5] was proposed.
The seismic demands of the structure were obtained by combining
the inelastic response of rst-mode pushover analysis with the elastic

156

F. Khoshnoudian, M.M.B. Kashani / Journal of Constructional Steel Research 72 (2012) 155167

Notations
[c]
Damping matrix of structure
[k]
Lateral stiffness matrix of structure
[m]
Diagonal mass matrix of structure
{i}
Unit vector
{Peff(t)} The effective earthquake forces
{s}
Spatial distribution of effective forces
{sn}
Modal inertia force distribution of the nth mode
{sn*}
Incremental lateral force distribution for the nth stage
of multi-stage analysis
{u(t)}
Displacement of NDOF system
{n}
nth mode shape of the structure
mi
Lumped mass of ith oor
M*
Total mass of the structure
Mn*
Effective modal mass of the nth mode
qn(t)
Modal co-ordinate
r
the peak response of the structure in the CMP
procedure
g(t)
Acceleration of ground motion
n
Effective modal participating mass ratio of the nth
mode
t
Target displacement of the roof
n
Natural frequency of the nth mode
n
Damping ratio of nth mode
n
Modal participating factor of the nth mode

solution. The CMP procedure includes two types of pushover analysis;

single-stage and multi-stage pushover analyses. Earlier the contribution
of each mode in multi-stage analysis of the CMP procedure was derived
from effective modal participating mass ratio of that mode.
Afterwards, the accuracy of the CMP procedure has been veried for
another lateral force-resisting system, dual system with steel concentrically braced frames and in addition for improving the CMP procedure to
estimate seismic demands of tall buildings without need of the fundamental period of structure for performing two-stage or three-stage
pushover analysis, consequently authors suggest modied consecutive
modal pushover analysis (MCMP) as it will be explained in details in
the following.
2. Basic theoretical concepts
Modal analysis principles are used in the proposed procedure;
therefore it would be useful to understand the denition of these
principles. Eq. (1) shows the differential equation governing the response of a multi-degree-of-freedom (MDOF) system under earthquake ground motion [10]:
g cfu_ g kfug mfig:u
g t
mfu

where [m], [c] and [k] are diagonal matrices of mass, damping and
stiffness of structure respectively and {i} is unit vector. Right hand
side of previous equation represents the effective earthquake forces,
{Peff(t)}, and can be written as:
n
o
g t fsgu
g t :
P eff t mfig:u

response of higher modes. Upper-bound pushover analysis [6] was another method which was able to overcome the invariant load pattern
limitation. Incremental response spectrum analysis (IRSA) [7] was developed by other researchers in which whenever a new plastic hinge
occurs, elastic modal spectrum analysis was executed. More recently,
an adaptive modal combination (AMC) [8] procedure was proposed,
in which the applied lateral forces are updated in accordance with the
changes in the dynamic characteristics during inelastic analysis, for
each mode.
Although nonlinear static analysis is basically developed for structures with the predominant rst vibration mode, many nonlinear static
approaches, such as MPA [4], MMP [2], PRC[4] are proposed to consider
higher elastic modes as lateral load pattern to take into account higher
modes effects. The accuracy of the mentioned approaches was demonstrated by comparing to nonlinear dynamic analysis as an exact solution
and it was conrmed the possibility to use nonlinear static analysis for
predicting seismic demands of tall buildings which are inuenced by
higher modes effects. However there are some errors in these approaches, the simplicity and time-consuming of these methods lead
proting them. In addition, the aim of new pushover analysis is to reduce these errors. Beside in this paper, the new proposed pushover procedure is more reliable than the previous pushover analysis.
To improve pushover analysis, Consecutive Modal Pushover (CMP)
procedure was proposed [9], in which the incremental forces were applied to structure continuously. This method was examined on buildings with moment-resistant frame system. The achievement of this
method was demonstrated by having better estimation of hinge plastic
rotations. The distributions of Lateral forces in Consecutive Modal Pushover (CMP) procedure are calculated by using mode shapes which are
obtained from Eigen-analysis of linearly elastic structure. Using elastic
modal properties are suggested for most the nonlinear static procedures
merely because of simplicity and reduction of time-consuming of analysis. The reason of how higher modes are going to be represented by
elastic modes after deterioration can be explained by comparison of
the results obtained from the proposed nonlinear static procedure to
those obtained from nonlinear time history analysis as an exact

Vector of {s} is the distribution of effective earthquake forces over

building's height and can be expanded as a summation of the modal
inertia force distributions, sn, as follows:

fsg mfig

N
X

fsn g

n1

N
X

 
n m n

n1

where, n is the nth modal participation factor and {n} is the corresponding mode shape. The displacement vector of a N degree of
freedom system is dened as:
fut g

N 
X

n qn t

n1

where qn(t) is the modal coordinates and the following equation

will be obtained as follows:
2
g t :
q n 2 n n q_ n n qn n u

qn(t) presents a property of a multi degrees of freedom system, which

is dened by an elastic single degree of freedom system. n and n are
natural frequency and damping ratio of nth mode, respectively. n is
obtained as follows:

n 

T
n mfig
T
 :
n m n

To solve Eq. (5) the substitution Dn(t) instead ofqn(t) could be useful
and the relation between them is described as follows:
qn t n Dn t

F. Khoshnoudian, M.M.B. Kashani / Journal of Constructional Steel Research 72 (2012) 155167

where Dn(t) is governed by the equation of motion for a singledegree-of-freedom system subjected to g(t):
2 D_ 2 D u
g t :
D
n n
n
n n n

Displacement vector for the elastic system is indicated as follows:

fut g

N
X

 
 
 
n n Dn t 1 D1 t 1 2 D2 t 2
n1
 
3 D3 t 3 ::::

It is notable that each mode shape in the CMP procedure is normalized relative to its roof component, rn. Therefore, total displacement of
roof is determined as follows:
ur t 1 D1 t 2 D2 t 3 D3 t ::::

11

Also there are some parameters which are used in the procedure
and they are dened in Eqs. (12) to (15):


Mn Ln n
9
n

Response history analysis (RHA) and response spectrum analysis

(RSA) could be used to estimate the peak response of the system, ro.
However, in the case of usage in pushover analysis, the simplicity of
RSA leads us to use it instead of RHA. Generally, in modal response
spectrum analysis, the peak response of the corresponding nth
mode is determined by Eq. (10), and total response of the system is
estimated by some combination rules such as SRSS and CQC methods. In
RSA the value of Dnocould be derived by using standard or design spectra.
r no

157

st
r n An

10

12

Mn
M

13

where;
 
Ln n mfig

14

M mj

15

where M* is the total mass of the structure obtained by summation of the lumped masses, mj, over all oor levels.
3. Modied consecutive modal pushover (MCMP)

Where rno and rnst are the peak response and equivalent static response of nth mode, respectively, and An is derived from pseudoacceleration response (or design) spectrum which is determined by
Tn and n.

Consecutive Modal Pushover (CMP) procedure was proposed to estimate the seismic demands of structures. The process in the CMP analysis was, so that loads in each stage were applied to structure at the

a) configuration of structure

b) frame position

Fig. 1. Conguration of structure and frame position (N: Number of stories).

Table 1
General specication of structures.
Periods (s)
T3

T2

T1

0.126
0.179
0.233
0.361

0.241
0.374
0.468
0.794

0.835
1.452
1.809
3.119

Distributed
live load
(kg-f/m)

Distributed
dead load
(kg-f/m)

Lumped
masses
(kg-f
s2/m)

b
(m)

h
(m)

No. of
stories

Structure
type

1000
1000
1000
1000

3250
3250
3250
3250

5538
5636
5693
5794

15
15
15
15

32
48
64
96

10
15
20
30

F1
F2
F3
F4

158

F. Khoshnoudian, M.M.B. Kashani / Journal of Constructional Steel Research 72 (2012) 155167

Table 2
Used section of 10-story structure.
Frame
type

Levels

Column type

F1

1
2
3 and 4
5
6
7 to 10

Interior col.

Exterior col.

C4
C4
C5
C5
C5
C6

C4
C5
C5
C5
C6
C6

Beam
type

Brace
type

B5
B5
B5
B6
B6
B6

Br7
Br7
Br9
Br9
Br9
Br9

Table 3
Geometric properties of columns.
t (cm)

d (cm)

Column type

2.5
2
1.5

35
30
25

C4
C5
C6

Table 4
Geometric properties of beams.

deformed shape state resultant of previous stage and initial condition

(stress and deformation) of each stage was the same as the state at
the last step of analysis in the previous stage. In the CMP procedure,
the value of roof displacement increment, uri, at each stage of the
multi-stage analysis was the factor of target displacement. This factor was determined based on the effective modal participating
mass ratio. According to previous research on utilizing CMP procedure for steel moment-resisting frames, the procedure had acceptable accuracy in estimating hinge plastic rotations. The CMP procedure
consists of single-stage and multi-stage pushover analyses. First
apply the gravity loads and then perform these displacementcontrol pushover analyses using consecutive pushover analysis;
however the number of stages depends on fundamental period of
structure [9].
In this paper authors propose to use the modal response analysis,
to estimate contribution of each mode in the CMP procedure. Accordingly, numbers of modes that are participating in the CMP analysis are
not limited to two or three modes any more. Theoretically, using
greater number of modes in the modal response analysis should
cause reduction of structural response errors. So here it is proposed
to use of all modes which the sum of their effective modal masses
should be at least 90% of the total structure mass.
Ns
X

tw (cm)

tf (cm)

bf (cm)

h (cm)

Beam type

0.8
0.8

1.5
1.5

20
20

32.5
30

B5
B6

Table 5
Geometric properties of braces.
a (cm)

Section

Brace type

2
2

2UNP140
2UNP100

Br7
Br9

Mi 0:9M

16

i1

where Ns is the numbers of stages (or numbers of modes) that they

should be used in the process of analysis. Some of the principles of pushover analysis are used in consecutive modal pushover analysis procedure.
In this approach at least two nonlinear static analyses with modal load
distribution are performed. Load distributions in modied consecutive
modal pushover are based on modal shapes derived from Eigenanalysis of the linearly elastic structure. It is noteworthy to imply that
changes in the modal properties of the structure are ignored when the
structure experiences nonlinear yielding under increasing lateral loads
during pushover analysis [9]. Roof displacement increment, uri, for each
stage of the multi-stage pushover analysis is the factor of target displacement (Eq. (17)). The coefcient value is calculated based on the modal
response spectrum analysis for elastic system. The displacement increment, uri, at the roof in the ith stage of multi-stage pushover analysis, is
therefore calculated as:
uri i t

17

where;
i

i Di
Ns
P
n Dn

18

n1

Fig. 2. Generalized load-deformation curve for hinges [13].

where t is target displacement and i is the contribution of ith

mode. Several approaches can be used to estimate the total target displacement at the roof level. This displacement can be determined by
using the capacity spectrum method [11], the displacement coefcient
approach [12,13], the N2 method [14,15], or dynamic analysis of the

Table 6
Used ground motions.
PGA (g)

Component (deg)

Station number

Station name

Magnitude

Date

E.Q. name

No.

0.134
0.239
0.147
0.621
0.196
0.204
0.174

E
90
270
45
271
315
180

1061
24,157
1498
6604
1028
5051
135

Lamont
LABaldwin Hills
Rio Dell Overpass, FF
Cerro Prieto
Hollister City Hall
Parachute Test Site
LA Hollywood Stor Lot

Ms(7.3)
Ms(6.7)
Ms(7.2)
Ms(6.4)
Ml()
Ms(6.9)
Ms(6.6)

1999/11/12
1994/01/17
1980/11/08
1980/06/09
1961/04/09
1979/10/15
1971/09/02

Duzce, Turkey
Northridge
Trinidad, California
Victoria, Mexico
Hollister
Imperial Valley
San Fernando

1
2
3
4
5
6
7

F. Khoshnoudian, M.M.B. Kashani / Journal of Constructional Steel Research 72 (2012) 155167

159

1.4
1.2

Duzce

Psudo-Accleration(A/g)

Northridge
Trinidad

Victoria
San Fernando

0.8

Imprial Vally
Hollister

0.6

Average Spec.

0.4
0.2
0
0

0.5

1.5

2.5

3.5

Natural Vibration Period,T(Sec)

Fig. 3. Pseudo-acceleration spectra.

structure [16,3,17,18]. i is derived from Eq. (6) and Di is the peak response of a single-degree-of-freedom (SDOF) system equivalent to ith
mode which can be calculated by Eq. (8) and also can be estimated by
standard (or design) spectra, based on the modal response spectrum
analysis.

In addition to multi-stage analysis also there is single-stage pushover analysis which is proposed to be performed by uniform distribution for the dual systems.
The details of the MCMP procedure for a single plane frame are
expressed as a sequence of the following steps:

a) Story Drift Ratios of the 10-Story structure

0.7%

Single-Stage
Multi-Stage
NLTH

0.6%

Drift ratio

0.5%
0.4%
0.3%
0.2%
0.1%
0.0%

10

Floor

b) Story Drift Ratio of the 15-story structure

0.9%
Single-Stage

0.8%

Multi-Stage

Drift ratio

0.7%

NLTH

0.6%
0.5%
0.4%
0.3%
0.2%
0.1%
0.0%

10

11

12

13

14

15

Floor
Fig. 4. Peak values of story drift ratios derived from pushover analysis used in the MCMP procedure and from NLTH for the 10 and 15-story structures.

160

F. Khoshnoudian, M.M.B. Kashani / Journal of Constructional Steel Research 72 (2012) 155167

applied till displacement increment of each stage reaches the value

of urn = nt in which n is calculated by Eq. (18).As it was demonstrated previously, initial conditions of each stage are the conditions
of last step of previous stage. It is notable that the nonlinear behavior
of the structure depends on the loading path, and separation of the
loading input and the structural response is far from real behavior
[19]. Consequently, modal pushover analysis must be carried out consecutively in the order of modes, from the rst to the higher ones.
6. Calculating the peak values of desired responses such as, displacement, drift ratios, plastic hinge rotations from last step (step 5). Responses of single- and multi-stage analyses, have been shown by rS
and rM, respectively.
7. Calculate the envelope, r, of the peak responses as follows:

1. Calculation of natural frequencies and modes shapes. These properties are determined by Eigen-analysis of the linearly elastic
structure. The mode-shapes should be normalized with roof component so that the roof component of {n} equals unity.
2. Compute {sn*} = [m]{n}, where {sn*} shows the distribution of incremental lateral forces over the height of the structure for the
nth stage of multi-stage pushover analysis.
3. Compute the total target displacement of the structure at the roof,
t.
4. Calculating coefcients for each mode based on the Eq. (18).
5. As it was mentioned before, MCMP procedure is included singlestage and multi-stage pushover analysis. First of all, gravity should
be applied to the structure and then according to the steps listed
below displacement-controlled pushover analysis should be performed. It is noted that the vertical loads are applied to the buildings according to the FEMA-356 regulations:
5.1 Single-stage analysis, with uniform distribution over height of
building, until the control node, gradually reaches the total
target displacement.
5.2 Multi-stages analysis, that the number of stages (Ns) is equal
to number of rst modes which summation of their effective
modal mass (Mn) reaches at least 90% of total mass (M*).

r Maxfr S ; r M g:

This shows that the seismic demands of the inelastic structure is

obtained by enveloping the peak responses in this procedure.
4. System and excitation considered
To evaluate MCMP procedure, the dual system considering steel
concentrically braced frames has been used. The conguration and
braces distribution over height of frame are illustrated in Fig. 1. Also

In this step, lateral force distribution in nth stage is {sn*} that is derived from {sn*} = [m]{n}. These incremental lateral forces are

a) Story Drift Ratio of the 20-story structure

0.8%
0.7%

Single-Stage
Multi-Stage
NLTH

Drift ratio

0.6%
0.5%
0.4%
0.3%
0.2%
0.1%
0.0%

10 11 12 13 14 15 16 17 18 19 20

Floor

b) Story Drift Ratio of 30-story structure

0.9%
0.8%
0.7%

Single-Stage
Multi-Stage
NLTH

Drift ratio

0.6%
0.5%
0.4%
0.3%
0.2%
0.1%
0.0%

19

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30

Floor
Fig. 5. Peak values of story drift ratios derived from pushover analysis used in the MCMP procedure and from NLTH for the 20 and 30-story structures.

F. Khoshnoudian, M.M.B. Kashani / Journal of Constructional Steel Research 72 (2012) 155167

the position of frame in whole building is illustrated in Fig. 1. The considered structures are three-bay frames with four different heights of
10, 15, 20, and 30 stories, covering a wide range of fundamental periods. All the frames have 5 m bays. A story height of 3.2 m was assumed throughout. Specications such as dimensions and natural
vibration periods are listed in Table 1. Assumption of rigid diaphragm
has been used for all frames, and lumped masses have been applied in
center of mass at each level. As an example, used sections of 10-story
building and their geometric properties of columns, beams and braces
are listed in Tables 2, 3, 4 and 5 respectively.
Allowable stress design has been used in order to design structures
[20]. The dead and live loads have been assumed to be 650 and 200 kgf/m2, respectively and loading width of beams was assumed to be equal
to 5 m. The dead load (total dead load) and 20% of live load has been considered to calculate the mass of each oor. The plastic hinge properties
were determined according to FEMA-356 [13] to consider the nonlinear
behavior of structure (Fig. 2) which are located at each end of members.
The post-yield slope (BC) was assumed to be 3% of the elastic slope.
P- (second-order) effects have been included, but the panel zone
size, strength, and deformation have been neglected. The effect of p-,
as the geometric nonlinear behavior of the members has been considered
which is effective to determine the member stiffness matrix and then followed by the stiffness matrix of whole structure.
To evaluate the MCMP procedure, nonlinear time history analysis
has been performed. Seven ground motions have been selected from

161

the strong ground motion database of the Pacic Earthquake Engineering Research (PEER) Centre (http://peer.berkeley.edu). The distance to
the fault (more than 12 km) and the soil at the site have been the criteria to select these records. The ground motion records are selected
to be far eld records; the soil site corresponds to NEHPR class C and
also covers large frequency contents of earthquakes. The seismic effects
were determined in accordance with the requirements of the Iranian
code of practice for the seismic-resistant design of buildings [21]. To ensure that the structures respond into the inelastic range when subjected
to ground motions, the records were scaled up to 0.7 g. More characteristics of the used ground motion records are given in Table 6. Pseudoacceleration spectrum with the corresponding the mean spectrum, are
presented, for 5% damping ratio, in Fig. 3. The mean spectra are shown
by a thicker line.
5. Types of analysis
To evaluate the MCMP procedure, Non-Linear Time History (NLTH)
analysis has been used to achieve the exact value of responses. The
mean value of the responses has been determined over the set of used
ground motions. Also the MCMP procedure has been compared to
MPA (Modal Pushover Analysis) procedure as a well-known nonlinear
static procedure. P- effects have been included in the MCMP and
MPA procedures for all modes. In the MCMP procedure, the target displacement at the roof has been derived by averaging the maximum

a) Story Drift Ratios of the 10-Story structure

Floor

10
9

NLTH

MPA

MCMP

6
5
4
3
2
1
0
0.000

0.001

0.002

0.003

0.004

0.005

0.006

0.007

Drift Ratio

Floor

b) Story Drift Ratio of the 15-story structure

15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
0
0.0%

NLTH
MCMP
MPA

0.1%

0.2%

0.3%

0.4%

0.5%

0.6%

0.7%

0.8%

Drift ratio
Fig. 6. Height-wise variation of the story drifts for the 10 and 15-story structures.

0.9%

162

F. Khoshnoudian, M.M.B. Kashani / Journal of Constructional Steel Research 72 (2012) 155167

These mentioned gures show which pushover analysis, single-stage

or multi-stage analysis, is dominated in the seismic demands of tall
buildings. Story drift ratios of all four Structures have been shown in
Figs. 6 and 7.
As it was mentioned previously single- and multi-stage analyses
are important to predict seismic demands of lower and upper stories,
respectively. Therefore, it seems better to separate stories of each
structure to two parts to compare the results of MCMP and MPA
with NLTH analysis. This separation is caused by experienced behaviors of structures in respect to how incremental forces are applied.
Both the procedures (MPA and MCMP) are predicting drift ratios
with sufcient accuracy. However, MPA procedure is more accurate
than MCMP especially in 10-story structure. Drift ratios over upper
oors, where multi-stage analysis is governing, estimated by MCMP
procedure are as accurate as MPA procedure; however, in 10-story
building results are overestimated using MCMP procedure because
of system characteristics behave rigidly.
The errors in story drift ratios are illustrated in Figs. 8 and 9. As
shown in these gures the MCMP procedure provides better estimations than the MPA procedure at some higher oors, whereas the errors from the MPA procedure are less than those obtained from the
MCMP procedure at some lower oors. Maximum error for MCMP
procedure happens in lower stories which are obtained and dominated by single-stage analysis and choosing adequate load pattern
for single-stage could omit these errors over lower stories. The errors
for MPA procedure are 9.65, 4.37, 32.29 and 4.23% for 10-, 15-, 20- and
30-story buildings in the same order, and also for MCMP procedure

values of roof displacements resulting from the NLTH (Non-Linear Time

History) analysis for the selected ground motions. The target displacements that have been calculated are equal to values 13.36, 24.13,
30.57 and 46.02 cm for structures F1, F2, F3, and F4, in the same order
according to Table 1. The nonlinear time history analysis has been performed using the Wilson- time integration method, in which the stability and accuracy characteristics are determined by the parameter . This
parameter has been assumed a value of 1.4. A damping ratio of 5% has
been considered for the rst and the mode which summation of its effective modal masses and lower modes reaches at least 90% of total
mass of the structure, in order to dene the Rayleigh damping matrix.
It is noted that the nonlinear version of the computer program
SAP2000 [22] was used to perform these analysis.
6. Discussion of the results
As it was described previously, the MCMP procedure uses two types
of pushover analysis. The peak values of story drift ratios for 10, 15, 20
and 30 stories structure outcome from single- and multi-stage analysis
and also nonlinear time history analysis has been illustrated in Figs. 4
and 5, separately. Generally, according to the obtained results, it could
be concluded that single-stage analysis and multi-stage analysis cover
the seismic demands of lower and higher stories, respectively. This conclusion is acceptable for the all seismic demands which have been considered in this study, except for displacement results single-stage
analysis has been governed for all stories. Actually, it was expected
that using higher modes strongly affects on upper stories results.

a) Story Drift Ratio of the 20-story structure

20

Floor

18

NLTH

16

MCMP

14

MPA(3 Modes)

12
10
8
6
4
2
0
0.0%

0.1%

0.2%

0.3%

0.4%

0.5%

0.6%

0.7%

0.8%

Drift ratio

b) Story Drift Ratio of 30-story structure

30
27
24
21

Floor

18
15
12
9
NLTH
MCMP
MPA

6
3
0
0.0%

0.1%

0.2%

0.3%

0.4%

0.5%

0.6%

0.7%

0.8%

Drift ratio
Fig. 7. Height-wise variation of the story drifts for the 20 and 30-story structures.

0.9%

F. Khoshnoudian, M.M.B. Kashani / Journal of Constructional Steel Research 72 (2012) 155167

163

a) Story Drift Ratios of the 10-Story structure

10
9
8
7

Floor

6
5
4
3
MCMP

MPA

1
-10%

-5%

0
0%

5%

10%

15%

20%

Error

b) Story Drift Ratio of the 15-story structure

Floor

MCMP
MPA

-40%

-35%

-30%

-25%

-20%

-15%

-10%

-5%

15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
0
0%

5%

10%

Error
Fig. 8. Errors in the story drifts for the 10 and 15-story structures.

are 26.47%, 14.24%, 36.94% and 16.97%. It is notable that 15-story building has large displacement and deformation at rst two stories that
caused by some used records in NLTH analysis. These deformations cannot be predicted by such nonlinear static analysis procedures.
Therefore the accuracy of MCMP procedure as well as MPA procedure in predicting story drift ratios of studied buildings is acceptable.
Hinge plastic rotations of structures are obtained by considering the
maximum plastic rotations of middle spans. The value of hinge plastic
rotation related to each oor is obtained from software. The results of
hinge plastic rotations and their errors are shown in Figs. 10 to 13.
Obviously, MCMP procedure has better accuracy in prediction of
hinges plastic rotations and however, the error of MPA procedure is
signicant. This result is not only valid for the medium-rise structures
(10 and 15 stories structures) but also the results for high-rise structures (20 and 30 stories structures) are more accurate. As shown in
these gures MCMP procedure has more accurate results in upper
oors. However, the small value of hinge plastic rotations obtained
from NLTH analysis in lower oors cause 100% errors in MCMP and
MPA procedures. In this case, at lower oors, MCMP even has better
prediction in comparing to MPA procedure. As it was demonstrated previously, effects of higher modes at higher oors have a signicant role to
achieve more accurate results. The main reason of this achievement
which could also be considered as an advantage of MCMP procedure
is the manner how incremental forces are applied. Applying incremental forces consecutively in MCMP procedure takes into account the effects of higher modes and nonlinearity simultaneously, whereas in
MPA procedure these higher mode effects are not considered and

usually the nonlinearity of the system is considered just for the rst
mode. The manner of applying incremental lateral forces in the multistage pushover analysis of the MCMP procedure, in comparison with
the MPA procedure, has signicant effects in prediction of hinge plastic
rotations at the middle and upper oors. However, the pushover procedure suffers from the limitation that it is unable to take into account the
cumulative rotation of hinges due to cyclic hysteretic behavior [19].
The errors of hinge plastic rotations for MCMP procedure are 70.84%,
39.22%, 23.34% and 72.81% for 10-, 15-, 20- and 30-story buildings in the
same order, while in MPA procedure these errors are 87.49%, 100%,
100% and 100%. These results conrmed the accuracy of MCMP procedure in comparison with MPA procedure.
In addition to the above discussion for elaborating the explanation
about the hinge plastic rotation errors obtained from the MCMP procedure, some details are required. Regarding to Fig. 10, for 10-story
building in 3th oor and 10th oor, these errors are signicant, because
the exact values obtained from nonlinear time history analysis are
so small for the mentioned oors. The same remarkable errors are
extracted from Fig. 11 for 15-story building from 3th oor to 6th oor
where exact hinge plastic rotations are small and consequently cause
enormous errors. 4th oor and 20th oor of 20-story building (Fig. 12)
demonstrated that the error of the MCMP procedure is 100%. Because in
these oors the exact value of hinge plastic rotations is almost zero,
therefore the errors in these oors reach to 100%. For 30-story high rise
building, the errors of the proposed procedure in 6th to 9th oors are
signicant due to small value of hinge plastic rotations obtained from
nonlinear time history analysis as an exact solution.

164

F. Khoshnoudian, M.M.B. Kashani / Journal of Constructional Steel Research 72 (2012) 155167

a) Story Drift Ratio of the 20-story structure

20
18
MCMP
MPA

16
14

Floor

12
10
8
6
4
2

-16%

-14%

-12%

-10%

-8%

-6%

-4%

-2%

0
0%

2%

4%

6%

Error

b) Story Drift Ratio of 30-story structure

30
27
24

MCMP
MPA

21

Floor

18
15
12
9
6
3
0

-30%

-25%

-20%

-15%

-10%

-5%

0%

5%

10%

Error
Fig. 9. Errors in the story drifts for the 20 and 30-story structures.

7. Conclusions
The article focus on the assessment of modied version of consecutive
modal pushover analysis procedure for estimating the seismic demands
of tall building with dual system considering steel concentrically braced
frames. In the MCMP procedure, the seismic responses are evaluated by
enveloping the peak responses obtained from the multi-stage and
single-stage pushover analyses. Linearly-elastic modal properties are
used in the multi-stage pushover analysis. The force distribution over
the height of the building in each stage of the multi-stage pushover analysis is determined by mass matrix and relevant elastic mode shape. The
lateral forces are incrementally applied during the stages of the multistage pushover analysis. For this purpose, the MCMP procedure applied
to structure and then maximum responses from single- and multi-stage
pushover are obtained and at the end the envelope of the peak responses
is obtained according to the mentioned details.
The MCMP procedure utilizes the basis of CMP procedure and
uses single- and multi-stage pushover analyses. The contribution of
modes to calculate incremental displacements of each mode is the
difference between the CMP and MCMP procedures. Actually, in the
MCMP procedure, unlike to CMP procedure, more than two or three
modes could be considered. It does not need to know the fundamental
period of structure.
The accuracy of the MCMP procedure was evaluated, on the seismic
response of tall buildings considering dual system with steel concentrically braced frames in this research and some changes were made in the
CMP procedure to gain more accurate and reliable results. The

advantage of the MCMP procedure is the manner of applying lateral

forces which are applied to structure consecutively in order of modes
number.
As it was mentioned previously, single- and multi-stage pushover analyses inuence over lower and upper story responses, respectively. Indeed, in single-stage pushover analysis the accuracy
of results extremely depends on applied load pattern and in the
MCMP procedure choosing adequate load pattern for single-stage
analysis is very vital. In this investigation, uniform load pattern has
been selected to utilize in the single-stage pushover analysis.
According to remarkable effects of higher modes over upper stories,
the manner of calculating modes contribution is as important as
choosing adequate load pattern.
According to performed researches by authors, choosing effective
modal participating mass ratios to calculate mode contributions in
multi-stage analysis have no accurate results. Using of the basis of Response Spectrum Analysis (RSA) is proposed to estimate the mode contributions. In this case, the procedure has more reliable results especially
for upper stories.
The MCMP procedure has a reliable prediction of seismic demands
on upper oors; especially in case of hinge plastic rotations the MCMP
procedure has better estimation in comparing to MPA procedure over
all stories. However, in order to estimate displacements and drift ratios, the results of the MPA procedure for lower stories are more accurate than those obtained from the MCMP procedure. In addition, the
assessment of the both procedure in evaluating of displacements
and drift ratios for upper stories in comparing to nonlinear time

F. Khoshnoudian, M.M.B. Kashani / Journal of Constructional Steel Research 72 (2012) 155167

a) Hinge Plastic Rotaions

10
9
8
7

Floor

6
5
4
3

NLTH
MCMP
MPA(3 Modes)

2
1
0
0.000

0.001

0.002

0.003

0.004

0.005

0.006

0.007

Hinge Plastic Rotaion (Rad)

b) Errors of Hinge Plastic Rotaions

10
9
8
7

Floor

6
5
4
3
MCMP

2
MPA(3 Modes)

1
-100%

-80%

-60%

-40%

0
0%

-20%

20%

40%

60%

80%

Error
Fig. 10. Height-wise variation of the hinge plastic rotations and their errors for the 10-story structure.

Floor

a) Hinge Plastic Rotations

15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
0
0.000

NLTH
MCMP
MPA(3 Modes)

0.001

0.002

0.003

0.004

0.005

0.006

0.007

0.008

0.009

0.010

Hinge Plastic Rotation (Rad)

b) Errors of Hinge Plastic Rotations

MCMP

Floor

MPA(3 Modes)

-105%

-85%

-65%

-45%

-25%

15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
0
-5%

Error
Fig. 11. Height-wise variation of the hinge plastic rotations and their errors for the 15-story structure.

165

F. Khoshnoudian, M.M.B. Kashani / Journal of Constructional Steel Research 72 (2012) 155167

a) Hinge Plastic Rotations

20
18
16
14

Floor

12
10
8
6

NLTH

MCMP

MPA(3 Modes)

0
0.000

0.001

0.002

0.003

0.004

0.005

0.006

Hinge Plastic Rotation (Rad)

b) Errors of Hinge Plastic Rotations

20
18
MCMP

16

MPA

14

Floor

12
10
8
6
4
2
-120%

-100%

-80%

-60%

-40%

0
0%

-20%

20%

Error
Fig. 12. Height-wise variation of the hinge plastic rotations and their errors for the 20-story structure.

a) Hinge Plastic Rotations

30
27
24

Floor

21
18
15
12
9
NLTH
MCMP
MPA(5 Modes)

6
3
0
0.000

0.001

0.002

0.003

0.004

0.005

0.006

0.007

0.008

0.009

0.010

Hinge Plastic Rotation (Rad)

b) Errors of Hinge Plastic Rotations

30
27
24
21
MCMP

18

Floor

166

MPA(5 Modes)

15
12
9
6
3

-120%

-100%

-80%

-60%

-40%

-20%

0
0%

20%

Error
Fig. 13. Height-wise variation of the hinge plastic rotations and their errors for the 30-story structure.

F. Khoshnoudian, M.M.B. Kashani / Journal of Constructional Steel Research 72 (2012) 155167

history analysis demonstrated the reliability of both of them. To sum

up, the time-consuming of the MCMP procedure for estimating seismic demands of tall building in comparing to the MPA procedure is
the another advantage.
Furthermore, the signicant errors obtained from the MCMP procedure in predicting hinge plastic rotations happen where the exact values
extracted from nonlinear time history as an exact solution are small and
therefore the application of the proposed method is on a safe side.
References
[1] Papanikolaou V, Elnashai AS, Pareja. Limits of applicability of conventional and adaptive pushover analysis for seismic response assessment. Report 2005/02. MidAmerica Earthquake Center Civil and Environmental Engineering Department University of Illinois at Urbana-Champaign; 2005.
[2] Sasaki KK, Freeman SA, Paret TF. Multi-mode pushover procedure (MMP) a method
to identify the effects of higher modes in a pushover analysis. Proceedings of 6th US
Nat. Conf. on Earthq. Eng.; 1998. Seattle (Washington).
[3] Moghadam AS. A pushover procedure for tall buildings. Proc. 12th European Conference on Earthquake EngineeringLondon (United Kingdom): Elsevier Science
Ltd; 2002. p. 395.
[4] Chopra AK, Goel RK. A modal pushover analysis procedure for estimating seismic
demands for buildings. Earthquake Eng Struct Dyn 2002;31:56182.
[5] Chopra AK, Goel RK, Chintanapakdee C. Evaluation of a modied MPA procedure assuming higher modes as elastic to estimate seismic demands. Earthquake Spectra
2004;20(3):75778.
[6] Jan TS, Liu MW, Kao YC. An upper-bound pushover analysis procedure for estimating
the seismic demands of high-rise buildings. Eng Struct 2004;26:11728.
[7] Aydinoglu MN. An incremental response spectrum analysis procedure on inelastic
spectral displacements for multi-mode seismic performance evaluation. Bull
Earthquake Eng 2003;1:336.

167

[8] Kalkan E, Kunnath SK. Adaptive modal combination procedure for nonlinear static
analysis of building structures. ASCE. J Struct Eng 2006;132(11):172131.
[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] Chopra AK. Dynamics of structures. Theory and applications to earthquake engineering, 2nd edEnglewood Cliffs (NJ): Prentice Hall; 2001.
[11] Applied Technology Council. ATC-40. Seismic evaluation and retrot of concrete
buildings. Vol. 12. Redwood City (California); 1996.
[12] Building Seismic Safety Council (BSSC). NEHRP guidelines for the seismic rehabilitation
of buildings. FEMA-273. Washington (DC): Federal Emergency Management Agency;
1997.
[13] Building Seismic Safety Council (BSSC). Pre-standard and commentary for the seismic
rehabilitation of buildings. FEMA-356. Washington (DC): Federal Emergency
Management Agency; 2000.
[14] Fajfar P. Capacity spectrum method based on inelastic demand spectra. Earthquake Eng Struct Dyn 1999;28:97993.
[15] Fajfar P. A nonlinear analysis method for performance based seismic design.
Earthquake Spectra 2000;16(3):57392.
[16] Fajfar P, Gaspersic P. The N2 method for the seismic damage analysis of RC buildings.
Earthquake Eng Struct Dyn 1996;25:3146.
[17] Mwafy AM, Elnashai AS. Static pushover versus dynamic analysis of R/C buildings.
Eng Struct 2001;23:40724.
[18] Tso WK, Moghadam AS. Pushover procedure for seismic analysis of buildings. Progress
in Struct Eng Mat 1998;1(3):33744.
[19] Kim S, D'Amore E. Pushover analysis procedure in earthquake engineering. Earthquake
Spectra 1999;15:41734.
[20] AISC-ASD. Manual of steel construction, allowable stress design. Chicago (IL):
American Institute of Steel Construction; 1989.
[21] Standard No. 2800-05. Iranian code of practice for seismic resistant design of buildings.
3rd ed. Building and Housing Research Centre, Iran; 2005.
[22] Computers, Structures Incorporated (CSI). SAP 2000 NL. USA, CA: Berkeley; 2004.